The Maritime Landscape of Roman Britain: Water transport on the coasts and rivers of Britannica 9781407309583, 9781407322322

Table of contents :
Front Cover
Title Page
Copyright
Acknowledgements
Contents
List of Figures
Introduction
Chapter 1: Changes in the coastal and riverine landscapes
Chapter 2: Roads and waterways
Chapter 3: Boats, craft, and ships of the period
Chapter 4: Supply and provisioning of the Roman army
Chapter 6: Roman sea power and the classis Britannica
Chapter 7: Naval operations in Britannia
Chapter 8: Harbours, ports landing places
Chapter 9: The development of the inland waterways
Chapter 10: Cargoes within the Province
Chapter 11: Imports and exports
Chapter 12: Patterns of demand
Conclusion
Appendix 1
Appendix 2
Appendix 3
Appendix 4
Bibliography
Index of places

Citation preview

The Maritime Landscape of Roman Britain Water transport on the coasts and rivers of Britannia

James Ellis Jones

BAR British Series 556 2012

ISBN 9781407309583 paperback ISBN 9781407322322 e-format DOI https://doi.org/10.30861/9781407309583 A catalogue record for this book is available from the British Library

BAR

PUBLISHING

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

 

Acknowledgements

I am grateful to the following for laying the foundations of my interest in this topic. The ‘Appledore Boys’ of the wartime 1940’s for showing me how to enjoy the sea, and William Quance, Master Mariner, for teaching me to respect its power. The Army Apprentices College at Beachley for giving me the opportunity to pass this knowledge on to young soldiers. The Officers and Men of the 1st Battalion of the South Staffordshire Regiment, for making me a soldier. The members of Weston Bay Yacht Club, for interesting times, in old wooden boats, on the muddy waters of the Bristol Channel and to Martin Stott for allowing me, in British waters and on overseas passages, to “skipper” the yachts I could not possibly afford to own. At the University of Bristol I owe a debt of gratitude to Toby Parker, Mick Aston, Richard Harrison and Mark Corney for their forbearance with a somewhat geriatric and self-opinionated student, and to my fellow undergraduates of the ‘The Class of 2000’, who accepted the most mature of mature students, and welcomed me to their parties! From the University of St. Andrews, Colin Martin was not only most generous in his comments on my doctoral thesis but, at an earlier time, unknowingly gave my flagging spirits the will to continue with my research. Regrettably, my first attendance at the “Roman Army School” at Durham did not take place until 2009 and so, until that time, I was not directly exposed to the wisdom of David Breeze and Brian Dobson. Lest the reader may form the opinion that my thinking is solely conditioned by members of my own generation, I must pay tribute to the influence of the “younger guns” – Birgitta Hoffmann, Adrian Goldsworthy, Nick Hodgson, Matt Symonds, Jorit Wintjes, and David Woolliscroft. Earlier recruitment to the ranks of the Hadrianic Society would have much improved my work.

i

JAMES ELLIS JONES

 

Contents Acknowledgements Contents List of Figures

i ii iii

Introduction Chapter 1 Chapter 2 Chapter 3 Chapter 4 Chapter 5 Chapter 6 Chapter 7 Chapter 8 Chapter 9 Chapter 10 Chapter 11 Chapter 12 Conclusion

1 8 16 33 46 53 61 69 78 86 97 108 117 120

Changes in the coastal and riverine landscapes Roads and waterways Boats, craft, and ships of the period Supply and provisioning of the Roman army Military operations in Britannia Roman sea power and the classis Britannica Naval operations in Britannia Harbours, ports landing places The development of the inland waterways Cargoes within the Province Imports and exports Patterns of demand

Ancient Sources

121

Appendices 1 2 3 4

Gazetteer of locations on the coasts, estuaries and rivers Admiralty tidal stream atlas – Irish Sea and Bristol Channel Stowage factors Relative usage of Antonine Itinerary roads

122 171 173 174

Bibliography

175

Index of places

207

ii

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

 

List of Figures

1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 1.10 1.11 1.12 2.1 2.2 2.3 2.4 2.5 2.6 2.7 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 4.1 5.1 6.1 6.2 6.3 6.4 6.5 6.6 7.1 7.2 7.3 7.4 7.5 8.1 8.2 8.3 8.4 8.5 8.6 8.7 8.8 8.9

Unloading on the River Avon at Portus Abonae (Sea Mills) during the 1st century Remains of the signal station at Scarborough Coastal erosion at Reculver Deposition at Richborough Coastal change since the Roman period Line of the successive Roman and medieval waterfronts at Gloucester Romney Marsh area – present day Probable area of Romney Marsh c.300 Diagram showing Roman London waterfront development from the 1stto 3rd centuries Elevation of the 3rd-century quay at Caerleon Changes in the River Dee at Chester The port of Lilstock in the mid-19th century The remains of the port at Lilstock Ermine Street approaching the Humber Estuary Relative usage of Antonine Itinerary roads Oxen towing cart laden with barrels (Trajan’s Column) Four-wheeled post-wagons (angaria) (Trajan’s Column) Roman roads in Britannia Barbarian incursions of the late 3rd-century AD City crest of La Coruña showing late 1st-century Roman Lighthouse Coracle on the River Teifi Reconstruction model of Ferriby 1 Reconstruction model of the Brigg “raft” Reconstruction of the remains of the Barland's Farm boat, as excavated Reconstruction model of the Barland's Farm boat Unloading from sea-going vessel to shallow-draught barge Cut-away reconstruction of the Blackfriars ship Unloading Kentish ragstone Reconstruction of the Gallo-Roman trading vessel wrecked at Guernsey Seventeenth century drawing of a large Irish sailing curragh Gold model of 1st-century BC Broighter boat Interpretation of obverse of bronze coins of Cunobelin Mules towing carroballista (Trajan’s Column) Nineteenth century reconstruction of the “Saxon Shore” type fort at Cardiff Transporting supplies during the Dacian Wars (Trajan’s Column) Illustration of 1st/2nd-century bireme Trireme/?bireme on the Danube (Trajan’s Column) Tombstone of an optio who died in a shipwreck – “naufragio periit” Tile-stamps of the classis Britannica Troops being transported by barge on the Furnes-Dunkerque canal - August 1917 Marching camps in Wales Alternative strategy for the conquest of West Wales by amphibious operations. Gold medallion depicting the guardian spirit of Londinium kneeling to Constantius I The Nydam No 2 oak boat exhibited at Flensburg in 1865 Remains of the watch-tower on Holyhead Mountain, Anglesey The pebble beach at Cold Knap at low tide Third-century quay at Caerleon during excavation Quay wall at Caerleon adjacent to the site of the fortress of Isca Section of the Roman “quay wall” at Chester Drawing (1885) of iron shoes of wooden piles Remains of the Roman lighthouse at Dover Artists impression of the colonia at Gloucester Silted-up inlet with rock outcrop Boat channels at Littlecote Roman villa

iii

Title page 8 8 9 9 10 11 11 12 13 14 14 15 16 16 18 18 19 22 24 33 34 34 34 35 36 37 38 39 40 41 42 48 59 63 63 64 66 67 67 71 72 75 76 77 78 79 79 79 80 80 80 81 81

JAMES ELLIS JONES

  Blackfriars type vessel beached to unload at 3rd -century London waterfront Dedicatory coins from ferry crossing at Sudbrook Fresco in the Vatican Museum Loading quarried limestone in the Avon Gorge The trow “Palace” beached to unload stone for the construction of sea defences Remains (c. 1998) of a jetty at Cone Pill, Woolaston The Old Passage Ferry at Aust in 1965 The ferry stage remains in 2009 River Severn at Caersws Watersheds and main rivers of Britain Admiralty chart of wrecks in Yantlet Creek The embanked Lincolnshire Car Dyke The “Pilrow Cut”, completed in 14th century by direction of the Abbot of Glastonbury The improved 13th-century cut of the River Clwyd at Rhuddlan Castle Average loads on some rivers during the medieval period Water transport on medieval rivers Areas more than ten miles from a navigable waterway in c.1600 Areas with access to “improved” waterways by 1760 Silted-up port at Parkgate on the River Dee Abandoned port at Sunderland on the River Lune Skinner’s Quay, Chester 1836 Mineral resources of Britain Lead ingot from Chester Artist’s impression of the legionary fortress, canabae and bridge at Chester Distribution of Dressel 1 amphorae in Western Europe Major rivers and roads of Gaul Stowing barrels on the Danube (Trajan’s Column) Trade routes from the Mediterranean Post-Roman/Dark Age trading site at Tintagel Appx 2 Admiralty Tidal Stream Atlas – Irish Sea and Bristol Channel Appx 3 Stowage Factors Appx 4 Relative use of Antonine Itinerary roads

8.10 8.11 8.12 8.13 8.14 8.15 8.16 8.17 9.1 9.2 9.3 9.4 9.5 9.6 9.7 9.8 9.9 9.10 9.11 9.12 9.13 10.1 10.2 10.3 11.1 11.2 11.3 11.4 11.5

82 82 83 83 84 84 84 84 88 88 89 89 91 91 92 92 93 93 95 95 95 97 98 101 109 109 110 116 116 171 173 174

The author is extremely grateful to the individuals and institutions noted below, who have given their permission for the reproduction of illustrations in their ownership. Chris Molan, Title page; Chrissie Milne, Figure 1.8; John Allen, Figure 8.15; Sir Barry Cunliffe, Figures 1.6; 1.7; 2.6; 11.1:11.2; Hugh Davies, 2.2; Sheppard Frere, Figures 2.3; 2.4; 4.1; 6.1; 6.3; 11.3; John Langdon, Figures 9.7; 9.8; David Mattingly, Figures 1.4; 2.5; 9.2; 10.1; Bryn Waters, 8.9; Roger White, Figure 12.5; Cambrian Archaeological Association, Figures 1.9; 8,2; Cambridge University Press, Figure 3.3; Cheshire and Chester Archives and Local Studies Service, Figure 9.13; Chester Archaeological Society, Figure 8.5; Chester City Council, Figures 1.10; 6.4; 8.4; 10.2; 10.3; Gloucester City Museum and Art Gallery, Figures 1.5; 8.7; Guernsey Museum and Art Gallery, Figure 3.9; Imperial War Museum, Figure 4.6; Institute of Archaeology, Oxford, Figure 3.12; Museum of London, Figures 3.7; 3.8; 3.6; National Maritime Museum, Figures 3.2; 3.3; National Museum of Ireland, Figure 3.11; Newport Museum and Art Gallery, Figures 3.4; 3.5; Oxford University Press, Figure 11.1; Pepys Library, Magdalene College, Cambridge, Figure 3.10; United Kingdom Hydrographic Office, Appendix 2.

iv

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  without hindrance, since we have no enemy in our rear” (B.Civ.4.12,100). Livy’s description of the base used during the Istrian campaign of 178 BC provides an interesting commentary on the role of water transport in support of the army, and the consequent effect on local development. “The ships were sent to the nearest harbour in Istrian territory with transports and a large quantity of supplies and ... the legions encamped about five miles from the sea. In a short time a market was established by the harbour, and from there everything was transported to the camp” (41.1.4–6). During the campaign in Germany, Tiberius realised that “His own troops were affected not so much by wounds as by long marches and damage to their arms. Gaul had been exhausted by supplying horses; a long baggage-train presented facilities for ambuscades, and was embarrassing to its defenders. But by embarking on the sea, invasion would be easy for them, and a surprise to the enemy, while a campaign too would be more quickly begun, the legions and supplies would be brought up simultaneously, and the cavalry with their horses would arrive, in good condition, by the rivermouths and channels, at the heart of Germany.” (Tac. Ann. II, 5; XIII, 53).

Introduction This volume had its genesis in a University of Bristol extra-mural essay on the River Severn (Sabrina Fluvis) during the Romano-British period, followed by an undergraduate dissertation on Roman water transport on the Bristol Channel, then a doctoral thesis extending the area to Wales and The Marches, and finally a British Archaeological Report with an expansion to cover the Atlantic facing coasts and rivers of Britannia. It probably therefore suffers from the fact that, like Topsy, it “just growed”! It is essentially a work of synthesis, gathering together various strands of evidence and, through systematic analysis, arguing the seas and rivers were a major factor in the Roman supply system and that, whilst the system of roads has received much attention, insufficient consideration has been paid to the role of water transport. It will be suggested that, on the coasts and rivers of Roman Britain, the role and extent of water transport was governed initially by military necessity, and later by economic expediency. To some extent it also diverges from the well-beaten track of Romano-British history as, for example, it argues that “the perils of Land’s End” and “the dangers of the west coast route” are much exaggerated, and that there is considerable evidence that voyages, under sail, round the coasts of Britain, and in particular the south-west peninsula, are not the difficult undertaking suggested by some writers (e.g. McGrail 1985, 16; Peddie 1997, 137– 450). In essence, it is a study of the maritime dimension of transport, to and from the military installations and civilian settlements on the coasts and rivers of Britain, where water transport was the means of conveyance of military and commercial traffic during the RomanoBritish period. Evidence will be drawn from a variety of areas, namely classical sources, geomorphological and archaeological evidence for change, on-site assessment of the navigable potential of locations, archaeological data on forts, settlements, ships and boats, and analogous data from later periods. In addition, the writer will draw on experience of sailing on the coasts and rivers of the Britain, as well as delivering yachts to the Mediterranean. However, it is clearly unwise to relate directly modern craft and weather conditions to those of the Roman period, and this temptation will, hopefully, be avoided.

However, in relation to Roman Britain, whilst the period marks the crossing over from prehistoric to historic Britain, “the light shed by conventional documents is dim indeed” and “…we face serious interpretative difficulties that force us to make choices about the shape of our accounts” (Mattingly 2006, 3). Whilst the story of Roman Britain is well endowed with later historical accounts and archaeological interpretations, it is somewhat less well provided with reliable original sources. For example, our only source of significant information on early Roman Britain is the historian Gaius Cornelius Tacitus (56/7 to after 112), whose major work was a history of Rome in the first century AD. The absence of documentary evidence seriously hampers a realistic assessment of the Romano-British economy, and plethora of modern distribution maps is no substitute for a few copies of invoices or bills of lading; a substantial trade in perishable commodities could take place without leaving any trace in the archaeological record (Fulford 1978, 62). Pottery is valuable to an archaeologist as an indicator but, “at best the pots are surrogates for trading other commodities of which little record survives” (Cunliffe 2001, 445) As an assessment of the transport of corn, wine and olive oil is a major theme in this volume, this absence of archaeological evidence, and the paucity of contemporary documentation, means that it has been necessary to assemble information from such diverse sources as a sixth-century Byzantine wreck from off the Turkish coast, and the 17th-century Port Books of the City of Bristol.

Roman writers were certainly aware the advantages of water transport, understood it was routinely used to supply provisions for the army at war and that, when possible, a fleet would accompany an army, carrying its supplies, not only grain, but many other types of foodstuffs, such as meat, fish, vegetables, wine, oil and salt (Polyb. 3.97); Tacitus emphasised supplies transported by sea were generally safer than those sent overland (Ann. 2.5). When describing the campaigns of 42 BC, Appian invented the following speech for Cassius. “Provisions, the supply of which is the chief difficulty in large armies…… must be carried to them and overland with severe labour… we have abundance, brought to us by stable sea without labour……. and

If the sea is to be considered as “a thoroughfare rather than an obstacle to creep around, then to some extent it will have to be treated like a piece of land…it is necessary to examine the “topography” of the sea, its resources, hazards, currents, tides and winds, and then match these to the ship technology we know from

1

JAMES ELLIS JONES

  in transport by water and, to a first approximation, the Roman market for bulk commodities extended only slightly beyond where ships could go, although high value goods could travel to landlocked destinations (Temin, 2001, 180). Duncan-Jones used the pricing edict of Diocletian to suggest a ratio of 1: 4.9: 28–56 for the relative costs of sea, inland waterways and road transport (1982, 366–9). However, when engaged in full-scale warfare it is unlikely that the question of cost, or costeffectiveness, would have played a major part in the planning of military operations. Water transport had a significant role in the movement of troops and the supply of equipment and rations throughout the whole of the Romano-British period, but the amount of traffic varied dramatically over time and space, in direct proportion to the number of army units in the area. During the lengthy periods of peace in Britannia, transport by water provided a viable alternative to road transport, initially for military activity and, from the later first century, for the establishment of urban communities. Cities and most major towns were on, or adjacent to, navigable waterways. In addition to the transport of major cargoes of grain, wine and olive oil, the products of the extractive industries of quarrying, tin, copper and lead were particularly suited for movement by river. The needs of the cities, towns and remaining forts continued the requirement into the second, third, fourth and possibly fifth centuries.

archaeology” (Carver 1990, 119). In order to follow this advice, the chapters within this volume range far and wide, in both time and space, covering a spectrum of apparently unconnected topics as diverse as the technical calculation of a ship’s cargo capacity and the Roman capacity for wine. With this in mind, this introduction is more detailed than is usual and is intended to act as a “taster menu”, in order that the reader may select those areas that are of particular interest. Changes in the coastal and riverine landscape Changes since the Romano-British period are important, as we must attempt to place any interpretation of Roman water transport in the context of what (probably) existed at that time, and not in the landscape that we see today. This is opinion is shared by Peter Murphy, Coastal Strategy Officer for English Heritage, who considered “Textbook maps of Roman Britain based on modern geography may therefore be very misleading” (2009, 34). For example, the east coast of Britain shows the effects of both deposition and erosion, the most dramatic being that, during the Roman period, the coastline of the area now known as The Wash was in some cases as much as 25 km inland of its present position. By contrast, antiquarian accounts from the 17th and 18th centuries record a masonry structure typical of a Roman Saxon Shore fort at Walton Castle, but any remains have now been lost to the erosive effects of the sea. The Ordnance Survey of Roman Britain shows a number of roads leading towards the coast, but indicates no evidence for a settlement in the direction to which they point. Notable examples are those heading in the direction of present-day Huntstanton, Holkham Bay, Southwold, Horsey Skegness and Berwick upon Tweed. The only reasonable conclusion must be that Roman sites in these areas have, like Walton Castle, been lost to the sea.

Prior to the Roman conquest, there is considerable evidence for Iron Age trade with Britain via the Western Seaways. For example, Strabo (1, 14; 5, 2; 3, 4), Ammianus Marcellinus (27, 8) and Caesar (B.Gall. 2, 4) provide much information. It is clear that the conquest of Gaul resulted in a shift in the direction of trade; the longer and more dangerous west coast of routes appear to have declined as internal routes and markets within Gaul became secure (Fulford 1991, 36). Following the opening of the Continental inland waterways, fewer ships used the Western Seaways en route to the Irish Sea than previously but, in some instances, this decline may have been exaggerated. The majority of Spanish oil was shipped through the Straits of Gibraltar to either Italy for direct consumption, or to Marseilles for onward shipment via the Gaulish river and road systems. Cunliffe (2001, 421) is undoubtedly correct in his statement “that Mediterranean trading fleets regularly braved the Atlantic seems, at least on present evidence, highly unlikely”. However, this does not necessarily detract from the validity of the argument that, in the case of nonMediterranean traffic, for much of the Romano-British period, there was a flourishing Atlantic coast trade in both wine and olive oil. It will be later suggested that it would have made sense, for at least the western British component, to be shipped direct via the Western Seaways; the case for this, as both a logical and costeffective alternative, will be made in some detail.

Similarly, changes to the navigability of rivers, as the result of both channel change and variable tidal heights, may take place over a comparatively short time, and occurred throughout some 400 years of Roman occupation. There is archaeological evidence from Roman London that the tidal level of the River Thames fell by as much as 1.5 m between the end of the first and the middle of the third centuries (Milne, 1995, 78–81; Brigham 2001, 15–49). In order to maintain a workable depth of water for lying alongside in order to load and unload, the bases and tops of successive quays were constructed at a lower level than their predecessors. Roman roads and waterways The Roman roads of Britain were initially constructed as essential lines of communication and routes of supply for the campaigning army, and this remained their primary function throughout the Romano-British period. Whilst civilian use was permitted, this was carefully regulated, as is demonstrated by the weight restrictions imposed by the Codex Theodosianus. Despite the (generally) high standard of Roman roads, there is a clear cost advantage

Merchant ships, coastal vessels, boats and local craft of the period A wide range of ships and boats was available during the 2

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  When on campaign the Roman army took a large quantity of equipment and supplies into the field. In addition to the individual soldiers’ personal weapons and equipment, this included artillery, missiles, portable fortifications, tents, medical supplies, cooking gear and much more. However, approximately 90% of the weight of supplies needed by the Roman army was made up by the requirements of food for the soldiers and fodder for the animals. The volume of consumables involved was significant, for example, based on a campaigning army of 20,000 soldiers (1 kg per day) and 7,000 military slaves (0.8 kg per day) there would be a monthly requirement for 768,000 kg of grain, and hard fodder for the 8,600 horses and pack animals (2.2 kg per day) would add a further 588,000 kg. The soldiers’ daily ration of wine (0.27 ltrs) and olive oil (0.07 ltrs) would amount to 204,000 ltrs). The basic grain ration was supplemented by meat, vegetables (especially legumes), cheese, oil, vinegar, salt and wine, indicating that the Roman soldier enjoyed a rich and varied diet. However, many of these commodities were not available whilst serving in the field; others such as meat would have followed the army “on the hoof”. Whether stationed in winter quarters, in times of peace, or later in permanent barracks, many of the minor items formed part of bulk consignments of wheat, or loads of amphorae containing wine or olive oil. These commodities would not create a significant addition to the volume of transport required, forming only a small part of a major consignment.

Romano-British period, and were clearly capable of meeting all water transport needs, both civilian and military, on the open seas, estuaries and inland waterways. We are fortunate that there are a number of excavated Roman wrecks in the British Isles and these are fully discussed in this chapter. Sea-going ships are represented by the late third-century wreck found off St Peter Port harbour in Guernsey in the Channel Islands and the Blackfriars ship of the second century, much of this vessel and its final cargo having been recovered from the bed of the River Thames in London. A smaller vessel, of the third century, capable of coastal and cross-Channel passages has recently been found at Barland’s Farm in south Wales, two miles inland from the present foreshore of the Bristol Channel. In London, the New Guy’s Hospital boat provides an example of the type of river barge in use during the late second-century, probably for trans-shipping from sea-going ships entering the Thames Estuary. Much has been written about the types, origins and methods of construction of vessels of the times but, interesting though this may be, it is not proposed to enter into any detailed discussion. In order to enable a reasoned assessment of the nature of voyages, on either the seas or inland waterways, the emphasis will be placed on factors such as sea-keeping and passage-making. Romano-British vessels display similar characteristics to those of much later periods, for example, the Blackfriars ship, the St Peter Port ship and the ships of the Veneti (B.Gall. 3.13), have much in common, in terms of design and construction, with the medieval cog, possessing similar sea-keeping, passagemaking and cargo-carrying capabilities.

Military operations in Britannia This chapter is intended to form the preamble to a later examination (Chapters 8 and 9) of the role of water transport in the Province. It is not intended to provide a comprehensive guide to the campaigns, but rather to indicate the scale of logistical operations needed to maintain the operational capability of the army in times of both war and peace. However, in some cases, military actions are described in some detail, in order to demonstrate their relevance to waterborne operations.

A wide range of “native craft” were also in use, for example, simple log boats, made by hollowing out a single log, and shaping the ends and the outside, are known from the Mesolithic, continuing in use in parts of Europe until the 20th century. Various forms developed, such as the Hasholme logboat of c. 300 BC, which could carry a crew of two steersmen and eighteen paddlers, or a cargo of 5.5 tonnes with a five-man crew (Millett & McGrail, 1987). Caesar (BGall. 1.12) commented that the Celts used log rafts to cross rivers in Gaul; two log rafts of the second century were recovered from the River Rhine near Strasbourg in 1938 (Ellmers 1978, 106, Figs. 83/4). It is probable that rafts were widely used on lakes, rivers and in estuaries; however, it is unlikely that they would have been used successfully on the open sea.

There were four occasions on which an invading Roman army landed on the coast of Britain, firstly in 55 BC when Julius Caesar led a reconnaissance in force followed, a year later, by the full-scale invasion, thirdly the Claudian invasion of 43 that resulted in the establishment of Roman Britain, and finally in 296, when Constantius Chlorus recovered the province from the usurpations of Carausius and Allectus. These events have received extensive coverage, perhaps most notably by Gerald Grainge in The Roman Invasions of Britain. This chapter therefore commences with the conquest of the south-west peninsula immediately following the Claudian invasion by an army of some 40,000 soldiers, and concludes with the final abandonment of the Province by an army that, by then, numbered less than 10,000. The change was not only in numbers, but in the character of military operations. With only a few exceptions, the conquest and consolidation of most of Britain presented few problems. The mural frontiers of Hadrian and

Supply and provisioning of the Roman army The establishment of an effective supply and replenishment system, and using it as a strategic and tactical tool, may be considered one of the major factors in Roman military success. Indeed, Vegetius (Mil. 3.26) emphasises this by quoting the military proverb, “Whoever does not provide for provisions and other necessities, is conquered without fighting”.

3

JAMES ELLIS JONES

  Scotland, then southward down the coast, crossed the North Sea and eventually reached the territory of the Frisii on the coast of Holland (Ag. 28).

Antoninus proved effective and, on the occasions when the northern tribes rebelled, the campaigning armies of, for example, Commodus and Severus, brushed aside any opposition. By comparison with other parts of the northern Europe, Britannia remained peaceful, its major problem being its ability to produce a crop of relatively successful usurpers. By contrast, the late fourth century saw the development of a siege mentality, with the remnant of the army either confined to the remaining forts of Hadrian’s Wall and its hinterland, a handful of Welsh coastal garrisons or the Saxon Shore forts.

Naval operations in Britannia There is no evidence from Roman Britain to compare with naval victories such as Caesar’s defeat of the Venetic fleet and, when the army operated over extended lines of communication, particularly in the Highland Zones, the more mundane role of naval forces was the supply and replenishment of the army. However, in his description of Agricola’s campaigns in Scotland, Tacitus (Ag.25) refers to units of the fleet being deployed in offensive operations in support of the army, with the war being “…pushed forward simultaneously by land and sea, the infantry, cavalry and marines often meeting in the same camps”, with the fleet being sent ahead “to plunder at various points and thus spread uncertainty and terror”. He describes how during the course of combined operations, the soldiers of the army and the sailors of the fleet often “matched the perilous depths of woods and ravines against the hazards of storms and waves, victories on land against the conquest of the ocean”. There is some epigraphic evidence to suggest the concept of naval forces, other than in the south east of Britain, operating under the direct control of the military commander, and as an integrated part of the land forces. Association of the legion with naval vessels is demonstrated at York, where there was an inscription dated c. AD 122, dedicated to Marcus Minucius Audens, described as gubernator (pilot) of VI Victrix (RIB I 653). Whilst not necessarily demonstrating membership of a naval unit, a memorial from Chester (RIB 544) shows that an optio of legio XX never achieved his expected promotion to centurion, as he lost his life in a shipwreck.

Roman sea power and the classis Britannica At the commencement of the third century BC, Rome was essentially a land-based power, but the Punic Wars with Carthage necessitated far-reaching changes and, after the defeat of the Carthaginian fleet at Mylae in 262 BC, Rome celebrated her new status as a naval power by erecting a column in the forum, decorated with the beaks of the captured vessels. Naval operations played an important role in the civil wars of the first century BC, but, after Octavian’s defeat of the fleets of Mark Antony and Cleopatra at the Battle of Actium in 31 BC, major naval battles within the Mediterranean were a thing of the past and Roman could rightfully lay claim to mare nostrum. Caesar’s defeat of the Venetic fleet, off the Gulf of Morbihan in 56 BC (BGall. 3. 9–13) eliminated any threat to naval supremacy on the Atlantic coasts - the creation of the classis Britannica with its headquarters at Boulogne and the classis Germanica, based on the Rhine at Altenburg, ensured continued control of the European waterways. The classis Britannica was “very much a Romano-British institution, rather than just another Roman fleet” (Milne (2000, 130), operating in a different manner to the fleets at Misenum and Ravenna, and to these might be added the classis Germanica and other (mainly) river based fleets such as those of Moesia and Pannonica. Throughout their histories, these fleets were actively concerned with the protection of the frontiers based on the Rhine and the Danube, and as such, their function was “naval” in the true sense of the word, rather than the cross-Channel transport of supplies, provisions and personnel, that became the major role of the classis Britannica in the post-Conquest period. However, in the conquest and consolidation phase, an important role was close support of the army, possibly under the control of the local military commander. The demolition of the squadron headquarters at Dover took place c.215, and brought an end to a major base for the classis Britannica, but not to naval operations in Britain.

Starr (1960, 153) commented that, after the time of Agricola’s governorship, “an almost unrelieved blackness hides any naval actions north of the English Channel for the next two centuries”. In fact, naval operations in support of the army occurred, for example, during the Severan operations in Scotland, but for the majority of the period the major function was the operation of seaborne replenishment of the Roman military, both men and materials, and the conveyance of the cursus publicus. A notable exception is Maximian’s attempt in 288 to recover the breakaway Province from of Carausius and Allectus. Despite the face-saving propaganda of a failure attributed due to bad weather, his invasion fleet was forced to withdraw without landing, perhaps following a naval battle. However, sea raiding and piracy became endemic, particularly in the North Sea, and operations were restricted to the use of isolated detachments of naval vessels, based at some of the remaining coastal forts and facing a losing battle in attempting to protect a long and vulnerable coastline. Defence of the shores of Britain against sea-borne raiders, became a near impossibility, whether attempted by land-based forces, or by ships of the local fleet. There is some indication of small naval units endeavouring to protect vital centres of settlement

Although a relief at Boulogne (CIL XII 3564) refers to a trireme, the “Radians”, it is probable that liburnae were the predominant warships in the classis Britannica. This type of vessel was used in a variety of roles including conveyance of dispatches, transport of fleet officers and as scouting and patrol vessels. Their sea-worthiness in northern waters is indicated by Tacitus who records how, in 83, three liburnae were “hijacked” by a cohort of Usipi, possibly in the Firth of Clyde; sailed north round 4

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  garrison. At one extreme are the harbours of the Solway Firth and the Clyde estuary that remained in operation from the Flavian period, until the early fifth century. At the other extreme, is the River Tay and its estuary which, in the first century was used for the support of the Gask “frontier” and the Agricolan campaigns, but then remained largely unused until briefly used to support the Severan campaigns.

from the threat of sea raiders, but it is clear that this activity was “too little and too late”. Harbours, ports and landing places An understanding of the level of competence of Rome in matters of port construction and inland water transport, in other parts of the Empire, is necessary for an understanding of the degree of exploitation of estuaries and rivers in Britannia. Roman military engineering was more than capable of improving and maintaining these rivers, thereby establishing an effective system of inland waterways. It will be argued that the waterways of Roman Britain were exploited to a greater extent than at any period prior to the Canal Age of the 18th century. This was made possible by a combination of engineering skills, military necessity, economic advantage, plentiful labour resources and suitable vessels.

The problems of locating archaeological evidence for Roman ports and harbours is well illustrated by a report in British Archaeology (1999, Issue 50), confidently stating that “A man-made Roman harbour linked to the River Severn has been discovered at the Roman town of Wroxeter in Shropshire”. Despite efforts to recover structural remains, six years later it was necessary to state that “No evidence has yet been forthcoming about the harbour facilities within the town, which must have existed given the level of imports evidenced by artefacts” (White & Dalwood 2005, 18). Similarly, the fort of Arbeia, lying at the mouth of the River Tyne, four miles east of the terminus of Hadrian’s Wall, is the most extensively excavated Roman military supply-base in the Roman empire but no Roman port at South Shields has been physically located and “it must be conceded that the port could have been located anywhere along the south bank of the river within a reasonable distance of the fort” (Neil Hodgson, pers. comm.).

There are significant differences between the characteristics of the western and eastern coasts of Britain, not only in geomorphology but, for this examination, more importantly, is the greater density of military installations on the western coasts. According to the Fifth Edition of the Ordnance Survey Map of Roman Britain, with the exception of the Yorkshire signal stations, Lease Rigg and the Saxon Shore forts, no Roman fort is shown either on, or within, 20 km of the coast between Hadrian's Wall and Dover. On the west coast, between the Wall and the Bristol Channel, there were at least 25 forts. It is probable that forts also existed on the east coast at Whitby, Bridlington, Skegness and Huntstanton, but have been lost to coastal erosion and also on the west coast, in the area of Fleetwood, in Lancashire, and St David’s, in Pembrokeshire, but these have yet to be located. When compared to the south of Britannia where, for example, the Great Ouse was navigable from the coast at King's Lynn to St Ives near Cambridge, and the River Severn was navigable to, at least, Pool Quay, some 250 km from the Bristol Channel, northern Britain was less well provided with navigable rivers that penetrated far inland. The Tay (Tava) was navigable as far as the fort at Bertha and, with some improvement, might have been made navigable as far as the short-lived fortress at Inchtuthil; its tributary the Isla provided access to Cargill and possibly to Cardean. The Firth of Forth provided a number of harbours and/or landing places at Inveresk, Cramond and Carriden, seagoing vessels were able to reach Stirling and the River Carron gave access to Camelon. From its estuary, the River Clyde gave access by seagoing ships as far as Old Kilpatrick, with possible barge access further upriver to the isolated fort at Bothwellhaugh. The River Dee in Galloway was navigable from Kirkcudbright Bay to Castle Douglas and the fort at Glenlochar; the River Nith gave river access to Dumfries and the line of forts on Nithdale. The Esk may have been navigable in the direction of Netherby and the Eden as far as Carlisle.

The development of the inland waterways There has been considerable research into the navigation of mainland European rivers, and the use of the Ruhr, Rhine, Garonne and Rhone as major trade routes is well attested, but the rivers of Britain have not attracted similar attention. It is clear that the degree of use, even on the largest of the English rivers such as the Thames and Severn, would be less extensive than on their European counterparts. Votive reliefs belonging to the Shipmasters' Guilds indicate that even small tributaries of the major European rivers were included in the Roman water transport system (Eckholdt 1984, 3–9). It will be shown that this also was applicable to some British tributary rivers, such as the Teme and the Lugg. The demand for an efficient system of water transport existed. Clearly, it would have been necessary to carry out a system of riverbed and bank maintenance, clearing the river of obstructions and maximising the depth of available water. Based on evidence from other parts of the Empire, it will be argued that Roman military engineering was more than capable of improving and maintaining these rivers; thereby establishing an effective system of inland waterways. Writing in the sixth century, the British cleric Gildas was clearly aware of the historical importance of river transport as he states that Britain “… has the advantage of two noble rivers, the Thames and the Severn, arms as it were, along which, of old, foreign luxuries were wont to be carried by ship” (De excidio et conquestu Britanniae, 3). It is clear that, throughout the Roman period, the

All these waterways served, for varying periods, as supply routes for the army, either on campaign or in

5

JAMES ELLIS JONES

  fabric of a period of almost four hundred years.Whilst the rate of change was less dramatic than that of recent times, there is clear evidence for major variations of supply and demand, with a knock-on effect on the volume of shipping. This was, to a large extent determined by the predominant economic effect of the garrison, ranging from some 40,000 soldiers plus dependents, in the late first/early second centuries to as little as 10,000 in the late fourth century. Whilst this was the most significant determinant, other factors have needed to be taken into account, for example, changes in the sources of supply or in style of living. Whatever the level of demand, it is only relevant if a suitable source of supply is available. For example, Iberia apparently ceased to export olive oil in the later third century and this was mirrored by an increase in a rise of imports from North Africa. However, the remains of amphorae from Tunisia found, for example, at Caerwent, Gloucester and Cheddar, are insufficient to compensate for the absence of the Spanish product, indicating that the use of olive oil in Britain declined during the later Roman period (Tyers 1996, 72). The taste for the Mediterranean (Roman) way of life may have varied, for example, at Vindolanda; the Batavian troops were complaining that they had run out of beer, not wine. This type of change may sometimes be detected in the archaeological record, for example, in the virtual absence of Oxfordshire Ware drinking vessels in fourthcentury town houses at Beeches Rd, Cirencester. The excavators point out the marked contrast to periods when beakers and flagons were much in evidence and suggest that, by comparison with earlier periods, this indicates a fall in wine consumption, (McWhirr et al. 1986, 175).

Severn was a river of considerable importance. Above Wroxeter, the river is fordable at a number of points and the siting of forts on the riverbank upstream of this point is probably a reflection not of military strategy, but rather an indication of the use of the navigable potential of the river. The Wye was navigable at least as far as Hay on Wye and the Usk, Avon, Parrett, Dee, Conwy, Mersey, Ribble, Eden, Esk, Annan and Clyde were also rivers with significant inland traffic. Forts were located on or near the estuaries of rivers such as the Seiont, Tywi, Dyfi, Llugwy, Ystwyth, Mersey, Ribble, Lune, Ellen, Nith, Urr and Clyde. By contrast with later periods, Roman law provided legally enforceable constraints protecting the rights of navigation. This compares favourably with the situation facing the ‘improvers’ of the 17th/18th centuries who, before any work could be commenced, needed to spend much time and effort in pursuing legislation for the removal of obstacles. The work of the jurist Domitius Ulpianus dealt, inter alia, with the protection of rights of navigation on inland waterways. The matter is dealt with in some detail in Chapter 7, and clearly gives support to the contention that river navigation, during the RomanoBritish period, was not impeded by the weirs etc. that bedevilled medieval waterways. The extent to which this legislation was enforced, or enforceable, cannot be ascertained. However, it does at least show that, if an obstacle needed to be removed, it could be done without recourse to the tedious legal processes that made the later improvement of English rivers such a drawn-out procedure.

Cargoes from within the Province

Patterns of demand

Britain was rich in minerals - Tacitus (Ag. 32) wrote of “gold, silver and other metals”, Pliny described lead “in abundance” (HN 34, 17. 164), Diodorus Siculus wrote of large quantities of tin. In addition, the Province was rich in iron, coal, copper, salt and grain was produced in quantity (Strabo Geog. 4. 5.2); all these commodities provided cargoes for water transport. The enormous volumes of material needed for the construction, and periodic reconstructions, of forts and fortresses led to periodic fluctuations in the volume of shipping. For example, it has been estimated that 35,000 cubic metres of stone were used in the construction of the walls of Londinium in the early third century (Marsden 1980, 126–7). This would have needed 1,750 voyages, by ships of similar size to the Blackfriars ship, to bring about 45,000 tonnes of ragstone from the source, some 112 km away at Allington on the west bank of the Medway. The army played a predominant role in the British economy and units at the frontier zones, and at strategic positions along the lines of communication, creating a series of micro-economies in the immediate environs. The establishment of a large standing garrison on Hadrian’s Wall required the regular long-distance shipment of food supplies from the grain-growing areas of the south and it is probably no coincidence that this is also the period at which the earliest instances of large-scale, long-distance,

In order to assess the pattern of cargoes, an attempt is made to examine the factors creating the level of consumer product demand and the consequent impact on the requirement for transport, whether by sea or by land. The number of consumers is a vital factor in any calculation of consumption, and therefore the need for transport, but the population of Roman Britain is notoriously difficult to estimate (for example, see Millett 1990, 181–6). This at first appeared to pose a serious problem, but it was eventually realised that the rural population, whatever its absolute size, was a very occasional and limited user of imported consumer products such as olive oil and wine, and was selfsufficient in terms of its basic food supply. The size of the military population, at various periods, is determinable from the archaeological evidence of occupied fortresses and forts; a reasonable estimate of military dependents may be made for canabae and vici. There is sufficient evidence of the extent of urbanisation to produce a population estimate for this sector and, although there are undoubtedly many villa sites still awaiting discovery, the size of this population is capable of derivation. The term “Romano-British period” is used throughout this study but does, in fact, tend to obscure the diverse 6

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  have been necessary to maintain a sea-transport capability able to respond to such a demand.

distribution of pottery to the north occur. Middleton argued there is no evidence that the civilian market generated long distance trade and suggests that parasitic growth was generated in areas “touched by the nozzle of the military vacuum cleaner, sucking supplies to the frontier” (1979, 91–94).

-o-o-o-o-o-o-oIt is perhaps appropriate to conclude with a quotation from the 1993 Annual Caerleon Lecture (In Honorem Aquilae Legionis II Augustae) given by the late Michael Jarrett.

Imports and exports Until the later days of the Romano-British period, the volume of exports was greatly exceeded by the quantity of goods imported, consequently return cargoes were often unavailable and voyages back to the Continent were often made “in ballast”. Although imported pottery, in particular samian, has received a considerable amount of attention, the volume of cargo was minute in comparison with that of wine or olive oil, and it has been argued that the importation of pottery was only a minor trade, “tacked on” to bulkier cargoes of perishables (Fulford 1984, 129–42). The valley of the Guadalquivir was a major source for the export of olive oil, often shipped in Dressel 20 amphorae. It seems clear that a few cases of fine ware were included with the major cargo, thus incurring no transport costs and permitting distribution to an area that would normally have been serviced from the major kilns at Lyon. The finds in Britain may have been the result of a similar type of piggy-back cargo loading. Fortunately, wine and olive oil were transported in amphorae and although the perishables themselves have been consumed, amphora sherds survive well in the archaeological record, so providing evidence of patterns and chronologies of distribution. The province often produced a significant surplus of grain, as indicated by the account of the interruption of shipments of British grain to the mouth of the Rhine in the mid fourth century (Libanius Ora. 18, 82–3). Ammianus Marcellinus records (18. 2, 3) that Julian “constructed granaries in place of those burnt, in which could be stored the supply of grain usually brought over from Britain”. However, imports from the Continent also occurred, for example, charred grain recovered from Rocester and South Shields has been identified as a bread-wheat, Triticum aestivum s., probably imported from north Gaul.

“We must recognise that our land-bound experience restricts our thinking. Railways and motor transport have virtually killed river, canal and coast-wise traffic for people and goods alike, so that substantial areas of water, like the Bristol Channel or the Irish Sea, now divide people whom they once linked. The point is doubly important because we easily neglect the possibility of “combined operations”, and because north and south Wales each had closer links with neighbouring parts of England than they had with the other’. Regrettably, Michael Jarrett’s death prevented his possible expansion of this topic, in itself an echo of the concept that “The land divides, the sea unites”. It is hoped that these efforts, in some cases highly speculative, may go some way towards developing this theme and may encourage others in its pursuit.

Strabo included slaves in his list of exports from the Province and this is demonstrated by a wooden writing tablet from London recording the purchase of a female slave for 600 denarii. The tablet is dated around 80–120 and, as the price paid for the slave was twice the annual salary of a comparatively well-paid legionary soldier, the export of British slaves could have been a highly profitable enterprise. Whilst not usually considered as imports or exports, for the whole of the Romano-British period, a major requirement was the transport of military and state personnel, both across the Channel, and around the coasts of Britain. In essence, there appear to have been at least fourteen occasions of significant troop transfer to the continent (often in the service of British usurpers), and at least fifteen major reinforcements (sometimes of returning vexillations) of the garrison of Britannia. Whilst this only amounts to some thirty major troop movements over a period of 350 years, it would 7

JAMES ELLIS JONES

  isostatic change. In the British Isles, the most dramatic effect was the submergence of Dogger Land in the North Sea, and the separation of England from France by the creation of the English Channel. However, the absolute height of the sea surface is not the only factor involved in coastal change, as the land surface is also changing in different directions, and in different parts of the country. It is therefore important to consider the relative influences of eustatic and isostatic change and their effects on relative sea level. If sea level is rising but, because of isostatic processes, an area of land is also rising, the relationship between sea and land will remain the same, as they both rise together. If an area of land is sinking but sea level remains constant, then the relative sea level will rise. As the result of isostatic rebound following the thermal expansion of the oceans and melting of the northern ice sheets at the end of the last Ice Age, the northern coast of Britain has been rising, as demonstrated by the “raised beaches” preserving the earlier coastline of Scotland. Consequently the south coast has been sinking, as shown by the “drowned” river valleys (rias) of Devon, Dorset and Cornwall. The process continues, with present day uplift rates of more than 1 mm/year in northwest Britain compared with subsidence rates of more than 1.5 mm/year in southeast Britain. Shorter to mid-term local and regional changes in sea level are caused by oceanographic factors, such as changes in ocean temperature and shifts in ocean currents. Meteorological factors include changes in air pressure (an increase of 1mb in atmospheric pressure depresses sea level by 0.99 cm and vice versa) and hydrological factors such as changes in tidal regime and river discharge.

Chapter 1 Changes in the coastal and riverine landscapes Introduction

Figure 1.1 The remains of the signal station at Scarborough It is of prime importance that we attempt to place any interpretation of Roman water transport in the context of what (probably) existed at that time, rather than in the landscape we see today. As the result of the processes of erosion and deposition, significant changes to the coasts and rivers of Britain have taken place during the past 2000 years but regrettably, “no survey has ever been undertaken to locate the coastline of England in the Roman period” (Allen, Bradley, Fulford et al.1997, 124). The dramatic effect of these changes are observable today (above) and, for example, at the fortlet at Martinhoe on the north Devon coast, where part of the outer rampart has slipped into the sea, and at Loughor in south Wales where part of the fort itself has been eroded. The opposite effect of deposition, has led to the silting of the estuary of the River Dee, leaving the once thriving Roman port of Deva (modern Chester), reachable only today by small craft. It is not intended to attempt a detailed study of the Romano-British coastline and rivers, but to select examples of changes resulting from natural processes, that may be historically, epigraphically, archaeologically or geomorphologically demonstrated. Changes may also occur because of human intervention, for example, deliberate activities such as the construction of riverine structures, or unintentional changes, such as the siltation of rivers caused by mining activities. These matters, in particular the early-medieval construction, often by the ecclesiastical authorities, of watermills, weirs and dams, are addressed in Chapter 10. Sea levels and coastal processes At the end of the last glaciation, some 10,000 years ago, the ice sheets melted and the volume of water in the oceans increased, resulting in a rise in sea level, termed

Figure 1.2 Coastal erosion at Reculver

8

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  In 1204 when King John visited the town, the port dues collected were about two thirds of those collected at London, at this time the town was granted its first royal charter. The quantity of shipping based at the port is illustrated by the fact that when Edward III sent an expedition to France in 1347, nineteen ships-of-the-line came from King’s Lynn at a time when London’s share was only twenty-four (Summers 1973, 35–9).

Even the most superficial study of Roman coastal and riverside sites will indicate that significant changes have taken place since they were built. The effects of coastal erosion may be seen at, for example, the Saxon Shore fort at Reculver (above) where, as late as 1600, the cliff top lay 600 m to the north, but now approximately one-third of the fort has been destroyed and lies underwater at high tide. The opposite process, that of deposition, is clearly demonstrated only 15 km distant, at Richborough (often considered as the beachhead for the Claudian invasion of 43), where the Wantsum Channel has silted up, leaving Richborough some 5 km distant from the modern coastline, and joined by land to the former Isle of Thanet. The subsequent deposition of a bank of shingle to the north of Richborough forced the River Stour to find a different course to the sea, via a substantial southern loop, which took it past the eastern edge of the fort at Richborough, eroding away part of the fort in the process.

Figure 1.4 Coastal change since the Roman period

Figure 1.3 Deposition at Richborough

Change offshore may present significant hazards to shipping. For example, the Goodwin Sands, the infamous “shippe swallower”, lie astride one of the shortest and busiest shipping routes between Britain and the Continent and are notorious for their constantly changing shape. The remains of a cargo of samian from central Gaul, dated to c. 170, have periodically been dredged up by fishermen from a wreck-site now known as the Pudding Pan (Smith 1909, 395–414); Hartley 1972, 27). More recent finds from the Goodwins and Thames estuary of mortaria, amphorae and potsherds indicate further Roman wreck sites in the approaches to Londinium (Dean 1984, 78-9)

The Fenland in East Anglia comprises the largest coastal lowland in Britain, with the majority of the ground level surface at or below 5 metres OD. As the result of drainage and reclamation, some areas lie below the zero altitude datum, and are therefore some 4-5 m below the altitude at which the Highest Astronomical Tide intersects the coast (Tooley 1990, 1–16). Modern systems of drainage and sea defence mean that the present coastal morphology and layout of Fenland represents a striking example of coastal change on a large scale, and it is clear that the extent of the Wash was far greater in antiquity. Find spots of Iron Age and Romano-British material indicate that the earlier coastline was, in some areas, as much as 25 km inland of the present coastline, following the marginally higher land from Chapel Hill, to the east of Sleaford in the direction of Bourne. However, isolated areas of slightly elevated land have produced plentiful Roman finds, demonstrating that these were tidal islands during the Romano-British period (Jones & Mattingly 1990, 11–12). The effects and extent of later siltation and consequent development of the present-day coastline are demonstrated by the growth of Lynn (now King’s Lynn).

River and stream processes Two patterns of river change may be distinguished, i.e. autogenic, which are those inherent in the river regime and involve channel migration, meanders, cut-offs, crevassing, etc., and allogenic, which occur in response to changes involving, for example, climatic fluctuation or altered sediment load or discharge, perhaps as a result of

9

JAMES ELLIS JONES

  Changes to the navigability of rivers, as the result of both channel change, siltation, deposition and variable tidal heights, may take place over a comparatively short time, and occurred throughout some 400 years of Roman occupation. For example, during the first century, the River Severn ran within 100 m of the defences of the fortress and colonia at Gloucester (Glevum), but the second to third-century riverfront was generally some 250 m west of the fortress (Hurst 1986, 1–3).

human activity. Man induced channel changes are of two basic types - direct and indirect. Direct changes are those brought about by some generally purposeful human action on the stream channel, generally related to engineering schemes intended to alleviate existing or impending threat of flooding, sedimentation or erosion. Indirect changes are brought about by the effects of human activity on the processes that control stream channel flow. These changes may be the result of many different causes; the various human environmental impacts on drainage basins include afforestation and deforestation, inter-basin water transfers and the effects of road construction and urbanisation. Others are nonintentional, but equally significant, as in the case of the vast quantities of crushed waste from china and tin ore processing discharged into Cornish rivers, resulting in a migration downstream of the tidal limits leading, for example, to the demise of Lostwithiel (the site of a Roman fort) as a port (Murphy 2009, 80). Rivers may be defined as straight, meandering, or braided, with some having a well-developed sinuosity, whereas others are only slightly winding. Although the course of a river may appear to be straight, the deepest points on the channel display a degree of sinuosity and, although the banks of a river may be straight, the flow of water between them is not. However, very few rivers have a straight-walled reach for any distance. A meandering river may flow within a winding channel defined by valley walls, or it may have a channel with very contorted bends sprawled across a flood plain. Braiding is the division of a single channel into two or more channels. Many rivers exhibit each of these three channel patterns (straight, meandering, or braided) somewhere along with their length. Siltation occurs when material carried in a body of water is deposited on its bed or margins. The quantity of material that may be carried or moved is controlled by the volume of water and its velocity. If either of these is reduced, assuming that the body of water is carrying silt at or near its capacity, deposition may occur. A reduction in the amount of water produced within the catchment area may result in deposition and a loss of velocity may occur when the down river flow of fresh water meets the tidal estuary, this loss of velocity affecting the incoming tidal flow as much as the downstream river flow.

Figure 1.5 Line of Roman and medieval waterfronts at Gloucester The location of structures such as quays, bridges, weirs and settlements indicates evidence of mobility in the flood plains of large rivers, such as the Thames, Trent and Severn, and suggests that there was limited channel change during the Roman and medieval periods. Clearly, there was some modification during the period, for example, alterations in the number and dominance of channels where there was a multi-channel system. There is evidence of some limited lateral channel shifting; for example, the Roman quay at the fortress of Caerleon is 150 m distant from the present course of the River Usk (Boon 1978, 24–36); at the villa at Park Farm near Lydney, the remains of the Roman quay are approximately 40 m to the east of the present channel. On Hadrian's Wall, the Willowford Bridge now lies some 72 m to the east of the present channel of the River Irthing and, as late as the visit by Camden in 1599, it is probable that most of the bridge would have still lain in the river and that the piers would have still been visible. After this time, the western movement of the river covered the piers with silt and, in the first edition of The Roman Wall, Collingwood Bruce (1851) records no trace of the bridge.

Extensive bars of pebble or sand now obstruct the estuaries of the Rivers Mawddach, Dyfi, Rheidol and Ystwyth, consequently causing changes to the river regimes, and affecting access to the known Roman forts at Brithdir, Pennal, Pen Llwyn, and Trawscoed. The site of the supply depot at Rhyn Park is on a flat plateau on the southern side of the deeply cut valley of the Afon Ceiriog (a tributary of the River Dee) with northern defences originally lying close to the lip of the valley. However, erosion of the valley edge has caused the entire northern corner, and about half of the length of the rampart and ditch to disappear (Arnold & Davies 2000, 8).

The River Severn appears to be significantly less laterally stable as the piedmont zone, close to the Welsh mountains, is reached, and significant channel change at Welshpool has been demonstrated (Lewin 1987, 161–70); the River Vyrnwy, a tributary of the Severn, also displays post-Roman incision and active migration over the past 200 years (ibid. 1992, 103–110). The entire lower Severn has been plotted, using 1st edition Ordnance Survey maps of the 1830s and 1840s, and these have been compared to 10

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  of which lay within easy reach of the rivers that flowed eastward into the area of Romney Marsh.

modern maps, using fixed ground control points; the conclusion reached was that only limited little lateral channel change has occurred during the last 150 years (Brown 1997, 225–9). Sedimentological change

and

archaeological

evidence

for

Brown (ibid.) has drawn attention to evidence from several areas of Britain that show an increase in flooding and alluviation during the Romano-British period. At Fengate, near Peterborough, there is evidence of illuviation, with the lower part of the site being covered by Roman freshwater clays; from the sedimentary evidence, it has been concluded that there was freshwater flooding in the second and third centuries. Roman alluviation has been identified at Braughing in Hertfordshire and, moving further north; there is evidence of late and post-Roman flooding in York, Brough on Humber and Watercrook. Figure 1.7

The evolution of the area now known as the Romney Marshes (259 km² of marshland and shingle, projecting into the English Channel between the towns of Hastings and Folkestone), has been comprehensively described by Cunliffe (1980, 37–56).

Probable area of Romney Marsh c.300

On the east coast, along an old channel of the Great Ouse, spring tides are generally supposed to have penetrated for a distance of approximately some 60 km and tidal influence was felt as far inland as Cambridge and St Neots. Many rivers followed different courses to the present time or even to those of the medieval period. At Earith, instead of turning east the Ouse flowed in a northerly direction via Upwell and Wisbech to the Wash. This ancient channel, the western branch of the Ouse, became known later as the West Water. The River Cam formerly reached Ely by a more devious route than at present and below Littleport its course was again different, for instead of flowing to Lynn it joined the West Water at Upwell. The channel between the Littleport and the sea was earlier known as the Well Stream, but later as the Wisbech Ouse (Summers 1973, 11–12). In the Somerset Levels, there is considerable evidence for a widespread marine transgression during the Late to Post-Roman period, and it is probable that flood-water levels extended up to 10 km inland, burying both reclaimed and un-reclaimed Roman landscapes by as much as c. 0.7 m of alluvium. Recent palaeoenvironmental work has established that the marine transgression resulted in the establishment of brackish conditions, over what had been a freshwater landscape (Rippon 2000, 138). He identified the creation of an extensive Romano-British flood defence system protecting the Somerset Levels, but Brunning and FarrCox (2005, 7–15) have recently suggested that no such defences were really needed. They draw attention to the recent identification (using LIDAR - light direction and ranging mapping techniques) of the relict channels of the River Siger and point out that its drainage potential would have protected the Roman settlements at Lake House Farm, East Brent and Rooksbridge, and also provided a water communication route to the Severn

Figure 1.6 Romney Marsh area – present day The present-day coastline, the result of both natural processes and the effects of man, is shown on Figure 1.6; Cunliffe's interpretation of the area during the third century is shown in Figure 1.7. He points out that the average annual rate of advance on the coastline of Dungeness is between 2.7 and 3.6 m per year (Lewis &Belkin 1940, 268–9), indicating that the evolution of Romney Marsh is far from complete. The Hythe Inlet gave access to Portus Lemanis (Lympne), where a shoreline of Roman date was found some 1.8 m below the present level of the marsh, overlooking the best natural harbour in the area in a sheltered lagoon, protected to the east by sand dunes and to the southeast by a shingle bank. It is probable that cargo could be carried along tidal creeks and rivers channels by river barges, as far upstream as Bodiham on the River Rother, linking with the iron-producing sites on the Weald, many 11

JAMES ELLIS JONES

  that this single island existed until at least the end of the Roman period and became separate islands between the 11th and 15th centuries as the result of a 4 m rise in sea level.

Estuary. On the south coast of Wales, the construction of some 28 km of sea wall, probably in the second century, accompanied by systematic drainage enabled the reclamation of some 30 sq. km of land on the Wentlooge Levels. An inscribed stone from Goldcliff on the Caldicot Level records the completion of 33 pedes of work by the century of Statorius Maximus of legio II Augusta. This may have been part of the reclamation of land in order to enlarge the territorium of the legion, without encroaching on the lands of the civitas at Caerwent. However, Allen (1990, 13–28; 1991, 485–94) has shown that, since the Romano-British period, a rise of c.1 m in sea level has resulted in much of the land reclaimed with sea defences being lost within the mud flats of the tidal foreshore.

The archaeological remains of ports and harbours that exist on coasts, estuaries and rivers may be dated by the associated ceramic, numismatic or epigraphic evidence, and provide a useful tool in determining the highest sea levels that existed during the Romano-British period. As accurate records of present-day levels are available from the Admiralty Charts produced by the Hydrographic Office, it is possible to establish changes that have taken place since that period, and therefore understand with greater accuracy how the coastal and riverine landscape has changed over time. For example, evidence from London indicates that significant tidal level change took place during the Romano-British period, and that this had an immediate effect on both navigability and waterfront development.

In the north-west of England there is evidence (Tooley 1980, 84) for a marine transgression (Lytham IX) during the Romano British period, which led to marine sedimentation being recorded at a height of 5.4 m above OD, i.e. at least a metre above present day MWST (mean high water spring tide) levels. The 19th/early 20th century coastal development of The Fylde took place on recently drained areas of wetland (Howard-Davis & Buxton 2000, 1). Speed’s 1610 map of the area shows extensive lagoonal formations where Blackpool and Southport now stand. The course of the River Ribble has changed and it is probable that, during the Romano-British period, the fort at Kirkham was accessible to sea-going ships (Singleton 1980, 8). There has been considerable discussion on possible changes to the Cumbrian coast and it is generally agreed that the increasing sea levels of the Roman period were followed by falling sea levels and probable aggregation. The medieval period was one of increased erosion and since that time there has been renewed aggregation. However, Clare (2004, 39–51) convincingly argues that coastal changes were localised and that, whilst there has undoubtedly been erosion in some places and aggradation in others, the coastline in general has remained much as it was in Roman times. At South Shields, before infilling during 1816, the Mill Stream flowed through a channel to the south of the fort and Collingwood Bruce (1884, 293) records that “the memory is still preserved of occasions when the tide had risen so high as to insulate the promontory”. In earlier times the channel was known as “The Gut” and a map from the time of Henry VIII shows that, at high tide this afforded an alternative passage from the estuary to the River Tyne; Dore and Gillam (1979, 1) believe that this was also the case during the Roman period.

Figure 1.8 Diagram showing Roman London waterfront development from the first to third centuries. (Latest at top)

Determination of tidal heights and sea levels by archaeological evidence

The earliest evidence for waterfront activity from London is the construction of a timber quay near to the area of the present City of London. The height of the quay indicates that the level the river was expected to reach at high water was c.1.2 m OD (Ordnance Datum - the Mean Sea Level as calculated by the Ordnance Survey at Newlyn in Cornwall). Following the Boudiccan sacking of Londinium in 60, reconstruction of the area was undertaken and a further timber quay of a similar height

Given that settlement sites were above the maximum reach of contemporary sea levels, it is possible to determine some of the most obvious indications of changes in sea level and coastal topography. From the presence of archaeological sites in the inter-tidal zone of the Scilly Isles, it is apparent that a rise in sea level resulted in the partial inundation of a single mass, resulting in the present group of islands. It is probable 12

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  therefore indicates a HAT at Caerleon of 5.6 m OD during the third century, and therefore suggests that tidal levels were some 1 m below those of today, and this is supported by the subsequent deposition of sediment on top of the hard standings of the quay (Waddelove & Waddelove 1990, 260). Disputing these conclusions, Rippon (2000, 28) has commented, inter alia, that “the HAT only occurs every 19 years: it is unlikely that a legionary was stationed by the River Usk for some 20 years in order to establish where the HAT was before construction started”. It is of course possible that this was intended as a semi-humorous comment, but it should be borne in mind that a legionary fortress had existed on the banks of the Usk for some 200 years prior to the construction of this quay and, during this period, Roman engineers would have become well aware of tidal regimes.

was built near to the present London Bridge. In an apparent response to a fall in the river level of the Thames, the tops of new quays constructed in the early second century were 0.5 m lower than the earlier structures. A progressive fall in river level in the second and third centuries required the continual extension of the waterfront in an effort to maintain a workable depth of water for shipping (Brigham 2001, 15–49). Excavations of waterfront construction from c.70 to 250, in the area of the present City of London, show that both the base and the top of each successive quay (Figure 1.8) were made at a lower level than its predecessor. This indicates that the level of the Thames fell by as much as 1.5 m between the end of the first and the middle of the third century (Milne 1995, 78–81). It is therefore probable that the tidal head had moved downstream and, at some point, sea-going ships would have been unable to use the London waterfronts as, even with the extended quays, insufficient water would have been available for them to dock and transfer of cargo to barges was therefore required (Figure 3.6).

The suggestion of a major marine transgression on the Somerset Levels in the third and fourth centuries AD was based mainly on the finding of Roman material at depths of up to 2 m at Highbridge, and the evidence of briquetage mounds found during the cutting of the Huntspill River in the 1940's. Hawkins (1973, 75–85) has argued that the present sea levels were attained by the end of the Neolithic period and that there was no major Roman marine transgression. He considered that the estuarine clay that seals the Roman material was the result of deposition to the maximum height of High Water Spring Tides in the equinoctial periods. He accounted for the varying depth of finds of Roman material as being the result of the consolidation of the underlying peat and/or the deposition of broken pottery and other debris in tidal pills. Leech (1977, 25–26) considered that the archaeological evidence strengthened Hawkins’s main points, but indicated that some of his detailed conclusions required modification. Estuarine clay levels revealed at Burnham-on-Sea and Badgworth led him to conclude that the accretion of the coastal clay belt was a continuing process occurring before, during and after the Roman period.

Excavations in 1963, at the legionary fortress of Caerleon, situated on the tidal River Usk, 14 km upstream of its confluence with the Severn Estuary at Newport, revealed a third-century stone-built Roman quay (Figure 1.9), lying some 230 m from the line of the present riverbank (Boon 1978, 24–36).

The problems of precise identification of tidal level change are well illustrated at Chester. During the construction of a gas-holder in 1890, “An ancient river wall was found, about 17 feet from the present line of the river, built of large blocks of sandstone, at an angle of about 45 degrees, starting from what was no doubt the original bed of the river. The wall had no foundation, and was built without mortar. A flight of steps led from the top of the wall to the bottom” (Hunter 1897, 81). Waddelove and Waddelove (1990, 253–66) calculate that the height of the wall was at 2.34 m OD and suggest that, when allowance has been made for a safety margin, the contemporary HAT appears to have been no higher than 2.03 m OD. Whilst commenting that this height “is considerably more than the figure of 1ft. 2ins actually recorded by the present authors”, they suggest that, from their calculated Roman HAT, there has been a rise in the Dee of at least 3.68 m. They go on to suggest, “A picture could be emerging of an early Flavian quay or wharf fronted with timber, adjoining a sloping stone revetment

Figure 1.9 Elevation of the third-century quay at Caerleon The height of the hard standings behind the quay were, on average, 6.56 m above OD, with the present ground surface lying at about 0.6 m above the quay. The maximum height to which sea level may be predicted to rise at any point is that of the Highest Astronomical Tide (HAT) and, at the present time, Newport Docks are at 5.81 m above OD, with Caerleon lying 10 cm higher, the HAT is 7.69 m OD at Newport and therefore 7.79 m OD at Caerleon. At any time, the top of a quay would need to be above HAT, to ensure that any goods stacked on it were not damaged, and to prevent vessels being washed onto it as a result of storm surges; it is probable that a height of 1 m above HAT would be needed. This 13

JAMES ELLIS JONES

  processes are interdependent. The site of an Iron Age/Roman emporium at Meols, still extant in the 19th century, has now been eroded and lost to the sea, with the resultant sediment being deposited in the Dee estuary. Similarly, in Yorkshire, to the north of the Humber estuary, the Holderness Coast is one of Europe's fastest eroding coastlines with an annual rate of erosion of around 2 m per year, and the port at New Hythe (from the Old English term h ð meaning landing place) disappeared completely sometime after1416. By contrast, the Spurn is an excellent example of the production of a spit as a result of longshore drift, with around 3% of the material eroded from the northern coastline being deposited here each year. Evidence from historical periods indicates how dramatic may be the effect of coastal change. For example, in Suffolk, the harbour at Dunwich suffered such inundation by the sea that, by 1600, the hitherto prosperous harbour had been rendered useless and by 1740, nothing remained of what had once been a rich city with 50 parish churches and religious foundations. Of Orford, Defoe (1991 Edn., 54) stated that it “was once a good town, but is decayed, and as it stands on the land side of the river, the sea throws up more land to it, and falls off itself from it, as if it was resolved to disown the place, and that it should be a sea port no longer”.

which was also the landing stage for a passenger ferry” and that sea-going vessels would have needed to off-load at Meols, for shipment by lighter to Deva. However, Ward (1996, 8) draws attention to the excavations during 1976-8 indicating that a site on the north-eastern bank of the river at Shipgate Street showed early Roman occupation at the level of 5.6 m OD, and that the latest Roman occupation level lay at 6.3 m OD. He points out that this is at variance with the interpretation of Waddelove and Waddelove, where the Roman water level is considered to have been much lower than the present one. Moreover, he comments that there is no evidence that the wall discussed in their paper was Roman in date (he considers that an early/post-medieval date might be preferable), nor that its surviving top, as recorded, represented its original total height. During the excavations of 1885, a grave was found, with a floor level at c. 3.5 m OD, that included the skeletons of an adult and a child, and amongst the grave goods was a coin of Domitian (81–96). Mason (2002, 68) points out that this must obviously have been dry land in the early Roman period and suggests that, together with the evidence from more recent excavations near the Old Dee Bridge, even the highest tide in the Roman period is unlikely to have reached a level over 4.5 m OD. Ordinary tides may have averaged around 4 m, receding to 1.5-2.0 m at the ebb. In any case, it was the general practice in Roman Britain, wherever possible, to site a permanent fortress within the navigable tidal limits of river, thereby enabling sea-going vessels to transfer their cargoes without the need for lighterage, and this alone may suggest that the Waddelove and Waddelove interpretation is open to question.

Figure 1.11 The port of Lilstock in the mid-19th century A more recent example occurs on the North Somerset coast, where Lilstock (Figure 1.11) was the landing place for the Saxon settlement of Stogursey. In1820 a pier and breakwater were built providing a small harbour, to which coal was imported from South Wales and from which pit props and lime were exported. Further protection was provided in 1860 by the construction of a stone pier from the north side of the harbour wall; warehouses, Coast Guard houses and an inn were built on the harbourside and a customs officer was in residence. An ambitious plan for the construction of a transpeninsula ship canal from Seaton in Devon to Lilstock

Figure 1.10 Changes in the River Dee at Chester Top left - Late Roman Bottom left - Dark Age Top right - Early Medieval Bottom right - Late medieval It has been shown that the effects of erosion and deposition are important, may make significant changes to coastal and riverine landscapes, and that these two 14

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  underlined, at the time of writing in 2010, by the discovery of a large building complex, lying between the amphitheatre and the River Usk.

was considered by the Board of Admiralty in 1888 but was, not surprisingly, rejected. Longshore drift led to the rapid creation of a shingle and cobble ridge (Figure 1.12), thereby denying access to the port, which was abandoned and the pier subsequently destroyed after the First World War.

It is therefore appropriate to repeat the opening sentence of this chapter when it was stated that it is of prime importance that we should attempt to place any interpretation of Roman water transport in the context of what (probably) existed at that time, rather than in the landscape we see today. Along an irregular coast (as are the Welsh or Scottish coastlines) the mean sea level surface is extremely sensitive to alterations in the form of the shore, the depth of the tidal channels and other shore features. These may result in local fluctuations in measured mean sea level ranging from fractions of a few millimetres to perhaps a metre (Everard 1980, 6). Therefore, it is clearly unhelpful to attempt a generalisation for either the 400 years of Roman Britain, or for its extensive coastline.

Figure 1.12 Photo taken from the same position, showing the shingle ridge and the few remains of the port at Lilstock -o-o-o-o-o-o-oIn 1997, English Heritage and the RCHME published a survey of “England's Coastal Heritage” with, as the front cover, the well-known aerial photograph of the remains of the Roman signal station on the cliff-top at Scarborough (Figure 1.1). It is a matter of considerable regret that the editors felt compelled to write that “Despite the ubiquity of artefacts of Roman date in the intertidal zone, there is little that can be written about the nature of coastal settlement, its interrelationship with the sea, and the exploitation of marine resources” (Fulford & Champion 1997, 217). Perhaps one of the more interesting suggestions on this topic, is that of John Haywood (1991, 27–30), who points out that, during the third and fourth centuries a major marine transgression resulted in the flooding of large areas occupied by the Germanic tribes. He suggests that this loss of land led, in part, to their raids across the North Sea to Britannia (see Chapter 7). In addition to changes in the geomorphology, rendering interpretation of the Roman period landscape a matter of some conjecture, a further consequence is the loss of archaeological evidence. This is of major significance in the case of most riverside structures, not usually of sufficiently massive construction (other than at major ports such as London or Dover) to withstand erosive effects, and therefore having long disappeared. On the other hand, the effects of channel change and deposition may bury such structures under layers of sediment, and render their discovery a matter of either chance or, as in the case of the quay at Caerleon, by determined archaeological investigation. This has been dramatically 15

JAMES ELLIS JONES

  cursor to the development of an integrated economy. Comparable in length to the combined total of the present-day motorway and trunk road systems, the average width of roads in Roman Britain enabled two wagons to pass in opposite directions, with room to spare; the quality of construction was not bettered until the days of Telford and McAdam in the 19th century. Hugh Davies (pers. comm.) concluded, “It would not have been possible for the Roman army to mount an invasion of Britain, and then occupy it for nearly four centuries without such a road system”. He investigated the Roman roads “from an engineering viewpoint” and used the Antonine Itinerary as a basis for assessing the comparative usage of those named therein. By calculating the number of routes following a particular stretch of road and drawing map lines with width proportional to that number, he produced a map of routes of the Antonine Itinerary (Figure 2.2) which dramatically demonstrates the relative importance of each section of road (2002, 9).

Chapter 2 Roads and waterways Introduction Hopkins (1983, xxi) considered “Above all we need to know the volume and value of trade in the classical world. We need to know what was traded, and the routes along which food, goods and metals flowed. How much did the volume, value, contents and direction of trade change over time and between regions during classical antiquity?” It is obvious that the absence of documentary evidence seriously hampers any assessment of the Romano-British economy, and a plethora of modern distribution maps is no substitute for the Roman equivalent of invoices or bills of lading. Transport routes by sea and river will be discussed and, in order to present a balanced view, Roman roads and road transport and the relative merits and demerits of road versus water transport will be considered. The use of the Western Seaways will be examined in some detail, and it will be suggested that the prevailing pre-occupation with crossChannel trade routes is not relevant to either the eastern or western coasts of Britain. On the roads

Figure 2.2 Relative usage of Antonine Itinerary roads. (Reproduced full-size at Appendix 4) As an added bonus, the road network provided a valuable geographical location system by noting the distance along a specific road and, for example, this method is used by Pliny (Ep. II. 17) to give directions for a visitor to his villa (see below).

Figure 2.1 Ermine Street approaching the Humber Estuary

Military roads

Prior to the arrival of the Romans, there were no engineered roads in Britain - only track-ways following natural routes across the landscape. Within a short time of the invasion the army constructed a comprehensive system of roads, with paved or gravel surfaces and bridges where necessary. This represented a far-reaching modification to the landscape and was a necessary pre-

Whilst it is probable that the army made use of preexisting Iron Age track-ways in the early stages of campaigning, these usually consisted of unmade surfaces that, whilst usable by pack animals, were totally unsuitable for heavy wheeled transport. The earliest roads in Roman Britain were constructed as an essential part of 16

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Roman roads in Britain the average width of metalling was 6.5 m (22 pedes), and, as Davies (op. cit. 73) has demonstrated, 15 pedes would be adequate to allow twoway wagon traffic, it is probable that the major Roman roads in Britain were adequate to deal with normal traffic flow. However, the broken mountainous terrain of much of Wales rendered aligned roads out of the question and that this led to the frequent construction of terraced roads, much narrower than normal, and only some 3 m wide (Margary 1967, 315).

the campaigns of the army, and have been described as following “routes of penetration” (Jarrett 1994, 19) or “line of march” (Hanson 1978, 61). Davies (op. cit.115) suggested that a better indication of a road’s purpose would be obtained by categories of “penetration roads”, “territory-holding roads” and “frontier-support roads”. A penetration road followed the campaigning army into the territory being invaded (e.g. Dere Street) and, once an area had been secured, a lateral road (e.g. the Stanegate) would be built behind the advancing frontline. After the construction of a formal frontier such as Hadrian's Wall, a frontier-support road was needed and Davies points out that, in contrast to the Stanegate, the Military Way follows the line and contours of the Wall itself, rather than attempting to follow a conventional “straight” line. Military roads were designed to enable supplies and reinforcements to be brought to the frontline, and to the rear supply bases (see Chapter 4). The Royal Engineers of the British army studied the resources needed by the Romans to build a road from the beachhead in Kent to the Thames at London (Davis 1984, 6–8; Peddie 1987, 188). The exercise was based on the theoretical clearance of a strip of land 8.6 m wide, with a 2.6 m running surface down the centre, flanked by wooden kerbs. It was calculated that, with this “light” construction, a team of 1000 men could have constructed the road from Richborough to Westminster, in some 15 weeks, ensuring that the road-head was no more than two days to the rear of the advancing frontline. In order to build a fullyengineered road over the same route it was calculated that a team of 3400 men would have needed three years to complete it. Davies (op. cit. 114–5) has pointed out that such a “light” construction road, without any drainage ditches or metalling, would not have been passable in wet conditions, even allowing for the use of some corduroy timber over marshy ground, and that this type of construction would easily break down, particularly under wheeled traffic, and quickly become unusable at any time of the year. He considers it implausible that such a road would have been suitable to cope with large-scale military movements for a period of years. He therefore suggests an “intermediate” type of road would have been constructed, with drainage ditches and light metalling, if possible by the end of the relevant campaigning season. This would enable any forward fort to be supplied during the winter, and provide an adequate road for the field army to advance quickly from its winter quarters, at the start of the next campaigning season.

In discussing the steep gradients of some Roman roads in Britannia, Collingwood Bruce (1863, 76) pointed out that the Military Way to the south of Hadrian’s Wall takes the easiest route that the Wall and vallum permitted, but challenges wagon drivers of his day to negotiate the steep inclines with laden vehicles. At first sight this might refer to the problems of ascending a steep gradient, however, the main problem faced by wagoners would be descending, rather than ascending steep gradients, because braking systems were primitive, or non-existent. Where a gradient exceeded 1 in 15, great care would be needed; and only on a surface composed of very loosely ground gravel could a slope steeper than 1 in 20 be descended. Gradients in excess of this exist on Roman roads, notably at Birdlip Hill on Ermine Street, between Cirencester and Gloucester, where the natural gradient is 1 in 2 but, by the use of zigzag diversion, the road has a gradient of 1 in 5, though there is a length of about 75 m with a slope in excess of 1 in 5 (Davies op. cit. 79). However, in the same way that additional horses, mules or hauliers could be used to ascend a steep gradient, similar forces, but in the opposite direction, could be used to descend the steepest of slopes. Local roads Although these may have originally been constructed by the military to serve a fort or supply depot, any new construction or repair and maintenance became the responsibility of the civilian authorities, usually working through the ordo (town council) of the civitas capital. For example, the military road originally built to serve the fortress at Caerleon declined in military importance by the early third century and became (within their respective geographical areas) the responsibility of the citizens of the civitates at Caerwent (Venta Silurum) and Carmarthen (Moridunum). Minor roads were usually built by landowners for private use, to either connect various elements of a villa estate, or to link into a local road system.

Fully-engineered roads The capability of a road to carry traffic is determined by the width of land that has been metalled, using stone or gravel to construct a firm and well-drained surface. An un-metalled road is thought by Pliny (Ep. II. 17) to be unsuitable for wheeled vehicles, as he comments that his villa at Laurentum could be approached on the main road from Ostia, but the final journey of six miles was over a side road, which he describes as “sandy for some distance and heavy and slow-going if you drive, but soft and easily covered on horseback”. For fully-engineered

Fords and bridges Fords were the first river crossings. To wade across a river either at low tide within its tidal limits or to brave the waters higher upstream was often a hazardous undertaking and at high tide or after rains the river could be impassible. Nevertheless, fords were simple and cheap and, when bridges were washed away, a ford would once again provide a means of crossing until a new bridge 17

JAMES ELLIS JONES

  The limits, in Roman pounds, imposed by the Codex on four types of draught vehicles were as follows;

could be built. As they provided a quick and economical method of crossing small watercourses, particularly in the early days of occupation, there were probably many fords on Roman roads. However, Davies (op. cit. 88) has been able to identify only 43 ford sites and even this small number includes references to excavations on roads that appear to be leading to a ford, with very few actual ford sites being known.

1500 1000 600 200

Bridges replaced fords at the busier and more vital crossing points, and ferries sometimes substituted for fords before a bridge was built. Bridges are known from 86 sites, but it is probable that this represents less than 5% of those needed within the 6000 miles of Roman roads. The bridges were usually constructed with a timber superstructure and, if more than one span was needed, supported on piers made of stone blocks or bricks. Stonearch bridges, some of which still survive intact in other parts of the Roman Empire, appear to have been rare in Roman Britain with the only known examples being at Piercebridge on Dere Street, and Willowford and Chesters on Hadrian's Wall (Dymond, 1961, 136–64; Davies op. cit. 90–98).

(490 kg) (328 kg) (200 kg) (65 kg)

Post-wagon (angaria) Post-carriage (raeda) Cart (vereda/carrus) 2-wheeler (birota)

Figure 2.4 Four-wheeled post-wagons (angaria) Of the vehicles specified in the Codex, an angaria was a large four-wheeled vehicle drawn by 6/8 oxen, a raeda was a large four-wheeled vehicle drawn by 8/10 mules or horses, a vereda/carrus was a smaller four-wheeled vehicle drawn by 4/6 mules or horses, and a birota was a two-wheeled vehicle drawn by 2/4 mules. The speed that could be regularly maintained by these vehicles is estimated at 3.2, 4.8, 7.0 and 8.0 kph respectively (Kendal 1996, 141–3).

Vehicles There are several references to military wagons in the Vindolanda tablets; in one the commander of the 9th Batavians is told that steps had been taken “that wagons may be given to you” (315), and was advised in another (316), that you should “make up in your mind, my lord, what quantities of wagons you are going to send to carry stone”. Another letter (309) includes a list of stores including “34 wheel-hubs; 38 cart-axles, one of them turned on a lathe; 300 spokes” (Birley 2002, 92–3).

These restrictions do not represent the full load-carrying capability of each vehicle and appear to have been imposed in order to minimise wear on road surfaces. In cases of military necessity, and subject to suitable road surfaces being available, these loads could be increased by up to 75%, resulting in figures of 850 kg for the postwagon, 600 kg for the post-carriage, 350 kg for the cart and 120 kg for the 2-wheeler. Langdon (1993, 1–11) considered that in terms of loads that could be hauled, there was little difference between the effectiveness of the transport in Roman Gaul and that which came later in the medieval period, and that a ton was a very respectable load for horse-hauled transport in the medieval period. As part of an assessment of the transport requirements for the construction of Hadrian’s Wall, Hill (2006, 85–94) arrived at loads of 2033 kg for the angaria (post-wagon), 1360 kg for the raeda (post-carriage) and 680 kg for the carrus (cart), but the loaded distances travelled on the Wall will have been very short, probably no more than one or two kilometres. Bearing all this in mind, and despite the strictures of the Codex, for long distance transport a load of 1200 kg for an angaria might be considered a reasonable assessment.

The edicts of the Codex Theodosianus date from the late fourth century, but it is unlikely that the range of freightcarrying vehicles had changed significantly from the earlier period. The edicts specified types of vehicles and imposed weight limits for their cargoes when travelling on the highways (8.5.30). An Imperial decree from 368 lays down the maximum loads, with the threat of severe penalties for exceeding the limits, for vehicles of the cursus clabularius, the system created by Septimius Severus charged with the task of carrying provisions for the army.

Restrictions on the size of loads are by no means confined to the Roman period. Charles II laid down that “no wagon, or waine, cart or carriage carrying goods for hire should be drawn up by more than seven horses or eight oxen or carry more than 20 cwt. between October 1st and May 1st, or more than 30 cwt. between May 1st and October 1st ”. In 1623, James I issued a proclamation

Figure 2.3 Oxen towing cart laden with barrels 18

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  accommodation for the couriers and, if necessary, a change of animals or vehicles.

stating that “no carter or other person whatsoever shall travel with any waine, cart or carriage with more than two wheels nor above the weight of twenty hundred; nor shall draw any waine, cart or other carriage with above five horses at once”. Transport animals The weight that may be carried by a packhorse is in the order of 100 kg; this compares with the ability of carthorses to draw a tonne. Obviously, a packhorse could travel on a track, whereas a wagon needed a road in reasonable condition. A mule has a greater load carrying capacity (c. 110 kg) than a horse, its hooves are harder, its skin is tougher, and it can stand extremes of temperature better than a horse During the Peninsular War of the 18th century, the weight of the load was fixed at 200 lb., and the length of journey, in mountainous country with bad roads, was from 10 to 12 miles loaded, 15 to 16 unloaded (Fleming 1902). There are skeletal remains of native ponies from Newstead and Corbridge and, with a load carrying capacity of 75 kg; it is probable that they were used as pack animals (Hill 2006, 87–8). The horse, the mule and the pony are capable of covering some 30 km per day when fully loaded. The ox is docile, surefooted and very powerful, and was the main draught animal used in Roman Britain. However, it is extremely slow and can only cover a distance of about 15 km per day when pulling a heavy load. Whilst this limited distance may occasion surprise, when Cato the Elder transported an oil mill from Suessa to his home in Venafrum, the journey took six days to cover a distance of 40 km, averaging only 6.6 km per day (Agri. 20.3.3). A workman is able to consistently carry loads of 50 kg (Trajan's column shows legionaries building bricks for the construction of a fort) and later building manuals suggest that “because of the effort required in loading and unloading, carts and animals should not be used for distances of less than 100 yards” (Rea 1902, 34)

Figure 2.5 Roman roads in Britannia On the seaways and coasts The early Western Seaways Prior to the Roman period, it is clear that there was significant trade along the Atlantic coasts and, particularly in the case of Cornish tin, there was a degree of maritime exchange between the Atlantic communities and the Mediterranean world. For example, during the Late Bronze Age (c.1200 to c. 600 BC), there is strong evidence for contact and exchange along the Atlantic coastal zones, archaeologically visible through widespread distribution of common metalwork forms. However, these networks came to an end at about 600 BC and, apart from some evidence of cross-channel trade between the southern Britain and Amorica in the first century BC, there is little evidence for the exchange of material goods between Atlantic communities throughout the Iron Age (Henderson 2007, 5–9). The similarity of cliff-castles in south-west England and Brittany also suggests extensive trading contacts between the southern Britain and Amorica during the late pre-Roman Iron Age. It has been suggested that short coastal journeys between communities must have been common and could have formed an interconnecting chain of contacts extending considerable distances along the Atlantic coasts of Europe (Waddell 1992, 29) and Cunliffe (1954, 354) envisaged a similar “series of interlocking systems of trade and exchange based on easy access to the sea”.

Reference has previously been made to the “seasonality” of transport by water; the following extract from the Vindolanda tablets suggests that this could also apply to transport by road. “I would have already been to collect them except that I do not care to injure the animals while the roads are bad” indicates that transportation by road, particularly in the winter, might also present problems (Tab. Vindol. II 343). The cursus publicus The messengers of the imperial postal service (cursus publicus) were usually drawn from the army and held a warrant entitling them to use the major roads and be provided with free food and accommodation. This was provided at well-equipped inns called mansiones or stationes placed at intervals between 40–56 km apart, as this was an average day’s fast travel by road. The system did not operate with relays, each carrier travelling the whole distance himself, usually in a light horse-drawn vehicle. The mansiones provided food and

19

JAMES ELLIS JONES

  both in coins and pottery, more widely represented. It is probable that the Veneti specialised in trading voyages along the Atlantic coast from the Loire to Ushant, across the western Channel route to Britain, and possibly to Ireland, and that the Coriosolites used a mid-Channel route from Alet and the Cotentin peninsula (McGrail op. cit. 41). Significant quantities of sherds of Dressel lA amphorae, together with distinctive Armorican pottery, have been found at Hengistbury Head on the Dorset coast. The headland protected a sheltered harbour, and the rivers Stour and Avon provided access to the prosperous areas of Wessex chalkland. In addition to surpluses of corn and wool from the Wessex area, the hinterland produced iron, salt and Kimmeridge shale. Evidence of lead, copper, silver and tin ore from southwest Britain has been found; it is clear that Hengistbury functioned at this time as a port-of-trade with crossChannel imports of wine, pottery, glass and tableware, together with exotic fruit such as figs. The existence of Hengistbury as a major port was comparatively shortlived, as Caesar’s conquest of Gaul led to a decline in volume of trade; the economic and political centre of Britain shifted to the east, where the Thames provided a convenient access to the established transport systems of the rivers of northern Gaul. By the early Roman period, Hengistbury Head had lost its economic importance and was, to all intents and purposes, abandoned as a port-oftrade.

Henderson (op. cit.) dismissed the long-held concept of the Phoenicians sailing to Cornwall for tin and then travelling on to supply all the ancient civilisations of the Mediterranean and the Near East (Penhallurick 1986, 123–31), describing it as “wholly erroneous”. However, the Ora Maritima, a fourth century poem by Avienus, described the routes, used by Tartessan and Carthaginian traders in search of tin and lead, along the western Atlantic coasts from southern Iberia to Britain (Oestrymnis). Mentioning insula Albionum (the island of the Albans) and gens Hiernorum (the race of the Irish), it is thought to include quotations from the Massilliot Periplus, an early sailing manual believed to date from the sixth century BC (Hawkes 1977, 19). McGrail (1990, 36) has extracted the following from Murphy’s translation (1977, lines 94–16); “The hardy and industrious peoples of the islands and coasts of the lands around Ushant or Ouessant were heavily involved in maritime trade, much of it in tin and lead. They used hide boats on these oceanic voyages. From Ushant/Ouessant it is two-days’ sail to Ireland and Albion is sighted on this voyage. Merchants from Tartessus, from Carthage, and from the vicinity of the Pillars of Hercules sailed to the Ushant/Ouessant region to trade”. The mere existence of the classical accounts is a fair indication of the trade may have been intensive (Cunliffe 1984, 247). Writing at the end of the first century BC, Strabo (HN. 4.5, 2) described four sailing routes between Gaul and Britain, “namely from the mouths of the rivers Rhine, Seine, and Loire and Garonne. Those who put to sea from the region around the Rhine do not, however, sail from the river estuary itself, but from the Morini, who are the neighbours of the Menapi and in whose territory lies Itium, used by the deified Caesar as a harbour when he crossed to the island”. He stated the distance of the crossing to Britain from the rivers of Gaul as 320 stades (57 km), “that from the mouth of the Seine it is less than a day’s run to Britain” (ibid. 1, 14) and that “People setting out, sail on an ebb tide in the evening, land about the eighth hour of the following day” (ibid. 3, 4). Ammianus Marcellinus (XXVII, 8) comments that Boulogne “is separated from the tract of land opposite by the narrows of the tidal sea, wont as it is still to rise in astonishing surges and then, without any harm to those plying it, to subside until it has all the appearance of a plain”.

The Roman conquest of Gaul not only had a major political influence on Britain but the new communications system of roads introduced during the reign of Augustus (Cunliffe 2001, Figure 9.21), had a dramatic effect on the volume of shipping along the Western Seaways. The network of major roads from Lyon (Lugdunum), located at the confluence of the Rhône and the Saône, made it possible to travel by road to the newly-established Channel port at Boulogne (Gesoriacum) where, in the early first century, a lighthouse was built to indicate the entrance to the estuary of the Liane. The overland routes through Gaul to the North Sea took the bulk of traffic, leaving the Atlantic sea lanes to shrink back to local cabotage, with Poole harbour taking over from Hengistbury Head as the most significant port-of-trade on the south coast of Britain. Trade routes in the Irish Sea Province

Pre-conquest cross-Channel trade Maritime archaeology has revealed evidence of preConquest Roman shipping in the west of Britain. Near the Porcupine Bank, 150 miles off the west coast of Ireland, a jar (olla) of first-century BC grey-ware, was dredged up by a trawler from 150 fathoms of water (Collingwood & Taylor, 1934. 220–1). An anchor-stock found off Porth Felen, a rocky cove on the tip of the Lleyn peninsula of north Wales, is of a Mediterranean pattern, obsolete by the time of the Roman conquest of Wales; from its decoration it may be assigned to the late second/early first century BC (Boon 1977, 10–30). An Arretine sherd of the Augustan period (c. 20 BC–AD 10) was recovered from

Before the Caesarean conquest, significant quantities of north Italian wine were transported to Bordeaux, along the Aude and Garonne rivers, for onward shipment to Britain. Caesar commented that “Of the whole seaboard in that region the Veneti exercise the most extensive sway, because they have very many vessels and in these they are accustomed to sail to Britain” (BGall. 2, 4). Despite Caesar's emphasis on the dominance of the Veneti, their presence in Britain is not well documented in the archaeological record, but Cunliffe (1984, 43–5) has pointed out that the Coriosolites from Brittany are,

20

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  the island of Steep Holm in the Bristol Channel (Boon 1987b, 375–6).

of La Coruña, and then turning to westward until the Roman lighthouse becomes visible.

Cunliffe's model for Iron Age trade between Britain and Gaul (1984, 4) envisages a system of redistribution through large maritime trading centres (emporia), in particular, Alet in Brittany, Hengistbury Head in Christchurch Harbour and Mount Batten in Plymouth Sound. He identified two major axes of trade, and has termed the first the Seine–Solent access and the second the Atlantic route, running up the western seaboard of Gaul, via the Channel Islands, to the south-west of Britain. Matthews (1996, 16) suggested that Cunliffe had not adequately considered the Irish Sea evidence, and points to further intermediate trading posts along the western coast of Britain. He suggests The Rumps on the north coast of Cornwall, one day’s sail from Mount Batten, St David’s Head a further days sail across the Bristol Channel, a possible site on Anglesey, with Meols, on the Wirral, a similar distance to the north. His estimates of the time taken to make each passage are, perhaps, somewhat optimistic, particularly the passage from Mount Batten to The Rumps which, in the writer’s experience, is unlikely to be made in much less than two days. However, alternative passage plans, taking account of wind and tide, and allowing for temporary anchorage at suitable sheltered havens, would undoubtedly result in a viable route through the Irish Sea.

The maritime trading site at Meols, on the northern tip of the Wirral peninsula, was situated on the coast of the Dee estuary, some 30 km from the Roman fortress at Chester. Discovered in the early 19th century, the site has now been eroded by the sea, but sufficient Iron Age material was recovered to indicate clear evidence of long-distance trade prior to the Roman occupation of Wales, together with artefacts of the Romano-British period (Griffiths, Philpott & Egan, 2007). Two silver coins of the Coriosolites, three Carthaginian drachmas, and one early Celtic gold coin of uncertain origin have been recovered. The coins of the Coriosolites date from the early first century BC, prior to any direct Roman interest in Britain or Gaul; the Carthaginians coins are similar to a type known to have been minted in the second/third centuries BC and, despite difficulties in identification, the gold coin probably dated to the last half of the first century BC. The absence from Meols of traded goods, such as pottery or metalwork, does not necessarily negate the concept of an emporium, as that the local population did not produce such goods, nor was acquisition and display of social value (Matthews op. cit. 20–21). Trade via the Continental river systems With an eastward shift in the direction of trade across the Channel, the longer and more dangerous west coast routes appear to have declined, as internal routes and markets within Gaul became secure (Fulford 1991, 35–6). However, with most of Britain firmly under Roman control by AD 85, the Atlantic seaways were rapidly opened up to Roman shipping. Richborough, protected by the Isle of Thanet, appears to have been the major transhipment centre for goods arriving via the River Rhine in the first century, with Dover and London developing in the late first and early second centuries (Philp 1981; Milne 1985; 1995). By the late third century, the London harbour was no longer being maintained; traffic along the Rhine axis seems to have diminished in the late third and fourth centuries, with a shift westwards along the Channel to reflect greater use of the River Seine route (Milne 1993, 11–15).

The most westerly route from the Britain to the Continent suggested by McGrail (op. cit. 46) is from Carnsore Point, Wexford, to the vicinity of Scilly, and then from the Lizard Point to the vicinity of Ushant. He considers that the likely maximum speed made good is unlikely to have been more than 5 knots and with the need “to take a departure from, say, Carnsore Point or a Lizard Point in daylight and to make a landfall at Scilly or Ushant also in daylight, there would have been a period of at least 10 hours out of sight of land even in midsummer”. With the greatest respect to a mariner of McGrail's undoubted experience, the writer would suggest that the landfalls suggested are fraught with some danger, and present the potential problem of sailing off “into the bright blue yonder”. For example, in anything less than good visibility, the Scillies (“the notorious Scilly Isles” {McGrail 1985, 16}) easy to miss (or to hit, as was the case of Admiral Sir Cloudesley Showell and his fleet!), and the island of Ushant, notorious for shipwrecks and a consequent a concentration of no less than six lighthouses, presents a target of only some 5 km, after a sea passage of some 120 km. Perhaps a more effective strategy is to “aim off” to the east in both cases, and therefore make a landfall on the northern coasts of Cornwall and Brittany. Having sighted land, a change of course to the west would permit Land’s End or the west coast of Brittany to be rounded and a departure point to be located with some certainty. Before the days of radar and satellite navigation, this strategy was often employed by small boat sailors when crossing the Bay of Biscay from north to south, by making a landfall well to the east

Whilst it is certainly true that, for much of the RomanoBritish period, the Continental routes were secure, it should not be overlooked that there is evidence that this was not always the case. For example, Herodian (I. 10. 3) recorded that, in the 180s peasants and deserters “overran the whole land of the Gauls and Spaniards, attacking the largest cities, burning parts of them and plundering the rest before returning”; there was a further widespread peasant revolt in 282–4. In 253, the Franks and Alamanni broke through the Rhine frontier and ravaged the country as far south as Arles; with the end of the Gallic Empire in 274, they launched a further massive attack on Gaul. By this time many of the cities were surrounded by defensive walls, creating what Cunliffe (2001, 431) described as a “siege mentality”. Frankish ships were, at this time active

21

JAMES ELLIS JONES

  Spanish olive oil and the Western Seaways

in the Channel and the North Sea and, in order to guard the river routes to the interior, the Roman response was to build extensive fortifications at Alet, at the mouth of the Rance and at Le Yaudet on the estuary of the Léguer.

Though the Continental river routes may have been effective for the transport of Mediterranean products, it is arguable this was not the case for Spanish olive oil from Baetica on the Atlantic coast (see Chapter 11). If the Continental river routes were employed, at least five transhipments would be required and the source to destination distance would also be increased by some 600 km. Funari’s (1996) study of Dressel 20 stamps found in Britain lends support to a suggestion of direct shipment via the Western Seaways. He structured four chronological divisions, i.e. pre-Flavian, FlavianHadrianic, Antonine and Third Century. Over the whole of Britain, amphorae from the pre-Flavian period accounted for 16% of the total; the Flavian-Hadrianic period for 36%; the Antonine period for 29% and the Third Century period for 19%; the importation of Spanish olive oil had virtually ended by 250. Funari also introduced three broad geographical categories covering the southeast (290 amphorae), Wales (51 amphorae) and Hadrian's Wall (90 amphorae) and was thus able to calculate patterns of consumption by period for each region. Not surprisingly, consumption in the southeast peaked during the Flavian-Hadrianic period, but both Wales and Hadrian's Wall showed the greatest level of consumption during the Antonine period, reflecting the period of the greatest density of military occupation. Dressel 20 amphora potteries were scattered throughout the three conuentus of Hispalis, Astigiti and Corduba. An analysis of their geographical share of the British market shows that the products of Corduba are most common in the southeast, being less frequent in the north and very rare in Wales. Hispalis dominates the amphora market in Wales, and Astigi accounts for the majority of the imports to Hadrian's Wall (ibid. 76–82). This might suggest some form of governmental control of the olive oil supply to the two military regions, but Funari (ibid. 86) suggests that this was not necessarily the case, and that there were probably three different trade routes to these areas. Perhaps a better case may be made for two routes, the major being the products of Corduba and Astigi via the Gaulish canals and across the English Channel to the south-east, where the majority of the amphorae remained; the supplies for the east of Hadrian's Wall being shipped up the east coast of Britain via the North Sea, probably to the supply depot at South Shields (Arbeia). This leads to the suggestion of direct shipment of the Hispalic amphorae from the Guadalquivir to the legionary fortresses at Caerleon and Chester, from where re-distribution to the forts and settlements area took place. This would have certainly provided a more effective alternative, and suggests the probability of greater use of the Western Seaways route than is usually put forward.

Figure 2.6 Barbarian incursions of the late third century AD It has been suggested that contributory factors in favour of the Continental river systems were the dangers of an exposed Atlantic coastline, delays caused by trying to round headlands with unsuitable winds and the limited trading market on the Atlantic coast compared with that on the inland rivers (Marsden 1994, 157). In the first century BC/first century AD, the majority of trade between the Mediterranean and the northern provinces was via the Rhone and Aude, and then to the Garonne, Loire, Seine or Rhine and, after loading on to local ships, then cross-Channel to Britain (Peacock 1978, 49–51; Parker 1980, 56; Cunliffe 1984, 3–23). Greene (1986, 40–1) suggests that the number of trading points along a particular route will have been an important cost consideration. Although the longest, the route from the Mediterranean via Gibraltar and the Atlantic coasts is the cheapest, however, via the Mosel and Rhine would have made contact with the greatest number of major cities and fortresses, and might have proved to be the most profitable. Peacock’s map of routes from the Mediterranean to Britain is accompanied by a graph demonstrating that the Atlantic route incurs a transport cost index of less than 80% of the Garonne/Atlantic route, and only 37% of the most expensive Rhone/Mosel/Rhine route. The graph illustrates the effects of the different routes and the means of transport costs, using the ratios established by Duncan-Jones (1982, 366-9).

Sailing the Western Seaways It is intended to firstly discus the viability of shipping via the Atlantic/Western Approaches/Irish Sea/St George’s 22

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Taking the shortest crossing from Gaul to Cornwall, at an average speed of 2.5 knots, a vessel would be out of sight of land for some 20 hours (McGrail 1983, 300). Passage making, at this or any other time prior to the production of accurate charts, called for a mental map of the voyage, identifying landmarks and the course and distance to sail between them (in conditions of good visibility, a coastline 30 m high can be seen from the deck of a ship at a distance of 11 nautical miles). Writing in the mid-16th century, William Bourne wrote that the first rule of a good navigator was “the chiefest that belongeth to a seafaring man, is to knowe the place and that he shall happen to fall with, whiche he must knowe by the beholding of the countrey, by taking some principall marke thereof, and the chiefest thing to do is, to beholde the hilles and dales of the land, that he may knowe them when he shall happen again”. In conditions of limited visibility “Steering by run of the sea”, (a practice still used by Shetland fishermen as late as the 19th century [Walton 1974, 10]), involved pointing the vessel in the direction of a well-known coastal landmark and steering a consistent angle across the prevailing swell pattern (Wooding 1996, 16–17).

Channel route and to then consider some opinions in favour of cross-Channel import from the Continental rivers, and subsequent transport by road to the point of demand. When considering Roman shipping in the Atlantic, later writers have perhaps been too heavily influenced by Caesar’s comments (BGall. III. 12) concerning the perils of the “vast and open sea”. Caesar was familiar with the confined waters of the Mediterranean and, whilst there are undoubtedly areas such as the Aegean that are noted for their strong winds and storms, wind directions are predictable and there are few appreciable tides. His loss of ships during his second invasion may also have been an influencing factor. This fear of the sea was echoed four centuries later when Ambrosius, Bishop of Mediolanium (337–97), commented “God did not make the sea to be sailed over, but for the sake of the beauty of the elements. The sea is tossed by storms; you ought, therefore, to fear it, not to use it. He who never puts to sea need fear no shipwreck” (De Elia 70–1). The Atlantic maritime approaches have recently been described as “notoriously difficult to navigate, even for modern mariners, as they are stormy, unpredictable, affected by strong, quickly shifting winds and very wide tidal ranges” (Henderson 2007, 48), but he rightly goes on to comment that “Despite proving difficult for the Romans, we can have little doubt that these conditions were seen as "normal" to Atlantic societies and that they had adapted their sailing techniques accordingly”.

It would be a mistake to think that sailors of the Roman period were without aids to navigation and seamanship. For example, the Periplus Maris Eythraei, a first-century manual for traders who sailed between Egypt and India, gave a detailed description of the west coast of the subcontinent. The sailing directions for entering a river called Lamnaios (41–5) are “This gulf which leads to Barygaza, since it is narrow, is hard for vessels coming from seaward to manage. For they arrive at either its right-hand side or its left-hand, and attempting it by the left-hand side is better than the other. For, on the righthand side, at the very mouth of the Gulf, there extends a rough and rock-strewn reef called Herone, near the village of Kammoni. Opposite in, on the left-hand side, is a promontory in front of Astrakapra called Papike; mooring here is difficult because of the current around it and because the bottom, being rough and rocky, cuts the anchor cables”. This may be favourably compared with the written instructions for a voyage from Norway to Greenland given in the 12th century Landnamabok; “From Hernor in Norway one is to keep sailing west for Hvarf in Greenland and then you will sail north of Shetland so that you can just see it in very clear weather, but south of the Faeroes so that the sea appears half-way up the mountain slopes, but on the south of Iceland so that you may have the birds and whales from it.” It was not until the 16th century that detailed instructions comparable with the Periplus Maris Eythraei were compiled for the coast of Britain when, in 1541, The New Rutter of the Seas for North Partes was published for circumnavigation of the British Isles, describing coasts and ports, anchorages, rocks, etc., following the format of present day sailing directions, giving tidal information and views of the coast as seen from seaward.

In 330 BC, the fear of the Atlantic did not deter Pytheas who, sailing from Massalia (Marseille), evaded the Carthaginian blockade, passed through the Straits of Gibraltar and headed north, perhaps in search of the source of tin obtained by the Carthaginians from an island in northwestern Europe. He wrote of the islands of “Albion” and “Ierne” and commented on several nautical peculiarities, notably the tides. After circumnavigating Ireland, he sailed north for six days and reached an island that he called “Thule” (probably Iceland). Cunliffe (2001, 91/2) suggests that, rather than passing through the Straits of Gibraltar in a Mediterranean vessel, Pytheas avoided the Carthaginian blockade by travelling over-land from Marseilles to the Gironde, joining a local ship, and then changing vessels as the voyage progressed. Reference to Homer’s Odyssey or Virgil’s Aeneid will readily confirm that mariners in the Mediterranean favoured coastal passages, although cross-Mediterranean voyages were made, for example the export of grain from Egypt to Rome. With an offshore wind, “coast hugging” is a dangerous practice, particularly in a vessel without the ability to sail effectively to windward. This was the case with ships of Romano Celtic ships of the period, and the sailors’ adage that “The trouble begins when you get near land” was even truer then, than it is today. However, sailing in sight on the coast does not necessarily mean hugging the coast, as the normal method of navigation is to sail from headland to headland, giving a suitably wide berth to each promontory and thus avoiding off-lying underwater hazards . 23

JAMES ELLIS JONES

  Sounding leads to measure depths, dating back to the second century BC, have been found around the Mediterranean coast of France. A lead weight, hollowed on the underside and with the recess filled with tallow, was heaved overboard to bring up sediment from the seabed. The nature of the bottom gave information regarding the ship’s position. The use of the lead-line by Roman navigators is described by St Paul in Chapter 27 of the Acts of the Apostles. Writing of his shipwreck on the coast of Malta he states, “but when the fourteenth night was come, as we were driven up and down in Adria, about midnight the seamen deemed that they drew near to some country; and sounded, and found to twenty fathoms: and when they had gone a little further, they sounded again and found fifteen fathoms”. Even in the recent past, the value of the lead has been emphasised because “Above all, the sailor’s attention is most earnestly directed to that simple but important instrument are saving lives at sea, viz. the Lead; the neglect of which may be said to have been the great cause of disasters to modern shipping” (King 1897, 166). Also used in Roman times was the “log”, a shaped piece of wood thrown overboard with a “stray” line attached, in order to measure the speed of a ship through the water.

Rounding Land’s End - “The Perils of the Sea” It is well known that sailors have a tendency to exaggerate the fury of the elements at sea, and to such an accusation, the writer would plead “guilty as charged”. He has been known to praise his wife's stoicism in a small boat when, on passage to Ireland, “we were caught in a full Force Eight”, dramatically reinforced by a description of “an electrical storm crashing down all around us”. Whilst it is certainly true that the female crew-member calmly retired below and continued with her knitting, it is equally true that the sea state was probably nothing more than a Force Six (a strong breeze), and that the occasional lightning strikes took place several miles away. The perils of Land’s End are frequently over-emphasised. McGrail (1985, 16) considers “the passage around Land’s End, even when the notorious Scilly Isles are left well to the westward, can be difficult and may indeed be impossible for several consecutive days until there is a favourable wind shift”. Holbrook (2001, 152) is equally concerned with the dangers of “Land’s End and the hazardous Atlantic coast of Cornwall”. Peddie (1987, 137–45; 155–8) dwells at length on the “hazardous weather conditions and dangerous tidal currents around Land’s End” and, in order to avoid “the dangers and dead haul of supply by ship round Land’s End”, repeats McGrail’s case for a trans-peninsula route. The 1897 edition of the Pilots Handbook for the English Channel is rather less pessimistic, pointing out that “Land's End is visible 26 miles off in clear weather”, that vessels may “use a channel half a mile wide, carrying 8 to 11 fathoms, between Longships and Land's End” and that “a vessel rounding Land’s End after leaving Mount’s Bay at halfebb can take advantage of a nine-hour tide up the Bristol Channel”. Commander King does however draw mariners’ attention to the fact that “the sea beats heavy in Northwest Gales, when its locality should be avoided, especially by small and heavily laden vessels” (King 1893, 22). Evidence from earlier periods also indicates that voyages round Land’s End do not appear to have been regarded as being the dangerous undertaking considered by some recent writers. For example, in 1683, the Port Books of London record 154 shipments from the west, namely 50 voyages from Liverpool, 35 from Bristol, 34 from Chester, 15 from Carmarthen, 8 from Blakeney, 4 from Chepstow, 4 from St Ives, 3 from Cardigan and 1 from Padstow (Willan 1967, 206).

The Roman lighthouse, still standing to a height of 33 m, at La Coruña (Flavium Brigantium) on the western tip of Spain, indicates that navigation in the area was of some significance, and direct passages northwards across the Bay of Biscay from that point to the vicinity of Ushant were certainly possible. Curchin (1991, 152) believed that the “traders braved the rough waters of the Bay of Biscay to bring their wares to Bordeaux and north to Brittany, where at least 20 sites with Baetican oil amphorae are recorded”. Trade from the north of Iberia is indicated by a Portuguese amphora found at Vindolanda, on Hadrian's Wall (Britannia 5, 467) and, in the sub-Roman period, there is evidence of direct trading contact between the eastern Mediterranean and the west of Britain, a distance of some 3,000 km from the source (Campbell 1996, 86). This demonstrates that, in the pursuance of profitable trade, ship owners were prepared to brave the so-called perils of the Atlantic coasts.

Similar views on the dangers of the Atlantic coasts are held, for example, Henderson (2007, 32) suggests that “the rough seas of the south of the Bay of Biscay and the strong north-westerly winds provide major problems to modern day sailors”, but on only one occasion has the writer experienced weather conditions in which it was it necessary to adopt the passive tactic of “heaving to”. By contrast, on one crossing, the absence of wind (from any direction) and a consequent flat calm, led to continuous use of the engine, and the need to refuel from a Spanish

Figure 2.7 City crest of Corunna showing late first century Roman lighthouse.

24

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  27th May until the 24th of September and that from then up to the 11th November “navigation is uncertain”. From that time until the 10th of March, “the seas are closed”, and the Theodosian Code (XIII.9.3) states that “from the month of November navigation shall be suspended; the month of April, since it is the nearest the summer, shall be used for the acceptance of cargo”. Military operations involving the fleet were undertaken outside the normal sailing season, as for example, the crossing from the continent to Britain by the Emperor Constans in the winter of 343, in order to deal with some, as yet unknown emergency (Salway 1993, 245).

fishing vessel in order to reach Santander. Colin Martin (2009, pers. comm.) has recently studied the Journal of Alexander Gillespie, a successful Scottish merchant skipper who traded throughout Europe from 1662 to 1685, each year making up to 5 voyages (extending from Norway to Bordeaux), without serious mishap and without once failing to reach his intended destination. “His technique was very simple. He stayed in port for however long it took for conditions to become exactly right for a swift and safe completion of his next leg”. The fact that “the River Exe is at all times difficult of access, unapproachable in a heavy sea, and must on no account be depended on for refuge” (King 1897, 65) did not deter the Romans from establishing a port at Topsham from which the legionary fortress, and subsequent civitas, were supplied by sea. In the same way, “Chichester harbour should not be attempted without a pilot as the entrance is difficult; the banks of it are constantly undergoing alterations” (ibid. 118), but it is probable that this area was chosen to land some part of the Claudian invasion force. The Goodwin Sands are “constantly in motion, and influenced by the prevailing winds” (ibid. 148), lie astride the approaches to the Thames Estuary and are notorious for the number of shipwrecks, as evidenced by the Roman shipwreck on the sands of Pudding Pan, but Londinium became a major port. It seems probable that Roman mariners also used the tactics of Martin’s Scottish skipper, and so avoided the “Perils of the sea” that seem to pre-occupy the minds of some modern writers!

Off the European coasts of the Atlantic, gales of force 7 and above are eight times more frequent during the winter months than in summer; rough seas may be expected one day in four compared with one in twelve in the summer months; cloud cover is greater in winter and the whole area may be overcast for several successive days. McGrail (op. cit. 171) suggested that the normal sailing season would therefore have been during the months from May to October, with winter coastal voyages being the exception. Nevertheless, in the 15th century, fleets from British ports bound for Lisbon crossed the Bay of Biscay in the summer, returning in December, bringing with them sweet wine and fruit, to be sold at a premium for the Christmas market at home. Hatcher (1993, 476–8) has argued that coastal trading in the North Sea was reduced “to a virtual cessation…..in December and January and a mere trickle in November and February”. However, it is probable that the more sheltered waters of the upper Bristol Channel and the Severn Estuary meant that comparatively short cross-Channel passages and estuary voyages could continue, albeit on a reduced scale, during the winter months.

This is not to underestimate some of the problems of the western coasts of Britain, where the Admiralty Pilot (1974b, 16) warns that sea rain can be expected on twenty days per month in winter and fifteen days in summer. The problem is not so much the rain as the restriction of visibility, making reliance on coastal marks unreliable. McGrail (2001, 171–2; 206–7) agrees but has also emphasised the competence of Celtic seafarers of the Roman period in exploiting the Western Seaways. Present day small boat sailors have a similar attitude, with voyages round Land’s End being commonplace by couples in their “seventies” (Malcolm and Anne Craven 2011, pers. comm.).

Against the Western Seaways route? The case for shipment to Britain via the continental rivers has a long and distinguished pedigree, but the case put forward is sometimes open to question. For example, the statement that “Even the abundant Dressel 20 oil amphorae were probably distributed via the Gaulish rivers system. The Claudian wreck at Port-Vendres (near the Franco-Spanish border near Perpignan) carried both Dressel 20 and Haltern 70 amphorae from Baetica, and the distribution of the amphorae stamps represented in the wreck demonstrates that trans-Gaulish routes were employed in preference to the Atlantic seaways at this period” (Colls 1977, 1–145). It might be argued that the only fact “demonstrated” is that olive oil was shipped from Iberia to the south of France!

It should perhaps be pointed out that the west coast of Britain is not alone in attracting predictions of seafaring doom and gloom. In the 17th century, the east coast of Britain was considered so dangerous that little coal was shipped from Newcastle to southern England during the winter months of December, January and February. Seamen are said to have declared that they would “rather run the hazard of an East India voyage, then be obliged to sail the winter between London and Newcastle” (Willan 1967, 25).

With considerably more justification, Paul Tyers (1996, 95) argued that the problems surrounding the identification of distribution networks are well illustrated by Peacock's discussion of the Pelichet 47/ Dressel 30 amphora (1978, 49-51). Despite any supposed cost advantage of the sea-route via the Straits of Gibraltar and the Atlantic to bring these amphorae from the south of France, the actual distribution suggests that the principal

Seasonality of shipping Vegetius (Mil. IV. 39) describes how, in the Mediterranean, sailing “is believed to be safe” from the

25

JAMES ELLIS JONES

  trade route was the Rhone Valley and the Rhine. He suggested that the British examples of this common second-century type were passed through marketing mechanisms centred on the Rhineland, the area of principal demand for the product. Tyer’s argument is clearly valid, as these amphorae were manufactured and filled in the lower Rhone Valley and therefore distribution via the Western Seaways was obviously unrealistic. Fulford (2007, 57, 69) considers there is good evidence for the continuation of a low volume, longdistance Atlantic traffic, but suggests that the continuation of these routes through the Roman period is “all but drowned out” by the volume of Imperially driven traffic between the first and fourth centuries. He suggests that whatever cost was passed on to the consumer, it did not include a transport cost and argues that, in this scenario, the choice of route is no longer important provided that there was an infrastructure to support the passing traffic.

was necessary to “save the water” during contrary tides, by anchoring firstly in Cardiff Roads, and then off Lynmouth itself. The passage of 80 km took 24 hours, a speed of three kph (about 1¾ knots), but the vessel had arrived in darkness and loading could not commence for another 12 hours. A 25 km passage from the River Yeo, north of Weston-Super-Mare to Kingsroad, off the mouth of the Bristol Avon, was completed on one favourable tide, i.e. less than 6 hours, giving a speed of c. 5 kph (just under 3 knots). However, the vessel arrived with the last of the flood, and needed to wait for the next incoming tide (some 12 hours) before passage could be made up the River Avon to Bristol. In the open seas, faster speeds were possible, but again, were totally dependent on wind and tide. For example, an outward passage to Ireland in the Garlandstone (a 25 m ketch), carrying 120 tons of coal from Portishead on the North Somerset coast, to Courtmacsherry to the west of the Old Head of Kinsale, a distance of some 225 nautical miles, (c.420 km) took “three full days”. This gives an approximate distance covered per hour of 5.8 km, or an average speed of just over three knots. Discharging the coal took a full working day, a return cargo was fortunately available at Bandon, six miles upriver from Kinsale, and the short passage there accounted for a further day. After cleaning the holds, the loading of 80 tons of oats took a further three days, as it arrived in individual cartloads from the surrounding countryside. After a late-evening downriver passage from Bandon, Kinsale was left in the early morning and, in gale force winds, the return passage to Portishead was made in 28 hours at an average speed of 15 kph (just over 8 knots).

The case for the irrelevance of transport costs may possibly apply to high value/low volume commodities such as pottery. However, there was clearly an understanding of the significance of transport cost in the Romano-British period. Writing in the fourth century, Libanius (Ora. 18, 82–3) states that “In the earlier times corn was shipped from the island (Britain), first over the sea and then up the Rhine, but since the barbarians had become a force to be reckoned with, they had blocked its transport and the cargo vessels have long been hauled ashore and had rotted away. A few still plied, but since they discharged their cargo in coastal ports, it was necessary to transport the grain by wagon instead of river, and this was a very expensive affair,” [added emphasis].

Headway within a river could still be made, albeit with considerable effort, even if wind or tide were not favourable, by the techniques of kedging or drudging. In kedging, an anchor, attached to the ships winch by a 6080 fathom line (c. 150 m), was towed ahead by the ship’s boat, and then dropped over the side. The vessel was then winched ahead, the anchor hauled up and the process repeated. Eglinton recalls, when as part of a three-man crew a vessel was kedged from Bideford to Appledore, a distance of about 5 km, against a strong headwind. Another method, with the tide and against a contrary wind, was for a vessel to be drudged (or dredged) stern first with the stream, using the anchor as a drogue, just a clear of the riverbed, and so maintain steerageway. The writer has employed the first method with limited success, but has been unable to use drudging, with any degree of result whatsoever.

Duration of voyages and passage times in British waters during later periods Whilst once again acknowledging the dangers of transfer of information from one period to another, we may obtain some idea of passage times and sailing conditions during the Romano-British period by considering some accounts by individual mariners of the voyages made in much later periods. Since that time, there have been some changes in vessel design, notably the use of the fore-and-aft rig, and these led to improvement in windward performance, but as will be seen, with sailing vessels it is wind, tide and loading/unloading cargo, not solely vessel capability, that are the dominant factors. Edmund Eglinton worked sailing trows and ketches in the Bristol Channel, Saint George’s Channel and the Irish Sea, during the early years of the 20th century and, in 1982, recorded his experiences in The Last of the Sailing Coasters. Passages within the confines of the Bristol Channel and the Severn Estuary, involved not only waiting for favourable conditions of wind and tide before setting sail, but also anchoring on passage during a foul tide, in order to avoid losing ground that had already been made. For example, during a passage from Newport on the River Usk, to Lynmouth on the north Devon coast, it

Sailing farther afield, from 1938-40, Tristan Jones served on the sailing vessel the Second Apprentice, similar to a Thames “spritsail” barge i.e. with a square shaped hull, a flat bottom (therefore able to take the ground), and equipped with lee-boards, but ketch-rigged with two masts, the type generally known as a “Dandy”. The ship was based at Rye and, during the time Jones was on board, the ship made voyages northwards as far as the River Tyne, across the North Sea and the inland

26

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  The degree to which a river was used varied greatly. For example, the River Severn served the fortress, and then the colonia at Gloucester; the fortress, later civitas capital, at Wroxeter; the forts at Caersws, Forden Gaer, Abertanant, and Pentrehyling; the small industrial towns of Worcester and Droitwich, and numerous villas and small settlements close to its banks. Above Wroxeter, the river is fordable at a number of points and the siting of forts on the riverbank upstream of this point is probably a reflection not of military strategy, but rather an indication of the use of the navigable potential of the river. By contrast, the Afon Wnion served only the fortlet at Brithdir. The Mersey, Ribble, Thames, Ouse, Humber, Tyne, Wear, Tay, Eden, Esk, Annan, Clyde, Wye, Usk, Avon, Parrett, Dee, Conwy and many others, were also rivers with significant inland traffic. Forts were located on or near the estuaries of rivers such as the Tyne, Humber, Tay, Seiont, Tywi, Dyfi, Llugwy, Ystwyth, Mersey, Ribble, Lune, Ellen, Nith, Urr and Clyde. Compared to the absence today of commercial navigation on many of these rivers, large or small, this level of use may be considered surprising. Of course, the fact that, during the Roman period, a river might have been naturally navigable, or might have been made navigable, does not prove that it was used as a waterway.

waterways of the Continent, to various ports along the South Coast and as far west as the Channel Islands. During this time, Jones comments that “The fact that I never saw Captain Tansey Lee use a chart, or that he never opened a book apart from the Bible, the accounts book, and the ship’s log, does nothing to detract from his intelligence and his experience, so hard-won over so many years of silent struggle and anxiety” (1982, 106-7). Jones describes a passage from Millengen, on the German/Dutch border to Remagen, a distance of some 240 km, with a favourable westerly breeze, but upstream against the current. When in the lee of the hills, the wind was lost and thirty-foot long sweeps were used to make some progress, under oars, against the current. The upriver passage took four days, 4500 empty bottles were then unloaded and a part-cargo of scrap iron and full bottles of Apollinaris mineral water were taken on board. Sailing by day and night, with the current in the vessel's favour, the German/Dutch frontier was reached in a day and a half, and from that point to the estuary of the River Scheldt took a further one and a half days. In the open waters of the North Sea, a favourable north-westerly wind aided passage to the Straits of Dover, where the ship was headed by a south-westerly wind and “meant a couple of tacks on either board for a whole day”. The entrance to Rye Harbour was finally made five days after leaving the dock at Remagen; the round trip, of about 900 miles, had taken Second Apprentice just two weeks (ibid. 142-5). Jones describes how, “During the winter time, we made quite a few passages, mainly up and down the English Channel”. One outward voyage involved the delivery of a cargo of house bricks from the Thames Estuary to St Helier on Jersey, in the Channel Islands. The return cargo was barrels of offal and cowhides, to take to Greenwich. Leaving in the early morning, with a favourable wind, by dawn of the next day the ship was off Beachy Head, and by dusk was anchored off Whitstable to await a favourable wind and tide. Weighing anchor at dawn, “we were soon moored off Greenwich a day before the hides were due to arrive”. On another occasion, the vessel encountered a westerly wind off Spurn Head, at the mouth of the River Humber, and “we rowed the boat day and night for a full twenty-four hours before we were alongside near Goole”. This was a distance of some sixty miles, against the wind, both with and against the tide (ibid. 119 – 225).

Basing their research on the river systems of the Thames, Wash, Humber and the Severn, Edwards and Hindle argued that the medieval rivers of England and Wales were navigable much further upstream than previously believed. They suggested that, since most towns were within reach of navigable waterways, water transport was much more significant than had been recognised (1991, 123–34). However, Jones has drawn attention to a decline in the extent of England’s navigable inland waterway network during the later Middle Ages. For example, the Foss Dyke was usable by Lincoln ships during the 12th century but by 1335, was “so obstructed that the passage of ships and boats is no longer possible” (Commission of 1335). The deterioration of the navigability of the Thames is illustrated by the fact that, at least until the 13th century, vessels laden with grain could pass downriver from Radcot (35 km upriver of Oxford), but by the 15th century the river was not suitable for navigation beyond Reading. Jones (2000, 60–82) suggests this may be the result of an increase in obstructions to river traffic, such as mill-weirs and fish-traps. The failure to maintain navigation may also be, in part, associated with a fall in late medieval urban demand for bulk produce like grain and timber.

On the rivers Aston (1985b, 11) has commented that “Roads, tracks, rivers and the small harbours and ports on rivers, estuaries and the coast, were vital for moving goods about in the past” and that “as well as a host of river jetties and trans-shipment points, almost any small cove, beach or inlet around the coast would have been used in earlier times for the loading and unloading of goods”. The Barland’s Farm boat and the multi-period small port at Magor Pill (Allen 1999, 45-60) confirm that this was the case during the Romano-British period.

Portage Portage refers to the practice of carrying watercraft or cargo over land to avoid river obstacles or between two bodies of water, a place where this carrying occurs is similarly called a portage. Strabo (Geog. 4.1) refers to portage from one navigable point to another in Gaul, when he comments that “The course of the rivers is so happily disposed in relation to each other that you may travel from one sea to the other, carrying the merchandise

27

JAMES ELLIS JONES

  use of this route (ibid. 195). If this were to be the case, it would represent a most exciting development, but it seems the slab was originally found built into a wall in Hailes House, Colinton, Midlothian, thus eliminating its value as evidence for a Roman portage point across the Great Glen. However, Phillips makes a convincing case for portages in Early Medieval Scotland and it must remain a possibility that they were also used during the Roman period.

only a short distance, and that easily across the plains, but for the most part by the rivers, ascending some and descending others.”. In the 8th to 10th centuries, the Vikings exploited the network of eastern European rivers with portages connecting the Volga, Western Dvina, Dneiper and Don, the four most important rivers of the region. In north America the extent and importance of portage is demonstrated by many site names such as Portage Bay, Washington; Portage County, Ohio; Portage-du-Fort, Quebec; Grand Portage, Minnesota and Portage la Prairie, Manitoba. There is no similar evidence from Roman Britain but, for example, in the Irish Sea, voyaging is characterized by “short-haul pottering” between beaches on rocky foreshores and islands, and there are numerous inshore lakes and narrow necks of land, “inviting portages” (Carver 2010, 254).

Seasonality of shipping Even on the inland waterways, occasional bad weather led to seasonal stoppages; in January and February 1695, severe frost and intermittent heavy snow impeded trade from clearing Gloucester and, in the winter of 1716–7, ice prevented the movement of Severn trows above Bewdley (Hussey 2000, 47). Pratt (1912, 158) commented, “It is certain that in the case of English navigable rivers of any type, magic might require to be done, and spent, in order to keep navigation open. With most of them it was a matter of carrying on an unceasing warfare with elemental conditions”. Flood waters in the River Severn sometimes rendered it impossible for large vessels to pass under the bridges, so that “a vessel may go up when the water is low, and a freshet may come, and a vessel may not be able to get back again for perhaps many days”(ibid. 155). By contrast, during drought conditions during the summer months, river levels could fall to a point where navigation was not possible. John Langdon drew attention to strong seasonal patterns in the use of the Thames during the medieval period and considers that it was only in winter, when water levels were likely to be high, was it considered a safe option to use the river. It was also then that the roads were most difficult (2007, 125).

The distribution of South-East Dorset Black Burnished Category 1 (BB1) pottery in Britain was studied by Allen and Fulford (1996, 223–81) who suggested, inter alia, a west coast distribution involving a trans-peninsula crossing from the Dorset coast to the Bristol Channel. One of the options they consider is shipment from Poole Harbour to Weymouth, then 50 km northwards by road via Dorchester (Durnovaria) to Ilchester (Lindinis), where Roman stone quays have been claimed (Cox 1950, 95) and Thew (1994, 134) has found geo-archaeological evidence for the canalisation of the River Yeo during the Roman period. They suggest that from Ilchester the pottery would be loaded onto boats for shipment, first down the Yeo and then the Parrett, to a trans-shipment port near the Bristol Channel. The total journey involves some 50 km by sea, a further 50 km by road, and then 30 km by river. The fort at Ambleside (Galava) is situated at the head of Lake Windermere and a mountain road (High Street) runs northeast to join the Manchester/Carlisle main route at Brougham. A westward road, via the notorious Wrynose Pass and the fort at Hardknott, joins the coast at Ravenglass. A depression in the ground marks the old course of the River Rothay, which in wet weather still contains standing water; there are also traces, now visible in summer when the water is low, of what may have been a Roman quay in the lake south of the fort. In the 19th century, sea-going luggers operated on Lake Windermere, transporting Cumberland roofing slate down the west coast, and in some cases as far as London. The River Leven is tidal to Hatherthwaite, and it is possible that Roman improvement of the 5 km stretch of the river, followed by a short portage near the point where it leaves Lake Windermere, enabled waterborne supply and replenishment of the fort. It has been suggested that portages between the lochs and rivers of the Great Glen may have enabled a maritime link between the Pictish and Scottish nations (Phillips 2006, 191–80). Citing Fort Augustus as the transition point from Loch Oinch to Loch Ness, and on the evidence of a Roman sculptured “three mother goddesses” slab now in St Benedict’s Abbey and a coin hoard found near Fort Augustus, she suggests that there may have been Roman

Post-Roman decline in navigability The Edwards and Hindle (1991, 123–34) investigation into the transportation system of medieval England and Wales, examined 700 references to navigation contained within the Calendars of State Rolls between 1290 and 1441, and extracted data on such matters as the granting of navigation rights and portage, enquiries over obstructions and disputes over tolls. In some 350 cases, there were there were references to specific heavy or bulky cargoes such as grain, wine, wool and stone. They contend that this information reflects both the availability of navigable water and the demands of trade, and that the sheer number of references for many rivers makes it possible to establish the head of navigation quite clearly, whilst a reasonable indication of the limit of navigation can be obtained for the rest (ibid. 126). Examples of the information extracted include that, in 1236, boats from Richmond on the River Swale were charged special tolls at York (Calendar of Patent Rolls 1377–81, 461–2), and that the Derwent was navigable beyond Derby, with barge-loads of lead from the mines at Wirksworth being sent to Nottingham in the 1320s (Calendar of Memoranda Rolls 1326–7, 43). Both these examples have

28

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  During the late-12th century, the Leges Edwardi Confessoris (Laws of Edward the Confessor) demanded that major rivers “along which ships transport provisions from different places to cities and burghs” and “minor waterways that carry ships along with things that are essential for burghs and cities”, should be maintained open for navigation. This led to petitions for the removal obstructions to water transport as, for example, when the city of Lincoln complained to the Crown when the Foss Dyke became blocked and, when the River Derwent was blocked by the erection of mills, the town of Derby petitioned for their removal (Jones op. cit. 61). A requirement to remove obstacles to navigation was often opposed by riparian landowners who, not unnaturally, desired to use it for their own interests in milling or fishing. However, regulations regarding these matters began to be enforced in the late 13th century, and numerous statutes relating specifically to the removal of weirs, jetties, mills, mill-dams, etc. causing obstructions to boats were passed (Pratt 1912, 121). The first Act for the improvement of an English river was a Statute of 1424 (2 Hen. VI.) appointing a commission “to survey, redress and amend all those faults of the River Lea”. In 1431, Parliament passed an Act for the protection of boatmen in the Severn Estuary against “many Welshmen and ill-disposed persons” who “used to assemble in manner of war and stop trows, boats and floats or drags on their way with merchandise to Bristol, Gloucester, Worcester and other places, hewing these craft in pieces, and beating the sailors with intent to force them to hire boats from the said Welshmen”. The Act declared the Severn a “free river for all the King's subjects to carry on within the stream of the River” (9 Hen. VI. c. 5).

particular relevance to the Roman period, with the reference to the Swale pointing to navigability as far as Catterick (Cataractonium, and that relating to the Derwent indicating the probability of water transport to Littlechester (Derventio). However, there is evidence that, as a result of an increase in the number and size of obstructions during the medieval period, many waterways suffered a reduction in navigability and that, despite repeated legislation, the Crown was unable or unwilling to prevent the accumulation of the obstruction of England's waterways that so hindered river traffic (Jones 2000, 69). For example, the Foss Dyke was open for commercial navigation during the 12th/13th centuries but became unusable during the 14th century and was not reopened until the 17th century. Similarly, navigation of the Derwent ceased after mills were built on the river c.1270 and was not restored until early in the 18th century. The ecclesiastical authorities were often at the root of the problem; the dam erected in the 13th century at Outwell on the Great Ouse, by Walter Langton, Bishop of Lichfield and Coventry, graphically demonstrates this problem when, as a result, traffic over a wide area was disrupted. Vessels which had previously travelled to Lynn along the Nene to Outwell and then via Well Creek and the Great Ouse, were now diverted at Upwell, down what was then known as the Wisbech Ouse. In addition to flooding two thousand acres of valuable arable land belonging to the Abbot of Crowland, this detour added over 50 km to the journey, “to the damage of the King and his liege people of those parts a thousand marks yearly” (Dugdale 1772, 305). The Prior of Barnwell was perhaps even more at fault when, in 1380, he reduced the breadth of the River Cam by at least 4 m, apparently to extend the area of his gardens, “to the hurt of the whole community of Cambridge passing there with ships and boats” (Flower 1915, 43).

Earlier reference was made to the Edicts of Diocletian and it is interesting to compare the Roman attitude to maintaining the navigability of the waterways with that the medieval and later periods. It becomes clear that the rivers of the Roman era had a far greater potential navigation than those with the medieval period, and the so-called “improvers” of navigation during the 17th to 19th centuries might therefore be more suitably renamed as “restorers”.

Legal constraints The imposition of legal constraints became inevitable, with earliest legislation applying to navigable rivers referring to the taking of salmon, or to restrictions on weirs and other hindrances to navigation. The construction of weirs caused conflict of interest by hindering river navigation, and Magna Carta called for all inland weirs on the Thames and Medway to be removed. Somewhat ironically, the agreement was signed in 1215 on Runnymede, a small island in the Thames, typical of those artificially created to enable the construction of fish weirs. By digging a second channel on the point of a slight inflection in the main channel, it was possible to create islands of this type. An alternative was to dig a bypass channel or “barge-gutter” around the weir and this practice was common on the River Severn, resulting in many small, thin islands known as bylets or eyots. A number of these have remained visible, because of the channel stability of the Severn, largely the result of its tall cohesive banks, although the barge-gutters have often silted up (Brown 1997, 259).

Land versus water transport In an address to the Royal Society in 1675, Sir Robert Southwell argued, “the principal use of the sea and rivers is for easier carriage of commodities”. Comparing the cost of land and water transport, he suggested that the ratio between the cost of carriage by sea and by wheeled carriage was 1:20, and between inland water carriage and wheeled carriage, it was 1:2. In addition, he divided land carriage into two forms, and claimed that carriage by horseback accounted for 60% of the total land carriage, even though it was near to one-third dearer than carriage by wagon (Birch 1675, iii, 203 – quoted in Willan 1976. 1). In his Wealth of Nations Adam Smith (1776, Book I: Ch. Xi., 20–1) made a comparison between the cost of land

29

JAMES ELLIS JONES

  transport and the cost of sea transport by writing as follows. “A broad wheeled wagon, attended by two men, and drawn by eight horses, in about six weeks’ time carries and brings back between London and Edinburgh near four ton weight of goods. In about the same time a ship navigated by six or eight men, and sailing between the ports of London and Leith, frequently carries and brings back two hundred tons weight of goods. Six or eight men, therefore, by the help of water carriage, can carry and bring back in the same time the same quantity of goods between London and Edinburgh as 15 broad wheeled wagons, attended by a hundred men, and drawn by four hundred horses”. He concluded, “Such are the advantages of water carriage it is natural that the first improvements of art and industry should be made where that conveniency opens the whole world for the market to the produce of every sort of labour”.

economically from Newcastle to London, (54,742 tons were shipped in 1591–2), but could not be carried for any significant distance by land, as the cost of land carriage doubled the pit-head price every 10 miles or so. At Bath, during the 18th century, Tyne coal coming 400 miles by sea and river, undercut Mendip coal that only travelled 12 miles by road. The cost of shipping a ton of salt from Droitwich to Worcester, a distance of 7 miles, was five shillings and the same price per ton was charged for shipment to Bristol, a distance of 77 miles (Willan 1964, 1–9; 38), suggesting that the major costs were loading and unloading, not distance travelled. On the other hand, weighty and bulky goods of high value could stand the cost of expensive land carriage even over long distances; the classic examples of this are cloth and wool (ibid. 1976, 1-9). Transport of wheat

Concerning the Roman period, all the major quarries provisioning the Saxon Shore forts lay within a day's sail and a vessel could complete a return journey, including loading and unloading, in three days (Pearson 2003, 94– 5). Using Marsden’s figures (1994, 89) for the Blackfriars 1 ship with a maximum cargo capacity of 50 tons, he calculated that, for Pevensey, this would require 1580 boatloads, requiring the continuous operation of a fleet of 18 vessels over a season of 280 days. Following Kendal (1996, 144) he further calculated that for movement by land in post-wagons (angaria), carrying a load of 0.850 tons at a speed of 3.2 km/h, some 250 post-wagons, drawn by 1500 to 2000 oxen, would have been needed to perform a similar task.

In the Edict of Diocletian (I.1 [wheat], XVII. 3 [transport]), the authorised charge per mile for a wagonload of 1200 lbs. was 20 denarii; 60 modii of wheat weigh 1200 lbs.; wheat is priced at 100 denarii per modius. A load of wheat would therefore double in price if carried on a journey of 300 miles [(60 x 100) = (20x300) = 600)]. As a further example, a modius of wheat could be carried from Alexandria to Rome, some 1,250 miles, for 16 denarii, a sum that would not pay for a land journey of 50 miles. It has been argued that grain, arriving in Rome from Egypt and Africa, would have undercut cereals grown in Italy itself, if they had to be transported by road (Yeo 1946, 241–2). In case of famine in an inland Lydian town, the local authorities had no resource other than to compel the local landowners to release their stocks; conversely, when the military units, which had hitherto consumed the local corn, were withdrawn from Lydia, the growers had no alternative but to let their crops rot on the threshing floor. Wheat seems, in fact, never to have been transported any distance by land, except by the Imperial government, for whom cost was not the prime consideration (Jones 1974, 37).

Whilst it might be suggested that some of these figures could its benefit from closer scrutiny, they do at least give an idea of the order of magnitude of the respective methods of transportation Cost comparison Duncan-Jones (1982, 366–9) used the pricing edict of Diocletian to suggest a ratio of 1: 4.9: 28–56 for the relative costs of sea, inland waterways and road transport. He considered several means of transport, and comparisons with the known costs of later periods indicate close agreement between the cost ratios between transport by sea, river and land from all periods. When operating beyond the Rhine, without the benefit of a developed road network, Kurnow (1980, 21–3) has argued that the ratios would rise to 1: 5.9: 62.5 and that goods were likely to have been taken by water, from the Rhine around Denmark to the coast of north Germany, rather than overland beyond the Rhine.

Transport of wine by road As a theoretical example, it may be helpful to consider the supply of wine to the legionary fortress at Chester if transported to Britain via the Continental river systems, the Channel and then by road. The consumption of wine by a legionary soldier has been estimated as 416 litres per annum and, based on a full strength legion in garrison, gives a total annual requirement of 2,080,000 litres. Dependent on the location of the production area, wine was shipped either in amphorae or in barrels, with Gaulish wine being carried in Gauloise 4 amphorae (capacity 37 litres with a container weight of 25 kg) and Rhenish wine shipped in barrels (capacity up to 950 litres and a container weight up to 200 kg). Assuming the maximum size of vehicle, and taking the load carrying capacity of an angaria as 1200 kg, either 19 amphorae with a total capacity of 703 litres or a barrel (or two

In the case of long distance transport, the relationship between the weight and bulk of goods and their value is critical. Transport of weighty and bulky goods of low value by land is extremely expensive, but the significantly lower cost of carriage by water makes the task economically viable. A classic example of this is coal that, in the 16th century, could be shipped 30

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  deliberately imposed feature”. Noting that, in some cases such as the approaches to Ilchester, Verulanium, Leicester and Silchester, road alignments are modified to conform with gates, street grids or boundaries, he suggests that these changes in direction are a deliberate feature, not an arbitrary correction to an overall design and concludes that existing features within the landscape, such as settlements were taken into account, as opposed to straight-line roads being the arbitrarily driven across the landscape (2002, 155). The aerial photograph (Fig. 2.1 above) of the approach of the Humber estuary clearly supports this contention, as the straight line of Ermine Street may be seen to bend to the east, in order to enter the site of the pre-existing ferry-terminal at Old Winteringham.

smaller barrels) containing 950 litres might be carried. This suggests that 2959 wagonloads of amphorae or 2190 wagonloads of wine in barrels were needed to meet the annual needs of a full-strength legion at Chester. The distance from London to Chester is c. 310 km and as oxen travel at 3.2–4 kph for 7–8 hours per day, but not more than 96 km in a week (Knight & Castle 1992, 28) the outward journey-time would be between 3 and 4 weeks; slightly less for the un-laden return journey. For example, on 2third November 1394 a carter from Northgate Street, Chester (close to the Portpool), agreed to transport a wagonload of goods to arrive in London by Christmas, and for this he was to be paid twenty shillings, half of it in advance. It was therefore clearly expected that a heavily laden cart would not reach London in the winter in less than a month. In the event, cargo did not arrive until the 1fourth January in the New Year (Laughton 1996, 70).

In comparison with carriage by water, land transport is expensive, but clearly, goods were extensively transported by land. Weighty and bulky goods of high value could stand the cost of expensive land carriage even over long distances, and classic examples of this are cloth and wool during the 16th century (Willan 1976, 1–9). The manufacture of cloth was the largest industry in the country, was widely dispersed, and dominated the export trade. One of the most impressive examples of land carriage is provided by the Kendal pack-horse men whose fell-ponies, carrying bales weighing c. 85 kg, took wool to Southampton and returned with such goods as raisins, figs, madder and woad, alum and canvas. The round trips took just over a month, and it is significant that they were undertaken in winter as well as in summer (ibid. 8). I am grateful to Sue Millard of The Fell Pony Society for confirming, from both reference to the work of Richardson (1990, 41–3) and her personal experience, that this journey time is achievable. Some cloth was sent by land when it could have gone by water. Norwich cloth could have gone down the Yare to Great Yarmouth for coastal transport to London; however, the more direct land route to the capital was taken. One of the advantages of land carriage is that it can be door-to-door. An alternative to land carriage of cloth from Halifax to London would have involved a transporting of the cloth by land to the River Ouse, by river to Hull, by coasting ship to London, and by lighter from the ship to the London wharves. In view of the difference in distance between the land and water routes, dressed cloth from Shrewsbury did not reach the London export market by going down the Severn and round the south coast. However, when Shrewsbury Corporation was buying grain for the relief of the poor, it brought wheat, barley and rye in Norfolk and arranged for it to be shipped by sea to Bristol, and then up the River Severn (Willan, 1976, 10-11).

In contrast to transport by sea (Mil. IV. 39) there appears to be no official “close season” for road transport, but the comment already quoted from Vindolanda (II 343) indicates, at the least, a reluctance to use road transport during the winter months. With travel being restricted to, say, the eight months from March to October, in each year a wagon would be able to undertake some five return journeys from London to Chester. Therefore, some 400 wagons would be required for wine shipped in barrels, or 500 if amphorae were used. Clearly, the problem would be much increased when, as was often the case, the majority of the legion was detached many miles to the north in military operations, or when engaged in the construction of the Walls. It is of course realized that the degree of precision implied by the above is not capable of absolute justification. For example, the wine “ration” has been based on the estimates of R.W Davies but others, notably Roth and Le Roux suggest lower figures. On the other hand, the fourth largest city in Roman Britain (Viroconium - present day Wroxeter) lay on the road from London, but no allowance has been made for civilian consumption along the route to Chester. However, as an order of guidance, it has some value when considering the demands placed on the road system if, as has been suggested, road transport from London was the principle method of dissemination. It is tempting to suggest that the volume of traffic generated by this method would, by comparison, lead to the modern-day traffic jams on the M25 paling into insignificance. -o-o-o-o-o-o-o-

The preponderance of forts and settlements on the coasts, estuaries and navigable rivers provides an excellent case for supply by water transport. Roman Britain lent itself well to this method of supply, because of its long coastline, the isolation of the forts and the lack of an effective road system during the early campaigns. For example, the case for the widespread use of water

The resolutely straight line pursued by many Roman roads, for example, the course of Ermine Street, shown in Figure 2.1 above, has often led to the concept of the imposition of a rigid geometric network on the preRoman landscape countryside. However, Hugh Davies considers that “straight lengths of road could be of consequence of the design method, rather than a 31

JAMES ELLIS JONES

  transport in Wales and the Marches is well supported by an analysis of Maps 4.33–38 in Jones and Mattingly (1990, 103–5). Of the 66 sites shown, 51 are on coasts and potentially navigable waterways; only 15 are unlikely to have been supplied directly by water transport. The export of Spanish olive oil from Baetica presents a strong case for the use of the Western Seaways. Vessels loaded with this commodity, carried in Dressel 20 amphorae, could have voyaged northwards via the Atlantic coasts, crossed the Western Approaches and made passage through the Irish Sea, discharging their cargoes at the required point of delivery, as far north as Hadrian's Wall. Alternatively, the cargoes could have been offloaded at the legionary bases at Chester or Caerleon, for trans-shipment by sea, estuary or river to the forts within the command structure of either legion. By contrast, shipment on the major continental river routes, during the Roman period, involved a long chain of transhipment centres, where “cargoes would be laboriously transferred from cart to river barge, or from river barge to sea-going ship, the process involving much double or treble handling of loads” (Milne 1990, 82). However, in order to counterbalance the admittedly biased views of the writer, the final words should perhaps belong to Hugh Davies who, after reading an earlier draft of this chapter, suggested, “You need good access to all parts of the land, not just those places close to the sea or navigable rivers. But the fact that all the Roman seaports, estuary harbours and riverside docks, had road access shows that land and water transport was seen as something integrated. For any particular journey, you used whatever form of transport was most efficient at the time you wanted to make it. Perhaps an ideal watercourse was nearby, but no boat was available, in which case you either waited or set off on the road. Alternatively you had a good road link available, wagons standing by, but there was an uprising of some kind along the route, in which case it was safer to take a water journey, even if that was long and tedious. The problem is that we don’t have any account of actual journey making in Roman Britain, so we can only speculate”.

32

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Chapter 3 Boats, craft, and ships of the period

Boats

Introduction

Coracles

This study is primarily concerned with the vessels themselves, the voyages made and cargoes carried, rather than the methods of vessel construction (varied and interesting though they may be). Therefore, in this chapter the descriptions of vessels, for which we have archaeological, epigraphic or written evidence, are divided into sections dealing with vessels that plied the “Estuaries and rivers” and those normally sailing on the “Seas and coasts”. This division results in what may, at first sight, appear as unlikely combinations, for example, the skin-covered native curragh appears alongside the Blackfriars ship of heavy planked construction. In a similar manner, the County Hall ship, of Mediterranean type construction, finds itself in company with the lightly built coracle. In some cases, such a broad classification is open to criticism; for example, a vessel similar to the Barland’s Farm boat could have travelled far up the Rivers Severn or Wye, conducted regular trips across the Bristol Channel, or made a successful fair-weather passage to Ireland.

Hornell (1946, 111–6) describes the coracle as “being the most suitable form of craft known to man, whether he be net fisherman or angler” and considers that it is a far older form than the curragh, being derived from the same common source of the ancient skin covered craft of Iraq, India and Tibet. The load-carrying capacity of such a small craft is emphasised by Waters (1947, 57–8), who recalls travelling as a passenger in a coracle, and comments that his twelve-and-a-half stone “added only a couple of inches to the draught of the vessel”. He also records the comments of a certain Thomas Powell who owned a shallow draught coracle with a cut-down gunwale, but still maintained that it once carried “a man, a boy, a hundredweight of coal and a gallon of water, to within an inch of the gunwale” (ibid. 94). The coracle had disappeared from the rivers of Wales and the Marches, but the tradition has been revived and coracles are again being made.

At one end of the spectrum are ships such as the St Peter Port ship, archaeologically demonstrable to have been engaged in trade along the French Atlantic coast and at the other, the numerous small craft on the coasts and rivers. These were employed not only for fishing and local transport, but also for transporting goods up-river from larger coastal and sea-going vessels. The wellexcavated and reported remains of some, such as the Hasholme logboat and the Brigg “raft”, survive in the archaeological record and are from an earlier period, but their descendants continued in use during the RomanoBritish period and are therefore described below. Others, such as the various forms of skin-covered craft, do not survive, but are recorded by early writers in sufficient detail to provide evidence of their form and function.

Figure 3.1 Coracle on the River Teifi Log boats and plank boats

The types of naval vessels used on the coasts and estuaries of Britain are discussed in Chapter 6 but, as no definitive archaeological evidence for a Roman naval ship has been found, written and epigraphic evidence are the main sources of information. The structure of the classis Britannica, and the probable methods of operation in the study area, is also considered in that chapter.

Boats of these types were in use both well before, and well after, the Roman period, and have been investigated and published in great detail, in particular by Sean McGrail (1978, 1981a, 1998). Therefore, this is not the place to attempt more than a brief description of some of the archaeological remains found in this country, in an attempt to create a picture of the types of craft in use on the estuaries and inland waterways of Roman Britain.

Throughout this chapter, extensive use is made of model reconstructions and illustrations from reliable sources, e.g. The National Maritime Museum at Greenwich and the Museum of London. Whilst it is realised that reconstructions, necessarily based on limited evidence, cannot be expected to display absolute accuracy, it is considered that their inclusion is justified, particularly because of the impression given of the size of vessels of the period.

Brigg logboat Probably used in the tidal creeks and rivers of the Humber estuary to ferry people and goods, a log boat dating from c. 850 BC was found in 1886 at Brigg near the River Ancholme, but the excavated remains were destroyed by fire during an air raid on Hull in 1941. Made from a log some 15 m in length and almost 2 m in diameter, 90% of the log had been hollowed out to produce a boat 15 m in length, 1.5 m in depth and about 1 m deep at the stern. The boat could have carried 2 men 33

JAMES ELLIS JONES

  remainder re-interred. Re-excavation in 1974 revealed a flat-bottomed boat, some 12 m in length with a maximum beam of c. 2.5 m and a depth of c 0.5 m, constructed from sewn oak planks with moss caulking. It might have carried loads varying from twenty-six sheep with four men to seventeen cattle with six men (McGrail ibid. 187– 8)

standing and 26 kneeling paddlers. With a cargo of 5.5 tonnes and crew of five, the draft would have been c. 0.6 m, with a freeboard of 0.3 m (McGrail 1981, 16–17). Hasholme logboat Similar in construction and design to the boat found at Brigg, the Hasholme logboat (Millett & McGrail, 1987, 69–155) was excavated from land at present below sea level, where there had once been a tidal creek of the Humber estuary. Dating from c. 300 BC, at 13 m the boat was slightly smaller than the Brigg long boat, but probably carried out a similar role in the area of the Humber. Iron Age or Roman boats have been recovered from Fiskerton near Lincoln, in the Witham Valley, the best preserved is a vessel 7 m long with a slotted transom board inserted into the hollowed tree trunk and is similar to the Hasholme boat. The Ferriby plank boats

Figure 3.3 Reconstruction model of the Brigg “raft” The remains of three sewn-plank boats were found in the inter-tidal zone of the River Humber at North Ferriby and were excavated in 1937 and in 1963 (Wright 1990). Probably used as a ferry in the Humber Estuary and its tributaries, Ferriby 1 (of which the most remained) was c. 15 m in length and c. 3 m at its widest part, with a depth of c. 0.7 m.

Craft of the estuaries and rivers Barland’s Farm Boat A late third-century Romano-British boat was found at Barland's Farm, near Magor, Gwent, south Wales; some 3 km from the present foreshore of the Severn Estuary, and 6 km to the east of the River Usk. When excavated, the vessel, constructed entirely of oak, lay with a list to port, with a maximum of 5 planks surviving on the port side and only 3 on the starboard side.

McGrail (2001, 187-8) has suggested that, with a maximum load of 5.5 tonnes, this boat was capable of being paddled or poled in calm waters with draft of 0.3 m and a freeboard of 0.26 m. Perhaps surprisingly, he has suggested that this boat “may be compared with the thirdcentury Barland’s Farm boat working in the Severn Estuary”. Reference to the description of this boat given above will make it quite clear that the only features in common are a flat bottom and a planked keel, and this perhaps reflects a classification by “construction” rather than by “function”.

Figure 3.4 Reconstruction of the Barland’s Farm boat as excavated The boat lay with its stern resting on the remains of a third-century stone and timber structure, near to a stream that, in the Roman period, had flowed south into the estuary. A wall constructed of rough slabs of sandstone, was revetted by two squared oak uprights, with two partially surviving horizontal oak timbers between them. It has been interpreted as a jetty/landing stage or timber bridge. The stern of the boat was incomplete, suggesting that it had been dumped in a derelict state, possibly to

Figure 3.2 Reconstruction model of Ferriby 1 The Brigg “raft” The Brigg “raft”, dating from c. 700 BC was found in 1888 and parts of the remains were lifted, but the 34

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  sailed in the Severn Estuary and Bristol Channel in winds of Force 4 on the Beaufort Scale (a moderate breeze of c. 15 knots, with wave heights of 1–2 m). When sailing on a broad reach with the wind on the beam, or possibly a little forward of it, the boat could have made 4–5 knots under sail. McGrail and Roberts (1999, 141–5) suggest that, in open waters, 3.5 m oars used by each of two oarsmen sitting on the crossbeams, and pivoted through a rope grommet, could propel the boat at a steady 3 knots in light winds and that, when laden, the speed would be reduced to about 1½ to 2 knots. Standing up to scull with an oar pivoted at the stern would be another option, but this would only be possible in calm waters. In very confined waters, such as the stream where the remains of the boat were found, a pole of 3–4 m in length would be suitable for propelling the boat.

stabilise the stream edge near the bridge/jetty (Nayling, Maynard & McGrail 1994, 599). A 1:10 scale model of the boat was made “as found”, but with displaced, fragmented and distorted timbers reinstated and was used as a basis for a theoretical reconstruction of the original vessel. As reconstructed, the boat had overall dimensions of 11.4 m by 3.1 m by 0.9 m, with the bow and stern being generally similar, but not of exactly the same shape. The boat displays many of the characteristic features of the Romano-Celtic boat building tradition, with sawn planks laid edge-to-edge and not fastened together, the vessel made watertight by caulking made of plant material mixed with tar or resin. There is massive and relatively closely spaced framing, the planks fastened to the frames by hooked iron nails, and the mast stepped one-third of the water-line length from the bow. Although smaller, the Barland’s Farm boat is therefore structurally similar to the mid second-century Blackfriars ship from the River Thames, and the late third-century St Peter Port ship from Guernsey. It also has features in common with the Rhineland barges and river craft of this period.

The sea-keeping capabilities of the un-decked Barland’s Farm boat would have enabled it to make passages throughout the length and breadth of the Bristol Channel; in particularly favourable conditions, a voyage to Ireland would have been possible. With her shallow draft, the vessel could also have operated in the local rivers capable of progressing well inland, transporting cargoes to and from many of the Roman forts and settlements. The excavators suggest that, lightly loaded, the boat could have carried 15 medium-sized barrels of wine (actual size is not specified); 90 sacks of grain (again not detailed), or 4.5 tonnes of salt/coal in sacks; or the same weight of iron, slate or stone laid on dunnage in the bottom of the boat. Livestock could have been carried, provided they were temporarily penned by hurdles or with their legs trussed, and possible loads could be up to 50 sheep or 8 cattle (Nayling, Maynard & McGrail 1999, 133–46). Only some 3 km from the site of the Barland’s Farm boat, an open boat of the type known as a “keel” has been excavated (Nayling 1998). Firmly dated by dendrochronology to AD 1240, the vessel is estimated to have a length of 13.2 m, a beam of 3.7 m and a depth of 1.23 m. The excavators suggest a cargo-carrying capacity of 3.75 tonnes and this ties in well with that of the Barland’s Farm boat.

The maststep was located approximately one-third of the waterline length from the bow and it has been suggested that a mast stepped in this position indicates that either a sprit or lugsail would give a better performance than a square sail. Although the majority of evidence for such sails is from a much later date, it has been suggested that vessels depicted on a second/third-century Rhineland mosaic, and on a gravestone, have leather lugsails fitted with battens (Ellmers 1978, 1–14, Figure 3) and the excavators of the Barland’s Farm boat consider a lugsail, fixed or dipping, as the most likely option. However, Palmer (2009, 94) considers that it is certainly not necessary to postulate a fore-and-aft rig to explain the mast position, going on to comment that “no rig would have been able to provide much windward ability for this little boat”.

County Hall Ship The wreck of a late third/early fourth-century vessel was found, during the building of London County Hall in 1910, on the south bank of the Thames, opposite Westminster, and the remains recovered and transported to the London Museum. Unfortunately, at that time techniques of preservation of wooden timbers were, to all intents and purposes, non-existent. Together with several attempts at restoration, this meant that there had been serious deterioration by the time it was decided to dismantle the remains of the exhibition, and place them in storage. Peter Marsden, who was given this task, found that very little of the hull had survived. He commented (1994, 109) that “most of what was on exhibition in the basement of Lancaster House comprised a variety of plasters mixed with wire, rags, small fragments of ancient planking, and even bone, all of which had been coated with tar or painted black on the inboard faces to simulate

Figure 3.5 Reconstruction model of the Barland’s Farm boat When loaded with 2.5–6.5 tonnes of cargo, giving a range of draughts from 0.3–0.5 m, the boat could have been 35

JAMES ELLIS JONES

  decaying, had been separated from the rest of the city by a riverside defensive wall (Brigham 1990a, 140–1) and that, at other locations, silted deposits had accumulated at the front of quays. As goods were still being imported into the city, some limited facilities for berthing and lading probably existed, but they have not, yet been located. It therefore seems probable that sea-going ships were moored in the river and unloaded their cargoes into shallow draft barges; Marsden (op. cit. 108) suggests that the County Hall ship might have been used in this way.

the ancient wood. Indeed, apart from the frame, very little else of the ship had survived in any meaningful form, and there was no guarantee that any of the timbers, apart from the hull, were in their correct positions”. However, a fullsize copy of the original large-scale plan of the vessel as excavated existed (although limited by today's standards), and the construction of a model for exhibition in the new Museum of London helped to clarify details of the vessel’s construction. The theoretical reconstruction of the vessel suggests a length of 19 m, a beam of 5 m and a height to the top of the gunwale of some 2 m. The ship was built with its planks fastened edge-to-edge by mortice-and-tenon joints without any caulking, a characteristic of shipbuilding in the Mediterranean region; the frames were fastened to the strakes by oak trenails. This is in marked contrast to the Romano-Celtic method then in general use in the Northern provinces, and typified by the St Peter Port and Blackfriars ships. Although there is no remaining evidence, the fact that the ship was of Mediterraneanstyle construction suggests that its propulsion was by a single square sail and that it was steered by two quarterrudders. In ancient wrecks, only the lowest parts of vessels survive, usually because they are buried under the cargo, but a deposit of alluvium over the collapsed side of the County Hall ship preserved some evidence of decking. Although the vessel is characteristic of shipbuilding in the Mediterranean area in classical times, a dendrochronological study shows a tree-ring pattern typical of southeast England, with a probable felling date of 287, and it is clear that it was built from trees growing in the region. The shipwright had probably been trained in the Mediterranean, as he was clearly experienced with the shipbuilding techniques of that area. Bronze coins found in the ship included a coin of Tetricus the Elder 270–3, two coins of Carausius 287–93 and a coin of Allectus 293–96.

It does seem somewhat surprising that a vessel of sophisticated Mediterranean-type construction should have been built, firstly at this time, and secondly for such a comparatively low-level task, and it is possible that this might have been a secondary use of the vessel. The most dramatic suggestion for the role of the vessel comes from Riley and Gomme (1912, 17–22) who consider that it was a warship and that it was sunk during the battles for the recovery of Britain in 296. In support of this, they point out that a rounded stone, weighing about 1.5 kg, was found partly embedded in a strake, as if it had been thrown from a considerable height. However, Marsden (op. cit. 128) draws attention to the fragile nature of buried water-logged wood, and suggests that an embedded stone need not have been a projectile, as the pressure of overlying sediments could well have pressed it into the waterlogged planking, long after the burial of the ship. New Guy's House boat This second-century boat, belonging to the RomanoCeltic shipbuilding tradition, was found in a silted creek, close to the south end of London Bridge, during the laying of foundations for extensions to Guy’s Hospital in 1958. Only one end of the boat was excavated and it has not been possible to determine whether this was the bow or the stern. It seems that the remains were part of a broad, barge-like vessel, constructed entirely of oak, with caulking of hazel shavings in resin used between the planks. As is usual in this tradition, the planks were attached to the frames by iron nails with flat heads and square shanks, the points being bent over their inboard faces. As only one end and part of one side were excavated, only a tentative reconstruction of the entire vessel has been possible but, by analogy with RomanoCeltic barges found on the Rhine, it is probable that both ends were pointed. It has been calculated that the vessel had a total length of at least 16 m, a beam of some 4.2 m and a height amidships of about 1 m. With such limited freeboard, it is clear that the vessel could not have been used in the open sea, and was probably a river barge designed for use in the shallows of the Thames and its tributaries. Later excavations in the creek in which the vessel was found abandoned revealed a low quay, and it is reasonable to assume that the barge was used there (Heathcote 1990, 193). It has been calculated that the barge could have carried a load of about 6 tonnes and, as there was considerable evidence of damage and repair, it is probable that the vessel had been abandoned due to old age. Pottery sherds found in the silt were dated to about

Figure 3.6 Unloading from sea-going vessel to shallow draught barge The decline during the fourth century in the importance of Londinium as a port, is evidenced by the fact that the second and third-century quays found there were 36

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  the hope that it would bring good fortune to the ship. The timbers cross-match with other London chronologies, indicating that the vessel was local in origin (Marsden 1994, 204). The ship was fully decked with possibly a cabin on the after deck; the main hold was located in the middle of the ship, lined with a ceiling of heavy oak planks that, at the time of the wreck, supported a cargo of building stone. Traces of Teredo and Limnoria borings were found in the planks, indicating that the ship had sailed at sea. It is probable that the last voyage of the vessel was from quarries near Maidstone on the River Medway, into the Thames estuary and then up the River Thames, using flood tides to reach London. The most likely cause of sinking was a collision with another vessel. It has been calculated that the vessel could have carried a cargo of about 50 tonnes, that the estimated 26 tonnes of ragstone found in the ship would have only partly filled the hold, and that there was capacity for a further 24 tonnes of cargo (ibid. 33–95). Clearly, highdensity cargoes such as ragstone place a considerable strain on the hull and explain why such heavy floor timbers and thick keel planks were used. It is even possible that, because of the huge quantities of stone required for building the walls of London, the vessel was specifically designed to carry materials for this task. The St Peter Port ship offers evidence for a deckhouse as living quarters for the crew, but no such evidence is available for the Blackfriars ship and the depiction of a deckhouse, as seen on a number of illustrations, is purely conjectural. As the vessel was clearly involved in making passages lasting several days, some form of living accommodation was probably provided, but this could have been in the after hold.

190–225, and dumps of rubbish in the creek just north of the boat date from a similar period (Marsden op. cit. 97– 104). Ships of the seas and coasts Blackfriars ship The wreck of the second-century Blackfriars ship lay in the bed of the River Thames, about 120 m from the shoreline, off the south-west corner of the city of Londinium and was excavated, in advance of building construction, in 1962–3. The wreck comprised the bottom, and parts of the collapsed sides, of a RomanoCeltic merchant vessel, about 18.5 m length and 6.5 m beam with a gunwale height amidships of about 2.8 m. The vessel was built entirely of oak and, instead of a keel, had two broad keel-planks, with the stem-post and sternpost remaining in situ. The frames were attached with 0.6 m long iron spikes to the edge-to-edge laid planks, which formed the shell of the boat. Each spike had been driven through the planks and frames, and then turned over on itself (known as clenching). Between the strakes was a caulking of hazel shavings in a pine resin.

Marsden (op. cit. 89) pointed out that different types of cargo take up varying volumes, and has calculated that possible cargoes for the Blackfriars ship could have been; 28 cubic metres of ragstone weighing 36.4 tonnes, 12 large barrels of wine, each weighing 1.2 tonnes (15.3 tonnes), or 18.3 tonnes of grain. Compared to a ragstone cargo, the lesser weight of cargo for wine is explicable by the voids surrounding the barrels, and that of wheat (whether loose or bagged) by the low density and high bulk of the product. If used in a military role (a modern analogy would be the ships of the “Royal Fleet Auxiliary”), a vessel similar in size and hold capacity, might carry 70 infantry or 15 cavalrymen and their horses (Grainge 2002, 45–51).

Figure 3.7 Cut-away reconstruction of the Blackfriars Ship

During the late first/early second century, ragstone was used extensively in public buildings such as the basilica and forum (Marsden 1987), public baths (Marsden 1976b), the amphitheatre (Frere 1978, 462) and the large fort (Grimes 1968, 15–40). However, the greatest single demand for stone was for the city wall, constructed in the early third century and requiring a massive amount of stone; Marsden (1994, 126–7) has estimated that 35,000 cubic metres was used in the construction of the wall. The carrying capacity of the Blackfriars ship has been estimated at about 20 cubic metres of stone weighing 26 tonnes, and it would have required 1,750 voyages, by ships of this size, to bring about 45,000 tonnes of stone

Dendrochronology indicates felling of the timbers between 130 and 175, but the absence of sapwood prevented a more precise dating for the construction of the ship. There was no evidence of any major repairs, little or no rot was observed, infestation by marine borers was minimal, and it is therefore clear that the vessel had been only in use for a very limited period before sinking. A bronze as of the Emperor Domitian, minted in Rome in 88–89, was found in a recess on the port side of the bottom of the mast socket, located about one-third of the length of the vessel from the bow. The reverse of the coin had a representation of Fortuna, the goddess of luck, holding a ship's rudder, and clearly had been deposited in 37

JAMES ELLIS JONES

  lowered required a mast pivoted in a deck-level tabernacle and a weighted heel” (2010 pers. comm.). The mast was stepped about one-third of the length of the vessel from its bow and, whilst Roman iconography makes it clear that the “square” sail was most commonly used on the larger ships throughout the Mediterranean region, McGrail (1981, 225–7, 234) has drawn attention to the steering problems of sailing to windward, using a square sail stepped so far forward. A spritsail or a lugsail is a possibility, although there is no archaeological or pictorial evidence to support either of these types in the Northern provinces at this time. Marsden (op. cit.) considers the square sail to be the most likely option, and submits that, by waiting for a following wind, the master would have minimised the steering problems, and points out that, in recent centuries, this practice was normal for sailing ships in northern European waters. It is probable that the ship was steered by two quarter- rudders.

from the source, some 112 km away The most likely location for the quarry supplying the stone is at Allington, 1.5 km north of Maidstone, where ragstone outcrops on the west bank of the Medway. Vessels were able to reach there on flood tides, and it may be significant that traces of Roman buildings have been found. It is possible that the quarry formed part of an Imperial estate and that the control of the production, labour force, and means of transportation were in the hands of the procurator. However, whether the vessel was part of the classis Britannica, and was acting as some form of fleet auxiliary (Milne 1996, 234–8), or whether it was in private ownership and under contract for the voyage cannot be known.

St Peter Port Ship A Gallo-Roman merchant vessel carrying a part-cargo of pine resin caught fire and sank in shallow water off St Peter Port on the east coast of Guernsey. During the fire, the pine resin melted, setting into a solid mass as the vessel sank, and holding the timbers together until the ship was discovered in 1982. Excavation took place from 1984 to 1986 and revealed that the ship is the largest, most intact sea-going vessel of the Roman period found outside the Mediterranean (Rule & Monaghan 1993). St Peter Port provides a natural anchorage, protected from the prevailing westerly winds and flanked by rocky reefs with the islands of Herm and Sark giving protection from easterly winds. The present harbour bears little resemblance to the shoreline during the Roman period; the natural reefs have been gradually absorbed into manmade breakwaters, as the harbour has been extended seawards. The ship was some 25 m in length, with a maximum beam of 6 m, and a height to the gunwale of at least 3 m and, although the bow structure had disintegrated, the excavators suggest that the ship was double ended (Rule & Monagham 1993, 13–70). The vessel was constructed entirely of oak, with caulked planks laid edge-to-edge, in a similar manner to the Blackfriars ship. A single sail was carried on a mast, stepped in a large floor-timber; situated approximately one-third of the ship’s length from the bow. The mast would have been at least 13 m tall and could have carried either a square sail or a fore-and-aft sail. There was evidence of deck planking at the stern, and an internal partition suggests that the ship had an after-hold and, although there is no proof for the existence of a forward hold, it is likely that one existed. Artefactual evidence of personal belongings suggests a crew of three, and that a galley area was positioned over the after hold, possibly in a deckhouse (ibid. 13–70). The discovery of a number of roofing tiles in the after part of the wreckage led to the suggestion that the putative deckhouse was roofed with tiles, but Gerald Grainge (pers. comm.) has “doubts about the appropriateness of tiles to roof a deck cabin on a seagoing vessel, if only

Figure 3.8 Unloading Kentish ragstone From a hypothetical reconstruction of the vessel a mast height of 12.7 m has been estimated and, since the deck of the Roman bridge at London is estimated to have been c. 5-6 m above the river level at low tide (Milne 1985, 44–5), the Blackfriars ship could not have passed beneath with its mast still standing. Either a drawbridge opening in the bridge permitted the passage of vessels with fixed masts, or the mast needed to be lowered. Marsden (1994, 70) considers that with a crew of three, the combined weight of the mast, yard, and sail, (some 3 tonnes), would have been beyond the capabilities of the crew whilst underway. He accepts the possibility of the mast being lowered downstream of the bridge and the vessel being towed upstream, but suggests the use of a drawbridge to be the most likely option. However, Grainge draws attention to “the impossibility of lowering masts which were stepped on the keel, as were those of craft like Blackfriars 1” and points out that “Masts designed to be 38

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Atlantic seaboard. Iconographic evidence is provided by a first-century Celtic coin from Canterbury and another found at Sheepen, Colchester, both depicting a high-sided ship with very bluff bows and stern, and possibly flat bottomed. The coins show a single mast with fore-and-aft stays and a square sail set on a single yard.

because of their being washed away in heavy weather and not offering resistance to the incursion of water from directions other than downward”. Eighty Roman coins were recovered, and fall into two chronological groups, the first comprising six regular aes issues of the second century, dating from the reign of Hadrian onwards, with the second group of seventy-four coins all being antoniniani of the late third century, indicating a probable date range for the sinking of the ship. (Rule & Monagham ibid. 71–74). The coins were of low value and representative of the range in everyday use at the time of shipwreck. It may reasonably be concluded that the money was not used for the purposes of trade and belonged to members of the ship's crew.

Quantities of plant remains, mainly of cereals, either preserved anaerobically or by charring, were recovered from the St Peter Port wreck, and it is probable that this material represents food stores for the crew of the vessel, probably stored within well-sealed amphorae. The after hold contained fragmentary remains of oak barrels or casks, all of which were charred to a greater or lesser extent. If these contained cargo, for example wine, a northern European origin is indicated, as products of the Mediterranean provinces tended to be transported in amphorae. Two Dressel 30 amphorae, thought to originate in Algeria in the third century, probably carried olive oil, although the excavators point out the possibility that they were being re-used for general liquid supplies, such as fresh water (ibid. 82). Tyers has drawn attention to the wide-ranging sources of supply in the small pottery assemblage recovered from the wreck (1996, 73–4). A large quantity of pitch blocks had been stored in the rear hold of the ship but during the fire, the pitch melted and spread across and between the floor chambers. The pitch was being transported in rectangular blocks, and the underside of each block was coated with sand to a depth of 3–4 cm. It is probable that during production, hot pitch was poured into square wooden frames placed in a sandy area, and that solidification of the pitch took place in situ, resulting in the incorporation of the sand into the base of the blocks. Study of the quartz grain surface-textures by scanning electron microscope has established that the sand is not native to the Channel Islands, where the sand is generally angular with high proportions of feldspars, mica, and some clay materials. The sand from the pitch blocks was dominated by well-rounded quartz grains, indicating an origin in a flat high-energy coastal environment and the most likely provenance for the sand, and therefore the pitch blocks, is on the coastal dune fields of Western France (Rule & Monaghan op. cit. 105– 7). The remains of marine organisms are characteristic of the Atlantic coasts of France, Spain, the English Channel west of Cherbourg and the south-west of Ireland. Interestingly, several of these species would not be found outside this area, for example, in the Channel east of the Isle of Wight, Wales, Northern Ireland or the Mediterranean (ibid. 113–121).

Figure 3.9 Reconstruction of the Gallo-Roman trading vessel wrecked at Guernsey No evidence for an anchor, or anchor stone, was found near the wreck and it is probable that, at the time of the fire, the ship was anchored (or beached), and drifted away from its original position. No ballast was recovered, and there was no indication of its use on previous voyages. It is probable that the weight of the bulky floor timbers lowered the centre of gravity sufficiently for the ship to have an acceptable degree of stability, particularly as it is likely that the vessel would have seldom sailed any distance without some kind of cargo in the hold. The flat bottom of the ship gave the advantage of being able to take the ground in a region of high tidal ranges and few developed harbours. This ability and the relatively shallow draft would have enabled it to enter the estuaries and rivers on which most contemporary ports were situated, thus reducing the need for cross-decking of cargoes. The underside of the keel showed a small amount of damage that may be attributed to abrasion when, in the absence of port facilities, the ship was deliberately beached to load or unload cargo (ibid. 15).

Whilst the evidence indicates that material of southern European origin had at some stage being taken aboard, it cannot itself indicate that the vessel or its cargo had originated in the Mediterranean and the excavators speculate that it operated primarily in the Channel, with perhaps occasional voyages as far south as Atlantic Spain or as far east as the North Sea, operating as a “caboteur”, i.e. a coaster moving cargoes of opportunity from point to point (ibid. 130).

In common with the Blackfriars ship, the vessel bears a striking similarity to the description of the Venetic ships given by Caesar in his description of the sea battle off Brittany in 56 BC (BGall. 3.13) and was clearly part of a shipbuilding tradition shared by many Celtic tribes of the

39

JAMES ELLIS JONES

  Hide-covered boats

In the body of the report, no reference is made to the cargo-carrying capacity of the ship, but similarity to the 25 m x 6.5 m x 2.5 m dimensions of the Dublin “keel” (McGrail 1993); suggest a cargo capacity of some 60 tonnes. An estimated capacity of the main hold is given as 130 cubic metres, and an additional capacity of the stern and bow holds as 30 cubic metres (Rule & Monaghan op. cit. 135). Unfortunately, no reconstruction drawings of the hull are provided, nor any indication of the method of arriving at these capacities, and so it is difficult to determine the accuracy of these figures. However, the beam of the St Peter Port ship (6 m) and the Blackfriars ship (6.1 m) are almost identical, and the height of deck above keel is again very similar at 2 m and 2.15 m respectively. The St Peter Port ship length of 25 m represents an increase of 35% over the 18.5 m length of the Blackfriars ship, and as both beam profiles are similar, a proportionate increase above the 28 cubic metres main hold capacity of Blackfriars (Marsden 1994, 195) to c. 37.8 cubic metres might be expected. Rule and Monahan’s estimate of 130 cubic metres seems therefore significantly overstated.

When faced with a difficult river crossing of the River Segre at Lérida in Spain, Caesar ordered the construction of “ships of the kind his experience in Britain in previous years had taught him to make. The keel and the first ribs were made of light timber; the rest of the hull was wattle and covered with hides.” (BCiv. 1.54). The similarity to the type of craft now called a curragh is clear and, as Caesar had only visited the southeast of England, it is obvious that the use of this type of vessel was much more extensive than it is today, when it is restricted to the west coast of Ireland. That this has been the case for some considerable time is indicated by a drawing of a curragh, made in 1685, by Captain Thomas Phillips of the Royal Navy, bearing a caption including the words “ordinarily used by the Wild Irish”.

Tugboats Whilst there is no archaeological evidence for their use in Roman Britain, it is probable that, in some situations, tugboats were employed in a similar fashion to the craft used for hauling vessels to their berths in the harbour at Portus. Known as lunuculi, they were true boats, not barges; carrying an effective spread of sail and able to be used for coastal as well as harbour and river work and were operated by the guild of the lenunculari tabularii auxiliari. Casson (1965, 31–39) suggests a modification of the barges known as codicariae, in order to provide an early form of tugboat as, even with the wind in a favourable direction, Roman merchant ships had only limited manoeuvrability when entering a harbour, being warped onto their berths or operating within the confined spaces of a river. He draws attention to a plaque from a tomb on the Isola Sacra, the burial ground for Portus, showing “well-muscled oarsmen… straining hard as they row a sturdy dory from the end of which a taut line runs outward and upward, a line that can only be a cable made fast to the bow of a lofty ship under tow” (ibid. 31–39). Instead of the customary pair of rudders mounted on the quarters, the boat is equipped with a single oversized oar mounted on the sternpost, presumably to give greater manoeuvrability when towing another vessel. A mast is stepped so far forward that it could only have carried some form of the fore-and-aft sail, a versatile type that allows maximum manoeuvrability. A further aid to manoeuvring is shown on a mosaic from Ostia that illustrates a capstan, which may have been used to heave in the towing line. In addition, by making one end of the towline fast to a fixed point upstream and then “winding in”, a vessel could be hauled up river by using the mechanical advantage gained by the use of a capstan, combines with the efforts of its own crew.

Figure 3.10 17th century drawing of a large Irish sailing curragh The sixth-century Massaliote periplus quoted by Avienus (Ora Maritima), is the earliest known reference to skin boats operating in the Atlantic, and this type of craft is also mentioned by Pliny (HN 4, 104) when describing the transportation of British tin on a six day voyage (probably from Cornwall) to the Isle of Wight (Vectis). Lucan (4. 136–8) described skin boats that “on expanded ocean the Briton sail” and Pliny (HN 7, 205–6) writes of boats with prominent keels, constructed by stretching hides over an osier or woven wickerwork skeleton. In his third-century Collectanea Rerum Memorabilium, Solinus describes voyages from Britain to Hibernia in “small boats formed from pliant twigs, covered with skins of oxen”. Written in the late seventh century, Adamnan’s Life of Saint Columba mentioned the construction of a curragh, describing it as a large craft with a sail set amidships and capable of carrying a crew of 17, together with sufficient provisions for 14 days (Marcus 1953, 315). The suitability of this type of vessel for ocean passages was shown by the crossing of the Atlantic in 1976 by Tim Severin in the “Brendan”. The construction 40

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  black curraghs, crammed with hirsute, half-naked warriors, bearing down on the British coastline, attracted by the prospect of abundant loot.

was based on information contained in the Navigatio Sancti Brendani Abbatis (Voyage of St Brendan) and a medieval image of St Brendan’s boat (Severin 1978, 25). Even though the vessel may not have been a completely authentic replica (for example, the addition of a fore-sail, leeboards and buoyancy tanks), its performance clearly demonstrated the sea-keeping capabilities of this type of vessel. McGrail (1990, 39) points out that hide boats are quickly built, are inexpensive compared to a similar-sized plank boat, and are readily repaired. They are excellent surfboats and can be operated from almost any shore without the need for formal landing places. He also suggests that they are more seaworthy and sea-kindly than equivalent plank boats, as their lightweight structure is buoyant and gives good freeboard, even when loaded.

Written, numismatic and epigraphic evidence The ships of the Veneti This chapter has, so far, been concerned with the archaeological remains of ships and boats found in the British Isles. There is, however, literary evidence from Julius Caesar and Strabo giving valuable information about the ships of the Veneti, a major tribe of sea-going merchants from the south coast of Brittany, whose sphere of influence stretched from Portugal to the shores of Britain, and possibly Ireland. Caesar describes them as the most powerful tribe on the coast, having the largest fleet of ships, in which they “trafficked” with Britain. After a year-long campaign, by both sea and land, Caesar eventually brought the Venetic fleet of 220 ships to battle off the Gulf of Morbihan on the southern coast of Brittany in 56 BC and, after a long and hard battle, only a very few escaped. It is interesting that Caesar comments that, before the battle, they “summoned reinforcements from Britain which faces that part of Gaul”, but it is not clear whether these were transported by Venetic or British ships (BGall. 3.9). Caesar states that “They [the Veneti] constructed their hulls with somewhat flatter bottoms than our craft, to make it easier to go through the shallow depths of low tide and over shoals; the prows and the sterns are rather high to handle the size of the waves when the sea is stormy; the ships are made throughout of oak to withstand any amount of violence and hard treatment. Beams are of timber a foot square made fast with iron nails an inch thick and anchors are held by iron chains instead of ropes. Their sails are of hide or softened leather instead of canvas, possibly because they have no flax or do not know how to use it, but more likely because they think canvas will not stand up to the storms of the ocean and the violence of the winds and will not drive such heavy ships efficiently” (ibid. 3.13). Strabo (V. 4.1) gives a similar description commenting that “because of the ebb tides, they make their ships with broad bottoms and high sterns, and high stems; they make them of oak (of which they have plentiful supply), and that is why they do not bring together the joints of the planks but leave gaps. These they caulk with seaweed, so the wood may not, for lack of moisture, dry out when the ships are hauled up, because the seaweed is naturally rather moist, whereas the oak is dry and without fat”. Regrettably, despite the number of ships involved, and with reasonable knowledge of the location of the battle, no remains of sunken vessels have been located.

It is clear that any evidence for a craft of this construction is unlikely to survive in the archaeological record. In 1896, a small gold model, dating to the first century BC, was found in a field at Broighter in County Derry, Ireland. The vessel carried a mast with a single yard and a steering oar positioned to the rear quarter, and may well represent an early version of the curragh. Equipped with 18 oars, equipment included 3 forked barge poles and a grappling pole or anchor. The proportions of the model, and the relative size of its fittings, suggest a vessel with a length of between 13 and 17 m. (Cunliffe 2001, 103–6).

Figure 3.11 Gold model of the first century BC Broighter boat

The ships of the Veneti were solidly built, rode high in the water, and had flat bottoms enabling them to move close inshore and take the ground in tidal waters. They were more seaworthy, but not as fast as Roman galleys. They were probably caulked with moss, had square leather sails on a single yard, on a mast stepped midships, and iron anchor chains to obtain catenary effect (a sag in the line, resulting in the lowering of the effective angle of

The peaceful missionary use of the curragh during the “Age of the Saints” has been mentioned above, but towards the end of the Imperial era, whole fleets were used in recurrent Irish raids against Roman Britain, in one of which, St Patrick is said to have been carried off (Marcus 1980, 8). In De excidio et conquestu Britanniae, written in the sixth century, Gildas tells of swarms of 41

JAMES ELLIS JONES

  major hazard to vessels plying between the Forth and Tay (Martin 1992, 22–3).

pull on the anchor) when anchoring. This combination is to be expected when sail-driven Channel ships are compared with Caesar’s oared warships that, even if built on the Loire, were to a Mediterranean specification. It will immediately be apparent that both the Blackfriars and the St Peter Port ships bear a marked resemblance to Caesar's description.

Evidence from coins Muckleroy et al. (1980, 439–44) drew attention to early first-century bronze coins of Cunobelin found at Canterbury. The obverses of these coins show the starboard side of a bulky, high-sided ship with a single mast and yard, with prominent stays and a steering oar. The stempost is straight, raking at an angle of 10 degrees from the vertical, the gunwale angles sharply upwards towards the bow and the sternpost is curved. The ship is clearly not a warship but is a merchantman, like the Venetic ships it so closely resembles. McGrail (2001, 196; 205–5) suggests that the depiction of ships on Cunobelin’s coinage probably reflects overseas trading voyages by his ships, from the Catuvelaunian tribal area situated to the north of the River Thames.

Shipwrecks A Romano-British vase, now in the National Museum of Wales, was dredged up by a trawler from a depth of 300 m on the Porcupine Bank, 240 km from the west coast of Ireland. On the bottom was a graffito, C PISCI FAGI (the property of Gaius Piscius Fagus?) and a crudely scratched drawing of an animal, possibly a bear (Nash-Williams, Journal of Roman Studies 24, 221). A group of 20 lead ingots found in the River Mersey in 1697 bear the marks of Vespasian, Titus and Domitian and are probably the remains of a sunken cargo (Britannia 22, 319). In the Thames Estuary, at least four mortaria with the stamps of Q. VALERIVS SE... have been brought up by fishermen in Herne Bay, but the wreck has not been located. Also in the Thames Estuary, six mortaria stamped CAVARIUS were recovered from the Ouze Deep but the precise location is not known (Dean 1984, 78/9). A cargo of mixed form samian from Lezoux, with the stamps of more than forty potters, was recovered from the Pudding Pan shoal, 10 km northeast of Whitstable. The remains of the vessel were not located, but Roman tiles and an intact amphora containing 6000 olive pits found in the area may have come from the same wreck (Smith 1909, 395–414; Dean 1984, 78/9). At Little Russel, east of St. Peter Port in Guernsey, several Beltran 2B amphorae were found protruding from the seabed; fragments of Baetican Dr 711 amphorae were brought up by a trawler (Keen 1979, 3). Several Haltern 70 form amphorae of first-century date were netted by French fishermen off the Isle of Wight and a number of Dressel 1A amphoras from Yarmouth Roads and the Ryde Middle Bank (Lyne 2005, 4). On the estuary of the River Tyne, the probable wreck of a Roman troopship on the Herd Sand is indicated by a shield-boss of Legio VIII Augusta, a helmet cheek-piece, a patera, 67 coins and other items (Hodgson, 2009, 71).

Figure 3.12 Interpretation of obverse of bronze coins of Cunobelin Duration of voyages and passage times By considering some of the recorded voyages made in later periods, we may obtain some idea of passage times of sailing vessels operating during the Romano-British period. Since that time there have been some changes in vessel design, notably the use of the fore-and-aft rig, and these led to some improvement in performance but, as will be seen, with sailing vessels it is wind, tide and loading/unloading cargo, not solely vessel capability, that are the dominant factors. Evidence for duration of London voyages in the 16th century

Wreck sites may survive as concentrations of artefacts and finds of Roman material on a coastal site may also be considered as potential indicators of a shipwreck. On the Isles of Scilly, Fulford (1989, 245–9) suggests that portable items and personal equipment deposited as votive offerings at temples on Nornour are distinctive, because of the narrow range of artefacts present, and points out that this is a characteristic of Roman ships' cargoes, assemblages from Roman waterfronts and of Roman ship deposits. He suggests that the most satisfactory explanation for the Nornour finds is that they derived from a second/third-century shipwreck from “a merchantman blown off course as it rounded Land's End on its way to serve markets in Wales and the northern frontier”. Six second-century amphorae from southern Spain in a cave close to Fife Ness may perhaps best be explained by a shipwreck on the adjacent Carr rocks, a

In January 1577, a commission appointed by the Lord Mayor of London produced a set of draft “Orders” that covered the majority of maritime insurance practices. The majority of these documents naturally relate to insurance matters, but there are some references to the maximum durations for trading voyages to and from various parts of Europe. These indicate periods of one to two months for journey times from the near Continent and Ireland, and five to six months from the Mediterranean (Kepler 1979, 265-8). These times do not, of course, reflect solely the minimum period of time on passage. If speed of sailing had been the only consideration, the estimated times would have 42

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  that a vessel be loaded so that neither freeboard nor stability are compromised; the two main factors to be considered are the total weight of cargo (affecting the freeboard), and its distribution within the hold affecting the stability. In the case of a single commodity cargo, e.g. amphorae or grain, the total weight of cargo is the significant factor, but with a mixed cargo (say the addition of a consignment of lead to either of these cargoes), the problem of effective distribution is added. At its simplest, the master would load the lead on the bottom of the hold, with the less dense grain or amphorae above. In the case of a single destination cargo this is relatively simple but, in the case of coastal trading, often multiple consignments, to and from several destinations, the problem is more complex. It is probable that, in particular, military cargoes comprised major elements of either grain, olive oil or wine, (stone or timber being confined to periods of construction and refurbishment), but it is equally probable that items of equipment and pottery were “piggy-backed”. Therefore, if several forts were to be supplied during the course of one voyage, the sequence of loading needed to reflect the most effective sequence of unloading, and therefore became similar to the problems of cabotage. In the case of merchant shipping, it was the aim to pick up new cargoes whenever possible, but as these would not usually be forthcoming at a military establishment (minerals under Imperial control, as at Caernarfon are a possible exception), return “in ballast” was more usual. In this case, high-density disposable commodities such as rubble, gravel, shingle or sand were loaded in the bottom of the hold to maintain stability, and this is the probable reason for the quantities of Pressili slate found near the third-century quay at Caerleon.

been much shorter, as the commission noted, in several orders, that a ship could reasonably be expected to sail at three miles (one Spanish League) per hour. Instead, the main part of each specified time reflected the commission's calculation of the extent of delays that could be expected during normal trading voyages. The nature of these delays may be summarised as follows; “interruptions by storms, contrary winds, grounding, fire, pirates, or the unability of ye shipp”; temporary internment by belligerent powers during wartime or by customs officials at any time; calling at various ports to load or discharge cargo; waiting in the Downs or off Dover for instructions for “right Discharge……. within this Realme, or to Saile from thence into France or Flanders”. Drawing comparison with the Roman period, Milne (1985, 31) has suggested that “Such features must have been just as prevalent in the first century as they were in the 1570s”, and “whilst one may argue that neither belligerent powers nor pirates were a threat at that time, they certainly were so in the late third to fifth centuries”. There seems little reason to doubt the general accuracy of the commission's estimates for the relatively long duration of trading voyages, as the insured’s desire for a maximum period would have been balanced by the insurer’s preference for a lesser period. This compromise between the interests of the insurer and the insured was, to some extent, regulated by the fact that many merchants acted in both capacities (Kepler 1979, 265–8). Whilst these times may seem unreasonably long by modern standards, it is worth noting that Braudel (1972, 363) stated that, on average, it took two to three months to cross the Mediterranean lengthwise in the late 16th century.

Vessel tonnage An old American customhouse formula, (length of keel × beam × depth of hold) ÷ 94, with all dimensions represented in feet, was used by Steffy (1982, 85) to calculate the tonnage of the Yassı Ada ship (see below). Applying this to the Blackfriars ship gives a tonnage of 55, compared to Marsden’s (1994, 89) estimate of 50, and to the St Peter Port ship gives a tonnage of 69 (Rule & Monaghan do not give an estimate of tonnage). When the formula used for Blackfriars and St Peter Port ships is applied to the Yassı Ada ship, [(keel 40 × beam 17 × depth of hold 10) ÷ 94], a 50 ton volume is indicated. From 1773, the tonnage of British merchant ships was calculated by a statutory formula of [(length overall × 60% beam) × 50% beam] ÷ 94 giving 51 tonnes for Blackfriars, 72 for St Peter Port and 58 for Yassı Ada. These differences illustrate the problems of applying varying formulae, but show sufficient similarity to suggest an indication of comparative tonnages.

Cargo stowage and ship stability A vessel’s cargo has to be stowed and distributed so that it has both a safe freeboard and adequate stability. Freeboard is the vertical distance between the highest watertight deck (or top of sides in an open boat) and the waterline. As cargo is loaded, draft is increased, and freeboard is lowered; if loaded to where the freeboard is below a safe minimum, a vessel may ship water and founder. On the other hand, if too lightly loaded, and therefore with excessive freeboard, ship-handling and steering capabilities are lessened, and the power to carry sail may be reduced. When a ship is transversely displaced from the upright (“heels”), it is essential that there is sufficient righting moment for the vessel to return to the upright. A balance needs to be struck between a “stiff” condition where the ship returns upright too rapidly, and a “tender” condition where the vessel will only slowly return upright. “Stiffness” leads to violent motion and the vessel becomes less sea-kindly; in cases of excessive “tenderness”, the vessel may roll to a point from which it is unstable and cannot recover to the vertical (McGrail 1989, 353–8). It is therefore necessary

Stowage factors The quantity of cargo that may be loaded to fill a hold of known capacity is calculated by the use of “stowage factors” (Appendix 3), and were used in Marsden’s

43

JAMES ELLIS JONES

  Type 2 amphorae are globular with the maximum diameter at the shoulder, again with a plain and rounded bottom, but because of differences in size, have been divided into “Small” and “Large” categories. Large Type 2 amphorae have an average capacity of 36.2 litres and a full weight of c. 46.6 kg. Small Type 2 amphorae have an average capacity of 15.3 litres and a full weight of 20 kg. The estimated number of amphorae of each type was 113 of Type 1 (weight 1,413 kg), 702 of Type 2 large (32,713 kg) and 85 of Type 2 small (1,701 kg), a total of 35,827 kg (Bass & van Doorninck 1982, 155–65).

calculations of the cargo-carrying capacity of Blackfriars. These factors take into account the volume and weight of cargo, as well as the interstices between items, space lost to containers and dunnage. Expressed in cubic metres per tonne, the factors range from tin ingots at 0.28, through iron scrap at 0.98 and earthenware in crates at 2.13, to bales of wool at 6.00. In the Roman period, a ship’s master would have made analogous estimates, based on experience and inherited rule of thumb (McGrail 1989, 356). Examples of the calculations for the Blackfriars ship are;

Taking Steffy’s (1982, 82-86) maximum 50 tonne loading, and proportionately increasing the quantity of Type 2 large to fill the hold (32,714 kg = c. 66% of 50,000kg), a full hold would accommodate c.1050 amphorae (702 + 50%), i.e. 21 amphorae per ton of cargo carrying capacity. The large Type 2 amphorae, containing 36 litres of wine, are sufficiently similar in shape, size and capacity to invite comparison with the Gauloise 4 wine amphorae, present in Britain from the late first to the early third century (Tyers 1996, 95). A Dressel 20 oil amphora containing 62 litres of olive oil weighs c. 84 kg (Peacock & Williams 1986, 52) and by comparison would lead to a cargo-carrying capacity of 595 amphorae, i.e. 12 amphorae per ton of cargo.

A cargo of 12 barrels of wine, based on a known barrel weight, when full, of 1.278 tonnes, a stowage factor of 1.78 and a hold volume of 28 cubic metres is 1.278 × 1.78 = 2.27 cubic metres = volume occupied by one barrel 28 ÷ 2.27 = 12.33 = barrel capacity of hold A similar calculation for large sacks of grain (weight 0.68 tonnes; stowage factor 1.5) is 0.68 × 1.5 = 1.02 cubic metres = volume occupied by one sack 28 ÷ 1.02 = 27.45 = sack capacity of hold and for the ragstone cargo 28 ÷ 1.3 (weight of 1 cubic metre of ragstone) = 21 = tonnage of ragstone.

Level of loading The seabed remains of cargoes of Mediterranean wrecks have usually been found to be less in weight than the hold capacity of theoretical reconstructions, and whilst this may be the result of non-recovery of the total amount carried, it is possible that loading to the maximum level was not normal practice (Parker 2004. pers. comm.). Whilst the maximum amount of cargo delivered produces the greatest return on investment, the sea-keeping capability of a vessel is of paramount importance. A major factor is freeboard, and this represents the safety margin, indicating to what depths a vessel may be loaded, dependent on the type of cargo, waters to be navigated, and the season of the year. Prudent masters and skippers could have followed this practice. For example, Marsden (op. cit. 81, 89) has calculated that the Blackfriars ship could have carried a cargo of about 50 tonnes, that the estimated 26 tonnes of ragstone found in the ship would have only partly filled the hold, and that there was capacity for a further 24 tonnes of cargo. It was not until the Merchant Shipping Act of 1875, instigated by Samuel Plimsoll, that a safe freeboard became a legal requirement.

Amphorae The stowage factors were calculated for modern ships and, not surprisingly, amphorae are missing from the list, so it will be necessary (and arguably beneficial, as archaeological evidence is used), to turn to other sources of information. As there are no significant remains of amphorae in wrecks from British waters, data from the well-excavated and well-published wreck of the Byzantine ship, wrecked at Yassı Ada, off the coast of Turkey, whilst carrying a cargo of c. 900 amphorae, will be used (Bass & van Doorninck 1982). Using a detailed 1:10 scale reconstruction of the vessel, Steffy (op. cit. 86) considered that a further 300 amphorae could have been stowed below decks, giving a total tonnage of slightly over 60 tonnes, and therefore in the same range as the Blackfriars and St Peter Port ships. It would, however, be unwise to transfer the maximum amphora loading of the Yassı Ada ship to a northern European context. Firstly, Steffy himself considers the number would have “loaded her a bit tightly” and, secondly, he suggests that the Byzantines “were probably fair weather sailors, keeping their ships in port when the strong winds blew from the north” (ibid.).

An average sea-going merchant ship? Based on the St Peter Port and Blackfriars ships, an average merchant ship of the period operating in British waters may have had a cargo-carrying capacity of some 60 tons and a hold capacity of c. 33 cubic metres. Whilst accepting that the sample is small, and may need to be amended in the light of future discoveries, this will be used as the basis for the calculation of volumes of traffic for sea-going ships. From shipwreck evidence, the

Detailed analysis of the Yassı Ada amphorae has been conducted, revealing two distinct types, defined by shape, weight and capacity. Amphorae of Type 1 have cylindrical bodies, tapering slightly to plain round bases, and are usually pinched at the waist to give an “hourglass” shape. With an average capacity of 8.2 litres, the weight, when full with wine, is calculated as c. 12.5 kg. 44

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  commonest type of merchant vessel found in the Mediterranean carried less than 75 tons of cargo (Parker 1992, 89) and this gives some support to the suggested size of vessel. Using a ship of 60 tons results in estimated single commodity cargoes of 32 tonnes of grain in sacks, 25,000 litres of olive oil in Dressel 20 amphorae, or a similar quantity of wine in Gauloise 4 amphorae. It is perhaps relevant that the average tonnage of coasting craft from Bristol in 1787 was 36.74 tons per vessel (Hussey 2000, 143). -o-o-o-o-o-o-oIn order to enable a reasoned assessment of the nature of voyages, on either sea voyages or coastal passages, the emphasis has been placed on sea-keeping and passage making, and it would have been noticed that little attempt has been made to introduce the topics of passage planning and journey times. The skipper of a Roman merchant vessel was primarily dependent not on the sailing characteristics of his vessel, but on the prevailing weather, tidal conditions and, to a lesser extent, the number and competence of his crew. If readers wish to study a realistic assessment of passage planning and journey times, they can do no better than to read The Roman Invasions of Britain (2005), in which Gerald Grainge draws on his extensive experience of sailing in the Channel and the Dover Strait to bring a seamen's-eye view to topic. Whatever the type of sailing vessel, not only the prevailing weather, but also the tidal effect is a primary consideration. For example, the Bristol Channel has an exceptionally high tidal range of some 13 m, second only to the Bay of Fundy, off Newfoundland. As a consequence, tidal streams may flow at up to 7 knots in certain areas, as indicated on the Admiralty Tidal Stream Atlas (Appendix 2), and even with a modern highperformance yacht, progress under sail against the tide is slow. The extreme range of rise and fall of the tide means that many hazards, such as the Gore Sands in Bridgwater Bay, are not only exposed for long periods, but also pose a significant hazard to navigation when covered. The majority of landing places do not have deep-water facilities and, for example, on the River Axe it is only possible to leave the river or to return to it, some two and a half hours before or after high water. In summary, the Roman merchant fleet included both shallow-draught vessels capable of carrying heavy loads, and efficient sea-going sailing ships capable of loading large cargoes. The average cargo capacity of Roman ships was not exceeded until the 15th century and that of the largest Mediterranean ships, not until the 19th century (Greene 1986, 25). At the time of the Claudian invasion, Roman naval and merchant mariners were clearly capable of dealing with Atlantic conditions, as were the ships in which they served. The wide-ranging variety of local craft, and their experienced sailors, enabled the rivers to be exploited for both military supply and replenishment, and for local user demand.

45

JAMES ELLIS JONES

  Whilst on campaign, the Roman army built and fortified a marching camp every day it moved, and this often served as its tactical base, providing functions similar to an operational base, but located in the direct vicinity of the enemy. The tactical base usually moved along with the army’s advance or retreat, and may have incorporated a similar function to the Forward Supply Depots of the British Imperial army.

Chapter 4 Supply and provisioning of the Roman army. Introduction John Mann (1996, 59) emphasised that “Roman history is essentially the virtually unique story of a nation trying to catch up with the situations produced by the incredible success of its army” and in an examination of the extent of water transport, it is apparent that the topic must be dominated by an examination of the requirements of the Roman army. Therefore, the purpose of this chapter is to consider the problems of supply and replenishment, with particular reference to transport by water.

Sources of Supply Middleton (1979, 91) argues that, in Roman Britain, long distance trade was dependent, indeed parasitic, upon official supply lines, and was almost certainly directed to the needs of the military market. He considers there is no evidence that the civilian market generated long distance trade and suggests that growth is likely to have been generated in areas “touched by the nozzle of the military vacuum cleaner, sucking supplies to the frontier”.

In the Republican period, armies were raised for specific military operations but, by the time of Augustus, the army had become a widely dispersed standing force, with garrisons spread throughout the Empire, having to be fed in both war and peace. In order to provide the essential supplies needed for this task a far-reaching and flexible logistical system evolved, using a blend of locally and distantly obtained resources and supplies, distributed from fortified central supply depots by an efficient transport system. The establishment of an effective supply and replenishment system; using it as a strategic and tactical tool, may be considered to be one of the major factors in Roman military success. Indeed, Vegetius emphasises this by quoting the military proverb, “Whoever does not provide for provisions and other necessities, is conquered without fighting”. (Mil. 3.26).

The army used various sources of supply, the major method being requisition or compulsory purchase at a fixed price; the supplies being obtained directly by the army or by officials, whom Strabo (Geog. 3. 4. 20) describes as “men of Equestrian rank, who distributed to the soldiers all necessities of life”. Military land (territorium or prata) extended for a considerable distance around each fort or fortress, and was sometimes farmed by the military or at other times by civilians to whom the land was leased. Contracts for the supply of food in bulk were also placed directly with civilian contractors. Tacitus’s eulogy on Agricola’s capability to root out corruption (Ag. 19) indicates that provincials were forced to supply the army with grain, and deliver it to military depots or forts. A complicated abuse of the system, described by Tacitus, is that a corrupt official required a large quantity of wheat from a farmer, preferably at a time when the farmer had no wheat to sell. The farmer was therefore forced to purchase from existing stocks in a granary, at an exorbitant price, which the farmer then had to re-sell to the official at the regulated price. In fact, the corn never left the granary and the official pocketed the difference. Tacitus cites a further abuse of the system, by an unscrupulous official, demanding that corn be delivered an inordinately longdistance from the place where it was grown, or a heavy payment made which, presumably, went into the official’s pocket. However, Agricola's predecessor was Sextus Julius Frontinus, one of Rome’s most able administrators, and it is unlikely that such abuses of power were common at the end of Frontinus’s governorship or that, if any existed, they were any less at the end of Agricola's first year of office. When corn for the army (frumentum in cellam) was requisitioned, i.e. was obtained by compulsory purchase, the payment was by no means ungenerous. The average market price per modius was 2.5 sesterces, but the requisitioned price was 4 sesterces and this indicates that, when requisitioning was carried out with a reasonable degree of fairness, there need have been little cause for complaint concerning the process.

Supply Lines The use of “lines of supply” (a continuous connection between the sources of supply, and the army in the field) is a major characteristic of any military logistical support system. Vegetius (ibid. 3.8) commented, “Among the things particularly incumbent upon a general is to see that the transportation of grain and other provisions is rendered secure from hostile attack. The only way to achieve this is to plant garrisons at suitable points through which our supply-trains pass. These may be cities or forts. If no fortifications are available, temporary forts are established in favourable positions and a number of infantry and cavalry stationed in them on outpost duty to provide safe passage for supplies”. An operational base is located at the point at which supplies are gathered, within the area of operations and often located on a port, containing storage and warehousing facilities to hold supplies to support the army through the campaigning season. The Roman army’s innovative approach to military problems was not followed by later European armies until the late19th/early-20th centuries, when it was mirrored by British Imperial army usage, with each Command, whether within the United Kingdom or overseas, being served by its own Base Supply Depot. A tactical base is usually located a short distance behind the army in the field, or within the marching camps.

46

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Because of the high cost of overland transport, bulk supplies of grain were obtained locally whenever possible. When the required quantity could not be provided by the surrounding countryside, there are indications of provisioning from considerable distances, including importation from the Continent. For example, a military warehouse at York contained barley, spelt wheat and rye, together with weeds that only grew far to the south of Yorkshire, with the occurrence of larkspur, a plant not native to Britain, suggesting a Continental source (Dark 2000, 83). In addition to locally provided food and standard military rations, many commodities, in particular wine, oil and relish, were shipped to military establishments, often from long distances. The army permitted soldiers to supplement their rations, with food purchased from sutlers or merchants who followed the army on campaign. However, the problems of transportation and storage limited the variety and quantity of food available, so that the diet of the soldier in the field would have been much more monotonous than whilst in garrison (Roth 1999, 16).

garrison. Whilst it would have taken several years to develop an effective transportation network, it is clear that it was relatively easy to order items that could not be obtained locally. However, a reference to a wagonload of hides, of which the writer says, “I would have already been to collect them except that I do not care to injure the animals while the roads are bad” indicates that transportation, particularly in the winter, might be difficult. There is evidence that some of the cereals consumed were produced from agricultural land under the direct responsibility of the fort, but private enterprise is indicated by the activities of local merchants and, in particular, a reference to a purchase of 5000 modii of grain indicates that significant quantities were purchased. An internal market existed within the fort, and records indicate the purchase of non-essential items including piper (pepper), sebum (tallow), sudaria (napkins or towels), coturnii (boots), sagaciae (cloaks) and superariae (overcoats).

Waateringe (1997, 261-6) has drawn attention to Caesar’s frequent comments on the importance of food supply for the army, e.g. “When the crops were beginning to ripen, Caesar set out through the Ardennes to fight Ambiotrix” (BGall. 6, 29) and “Before the attack, Caesar had given orders that nothing was to be set on fire; this was to ensure adequate supplies of grain and fodder if he chose to advance further” (ibid. 8, 3). He has analyzed Caesar’s mentions of corn, fodder, water, wood and cattle, in his writings on the campaigns of the Gallic and Civil Wars, with interesting results. For example, Gaul has 59 mentions of corn, Spain has 38 and Egypt (not surprisingly) has only 2; Spain has 16 comments on water, Africa has 8 and Italy has 0. Fodder seems to have been a problem in Gaul (16 mentions), Spain (14 mentions) but Italy has 0. He draws the conclusion that the number of times a product is mentioned, and the scarceness of it in a specific situation is more significant than the importance of the product per se.

When Agricola was Governor, every fort in Britain was ordered to be provisioned with sufficient supplies to last for one year (Ag. 22. 2–3). Each Roman soldier consumed approximately one third of a ton of corn per year and, assuming that in the third century the forces of Roman Britain numbered 25,000, the daily consumption of corn would have been 33 ½ tons (Webster 1969, 97–8, 216). Examination of the granaries (horrea) of legionary and auxiliary forts shows that they were capable of sustaining this demand (Haverfield 1912, 127–42; Birley, 1961, 272); the legionary fortress of Inchtuthil had a timber granary with an estimated capacity of 6300 cubic metres. A granary did not only store grain; the lowtemperature and good ventilation helping to preserve fresh and salted meat, vegetables, cheese, lard, olive oil and wine; several horrea in Britain, for example Ilkley and Balmuildy, have revealed amphora fragments (Gentry 1976, 26). The third-century granaries at South Shields contained seed assemblages suggesting the possibility of long distance importation of grain. Anderson (1992, 60–63, 81, 101–2) has suggested that, during the third and fourth centuries, the granaries were supplied with bread wheat that had been imported by ship, probably from the Netherlands.

Storage of supplies

Roman towns often developed near to the site of earlier settlements. For example, Cirencester is only two miles from the earlier oppidum at Bagendon, and it would seem reasonable to assume that, in the civilian context, only minor modification to the existing transport system was necessary. By contrast, supply to the army could not have developed organically from the Iron Age systems of distribution, firstly in terms of the quantities required, and secondly, because forts were usually located in areas without large resident populations, and therefore no infrastructure for production or transport existed.

Logistics of the army in the field When on campaign the Roman army took a large quantity of equipment and supplies into the field. In addition to the individual soldiers' personal weapons and equipment, this included artillery, missiles, portable fortifications, tents, medical supplies, cooking gear and much more. However, approximately 90% of the weight of supplies needed by the Roman army was made up by the requirements of food for the soldiers, fodder for the animals and firewood for cooking. Roth (op. cit. 2) goes so far as to suggest that “all military decisions from the basic strategic concept to the smallest tactical movement

The discovery, and subsequent decipherment, of the Vindolanda tablets (Bowman & Thomas 1984; Bowman 1994), has made a significant contribution to our knowledge of the problems of supply and demand in northern Britain. We now have clear evidence that a combination of purchase, requisition and direct production met the basic needs of the Vindolanda

47

JAMES ELLIS JONES

  consisted entirely of pack animals, giving greater flexibility and increased speed of movement, and that supply wagons were only used between campaigns. The legionary artillery, comprising the light artillery (catapultae), firing bolts with iron points, and the heavy artillery (ballistae), firing hewn stones, was transported by wagons and pack animals. A requirement of 1300 bolts or stones for the legion's artillery and, at an average weight of 2 kg per missile, approximately 20 pack animals, or five wagons, would have been needed to carry a load of 2.5 tonnes of reserve ammunition. Some 70 wagons, together with 160 pack animals, were therefore needed to transport the artillery of a legion.

were affected by, and often determined by, the need to provide these supplies to the army”. Whilst obtaining supplies by foraging from the local landscape played an important part, it was a limited resource. Although from a much later period, the comments of General Henri Bonnal in his account (1907) of the battle of Konnigratz in 1866 are significant. He gives 36 square miles as sufficient for the maintenance of an army corps (30,000–35,000 soldiers), some 1,000 men to the square mile, but only on condition of supplementing local resources by supplies from a base. British Field Service Regulations state that ordinary agricultural districts of Western Europe, not previously traversed by troops, will support a force of twice the strength of the population for a week at a maximum. Supply Trains The supply trains of medieval and early modern European armies were largely unorganised and comprised servants, batmen, camp followers, sutlers, wives and washerwomen. In effect, this provided much of the logistical support, and the soldiers generally carried only their arms and personal equipment. The Roman army bore a greater resemblance to modern European armies, with a limited amount of equipment and supplies carried by each individual soldier, with supply trains responsible for the transport, provision and supply of food and equipment. However, an important difference in the Roman army was that, unlike modern armies, support troops were not used, but reliance was placed on free and slave military servants attached to the army (Fortescue 1930, 6–7).

Figure 4.1 Mules towing carroballista (Trajan’s Column) The soldier as a load carrier It is important to consider the potential of the legionary soldier as a load carrier, as this determines the amount to be transported by other means, whether on land, or by water. Clearly, a balance had to be struck between the amount carried by the individual soldier, resulting in a reduction in the amount to be carried by the train, and his mobility and endurance in the field. The Roman soldier carried at least some of his rations on his person, but there is some difficulty in using written sources to establish the precise amount carried (Fortescue 1930, 6– 7). For example, Livy (Per. 57) suggested that each soldier was compelled by Scipio Aemilianus “to carry 30 days of grain” and, using a ration of 1 kg per soldier per day, this would mean a burden of 30 kg for each individual soldier; to this must be added his personal equipment and weapons. On the other hand, and perhaps more reasonably, Cicero praises the Roman soldier for carrying more than half a month's rations (cibaria), some 15 kg, along with his other equipment (Tusc.2.37). The reforms of Gaius Marius (157–86 BC) increased the amount of supplies required to be carried by soldiers, rather than by pack animals. Plutarch, in his biography of Marius comments that “Setting out on the expedition [to Numidia], he laboured to perfect his army as it went along, practising his men in all kinds of running and in long marches, and compelling them to carry their own baggage, and to prepare their own food. Hence, in after

Pack animals and wagons accompanying the army and carrying its supplies are collectively known as its train. The size of the train used by an army is an important factor in its fighting efficiency and mobility, as a lack of carrying capacity reduces the army's combat capability, but too large a train restricts movement (Van Creveld 1989, 47). There were four types of train; firstly, a “‘troop” train attached to an individual unit, which carried its gear and supplies; secondly, an “army” train containing equipment and provisions common to the entire force; thirdly, an “officers” train, transporting their personal equipment; and finally a “siege” train for the specialised equipment needed to take a fortified defensive position (Roth op.cit. 79–91). The invading army of 43, under Aulus Plautius, was probably accompanied by all these elements but, after that time, the major operational force was often a single legion accompanied by auxiliaries. A conjectural estimate of the number of pack animals in the legionary train is as follows. With one animal per contubernium giving 10 per century and therefore 600 for the six cohorts, 60 to carry animal fodder, plus a reserve of an extra 5% (40), plus some 250 animals to cover the demands of staff officers, the ambulance and supporting services giving a total of 950 pack animals per troop train. It is probable that the legionary troop train 48

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  remounts), mainly used for communication purposes but, when necessary, operating in a cavalry role (Brian Dobson pers. comm.). A 500-strong ala quingenaria was supported by 275 pack animals and the 1000-strong equivalent, an ala milliaria, required a train of 550 animals. When operating independently the most effective auxiliary unit was mixed cavalry and infantry, either a cohors equitata quingenaria, comprising 480 infantry and 120 cavalry or a cohors equitata milliaria with 800 infantry and 240 cavalry, requiring 230 and 450 pack animals respectively.

times, men who were fond of toil and did whatever was enjoined upon them completely and without murmur were called Marian Mules” (Vit. Mar. 13.1). Vegetius stated (Mil. 1.19) that, in addition to his arms (arma), the Roman soldier normally carried his personal equipment, including cooking gear (vas), tools (instrumenta) and rations (ciba). A number of modern writers (including Watson 1969, 62–3; Keegan 1994, 301–2; Fuentes 1991, 76–7) have attempted to determine the weight of rations by calculating the weight of arms and equipment, and deducting this from the total weight carried. An experiment was conducted by Junkelmann in 1984, with the aim of demonstrating the possibility that a Roman soldier could carry a load of up to 45 kg. German civilians were selected, none of them were athletes or chosen for their stamina, although they did train for the march on an individual basis. In the uniform of firstcentury legionaries, led by Junkelmann and using his reconstruction of the Roman pack, the group carried weights totalling at least 43 to 46 kg for 500 km over the Alps, averaging 25 km per day. Whilst re-enactment is not necessarily good evidence, it seems probable that if 20th century German civilians could handle such loads, trained first-century Roman soldiers could certainly have done at least as well, or even better (Junkelmann 1986, 43-58, 200). Jorit Wintjes (of Universität Würzburg) points out that this “was re-enacting a peacetime march through a pacified province at a somewhat leisurely pace” and he is “highly sceptical that, when going straight into bandit country one would carry around as much” (pers. comm.). After a detailed examination of the opinions of various authorities, Roth (op. cit. 68-77) considers that Vegetius' statement that the legionary soldier carried 60 Roman pounds (20 kg), in addition to his equipment, is plausible.

Fodder for the animals In order for an ancient army to fight and manoeuvre effectively, it was necessary for fodder for the animals to be available in sufficient quantity and quality. The Roman army generally did not usually campaign during the winter months of December, January and February (e.g. Ag. 18.1–2). If a war had not ended by the onset of winter, the army would retire into winter quarters (hiberna), primarily due to the lack of fodder available for animals, but also because of the difficulties of transporting supplies from a distance. “Hard” fodder, usually either barley or oats, was provisioned and distributed in a similar manner to wheat for the soldiers. Crops such as hay or straw, grown on farms specifically for the use of animals, are referred to as “green” fodder, and “pasturage” is the grass and vegetation on which animals can graze directly from the countryside. It is probable that Roman cavalry horses were slightly smaller than their modern equivalent, and a reasonable estimate of their requirements is approximately 2 kg of hard fodder and around 7 kg of green fodder per day. Whilst a significant part of a horse’s dietary requirement could be obtained by grazing, a Roman warhorse needed to be fed some grain in order to remain in good condition (Roth op. cit. 62–5). For example, Adrian Goldsworthy relates that when one of Wellington's exploring officers was captured by the French, he pretended to be a groom and was put in charge of their horses; he fed the best one on grain, and the remainder on grass. When he stole the best horse and made his escape, the pursuing French could not keep up with him (pers. comm.). In contrast to the dietary requirement of horses, some pack animals can function on poor quality pasturage and fodder, and are able to survive on leaves, thorns and thistles. Caesar noted that during the African campaign of 46 BC, when his forces were prevented from foraging, pack animals were even fed on seaweed, washed in freshwater and mixed with a little grass (BAfr. 24). In contrast to the requirement for war-horses, pack animals rarely need to run as far or for as long as cavalry horses, so the conditioning is likely to be different – “more to do with stamina than speed” (Goldsworthy, pers. comm.). Because of their large size, oxen require large quantities of food, approximately 7 kg of hard fodder, 11 kg of green fodder or, because oxen can obtain a large percentage of their nutritional requirement through grazing, approximately 22 kg of

The eight soldiers of a contubernium were allocated a mule to carry the two heaviest pieces of shared equipment. As un-ground grain was issued, each unit needed to have its own hand-mill (mola) of two round stone discs, a vat and wooden crank, weighing some 27 kg in total. A tent (papilio), weighing an estimated 40 kg, was made of leather or goats' skin, and was large enough to accommodate the entire contubernium. Eight men and a single mule could have carried the equipment of the squad, plus five days’ rations, giving a significant degree of self-sufficiency in the field. The addition of a second mule would have provided carrying capacity for a further twelve days’ rations. Auxiliary units Although auxiliary armour and weapons were, in some cases, lighter than those of the legions, the weight of many items of equipment, for example, provisions and cooking utensils, were similar. A 500-strong infantry unit (cohors quingenaria) had some 160 pack animals, and the 1000-strong equivalent, a cohors milliaria, needed a train of 320 pack animals. The auxiliary units supplied the majority of the mounted units of the Roman army, but each legion had a complement of 150 horses (including 49

JAMES ELLIS JONES

  barrack blocks at Caerleon indicate the same system of food preparation, but at century level, within garrison quarters. Whilst in the field, the soldiers’ food was cooked on open hearths, thus increasing the Roman army’s flexibility, as the need to provide portable ovens in the train may present a serious logistical problem. The grain portion of the soldiers ration could be eaten either as puls, a porridge or mush, made with water, salt and often with fat, oil or milk; similar to modern Italian polenta, or as “military camp bread” (panis militaris castrensis), a form of coarse black bread. Before grain can be made into bread, it needs to be threshed in order to remove the inedible husks. When obtained from supply depots, the threshing process was carried out at the point of collection, thus reducing the carrying load by the amount of the discarded chaff. However, when relying on locally collected grain, threshing was carried out within the camp; in order to reduce the time the troops remained in the open, and therefore vulnerable to attack (Roth op. cit. 44–60).

pasturage (Roth op. cit. 66–7). During the Peninsula War, Wellington's mules were given 2.3 kg of barley and 4.5 kg of straw per day and they performed very well on this diet (White 1984, 128–9). In the immediate post-1939/45 war period, the British army retained mule transport (the writer briefly served with a Pack Transport Troop), and it was found that a diet of only barley, without grazing, made the animals too fat and high-spirited. It has been generally considered that, because of the sensitivity of horses concerning fodder, they were rarely used as pack animals by the Roman army, mules, donkeys and oxen being preferred for this role. However, recent studies by Cluny Johnstone (2008, 128–145) may cast doubt as to whether this was the case in the Britain. Her concern that, on an Empire-wide basis the number of reported finds of mule bones is very small, led her to question whether a “discrepancy between the archaeological and historical records is real or whether it reflects problems in identifying mule bones”. A methodology, using discriminate function analysis, was developed to differentiate more accurately between horse, donkey and mule bones, and this demonstrated that just over 40% of the bones previously identified as coming from a horse were in fact mule bones and resulted in a ratio of 14 horses to 10 mules to one donkey (55%, 41% and 4%, respectively). The results were further broken down by area and the Rhineland, Danube and Balkan areas showed similar results to the overall figures, with horses forming about 56% and mules about 40%. The proportion of horses in Gaul was somewhat higher (64%) and at 82% was significantly higher in Britain (ibid. 133– 4). Johnson addressed the question of the proportionately smaller numbers of mules found in Roman Britain and accepted that the logistical difficulties of transporting mules across the sea (from the continent) must have been substantial for that cost to be offset by the increased expense of having to use larger numbers of horses as pack animals, or the extra time involved in using carts drawn by oxen. Of particular relevance to this study is her recognition that, in Roman Britain, the option to transport products by sea meant that the distances required for transport overland were considerably reduced and that therefore the slower pace of oxen carts and/or higher transport costs of horse pack trains were not problematic (ibid.141).

Every eight-man contubernium used its own cooking fire, on which to prepare bread from the uncooked grain that had been supplied, together with meat and vegetables that also needed to be cooked. The army in the field therefore had to collect a large amount of firewood or other fuel daily. Caesar considered the lack of firewood to be as significant as a shortage of water, fodder or grain (BCiv. 1.84.). In addition to the large amount of wood needed for firewood, the building of siege works and entrenchments often led to significant deforestation; for example, Josephus states that during the siege of Jerusalem in 70, all timber within 15 km of the city had been consumed by the besieging forces (BJud. 5.263–4). The soldiers’ rations In contrast to other ancient armies the Romans issued regular rations, with the cost of the supplies being deducted from the soldiers' pay, the normal rations for a soldier comprising bread, meat (either fresh veal, pork or salt pork), wine and oil. On active service, biscuit (bucellatum) was partially substituted for bread (on one occasion, Julian took provisions from forts to supply his field force with biscuit for 20 days [Amm. Marc. Hist. 17.8. 1–2]), and sour wine (acetum) substituted for wine, while the proportion of salt meat was increased. The diet laid down in 360 was two days of biscuit and one of bread, wine and vinegar on alternate days, and two days veal and one day’s salt pork. A sixth-century Egyptian papyrus contained the following rations scales; three pounds of bread, two pounds of meat, two pints of wine and an eighth of a pint of oil per day (Jones 1964 III, 191–2). This is significantly more than the daily ration of two pounds of bread, one pound of meat or cheese, and one pint of beer issued during the English Civil War. Grain (frumentum) accounted for the major part of the Roman civilian diet; this was equally true for the Roman soldier, and indeed the word “frumentum” was often used to refer to food supplies or provisions in general. A nongrain ration (cibaria) was issued in order to provide

Preparation of the soldiers’ food Modern armies generally use central facilities for the preparation of food, using specialist cooks to prepare the soldiers' meals; distributed to the troops in ready-to-eat form. In the British army, until its recent absorption into the Royal Logistic Corps, the Army Catering Corps was responsible for the preparation of food at all unit levels. By contrast, there is clear evidence for Roman soldiers preparing their own meals, at the contubernium level, whether operating on campaign or living within barracks. Herodian (Hist. 4.7.5) refers to Caracalla carrying his own grain and baking his own bread on campaign “like a common soldier”. The range of ovens located close to the 50

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  sufficient proteins, vitamins and other nutriments to maintain fighting fitness. The basic grain ration was supplemented by meat, vegetables (especially legumes), cheese, oil, vinegar, salt and wine; the Codex Theodosianus (7.4.6) confirms this by stating that the soldiers’ rations, in addition to biscuit or bread, included salt pork or mutton, wine, vinegar, oil and salt.

(Wright 1966, 224 no. 51). An amphora discovered at Carpow (Wright 1963, 166 no.51) had contained a wine flavoured with hore-hound; one discovered on the Cumberland Coast had come from Esuris in Lusitania (Wright 1955, 148 no.25) and probably contained olives (olivae) and a third, from Brough-on-Noe (Wright 1963, 166 no.50), had contained plums (pruna).

Davies (1971, 136) suggested that, in peacetime, the diet of the Roman soldier resembled that of the local population, as the surrounding area usually provided most of the garrison’s food. King (1984, 187–217) argues that there was a distinct Roman military diet influencing the surrounding civilian diet, rather than vice versa, pointing out that, at Augusta Raurica, the military diet was remarkably similar to that of civilians. There was no difference, for example, between the meat from domesticated and wild animals, poultry, oysters, edible snails and fish eaten by the legionaries at Vindonissa, and that eaten by the civilians. The evidence of tituli picti for various other items of food is remarkably similar (Callender 1965, 37–41). The diet revealed by the archaeology of British forts is on a par with that eaten in the towns, villas and Romano-British settlements, and often superior, with food being adequate in quantity and quality.

Although this indicates that the Roman soldier enjoyed a rich and varied diet, many commodities would not have been available whilst serving in the field; others such as meat would have accompanied the army “on the hoof”. Whether stationed in winter quarters, in times of peace, or later in permanent barracks, many of these items could have been “piggy-backed” on bulk consignments of wheat, or loads of amphorae containing wine or olive oil. These products would therefore not have created a significant addition to the volume of transport required, forming only a small part of the major consignment. Provision of grain for a campaigning Roman army in Britain Polybius (6.39) stated, “The infantry receive a ration of wheat equal to about two-thirds of an Attic medimnus a month, and the cavalry seven medimni of barley and two of wheat. Among the allies the infantry receive the same and the cavalry only one and one-third medimni of wheat and five of barley”. Based on this, and such other sources as the sixth-century Egyptian papyrus, Pliny’s Natural History (HN 18.12. 67-8) and the Codex Theodosianus (7.4.6), differing authors have produced a variety of calculations of the daily grain ration of a Roman soldier. Erdkamp (2007, 102) favoured 830 gm, Duncan-Jones (1982, 146) and Manning (1975, 112) 880 gm; Roth (1999, 24) 830 gm. Gentry (1976, 25) and Junkelmann (1997, 91) decided on 1 kg; Engels (1978, 123) and Goldsworthy (1996, 291) advocated 1.4 kg. Davies (1971, 122–42), who was considered by Webster (1985, 262) to have produced the most thorough account of Roman army diet, concluded that three pounds (1.33 kg) was the correct amount. Faced with this plethora of conflicting opinions it has been decided to use 1 kg as a standard daily ration of wheat for each soldier. It has to be confessed that there is no good evidence for selecting a figure from “somewhere in the middle”, but it does have the advantage of making somewhat easier the calculations that follow!

Examination of the bones of domestic animals from Roman military sites in Britain shows consumption of ox, sheep and pigs at every site, and goat on many. Some of the bones were from immature animals, indicating the consumption of suckling-pig, veal and lamb. The various livestock kept on site provided a supply of milk, from which cheese could have been made; cheese-presses have been found at Corbridge and Holt. Shellfish were a favoured part of the diet and included oyster, mussels, whelks and limpets, and whilst these would have been readily available for units stationed on the coast, e.g. Ravenglass, Caernarfon or Pennal, oyster shells have been excavated at inland locations such as Brecon, Chester, Hod Hill, Holt and Caerleon (Davies 1972, 122– 42). In a storehouse at Caerleon, the carbonised remains of barley, spelt, rye, wheat and oats were found together with lentils, horse beans and various weeds. Most of the weeds were not indigenous to Britain, and therefore were imported with the cereals. Helbaek (1964, 158–64) has concluded that the Romans probably intended to sow the lentils and rye, but that the grain was to be malted to make beer (cervesa).

The task of supply and replenishment required a high degree of organisation, and a logistical system capable of handling large volumes of products. For example, the demand for grain for the 20,000 fighting troops of a campaigning army comprising two legions, sixteen infantry and four cavalry regiments is some 3,500 tons. On active service, biscuit (bucellatum) was partially substituted for bread and sour wine (acetum) substituted for wine but, using the evidence of appreciable quantities of wine amphora fragments from various sites, Hilary Cool (2006, 132–3) suggests “that the soldiers had access to wine, even when on campaign”. Similarly, at Carpow,

Graffiti and tituli picti on amphorae give an indication of the food and drink consumed on military sites. In Britain, graffiti written on the neck of an amphora discovered in a cellar at the military stores depot at Richborough describes lympa, a wine from Mount Vesuvius and presumably produced before 79 (Bushe-Fox 1949). One of the handles of an amphora at Newstead was scratched vin(um) (Curle 1911, 268), and at Mumrills (Wright 1964, 184 no.40) a vessel with the inscription glukus (oinus), had contained sweet wine, and another at Caerleon had held amine, a high quality white wine

51

JAMES ELLIS JONES

  This commemorated the 57 years of continuous service by the 38th Foot in the West Indies from 1707 to 1764, and recalled the fact that their uniforms became so threadbare during their service in the tropics that they had to be repaired with pieces of sacking.

“A very large quantity of amphora fragments was recovered… and represents the importation of olive oil from Southern Spain” (Birley 1963, 202). Military slaves (calones) performed most of the non-combatant roles within the army, including assisting in foraging, milling and cooking, carrying gear and supplies and, perhaps most importantly, acting as muleteers (muliones). A legion was supported by some 1200 slaves and, depending on the composition of the unit, 120 to 240 military slaves for each auxiliary unit (Roth 1999, 114– 5), giving a total of some 7,000 calones (military slaves). At a daily ration of 0.8 kg per man (Duncan-Jones 1982, 146), this adds a further 1,000 tons, giving a total of 4,500 tons of grain. The demand for hard fodder for the army’s animals (each legion 170 horses, 1200 pack animals; each cavalry ala 550 horses, 275 pack animals; each infantry cohors, 160 pack ration animals) is sometimes overlooked, but with some 8,600 animals at 2.2 kg per day (Tomlin 1998, 31–84) requiring some 3,400 tons, it is equivalent to 75% of that required for the soldiers and slaves (Roth 1999, 62–5), i.e. a further 2,700 tons. The total of 8000 tons of grain required approximately 160 fifty-ton shiploads and, in addition, a daily ration per man of one-half of a sextarius (0.27 litres) of wine and two cyanthus (0.07 litres) of olive oil (Davies 1989, 187–208, Le Roux 1994, 404; Roth 1999, 40), add a further 36 shiploads. A supply fleet did not need to transport all the provisions for the entire campaign at one lift, as ships could make several roundtrips in order to meet the demand and, in logistical planning the speed of a vessel is as important as its capacity. Casson (1995, 140–3) suggests that the average speed of a merchant ship could be as low as 2 knots, but considers that this speed might be substantially increased if circumstances were favourable. The 196 shiploads of supplies for the campaigning army might therefore have been shipped to the north by 25 ships, each making eight trips, over the six-month period. It must be emphasised that this apparent degree of precision is not justifiable, and I am grateful to Adrian Goldsworthy for reminding me that this is a complex subject, “with almost everything we say being largely conjecture when it comes to numbers” (pers. comm.).

The army secured its food supply by various means, the major method being requisition or compulsory purchase a fixed price (Ag. 19.4). In the west of Roman Britain the major grain production area was the Cotswolds with, possibly, a smaller but significant (because of its proximity to the fortress of Deva) contribution from the Cheshire Plain (Carrington 2008). As the major areas of military activity were located in the mountainous terrain of firstly Wales and then later, and to a lesser extent, in the Highlands of Scotland, it is a reasonable assumption that the bulk of the campaigning army's grain supply would have been provided by water transport (see Libanius [Ora. 18, 82–3] and Appian [BCiv. 4.12, 100]), with connecting land transport being supplied as required. Many forts and settlements were located on the coasts, estuaries and navigable rivers, and therefore provided an excellent case for supply by water transport. In particular, the coast of Britain lends itself well to this method of provisioning because of its long coastline, the isolation of the forts and, during the early campaigns, the lack of an effective road system. The role of units of the Roman fleet, both in offensive operations, and in supplying logistical support for the army, will be discussed in Chapter 7, and the types of merchant ships and boats that may have been in use has been discussed in Chapter 2. Clearly, water transport played a significant role in the movement of troops and the supply of equipment and rations, throughout the whole of the Romano-British period. However, the amount of traffic varied dramatically over time with, for example, no military installations existing from the late second century on the coastline of Wales between Caernarfon on the north coast and Cardiff on the south coast.

-o-o-o-o-o-o-oFuller (1965, 315-6) has argued that Julius Caesar’s defective system of supply frequently compelled him to change his area of operations to his disadvantage. He contends that Caesar was not an organiser and that careful preparation, adequate supplies and the many other requirements needed to ensure the success of a campaign were distasteful to him or lost to sight by reliance on his genius to solve all difficulties. For lesser mortals, it is worth remembering the opinion of Vegetius that whilst “Time and opportunity may help to retrieve other misfortunes, but where supplies and provisions have not been carefully provided, the evil is without remedy” (Mil. 3.71). However, the problem has also been recognised in modern times as, for example, when the South Staffordshire Regiment was granted the distinction of a badge backing of buff-coloured Brown Holland material. 52

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Cudmore Farm, Tiverton, Cullompton, Hembury, Gittisham, Killerton, Bury Barton, North Tawton, Okehampton, Stoke Hill, Ide, Clayhanger, Lapford) and more recently at Rainsbury (Riley & Wilson North 2001, 77–9). Finds of first-century military character are known from the Mount Batten area of Plymouth Sound, and it would be surprising if a fort was not established at this sheltered coastal location. A fort in the Barnstaple area, on the Taw/Torridge estuary is probable, but this has not yet been found, although a possible fortlet is suggested at Newton Tracey (Griffith 1984a, 19). Nantstallon is situated at the highest navigable point of the River Camel and was, until recently, the only known fort in Cornwall (Fox and Ravenhill 1972, 56–111). However, in 2005, a fort of similar size to Nanstallon was located at Lostwithiel, overlooking the highest navigable point of the River Fowey and finds suggest occupation from the mid first to early fourth century. A further fort, showing possible evidence for silver processing, has now been located at Calstock, again situated on a navigable river. It is interesting that The Channel Pilot states that nearby St Austell Bay and Mevagissey Bay “provide shelter and good anchorage in south westerly gales” and that Tywardreath Bay provides similar protection from gales from the east (King 1897, 35).

Chapter 5 Military operations in Britannia This chapter is intended to form the preamble to a later examination (Chapter 7) of the role of water transport in the Province. It is not intended to provide a comprehensive guide to military campaigns, but rather to indicate the scale of logistical operations needed to maintain the operational capability of the army in times of both war and peace. However, in some cases, military actions are described in some detail, in order to demonstrate their relevance to waterborne operations. The conquest of the southwest peninsula The future Emperor Vespasian, commanding a battle group headed by legio II Augusta, advanced along the south coast and, according to Suetonius (Vesp. 4), fought forty engagements, subdued two tribes, took twenty strongholds, and captured the Isle of Wight. There is evidence that coastal supply bases at Fishbourne and Hamworthy supported Vespasian’s advance into the West Country and the presence of pre-Flavian pottery, and a well-chosen position, suggest that Topsham, on the Exe Estuary, also fulfilled this function. Recent excavation appears to have confirmed the existence of a fort/supply base (Britannia 32, 370).

It is surprising that, bearing in mind the evidence from other areas, e.g. the Mendips, north Wales, and Derbyshire, we have not found evidence for military exploitation (other than limited evidence from Nantstallon) of the mineral resources of Cornwall and this problem may be eased by the results from Calstock. The apparent absence of roads in the county suggests a dependence on coastal sea-routes; the location on navigable rivers of the forts at Nantstallon, Calstock and Lostwithiel gives support to this opinion.

The southwest peninsula was occupied by the Dumnonii, and in a Presidential Address to the Devonshire Association in 1891, R.N. Worth commented that they were “a race numerous and skilful, civilised, well capable in numbers and in natural resources of self-defence, habituated to strangers, profiting by commerce. That race was never conquered [by the Romans] …There would be an acknowledgement of [Roman] suzerainty - little more than nominal; a certain rendering of tribute… the periodic visit of a Roman trader - a welcome guest - or his constant presence in some favoured locality”. At the 1969 Roman Frontier Studies Conference, Aileen Fox (1974, 84) commented that, “For a long time it has been generally believed that the south-western Peninsula was neither conquered nor occupied by the Roman army” and then proceeded to present her evidence to the contrary. It is interesting to note that a similar view has also been held regarding the tribe of the Demetae in Wales, and a contrary opinion will be suggested later in this chapter. The presence of a considerable force in Dumnonian territory for several decades, suggests that the tribe resisted conquest and that, in the consolidation period, a network of forts was required to impose control. There is little evidence to indicate whether the conquest of the peninsula was achieved during the course of a rapid campaign, or by successive advances over the years. There is little evidence for military operations in the territory of the Dobunni, whose territory was probably the area of present day Gloucestershire, Herefordshire, Worcestershire and north Somerset, and it is possible that the tribe was actively pro-Roman.

The conquest and occupation of Wales. There is only limited documentary evidence available (e.g. the brief comments by Tacitus (Ann. 14, 29–30 and Ag.19) on the Anglesey campaign of 60 and the final capture of the island by Agricola in 77, and any attempt to reconstruct the history from that period must rely on archaeology. Marching (or temporary) camps are the primary indicators of Roman offensive operations but, when compared with northern England and Scotland, only a small number have been found in Wales and The Marches (Arnold & Davies 2000, 5–7). It will be later argued that this paucity of sites, particularly in west and north Wales, may be the result of the conquest of that area being achieved by amphibious operations, rather than by cross-country penetration over difficult terrain. Typology is of little help in identifying specific campaigns, as camps with tituli (an exterior bank and ditch protecting a gate) or claviculae (an internal arc of the rampart), once considered diagnostic of period, appear to be equally early. Morphology is equally of little assistance as camps thought to belong to the same series, such as Whittington in Shropshire and Penrhos in Denbighshire differ in shape. Whilst it is not possible to

Archaeological investigation has revealed the presence of many forts in Somerset and Devon (e.g. Wiveliscombe,

53

JAMES ELLIS JONES

  period in 138, it is probable that only Castle Collen, Neath, Caernarfon, Brecon Gaer, Caersws and Forden Gaer were still garrisoned. This pattern probably remained until the fourth century, with only the addition of the Saxon Shore type fort at Cardiff, a small fort/harbour at Caer Gybi and a signal station, on Anglesey (Jarrett 1969, 135–37).

assign camps to specific events, there does seem to be an indication of a strategy designed to drive a wedge between the Silures in the south and the Ordovices in the north; Jones & Mattingly (1990, 79–81) have produced a valuable map illustrating the possible early campaign routes. Jarrett (1994, 20) commented that the system of preFlavian forts in Wales looks like “a half-finished project” and this is probably the case, as the final conquest was delayed by some 15 years because of firstly, the need to cope with the Boudiccan revolt in 60 and, secondly, more pressing problems in northern England. The concept of Wales as a political entity had no significance in Roman times, and military campaigns were therefore directed against individual tribes within the area. Some evidence of Flavian or pre-Flavian occupation is indicated at the forts at Cardiff, Usk, Llandovery, Clyro, Wroxeter, Leintwardine, Whitchurch, and at Caersws. The distribution of forts appears to indicate the occupation in strength of the territory of the Silures, Cornovii and Ordovices, and it has been suggested that these tribes were anti-Roman, while others, such as the Demetae and the Deceangli have been considered as probably prepared to co-operate with the invaders (e.g. Frere 1978, 70: Jarrett 1969, 23–6). However, a 1.75 ha. auxiliary fort, probably founded during the campaigns of Sextus Julius Frontinus in 75, has been identified at Carmarthen (James 1992, 8–14), and a road running westward for 40 km from the fort, in the direction of Milford Haven and St David’s, is now known (Driver 2002, 60–1). This probably led to a fort, or forts, in Pembrokeshire (Arnold & Davies 2000, 15), and this suggestion receives some indirect support from the evidence of ships from Caerleon visiting that area (Boon 1978, 24–36). Most recently, a large fort has been recorded at Dinefwr Park, Llandeilo. Covering an area of some 4 ha. the fort was originally constructed in the mid-70s, with a smaller fort being built at some later date (Hughes 2003, 144–7; British Archaeology, September 2003). This new evidence must cast further doubt on whether the Demetae submitted to conquest without some form of resistance.

Towards the middle of the fourth century, the defences of the civitas of Caerwent were modernised by the addition of projecting towers or bastions, so that covering fire could be provided along the face of the wall, but it is difficult to accept earlier suggestions that these were designed to mount artillery (ballista/onager/catapulta). It is possible that this could have been the case at the Saxon Shore forts but artillery is a specialised function, and it is unlikely that the citizens of a civitas such as Caerwent would have had the necessary expertise. In any case, the use of catapults against small bodies of lightly armed invaders would seem inappropriate. Nevertheless, the external bastions would have given an impression of great defensive strength and, if manned by a local militia, possibly equipped with “army surplus” from Caerleon, would provide an illusion that might lead potential attackers to “try their luck” elsewhere. The occupation of the North of England Barri Jones (1968, 6) suggests, “Agricolan is an overworked adjective” and Shotter (2004, 24) considers that both Bolanus and Cerialis “may remain victims of the literary genre in which Cornelius Tacitus chose to write”. Tacitus may well be correct in considering Marcus Trebellius Maximus (63–9) as an elderly man with little inclination for launching military initiatives and Marcus Vettius Bolanus (69–71) is dismissed by him (Ag. 8.1) as “too mild a man for a dangerous province” but, writing in the reign of Domitian, Publius Papinus Statius commented that Bolanus penetrated Caledonia (beyond the line of the Forth and Clyde), where he constructed roads, forts and watchtowers and stripped a British king [possibly Venutius] of his armour (Silvae 5.2. 140–9). In a similar manner, Tacitus devotes only three sentences to the governorship of Petillius Cerialis (Ag. 17.1). He did, however, state that he “straightaway struck terror into the hearts of Brigantes. There were many battles, some of them not uncostly, as Cerialis embraced the greater part of Brigantian territory with victory or, at the least, warfare”.

Consequent upon the visit of Hadrian and the decision to construct a permanent northern frontier, there was a change of emphasis in the role of the legionary fortresses at Caerleon and Chester. Both legions were involved in the building of Hadrian's Wall and the Antonine Wall and, during the building season, the fortresses probably held only minimal garrisons. The switch from offensive deployment in the pre-Flavian period to the establishment of a policing garrison meant that security became the responsibility of the auxiliary units. The success of the policy of pacification is demonstrated by the scale of reduction in the size of the garrison achieved by the mid second century. Major reductions in the Welsh garrisons began under A. Platorius Nepos, governor from 122 to 125; the addition of garrison forts to Hadrian's Wall would certainly have required the redeployment of considerable numbers of troops. Many of the Welsh forts were de-commissioned and, by the end of the Hadrianatic

A legionary fortress had been established in c.58 at Wroxeter by legio XIV Gemina, and, with the decommissioning of the fortress at Wroxeter, a new fortress was built at Chester in c.75 by legio II Audiutrix. At Lincoln, a fortress was founded in c.60 for legio IX Hispana, with this legion subsequently moving its fortress to York c.71. Vexillations fortresses are known from Osmanthorpe, Broxtowe and Newton-on-Trent. It has previously been thought that the series of timber and turf forts along the converging routes of penetration to the north (Northwich, Middlewich, Ribchester, and Carlisle

54

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Tyne-Solway line performed a similar frontier-control function.

in the west; Roecliffe and Bowes to the east) were founded during the governorship of Agricola. However, the dendrochronological date of 72/3 for construction at Carlisle means that it is probable that they were established during the governorship of Cerialis. The known forts probably represent only a fraction of those that once existed, as forts of turf and timber materials that were in use for a short time in the Flavian period are archaeologically difficult to see (Bidwell & Hodson 2009, 10). The fort at Manchester was founded c.78 and became a hub of the regional transport network and, as part of the process of controlling the Pennines and its passes; a fort was established at Castleshaw, approximately a day’s march to the east (Philpott (ed.) 2004, 12).

For at least 30 years prior to the construction of Hadrian's Wall, the army had been involved in military operations across the Tyne-Solway isthmus, with Corbridge, Nether Denton, Vindolanda and Carlisle forming strong points in a recently subjugated area, and acting as part of a control and communication network in support of the operation to the north. At some time after c.86 the Roman army withdrew to the Tweed and south-west Scotland, and then to the Tyne-Solway isthmus at some time after 105. The forts at Corbridge, Carlisle and Vindolanda were all rebuilt c. 105 (Bidwell 1999, 12–13) and supplemented during the Hadrianic period by additions at Carvoran and Old Church Bampton. A major problem is that there is an apparent absence of an extension of the Stanegate to east of Corbridge and, although it is probable that other sites await discovery, the forts at Ebchester and Washing Well, Whickham, appear isolated from both communication and supply and replenishment. The suggestion by Bidwell and Holbrook (1989, 150 –153) of a pre-Military Way road running eastwards from Corbridge to the north of the River Tyne was supported by Dobson (1990, 237), who thought the notion of a road east of Corbridge north of the Tyne, before the Military Way, a good one but perhaps even earlier than the Wall, and that the Stanegate continuation east of Corbridge may well have run north of Tyne rather than south of it. The absence of a communication route from the Tyne to Dere Street and the rest of the frontier was questioned by Bidwell and Snape (2002, 257), who considered it “…scarcely credible that supply by sea did not play an important part in the conquest and occupation of north-east England”. It is possible that an early bridgehead fort existed at Gateshead, where there is evidence for a settlement lining the road leading towards the Roman bridge (Pons Aelius) across the Tyne, and there is also the possibility of a preHadrianic occupation at South Shields (Bidwell & Speak 1994, 14).

Following the withdrawal from northern Scotland c. 87, and the abandonment of all forts north of Newstead on the Tweed, there was a period of stabilisation and reorganisation with the rebuilding or alteration to some existing forts and the abandonment of others. The Pennine area was reinforced by new forts such as those at Bainbridge, Wensley Elsack, Ilkley, Slack, Castleshaw, Melandra Castle, Buxton and Brough-on-Noe and the trans-Pennine route, via the Stainmore Pass, saw the construction of forts at Carkin Moor, Brough-underStainmore, Kirby Thore and Brougham. To the west a road led from Ribchester to a new fort at Kirkham and to another, as yet un-located, in the Fleetwood area; in the north Lake District area forts were established at Caermote, Blennerhasset and Old Carlisle and the first Solway Firth fort was built at Kirkbride. Stability to the east of the Pennines is indicated by the abandonment of Long Sandall, Roall, Hayton, Broxtowe, Osmanthorpe, Newtown-on-Trent and Rossington. A further flush of construction took place around the turn of the century with new sites on King Street at Wilderspool, Wigan, Walton le Dale and Lancaster, the consolidation of the Lake District with forts at Watercrook, Ambleside, Troutbeck, Wreay and Old Penrith.

Hadrian’s Wall

The Stanegate frontier

A serious problem greeted Hadrian on his accession in 117 as “the Britons could no longer be held under Roman control” (SHA 5. 2) and, whilst there is no direct indication of location, insurrection probably occurred on the western end of the frontier, involving the tribe of the Novantae of Dumfries and Galloway. A coin proclaiming victory was issued in 119, and it seems that Trajan’s governor, Quintus Pompeius Falco, successfully dealt with the outbreak (Shotter 2004, 75). Hadrian came to Britain in 122, accompanied by a new governor (Aulus Platorius Nepos) and a new legion, legio VI Victrix from the Rhine, replaced legio IX Hispana at York. Possibly basing himself at Vindolanda, Hadrian initiated the building of a wall covering a distance of 117 km from Bowness on the Solway Firth (Ituna Aest) to Wallsend, on the estuary of the River Tyne and the combination of forts, turrets, watchtowers, mile-castles, fortlets and vallum, make it the most elaborate of all Roman frontier

The road running between Corbridge and Carlisle, known from medieval time as the Stanegate, has been seen as a basic component of a linear frontier of regularly spaced forts and fortlets (Kirkbride, Burgh-by-Sands, Carlisle, Brampton Old Church, Castlehill Boothby, Nether Denton, Throp, Carvoran, Haltwhistle Burn, Vindolanda, Newbrough and Corbridge) crossing the Tyne-Solway isthmus (Collingwood & Myres 1937; Birley 1961). However, Dobson (1986, 2) considered that there is no firm evidence that a frontier system was created on the Tyne-Solway isthmus between the abandonment of the Lowlands and the building of Hadrian's Wall. More recently, Hodgson (2000, 11–22; 2009, 11–15) has drawn attention to contemporary military sites on the German “frontiers”, where sites were irregularly sized and spaced, in order to use local topographic circumstances, and has argued that the pre-Hadrianic

55

JAMES ELLIS JONES

  A western extension of the Wall was established, continuing down the Cumbrian coast with forts at Kirkbride, Beckfoot, Maryport, Ravenglass and Moresby, as well as a series of fortlets and towers. It is presumed this was contemporary with the first plan for Hadrian’s Wall, which consisted of the Wall, together with the milecastles and turrets. No Solway towers and only a few fortlets have produced any pottery other than Hadrianic and they all appear to have been abandoned, and several towers demolished, when the Antonine Wall was built. Some fortlets were re-occupied following the withdrawal from Scotland, but appear to have been abandoned before the end of the second century and, with the exception of Ravenglass, many of the forts were run down in the early third century, or evacuated, as was Watercrook. Bidwell and Hodson (op. cit. 22) comment that, whether the evidence of damage at Bowes, Bainbridge, Ilkley and Ravenglass represent planned acts of destruction prior to re-building or are the consequences of enemy action, “must remain uncertain until the relevant levels can be examined in large-scale excavations”.

works. Major additions were made to the plan at an early stage with structures altered, rebuilt, or, in some cases demolished, throughout the second and third centuries. It is probable that the Military Way was not an original feature, but was built on the reoccupation of the wall following the withdrawal from Scotland. To the east of the River Irving, construction of the wall was entirely in stone but to the west, turf and timber was used, except for the turrets. Originally designed with mile-castles separated by pairs of turrets and fronted by a ditch and obstacles (e.g. lilia), with auxiliary units stationed in forts to the rear, forts were added to the wall itself, firstly at Benwell, Haltonchesters and Rudchester, followed by Chesters and Housesteads and then Carrawburgh, Greatchesters and Wallsend. Following the decision to build forts on the Wall itself, some of the Stanegate forts (including Kirkbride, Old Church and Nether Denton) were deactivated c.124–5 (Hodgson 2009, 15). It is probable that the construction of the Wall and its forts was completed by the end of Hadrian’s reign in 138. Building stones refer to work on the Wall by the classis Britannica and three southern civitates, the Catuvellauni, the Dumnonii and the Duritroges, probably as part of the original construction or the replacement of the Turf Wall with stone in the second century (Breeze & Dobson 2000; Bidwell 2008; Hodgson 2009).

By the late third century, west coast and river accessible forts such as Watercrook and Ribchester, display large volumes of coins compared to inland sites, indicating reoccupation and suggesting reconstruction of the defences as a response to the threat of coastal raiders. During the late third/early fourth century, an extensive programme of reconstruction and refurbishment was undertaken at many of the northern forts, for example Chester-le Street, Catterick, Castleford and Old Penrith. Coastal defences were further strengthened, most dramatically at Lancaster, where c. 330 a completely new fort, similar in design to the so-called Saxon Shore forts in the southeast of England, was constructed. The evidence of coin loss from the Cumbrian forts indicates that the defence of the north-west coast remained a priority into the late fourth century. For example, the latest coin from Bowness is an issue of Gratian of 367–75 and other Valentinianic coins have been found at Ravenglass, Maryport and Lancaster. Even later Theodosian issues are recorded from both Lancaster and Maryport and suggest that there was a continued military presence until the fifth century (Shotter, 2004, 195–204).

Hadrian died in July 138 and an inscription at Corbridge recording the construction of a building in 139 may indicate preparation for renewed activities in Scotland. The construction of the Antonine Wall in the early 140s (see below) led to a major change in strategy for the Hadrian’s Wall forts and the hinterland; some forts were evacuated, others retained but with heavily reduced garrisons, the vallum was slighted and the mile-castle gates were removed. The Tyne-Solway gap became almost devoid of troops, with the forts held by small detachments and the turrets and mile-castles being empty but, to the south of the Wall, Carlisle and Corbridge and South Shields (?) were still held. On the abandonment of southern Scotland in about 160, forts were refurbished, the mile-castle gates were restored and the rebuilding in stone of the remainder of the Turf Wall probably took place. Outpost forts remained to the north of Hadrian’s Wall at Netherby and Bewcastle to the west, and Risingham and High Rochester on Dere Street to the east, presumably to allow the army to maintain watch over the lands well to the north. It is possible that the threat of seaborne attacks from across the Irish Sea by the Scotti led to limited re-occupation in the third and fourth centuries. There is evidence of repairs being made to the wall in the early fourth century and some of the milecastles were occupied in the later fourth century. Geophysical surveys have revealed substantial civilian settlements at Birdoswald, Carvoran, Castlesteads, Chesters, Haltonchesters, Housesteads and Maryport and the vicus at Vindolanda has been extensively excavated. Wallsend was defended by a wall and ditch and, at Maryport and Castlesteads, a ditch may have defined the limit of the civilian settlement.

The Notitia Dignitatum (now generally accepted to date to c.395), lists 18 forts in northern England, some of which are yet to be identified. The Notitia represents the only written source of information on military dispositions at that time, but there are problems in matching the forts and their archaeology with the document (Bidwell & Hodgson 2009, 28–9), for example, the known forts at Watercrook and Ebchester are not mentioned. Incursions into Scotland Describing the Roman period in Scotland, Wilson (1863, ii, 26) commented that “It affects only a small portion of the country, and constitutes a mere episode, which might be omitted without a very greatly marring the integrity

56

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  have been a Roman bridgehead, used as a base for military campaigns inland, whereas others argue the site was simply a native Celtic settlement with evidence for trade with Roman Britain (British Archaeology, 12, 1996).

and completeness of the national annals”. This found an echo over a hundred years later; when Keppie stated that the Roman invasions of Scotland “amounted to a mere interlude (or even a hiccup) in the Scottish Iron Age.” (1986, 5). However, writing in 1894 on the subject of the Antonine Wall, General Pitt-Rivers considered that, as the most northerly boundary of the Roman Empire, it was “of interest to the whole civilized world”. He was undoubtedly aware that Tacitus had written (Ag. 23) “A good place for halting the advance was found in Britain itself. The Clyde and Forth, carried inland to a great depth of the tides of opposite seas, are separated by only a narrow neck of land. This isthmus was now firmly held by garrisons, and the whole expanse of country to the south was safely in our hands”. Pitt-Rivers warned, “The nation that allows the last vestiges of such work to perish will stand badly by the side of others that take a more enlightened view of such matters” (quoted in Hanson & Maxwell 1983, xiii).

The statements that “Agricola started his fifth campaign with a sea passage, and in a series of successful actions subdued nations hitherto unknown” and “The side of Britain that faces Ireland was lined with his forces” (Ag. 24) might suggest military action to the north of the Clyde, but the lack of archaeological evidence for Roman activity in the western Highlands indicates that this is unlikely. In Galloway, Flavian forts are known at Ward Law, Glenlochar and a fortlet at Gatehouse of Fleet, together with marching camps at Annan, Waterfoot, Glenluce and Girvan, pointing to a seaborne crossing of the Solway Firth being a more realistic option. This could then lead to inland advance up the valleys of the rivers Nith, Annan, Urr, Dee and Cree. Hanson (1980a, 23) has suggested that, because of the lack of evidence for contemporary Roman sites on the south coast of the Solway Firth, the point of embarkation is likely to have been further south. The large supply depot on the River Ribble at Walton le Dale, or the fort at Lancaster on the River Lune, might be considered as staring points. However, it is possible that the legions (as opposed to auxiliaries) set out from Chester, Wroxeter or even further south from the legionary fortress at Caerleon.

Writing in 77, Pliny stated that “for nearly 30 years now Roman armed forces have extended knowledge further than the Caledonian forest” (HN. 4, 102), suggesting that Cerialis and Frontinus carried out military operations in Scotland prior to the well-publicised invasions of Agricola. During the governorship of Agricola (c.77–83) there were several major advances into Scotland, by an army with a probable size of over 21,000 soldiers, into an area that contains the most rugged and broken terrain anywhere in Britain. Covering seven seasons of campaigning, the eastern coasts of Scotland were the scene of most action, with the Roman forces penetrating as far north as the Murray Firth. The west of Scotland was the focus of Agricola’s fourth and fifth campaigns, with the line of penetration to the north usually assumed to be through Allendale to Clydesdale. Agricola’s seventh, and final, campaign culminated in the battle of Mons Graupius (at a so far unidentified location), resulting in total defeat for the native forces. Whilst the site of the battle has been the subject of considerable debate, Alison Grant’s (2007, 104) observation that the battle “took place reasonably close to the sea” has much to commend it.

Founded under Agricola in c.83 the legionary fortress of Inchtuthil occupies 22 ha. of a large, steep sided plateau standing some 30m above the north bank of the River Tay. Geomorphology indicates that the Tay has changed its course in the past and, during the Roman period, may have flowed to the north of the fortress. Construction was undertaken by Legio XX Valeria Victrix and some three years later the original turf and timber defences had been modified by the addition of a stone wall. A 1.65 ha. enclosure, with no internal buildings, lies near the easiest access up to the plateau and could have served as the access point for heavy supplies brought up from the river. Extending to the edge of the plateau a further enclosure, usually known as the "officers’ compound" was defended by a ditch and rampart and contained not only standard timber barracks but also a somewhat palatial structure with hypocaust under floor heating which has been interpreted as an interim proprietoria, to accommodate the legate whilst his permanent quarters were under construction. By late 86 or the beginning of the 87 the interior was almost complete, with bath-house, sixty-four barrack blocks, four houses for tribunes, six granaries, a large hospital, a drill hall, a workshop and a temporary headquarters building, having all been completed. The transfer of legio II Adiutrix to deal with problems on the Danube frontier led to a decision to withdraw from Scotland abandon Scotland and, in a process of thorough destruction, the fortress was deliberately and systematically dismantled in order to prevent anything of value falling into enemy hands.

Tacitus stated that Agricola “drew up his forces along that part of Britain which faces Ireland, not in fear but in hope” (Ag. 24). There are two marching camps on the Ayrshire coast at Girvan, one of 14.9 ha. and another nearby of 6 ha. These are of sufficient size to justify the possible location of forces landed in a major seaborne landing to control that part of the Scottish coast and, possibly, also the place from where Agricola contemplated an invasion of Ireland. Tacitus comments that he had “often heard Agricola say that Ireland could be conquered and garrisoned with a single legion and few auxiliary troops” (Ag. 24). The recent discovery of a possible Roman fort at Drumanagh, 23 km miles north of Dublin, has led to suggestions that the Romans may have, after all, invaded Ireland The heavily defended, 20 ha. coastal site has produced first and second-century Roman coins, but its significance is disputed. Some claim it may

57

JAMES ELLIS JONES

  evidence” for the construction of the Antonine Wall, pointing out that whilst the best evidence for labour camps is from the east, most of the distance slabs have been found to the west. The Military Way, constructed of stone and gravel, was located between 15 to 40 m to the south of the rampart, was a little over 5 m wide and with drainage ditches on each side.

By the late-80s there was a withdrawal from the conquered territories apart from a few outliers (e.g. Newstead), being maintained in southern Scotland. Tacitus commented “Britain was completely conquered and immediately thrown away” (Hist. I, 2, 1) or, as more recently expressed by Woolliscroft and Hoffmann (2006, 15; 202), “The pull-out appears to have been a unilateral and entirely voluntary act. The Roman army simply gave up everything it had gained or built, and withdrew”. They have also suggested that we should use “a healthy dose of scepticism” when considering Tacitus as a wholly reliable source for Agricola’s campaigns and achievements.

The occupation and operation of the Antonine Wall led to the creation of outpost forts up to the River Tay (Camelon, Ardoch, Strageath and Bertha), some forts reused Flavian sites but others, such as Inveresk were new foundations or constructed adjacent to a Flavian precursor. The identification of temporary camps of likely Antonine date has proved more difficult than those of potentially Flavian or Severan date, so the extent of campaigning remains speculative, though the objectives appears to have been more limited (Rebecca Jones 2009, 21–28).

The Antonine Wall It seems probable that, as Antoninus Pius had no military experience, he followed the precedent set by Claudius 100 years before, and set out “win his spurs” in Britain. If this was the case, it was successful, as he retained the loyalty of the army and died, unusually for a Roman emperor, peaceably in his bed some 20 years later. Breeze (2006, 14) suggests that “it may seem strange to us that the Antonine Wall was constructed merely in order to boost to the authority of the Emperor over 1000 miles away in Rome, but stranger things have happened in the world of politics”. The army was under the command of Quintus Lollius Urbicus, a career soldier, who went on to become proconsul of Asia and Prefect of the City of Rome.

The timetable for the abandonment of the Wall has been subject to much debate (Breeze 2006, 99-102). An inscription from Hadrian’s Wall recording rebuilding in 158 (RIB 1389), suggesting an intention to abandon the Antonine Wall and reoccupy Hadrian’s Wall; this is further supported by a continuing rebuilding programme on Hadrian’s Wall through the 160s (Breeze & Dobson 2000, 131-3). It is unnecessary to envisage an abrupt abandonment of the Antonine Wall. As the latest dated coin from an archaeological context is a worn coin of Lucilla from Old Kilpatrick, struck between 164 and 169 (Abdy 2002, 196, 211), it is probable that both Walls were at some level of occupation, with the Antonine being de-commissioned whilst Hadrian’s was repaired/rebuilt.

Lying 160 km to the north of Hadrian’s Wall, the Antonine Wall, with 18 forts, covered a distance of 63 km from Old Kilpatrick on the Firth of Clyde (Clota Aestuarium) to Carriden on the Firth of Forth (Bodotria Aestuarium). With a garrison of some 7,000 soldiers it had a greater concentration of men than on Hadrian's Wall which had a complement of only 8000-9000 despite being roughly twice as long as the Antonine Wall (Hanson & Maxwell 1986, 169). Breeze draws attention to significant developments in the concept of frontier defence being incorporated, with the forts placed closer together, incorporating small enclosures, beaconplatforms and linked by road so providing a closer density of military installation than before. He graphically commented, “The Antonine Wall was no more military in purpose than was the Berlin Wall…..its purpose was frontier control”, likening the entry points to Roman “Checkpoint Charlie” (2006, 147; 2008, 13, 15).

From the end of the Antonine Wall to the Severan invasions In response to attacks from the north during the reign of Commodus, there was renewed campaigning in Scotland in the early 180s. Cassius Dio records that “the tribes in the island crossed the wall which separated them from the Roman legions, did a great deal of damage, and cut down a general and his troops; so, in alarm Ulpius Marcellus was sent against them. Marcellus inflicted a major defeat on the barbarians”. There is debate over whether some sites continued in occupation through this period, e.g. Cramond, where Holmes (2003, 154–5) argues that the excavated evidence shows no clear break in occupation. The legionary vexillation fortress at Carpow has long been seen as a Severan foundation, but reappraisal by John Casey (2011, 225–235) argues “… far from Severus building Carpow, he actually abandoned a pre-existing fortification and in so doing abandoned a system of frontier control that had been established by Commodus”. Whether Carpow was a Commodan, Severan, or even an Antonine, foundation, there was clearly an intention of long-term occupation. Its excavator, Robin Birley, wrote, “The nature of the buildings at Carpow does not indicate temporary occupation. The massive walls, the extensive

The annexation of a large area of territory prior to the construction of the Antonine Wall placed considerable demands on the army, and probably influenced the decision to build the wall in turf, with only the major buildings of the forts, and the ramparts at Balmuildy and Castlecary, being constructed in stone. This enabled the more limited construction skills of the auxiliary units to be used, and distance slabs indicate that, in contrast to the involvement of all legions in the building of Hadrian’s Wall, only the Second, with vexillations of the Sixth and the Twentieth, was employed on the Antonine Wall. Breeze (2006, 68) discusses “the haphazard nature of the

58

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  operations were carried out. Constantine possibly visited again later; and his grandson Constans came to Britain in 342/3 probably because of the northern frontier.

use of concrete, the deep hypocaust channels, all demonstrate the work of men who perhaps grudgingly intended to make their long stay comfortable” (1963,132).

The “Saxon Shore” defences In 197 there was again trouble on the northern frontier where the Caledonians broke a treaty and Severus at that time was concentrating on the Parthian War; so Virius Lupus (Governor of Britain) had no choice but “to buy peace from the Maeatae for a considerable sum of money, recovering a few captives” (75.5.4; 78.8.5).

It is perhaps unfortunate that our historical knowledge of these forts, constructed in the third century, is a result of their listing in the Notitia Dignitatum under the command of the Count of the Saxon Shore (Litus Saxonicum - the shore “facing” the Saxons, the short “settled by” the Saxons or the shore “attacked by” the Saxons?). This document was compiled in the late fourth or early fifth century, by which time the fort at Lympne, and quite possibly several others, had fallen out of use. A second problem is that, rather than all the 11 forts being constructed in the same timeframe, the archaeological evidence indicates that Reculver, Caistor and Brancaster were constructed in the earliest years of the third century, but that the remainder of the forts were not built until the 260s or, in the case of Pevensey, not until the 290s. The forts at Burgh Castle, Walton Castle, Bradwell, Richborough, Dover, Lympne and Portchester do not appear to have a specific association with the existing road network and their coastal location implies a naval presence at some, or all, all of the sites.

The Severan Campaigns In 209 Septimius Severus together with his family and the Imperial entourage, moved to York. At this time Britain was split into two provinces, Superior in the south and Inferior in the north, and York was upgraded to the status of a colonia. The Severan campaigns were concentrated on the east side of Scotland, with the main theatre of operations lay in the territory of the Maeatae to the north of the Forth/Clyde isthmus and limited use appears to have been made of the route northwards from Carlisle. Shotter suggested that the Severan campaigns probably made very limited use of existing - or more precisely, previously abandoned-sites, presumably relying on naval transport and avoiding the overland routes (1975, 88). The importance of naval involvement in the Severan campaigns is indicated by an inscription (CIL vi, 1643) recording a classis praefectus whose command, unusually, included not only the classis Britannica, but also elements of the Rhine and Danube flotillas, i.e. the classes Germanica, Pannonica and Moesica, all apparently combined for the Severan expedition. The evidence for the Severan campaigns (Birley 1999, 170-187; Breeze 1982, 128-136) comes primarily from literary sources, propaganda coin issues and the distribution of temporary camps, although the dating of these latter is less secure than often claimed. A series of denarius hoards of this general date up the north-east coast was traditionally linked to the progress of the Severan army, but is better seen as a broader diplomatic phenomenon in the decades before and (to a lesser extent) after the invasion (Hunter 2007c).

Figure 5.1 19th century reconstruction of the Saxon Shore type fort at Cardiff A number of scenarios for their purpose have been suggested, including as a defence against seaborne raiding, as trans-shipment ports or as a defence against an invasion by Imperial forces. Johnson (1976) considered that they served both for units of the army and also as naval bases from where units of the fleet, patrolling the most obvious routes, would intercept raiders before they could attack British shores. Frere (1978, 329–13) and Faulkner (2000, 90–1) also thought that their purpose was to counter piracy and seaborne raiding, and that those raiders eluding the naval patrols would be dealt with by high-mobility land forces. Esmonde Cleary (1989, 43) agreed, suggesting that Britain was under threat of raiding and piracy from peoples along the north sea littoral outside the Roman frontier on the lower Rhine, the forebears of the Anglo-Saxons. More recently, Andrew Pearson (2006, 345–50) comments that "Much of the

The main theatre of operations lay in the territory of the Maeatae to the north of the Forth/Clyde isthmus. The navy were clearly a key element in the supply chain, as indicated by the rebuilding of South Shields as a massive grain store, the reoccupation of Cramond and the use (new or continuing) of the coastal fortress at Carpow. Because of the Emperor's increasing infirmity a second campaigning season was entrusted to Caracalla, with somewhat inconclusive results. Severus himself was preparing to take the field again, when he died at York on February 4, 211 and the problems of the dynastic succession brought an end to this period of Imperial involvement. There is some evidence that, in 305, Constantius Chlorus, with his son Constantine, carried out punitive military operations in Scotland, before dying at York in the following year. We have no clear knowledge of where, or with what success, these 59

JAMES ELLIS JONES

  of a yet to be discovered fort and naval station (Mattingly 2006, 244). Wilson (1991, 142–7) queried the function as “signal stations”, and suggests they could have functioned as fortified strong points, in which the coastal population could take refuge when threatened by seaborne raiders.

literature concerned with late Roman Britain conveys a sense of an island under siege in which Germanic incursions were frequent, severe in character” but points out that archaeological evidence has proved elusive and makes a convincing case that attacks were sporadic, small-scale, and probably disruptive rather than destructive. White (1960, 9–13) argued that the scale of the defences suggested a perceived enemy with a capacity for siege warfare and not as a defence against lightly armed raiders. He suggested that the only large-scale threat was from the Roman army itself, in the course of an attempt to recover the province from Carausius and Allectus, whose usurpation of power lasted from the mid280s until 296. He thought that the forts were built to defend the coast against the forces of Maximian, but this is unlikely, as they are too thinly spread to pose a serious obstacle to a determined invasion force. There is also little to commend the suggestion that the so-called Saxon Shore forts in the west of Britain, at Lancaster and Cardiff, and the fortified base at Caer Gybi on Holyhead, were built for a similar purpose. The theory of an entirely Carausian/Allectan origin is no longer tenable, as the foundation of Brancaster, Caistor, Reculver, and possibly Richborough, are now known to predate the initial usurpation of 285. Recent opinion has tended to emphasise that the Saxon Shore forts fulfilled a primary role in the economic and military supply system, functioning as secure bases, where goods and supplies in transit could be held in temporary storage, protected by massive defences, and with a garrison of only moderate size. Gustav Milne (1990, 84) suggests that the function was not to protect the interior, but rather to facilitate access for commercial shipping and that the forts functioned as defended trans-shipment harbours. Perhaps this view is supported by their siting at the mouths of navigable rivers, rather than directly on the coast. Adrian Goldsworthy (2009, 343) notes that whilst the Saxon Shore forts could not have prevented every attack, they restricted access to the major rivers, thus making it harder for raiders to reach deep into the countryside.

Late Roman activity From the Notitia Dignitatum it is reasonable to infer that the Dux Britanniarum was probably based at York, with a responsibility for the security of the northern frontier and that the Comes Litoris Saxonici commanded from the Wash to the Solent. The sections of the Notitia referring to the west and Wales are absent, so we have little firm evidence for the military organisation in those areas at that time. The Comes Britanniarum probably commanded a small detachment (c.2000 men) of the field army stationed in Britain at an uncertain date. At this time, Britain was protected by an army (we do not know how large, but c.12,000 may be a reasonable estimate) that was effective for the time at which it existed. The later Roman army is sometimes compared unfavourably with the first-century troops of say, Augustus, Claudius or Vespasian, but this view is perhaps coloured by the undoubted differences in organisation, status and equipment. A parade of a first-century legion would have been impressive, with some four thousand heavily armed infantry, together with artillery, engineers and cavalry, dressed in parade uniform, drawn up by cohort and century, and drilling with a precision which would have been comparable to that of today's Brigade of Guards. By comparison, a fourth-century legion whilst possibly retaining the same title, consisted of perhaps one thousand legionaries, similarly but not identically uniformed; bearing more than a passing resemblance to Normans at the battle of Hastings. Whilst the later army lacked the ability and confidence of the early army with regard to pitched battles, it was still significantly more effective in this type of engagement than its tribal opponents and specialised in fighting lower scale conflict because this was the task most frequently required of it. However, it has been suggested that “it may be doubted whether in the course of the third century more Roman soldiers were killed by other Roman soldiers than by foreign enemies” (Goldsworthy 2008, 30).

Each of these scenarios still has its supporters, but it might be argued either that one or more function could have been served simultaneously, or that the function/functions may have evolved over time. Perhaps the most perceptive comment that may be made is that “The coarseness of the archaeological chronology is such that we are free to use (or misuse) the data to fit a number of scenarios” (Pearson 2003, 28). Whatever may be the case, it is the prime concern of this study to determine the extent to which transport, whether by water or by road, was employed during the construction of the forts and this is discussed in some detail in Chapter 10.

       

During the second half of the fourth century a chain of “signal stations”, with projecting bastions and a substantial watchtower, were established at Filey, Scarborough, Ravenscar, Goldborough, Huntcliff and possibly Sunderland. There is, however, no clear indication of the locations to which any signals were passed. Perhaps a significant cluster of Roman finds in the area of the natural harbour at Whitby is an indication

     

60

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Rome” (BJud 2. 382–3; 385–6). At the port of Ostia, cargoes of a hundred tons or more were transferred into either small boats (lintres) or river barges (codiciariae), for onward towage for 30 km to Rome, pulled upstream by teams of men or oxen along the winding course of the River Tiber. Whilst the major volume of traffic was the vast amount of grain needed to feed the city, some indication of the amount of other cargo passing through the port of Ostia is indicated by Rome's annual consumption of wine. In the first century, this has been estimated at over 25 million gallons, 90% of which arrived by boat from Campania and southern Latium, but by the end of the second century the majority was shipped in from Spain. In the case of olive oil, a “mountain” of 1,334,000,000 sherds, known as Monte Testaccio, accumulated over a period of two centuries, has been calculated estimated to contain fragments of 53,000,000 whole Dressel 20 amphorae, most of which seem to have carried imported Spanish olive oil (Johnson 2003, 109). This was used mainly for cooking and table use, but also for lighting and hygiene; it has been estimated that 1.5 million gallons were used each year for food alone (Franks 1940, 220/1).

Chapter 6 Roman sea power and the classis Britannica Introduction At the commencement of the third century BC, Rome was essentially a land-based power, but the Punic Wars with Carthage, the dominant naval power of the Mediterranean, necessitated far-reaching changes. In the winter of 261 BC, the Roman Senate gave orders for the construction of a hundred quinqueremes and twenty triremes which, constructed in some sixty days, enabled the defeat of the Carthaginian fleet at the Battle of Mylae in the following year. Rome celebrated her new status as a naval power by erecting a column in the forum, decorated with the beaks of the captured Carthaginian vessels. Naval operations played an important role in the civil wars of the first century BC, for example, in 79 BC a fleet of 1200 ships was used to transport Sulla’s army across the Adriatic from Greece to Italy (Plutarch Sulla, 27.1). Two years later, Scipio needed a fleet of 800 vessels to transport his army of seven legions, and its supplies, from Sicily to Italy (Plutarch Pompey, 11.2) and Pompey’s defeat of the pirates was aided by a fleet of 500 vessels. Caesar’s defeat of the Venetic fleet in 56 BC (BGall. 3. 9–13) had eliminated any threat to naval supremacy on the Atlantic coasts. After Octavian secured control of the Republic, by his defeat of the fleets of Mark Antony and Cleopatra at the Battle of Actium in 31 BC, major naval battles within the Mediterranean were a thing of the past and Roman could rightfully lay claim to mare nostrum. However, Jorit Wintjes draws attention to the naval activity during the civil war following Nero’s death, with the Imperial fleets “doing exactly the same thing, apparently quite effectively, they did during the earlier civil wars - that is laying waste to coastal areas” (pers. comm.).

The classis Britannica The first mention of a British fleet by Tacitus recounts one of the less glorious events in its history. He comments (Hist. 4. 79) that legio XIV Gemina was brought from Britain to the Rhine by the “Britannica classe”, but the fleet became separated from the army, was attacked by a tribe called the Cannifantes and most of the ships were sunk or captured (classem ultro Cannifantes adgressi sunt maiorque pars navium depressa aut capta). The command of a fleet was vested in a praefectus classis and was held by an equestrian, usually after the command of an auxiliary cavalry unit. The senior Roman fleet was based at Ravenna and its commander was paid an annual stipend of 300,000 sesterces. Next in seniority was the Misenum fleet whose prefect drew 200,000 sesterces per annum. The command of both the classis Germanica, based at Altenburg near Cologne, and the classis Britannica with its headquarters at Boulogne, attracted a salary of 100,000 sesterces. The prefects of the other provincial fleets, those of Pannonica, Moesica, Pontica, Syriaca, Nova Libica and Alexandrina, were paid only 60,000 sesterces, perhaps reflecting not only the relative importance of the British and German fleets, but also the need for securing the loyalty of their commanders (Starr 1960, 107). Spaul (2002, 45) suggests that the Prefect did not command a fleet in the sense that he was an admiral, but that his duties lay in the administrative role, and that he should probably be thought of as a “Minister of Marine Affairs”. This may not be true of the Mediterranean fleets, but as the last recorded Roman fleet action in northern waters was Caesar’s defeat of the Venetic fleet in 56 BC; this may well be the case in the classis Germanica and the classis Britannica.

The conquest of the northern provinces was eased by the development of fleets capable of close support of the army in offensive operations and equally competent in the non-combatant transport and supply of the military by waterborne transport. Up until the time of Augustus, naval forces were assembled on an ad hoc basis, usually for a specific campaign, raised by the military commander of the operation. Augustus organised a series of permanent fleets, within a formalised organisational and control structure and, as will be demonstrated later, naval power had a vital role during military operations in Britain. The major demand now became a requirement for a mercantile marine and, by the time of Augustus, the capability to deal with large volumes of waterborne traffic had become a necessity. The city of Rome, with perhaps one million inhabitants, had become dependent on the import of food and raw materials. A fragment from the fourth-century Epitome de Caesaribus states that in the time of Augustus “twenty million modii [c.130,000 tonnes] of grain were imported each year to the city”; this was in addition to the supply from Africa that Josephus stated fed “for eight months of the year the populace of

61

JAMES ELLIS JONES

  separate Emperors were proclaimed, Septimius Severus in Pannonia, Pescennius Niger in Syria and the Governor, Decimus Clodius Albinus in Britain. Civil war lasted for four years during which time Albinus removed part of the British army (the number is not known) to support his campaigns on the Continent, but after his defeat by Severus at Lyon, Britain was recovered in 197. In 208 Severus arrived in Britain with large reinforcements including legionary vexillations (ILS 9123), together with the fleets of Germany, Pannonia Moesia to combine with the British fleet. After campaigning in Scotland, probably including close support operations by the combined fleets, Severus died at York in 211.

The transport of personnel A major requirement was the transport of military and state personnel, both across the Channel, and around the coasts of Britain. The scale of movement ranged from individual soldiers on posting, to the transfer of complete legions. The range of activity undertaken is illustrated by tasks such as the carriage of members of the cursus publicus, the movement of vexillations, transport of units on a routine change of garrison, and major operations, as in the case of the Sarmatians, when 5000 heavy cavalry, plus dependants, were transported from the Continent. Aside from these somewhat mundane tasks, was the carriage of high status officials; one of the most eminent of these personages was undoubtedly Claudius who, when undertaking his journey from Rome to formally complete the conquest of Britain “first had sailed down the river to Ostia and from there he sailed along the coast to Massalia. Thence he travelled partly by land and partly along the rivers until he reached the Ocean. He crossed into Britain and joined the troops who were awaiting him by the Thames” (Dio.Cass. Ep. 69. 19). In order to minimise the Emperor’s discomfort, during the Channel crossing, probably in a trireme, the maximum possible speed would have been striven for, as earlier “on his way there (Massalia) by ship from Ostia he was twice almost drowned in storms blowing from the north off the coast of Liguria and near the Stoechades Islands” (Suet. Claud. 17).

Cross-channel troop movements continued after the demise of the classis Britannica, with legionary vexillations from Britain, probably accompanied by auxiliary units, being deployed against the Alamanni and the Franks in the campaigns of the mid-250s (CIL XII 6780). Troops were sent to join the Danubian expeditions under Gallienus, but did not return, as the establishment of the so-called Gallic Empire removed Britain from the influence of the central government in Rome. The integration of Germany, Gaul, Britain and parts of Spain to form the Imperium Galliarum, resulted in the further movement of troops to the continent, but we have no records of the quantities involved, the defeat of Tetricus by Aurelian in 274, ending the breakaway empire. Following the accusation of misappropriation of recovered booty and informed that his execution had been ordered by Maximian; Carausius took refuge in Britain in 286 and established himself as Emperor. His power base was the Roman fleet and he continued to hold Boulogne and significant parts of northern Gaul and was probably the victor in a naval battle with Maximian in 289. Boulogne was re-captured by Constantius in 297, Carausius was murdered and replaced by Allectus, his finance minister who, after a full-scale invasion across the Channel (Mason [2003, 153] suggests by 500 vessels) was killed in battle and Britain was returned to the Empire. Constantius again came to Britain in 306 to campaign in Scotland against the Caledonians and Picts and, after achieving victory, he died at York in 310. His son, Constantine, was proclaimed Emperor and withdrew troops from Britain to support his successful claim to the Purple (Frere 1978, 376–86). Following the death of Constantine in 337, friction between Constantius and Constans escalated into war in 340 and the consequent removal of troops from Britain to fight in the abortive invasion of Italy. In 359/360, Lupicinus, the magister equitorum, was sent by Julian to take two bodies of auxilia palatina and two numerii across the Channel during the height of winter (Salway 1993, 256). The socalled Great Barbarian Uprising of 367 brought a response in the person of Count Theodosius, who crossed the Channel with four regiments of the field army and restored a semblance of order to Britain (ibid. 270–1). Yet another usurper, Magnus Maximus, probably holding the office of either Dux Britanniarum or Comes Maritimi Tractus, led a considerable number of British troops on a

Starr (op. cit. 153) considered that the “transport of men and materials from the continent to Britain always remained the chief function of the British squadron”. That he should have perhaps phrased this as “to and from the Continent” is demonstrated by the number of recorded major movements that took place in both directions. For example, legio XIV Gemina was withdrawn from Britain for Nero’s campaign in the Caucasus in c. 66/7, returned in 69 following the battle of Bedriacum, and were transferred to Germany in 70. Legio II Adiutrix was posted to Britain in 71 and transferred back to the Danube in 85–92 and sometime before 130, legio IX Hispana was withdrawn from Britain. Revolt in the north c. 117–20, in which a centurion from Vindolanda was killed, resulted in the reinforcement of the British garrison by the despatch of 3,000 troops from Germany and Spain to the Tyne (ILS 2726; 2735). In the mid-140s, troops were sent to a campaign against the Moors (AE 1960, 28) and by contrast, reinforcements were sent from the continent to Britain in the same year (RIB I. 1322). The peace settlement c.175 following the conquest of Sarmatia, resulted in the transfer to the Roman army of 8,000 heavily armoured cataphract cavalry, of whom 5,500 (plus horses and dependants) were shipped across the Channel to Britain (Salway 1993,155–6; Dio. Cass. Ep. 71.16.2). Vexillations of two British legions were transferred to Gaul to aid in the suppression of a revolt in Brittany (CIL 3 1919, 8513, 12813; ILS 2770). The last decade of the second century saw major troop movements, back and forth across the English Channel as, following the murder of Commodus in 192, three

62

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Coates, quoting Dio Cassius, state that the Roman warships in the sea-battle against the Veneti were “fast ships from the Mediterranean”, but Parker (1999, 477–8) believes that there is no evidence that Roman warships made such passages He suggests that the requirement to put ashore each night, in order that the large crew might be fed, watered and rested, over the course such a long voyage makes a passage from the Mediterranean unlikely and that Dio was referring to ships built on the River Loire. This view receives some support from Marsden (1994, 175–6) who considered it rare for any Mediterranean shipping to use this route. Warships of Mediterranean type, for example, triremes and biremes (see below), were certainly used in the northern waters. The fact that the warships of this type were built locally (see County Hall ship -Chapter 3), should not surprise us. In a later period, many of the warships of the British East India Company were, in armament and design, virtually indistinguishable from the frigates, brigs, sloops, and even “ships of the line”, of the Royal Navy. However, they were built in local yards in India, and were colloquially known as “country-built ships”.

successful invasion of the continent in 383. In 398, the Vandal general Stilicho brought/sent? troops to Britain to counter an incursion from Caledonia, but in 401 troops were withdrawn to the continent in order to counter a Visigothic threat to Italy (Frere 1978, 406–7). Some of these troops probably returned to Britain because the last British usurper, Constantine III, was able to muster sufficient force to launch a successful assault on the continent in 407, controlling all Gaul and Spain until his eventual defeat by the forces of Honorius in 411 (ibid. 408–9). These events are important not only in the demand for shipping, but also because of their effect on the “ration strength” of the army in Britain. In essence, there appear to have been at least fourteen occasions of significant troop transfer to the continent (often in the service of British usurpers), and at least fifteen major reinforcements (sometimes of returning vexillations) of the garrison of Britannia. Whilst this only amounts to some thirty major troop movements over a period of 350 years, it would have been necessary to maintain a seatransport capability able to respond to such a demand. It is of course probable that contracts for the transport of troops were placed with civilian contractors. The “minor” fleets of the Mediterranean were primarily involved in the movement of officials and personnel, and were maintained for that purpose in areas where offensive operations were only a remote possibility. By contrast, throughout the Roman period, the northern provinces were seldom entirely peaceful, and it is therefore improbable that total reliance would have been placed on non-military sources.

Casson (1974, 7.6) writes that “Roman writers… consistently use layman’s language when referring to warships, instead of naval jargon, biremis for bireme, being but one example. They almost always speak of triremes, quadriremes and quinqeremes, whereas the official terms were trieres, quadrieres and penteres”. However, Gerald Grainge points out that it is not surprising that Roman writers used the Latin rather than the Greek term (pers. comm.).

Figure 6.2 Illustration of first/second-century bireme Other than in the description of a mock sea fight in the Fucine Lake, Tacitus makes no mention of a type of warship larger than a quadrireme (a vessel with two banks of oars on each side, with two rowers on each oar). This is not surprising, because the battle of Actium had demonstrated the value of the trireme (a vessel with three banks of oars on each side, with a single rower on each oar) and, in any case, there was no longer another major naval power in the Mediterranean to provide the need for large vessels in a fleet action. A major function of the navy was now to serve the Roman armies on the Rhine and Danube frontiers; Tacitus makes a number of references to triremes, for example, in describing how Corbulo “brought up his triremes by the Rhine channel, and the rest of his vessels according to their draughts, by the estuaries and canals” (Ann. 11.18.4). There is a

Figure 6.1 Transporting supplies during the Dacian Wars (Trajan’s Column) Warships and transport vessels Warships (Naves Longae) There is no archaeological evidence in Britain for ships of the Roman fleet, but in his commentary on the Gallic Wars, Caesar (e.g. BGall. 3. 9–13) gives considerable detail concerning both warships and transports, and it is unlikely that any significant changes had taken place by the time of the Claudian invasion of 43. Morrison and 63

JAMES ELLIS JONES

  (ibid. 263–4). Their sea-worthiness in northern waters is indicated by Tacitus who records how, in 83, three liburnae were “hijacked” by a cohort of Usipi, possibly in the Firth of Clyde; sailed north round Scotland, then southward down the coast, crossed the North Sea and eventually reached the territory of the Frisii on the coast of Holland (Ag. 28). It is arguable that Tacitus deliberately diminished the achievements of the Usipi in order not to detract from the subsequent circumnavigation of Britain ordered by his father-in-law Agricola. To have sailed a liburna northwards along the uncharted waters of the west of Scotland, rounding Cape Wrath, then negotiating the difficult Pentland Firth on the north coast of Scotland and then southwards down the east coast of Britain, before finally crossing the North Sea is in itself a remarkable achievement. When it is remembered that after one of the liburna's professional helmsman had escaped, the other two were murdered, it becomes clear that though the rebels were soldiers not sailors, at least some of their members must have been skilful skill seafarers, used to disciplined oarsmanship and sail handling and with a considerable knowledge of coastal navigation. The fate of many of the undamaged vessels of the Spanish Armada whilst attempting the same voyage, albeit from east to west, in vessels with superior seakeeping abilities and crew accommodation, serves to illustrate the achievements of the Usipi.

possible depiction of a trireme with three banks of oars on Trajan’s column.

Figure 6.3 Trireme/?bireme on the Danube (Trajan’s Column) Although a relief at Boulogne (CIL XII 3564) refers to a trireme, the “Radians”, it is probable that the liburna was the predominant warship in the northern provinces. This type of vessel was used in a variety of roles including conveyance of dispatches, transport of fleet officers and as scouting and patrol vessels. The origin of this type of vessel is described by Appian (Hist. 3), who writes of the Liburnians as a maritime people, “an Illyrian tribe, who practised piracy in the Ionian Sea and the islands, with fast, light ships, so that still now the Romans call their light fast two-level warships liburnae”. The term came to cover a wide variety of two-level oared warships, in some cases equipped with rams, the largest of which rowed twenty-five men on each side, and the smallest had eight oarsmen a side. On Trajan’s Column, three biremes are depicted with large crews, rams, steering oars, outriggers and two banks of oars. Two have large stern cabins and one has a prow richly decorated with tritons and a swan figurehead. There are several other depictions of biremes, one with a spritsail, and one with a lowered mast. Another bireme displays a standing mast (usually lowered before battle) with furled sail, mast-stays, hanging ropes, a raised and catted anchor, and a very wide steering oar blade (Coulston 2001, 119–23). As reconstructed, a typical liburna is 18 m on the waterline with a beam of 3 m and displacing about 15 tonnes. Sail were used whenever possible, but Morrison (1995, 317) considers that a sprint speed under oars of just over 7 knots should be attainable. Whilst under passage, not all the oars would be manned at the same time, with the rowers (remiges) being divided into squads and rowing in shifts. They were used in a variety of roles including conveyance of dispatches, transport of fleet officers, used as scouting vessels and, when cataphract (decked and boxed-in), were able to engage enemy vessels in battle

The ability of the fleet to operate in north Atlantic waters was demonstrated when, after the battle of Mons Graupius, Agricola gave orders for the circumnavigation of Britain (Ag. 38.4). The voyage began and ended at Portus Trucculens, but this is otherwise unattested and its location is unknown, although the subject of many suggestions, ranging from Carpow-on-Tay or Cramond on the east coast of Scotland, to the less likely Richborough on the south coast of England. Hind (1974, 285–8) has argued a reasonable case for Moresby on the Solway Firth and, most recently, Wolfson (2008) has made a somewhat less acceptable case for Lerwick on the Shetland Islands. Tacitus records that the Orkney (Orcades) Islands were discovered and conquered and that Thule (Shetland or Iceland?) was sighted, but not landed because “winter was approaching” (Ag.10). Whilst Vegetius is not a wholly reliable source, his description of camouflaged scouting craft is of considerable interest. “With the larger liburnae may however be associated light scouting boats (scafae exploratoriae), which may have about twenty rowers in a single oar-room. With these, it is customary to surprise and sometimes intercept the supplies of hostile ships, reconnoitre their movements and discover their plans. But as their whiteness would show up at a distance, their sails and ropes are tinged a blue colour, which is like the colour of the waves, and even the wax composition, with which their hulls are dressed are similarly coloured. The seamen and soldiers are also clothed in the same colour, for better concealment on scouting duty, not only at night but also by day” (Mil. 4, 37). There has been some discussion on the exact rendering of this passage (Dove

64

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  carried on each ship. This would give a requirement of 833 transports, and Peddie (1987, 40) arrives at a similar figure. Salway (1993, 61–2) assumes that the whole invasion task force sailed at the same time, “to make opposition to the landing less easy” and that “one section of the Roman army almost certainly made its base at Richborough…one probably landed near Chichester… the position of the third landing is quite unknown”. However, the statement by Dio Cassius (15, 19) that the expedition “sailed in three divisions” might equally apply to a single landing place, but with a first wave carrying out the initial landing, a second wave securing the beachhead and the final wave providing reinforcements for the breakout into the neighbouring countryside. Whatever may be the fact, (the writer’s opinion is that the most convincing case is made by Grainge), it is certain that a large number of vessels would be available on this side of the Channel for future operations, both in the initial conquest and for further operations on the coasts of Britain. In Chapter 7, it will be suggested that some were later deployed in campaigns on the coasts of Wales.

1970, 15–20; Gauld 1990, 402–406), but it is perhaps surprising that the effectiveness of either green or blue as a camouflage, for vessels operating in the grey waters of the North Sea, has received no comment. From the medieval period, vessels usually were brightly coloured, but when various schemes of camouflage have been used by modern European navies, a shade of grey, to blend in with the colour of the sea, has generally been considered the most suitable. Transport vessels The Roman merchant fleet of the Mediterranean included both shallow-draught vessels with a capacity for carrying heavy loads, and efficient sea-going sailing ships capable of loading large cargoes. The average cargo-carrying capacity of Roman ships was not exceeded until the 15th century, and that of the largest ships, not until the 19th century (Greene 1986, 25). The most common type of vessel, found in all periods, carried less than 75 tons of cargo or, for example, approximately 1500 amphorae (Parker 1992, 26). From the first century BC to the third century AD, vessels of 75–200 tons, capable of carrying 2000 to 3,000 amphorae became common in the Mediterranean. Parker (1990, 335–46) has shown that there is a general tendency towards smaller ships in the later Roman period and also that, by the fourth century, ships were being built with a greater reliance on iron-bolt and nail-fastenings, rather than the traditional morticeand-tenon jointing. He sees this trend towards smaller and less expensive vessels as the result of a decline in volume, and therefore profitability of trade, although accepts it may alternatively indicate changes in the organisation of trade, to conduct more and shorter journeys.

It was previously commented that, in northern waters, a Roman skipper was largely dependent not only on the sailing characteristics of his vessel or the competence of his crew, but also on the prevailing weather and tidal conditions. However, when operating in direct support of the military, there would be occasions when maximum operating speed, and therefore minimum passage time, would be the prime consideration. Marsden (1994, 197– 8) has suggested a theoretical maximum speed of between seven and nine knots for the early secondcentury Blackfriars ship excavated in London (see Chapter 3). Grainge (op. cit. 53-4), an experienced amateur sailor, considers that, in normal conditions of wind and weather, this type of vessel was unlikely to have achieved anything like this speed. Operating in the equivalent of a Beaufort Scale Force 4, he considers that the maximum speed on a beam reach (a sailing vessel’s position of maximum performance) was five knots, and that downwind it would have been even slower, of the order of three to four knots. As the “the speed of the convoy is the speed of the slowest ship” it is reasonable to assume that, when sailing in company, this would be the maximum speed attained.

Oneraria/actuaria/lembos/ceraerus/cybea For his first expedition to Britain, Caesar did not build transport ships, but requisitioned them in Gaul (BGall. 4.21); for his expedition the following year he had some 600 new transports built. He specified that they were to be beamier and of lower freeboard than those used the previous year, and were to ship oars, as well as sails (ibid. 5.1–2). For the naval expedition of Germanicus, from the Rhine to the Emms in 15, “1000 vessels were considered enough, and these were built at speed. Some were short craft with very little poop or prow, and broad-bellied, the more easily to withstand heavy seas: others had flat bottoms, enabling them to run aground without damage; while still more were fitted with rudders at each end, so as to head either way the moment the oarsmen reversed their stroke. Many had a deck-flooring to carry the military engines though they were equally useful for transporting horses or supplies” (Ann. 2.6). Although he points out that it is an assumption, Grainge (2002, 45–51) considers that the transports used in the invasion of 43 were similar in design and carrying capacity to the vessels used in Caesar's second expedition, and to those built for Germanicus. He estimates that some 60 infantrymen, or 10 cavalrymen and their horses could be

By using the number of ships (three) and the probable strength of the unit of rebellious Usipians (Ag.28) as a yardstick, Martin (1992, 9–12) has estimated the cargo carrying capability capacity of a modified liburna. Based on a complement made up (say) of 60 oarsman and 60 supernumeraries, in a cargo-carrying role, the supernumerary element could be replaced by a minimum disposable capacity of 5 tons. As reconstructed, a liburna has a length of 18 m, a beam of 3 m and a height amidships of 1.75 m; this is similar to the dimensions of the New Guy’s House boat. Whilst taking into account the respective roles of sea-going naval vessel and cargocarrying river craft, there is still sufficient correspondence to suggest that Martin’s estimate is soundly based. He suggests that by removing the lower

65

JAMES ELLIS JONES

  bank of oars a significant increase in capacity might be achieved. In a cargo-carrying role, the loss of maximum speed under oars would not be of great importance, the vessel being sailed whenever possible, but the assets of manoeuvrability and ability to progress in light airs would be retained. Using Martin’s figures as a basis, the reduction in the number of oarsmen might result in an additional cargo-carrying capacity of at least 2 tonnes.

never achieved his expected promotion to centurion, as he lost his life in a shipwreck.

Area of operations Starr (1989, 175) thought, “The activity of the German fleet at the mouths of the Rhine was intended not so much to protect Gallic coasting trade, as to guarantee the maritime connection of the German and the British armies”. It has recently been suggested “The classis Britannica must have been operating in strength as far north as the Solway Estuary and Tynemouth” (Mason 2003, 98), but Starr (1960, 175) considered that naval operations in the north were probably by local flotillas under independent commanders. He pointed out that the presence of the classis in the Irish Sea is suggested only during the wars of Agricola, and that no naval station on the west coast had been discovered. Fulford (1996, 22) was of the opinion that the lack of epigraphic evidence for the classis in the Irish Sea might suggest that the role of supply to the northern frontier was placed in the hands of legio II Augusta.

Figure 6.4 Tombstone of an optio who died in a shipwreck – naufragio perit (periit?)”. His name does not survive.

There is evidence for specialist soldiers/seamen at South Shields, on the mouth of the River Tyne; the fourthcentury garrison was a detachment of Tigris boatmen (numerus barcariorum Tigrisiensum), and Wilson (2002, 448) suggests that the Roman name Arbeia was possibly derived from the Aramaic word for Arabs (Arbaya). From the start of the second century, specialist units accustomed to dealing with the river conditions may have been stationed there. An undated inscription from Lancaster (RIB 601) refers to a numerus Barc(ariorum) under a p(rae)p(ositus) named Sabinus. A bronze discharge diploma dated AD 158, found on the foreshore at Ravenglass, records the presence of Cohors I Aelia Classica. The Cohors I Baetasiorum is recorded on an altar at Maryport on the Solway Firth in the late second century (RIB 830,837,838,842,843), and under the command of the comes litoris Saxonici at the Saxon Shore fort at Reculver during the third century (Notitia Dignitatum xxviii.18).

A classis Ivernica? Whilst accepting that there is no direct evidence for the existence of an independent western British fleet, Mason (2002, 65) draws attention to the fact that the volume of naval activity in the Irish Sea area must have been considerable, but that there are no bricks or tiles with the familiar “CL BR” stamp. He goes on to ask whether a tile found at Chester, bearing the stamp “CLIV”, was merely a batch number or, as Oceanus Ivernicus was the Roman name for the Irish Sea, whether it was an abbreviation for Cl[assis] Iv[ernica]. He further suggests that when, in the fourth century, so-called Saxon Shore type forts were constructed at Cardiff, Lancaster and Holyhead, Chester would have been “the obvious headquarters for such a separate naval command”, and that at this time “there could well have been a flotilla known as the classis Devae at Chester”. At first sight, these are attractive suggestions but, as the illustration of classis Britannica stamps shows, naval stamps were quite distinctive, with a period used as a separator, a distinctive surround or the use of a “lazy L”. It is therefore more probable that a stamp mark showing “CLIV” simply means 154 and that any other interpretation is unlikely. In any case, the legionary tile and brick production depot at Holt, just 12 km south of Chester, would have rendered any independent naval production both unnecessary and uneconomic. Furthermore, Chester is tidally dependent and does not lend itself to function as a major base, and its remoteness from the Bristol Channel (in a different legionary command) makes little sense in the area of command and control. It is far more likely that the army

Ulpian stated that “in the fleets all rowers and sailors are soldiers” (Digest 37, 13) and there is some slight epigraphic evidence to support the concept of naval forces, other than in the south east of Britain, operating under the direct control of the military commander, and as an integrated part of the land forces. Association of the legion with naval vessels is demonstrated at York, where there was an inscription dated c. AD 122, dedicated to Marcus Minucius Audens, described as gubernator (pilot) of VI Victrix (RIB I 653). Whilst not necessarily demonstrating membership of a naval unit, a memorial from Chester (RIB 544) shows that an optio of legio XX

66

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  In 1916, it was decided to introduce a cross-Channel service operated by vessels of the Royal Engineers and, perhaps ironically for any student of Romano-British history, a military port and depot was established at Richborough. By 1918, this had become a major seaport capable of handling 30,000 tons of cargo per week, with over 200 barges crossing the Channel to French ports (ibid. 24). In the war against Turkey, the scarcity of roads and railways meant that during the campaigns to capture Baghdad, water transport on the great rivers of the Tigris and Euphrates was the only practical means of moving large bodies of men and supplies. To a lesser extent, the Western Front was served by inland water transport, in a manner recalling the campaigns of Tiberius in the early first century AD, with vessels of a similar size, albeit with more modern methods of propulsion.

commanders at Caernarfon (or possibly Holyhead) and Cardiff exercised local control of the naval units that were stationed in the areas of legio XX and legio II respectively.

Figure 6.5 Tile stamps of the classis Britannica (after Peacock) Later British fleets under military control In Britain, the concept of a fleet under the control of the army was not confined to the Romano-British period as, from the Tudor period until the recent past; the military were in control of some form of naval activity. This offers some interesting comparisons with the Roman period as, until the Second World War, the “owned” vessels were crewed by civilians, with at other times reliance being placed on ships being requisitioned from civilian sources.

Figure 6.6 Troops being transported by barge on the Furnes-Dunkerque canal during August 1917 The Royal Fleet Auxiliary (again a civilian-manned operation, but under the overall direction of the Royal Navy), was constituted in 1905 and began by carrying coal and other stores, acquiring a tanker fleet when British warships became oil burning during World War I. The RFA saw service in every naval theatre of operations (from the Arctic to the Pacific) in the Second World War, including the Maltese, Russian and other convoys. During the Falklands Conflict in 1982, the RFA spearheaded logistic support for the Task Force, losing RFA Sir Galahad, with heavy casualties, to Argentinean air attack at Fitzroy (Villar 1984).

Under Henry VIII, vessels engaged in the transport of military stores and cargoes were administered by the Office of Ordnance, headed by a General and based at the Tower of London, which also served as the Royal Arsenal at that time (Habesch 2001, 1–8). However, a Royal Commission in 1597 found that “corrupt and fraudulent practices had become rife” and the Office was replaced by a new body entitled the Board of Ordnance. For the first 150 years, chartered vessels were employed by the Board, and the early surviving record of shipowning activity is of repairs to a vessel built in 1746. Because of general mismanagement and incompetence during the campaigns of the Crimean War of 1854–5, the Board of Ordnance was dissolved; its fleet, however, survived under a new name, the War Department Fleet. The Army Service Corps was founded in 1888, with a responsibility for supply and transport, including the War Department Fleet. A classic example of the use of rivers for the movement of large bodies of troops was provided by the Sudan campaign of 1896–8. Following a failed land-based attempt to relieve General Gordon at Khartoum, large numbers of river vessels, including all of Thomas Cook's Nile steamers, were used to ferry an Anglo-Egyptian Army in a punitive expedition that ended in the defeat of the Dervish army at Omdurman (ibid. 13– 24).

-o-o-o-o-o-o-oMilne (2000, 130) described the classis Britannica as “very much a Romano-British institution, rather than just another Roman fleet”, operating in a different manner to the fleets at Misenum and Ravenna, and to these might be added the classis Germanica and other (mainly) river based fleets such as those of Moesia and Pannonica. Throughout their histories, these fleets were actively concerned with the protection of the frontiers based on the Rhine and the Danube, and as such, their function was “naval” in the true sense of the word, rather than the cross-Channel transport of supplies, provisions and personnel, that became the major role of the classis Britannica in the post-Conquest period. However, in the conquest and consolidation phase, an important early role

67

JAMES ELLIS JONES

  was close support of the army, probably under the direct command of the local military commander.

   

It will have been observed that only brief reference has been made to the river-craft known to have operated on the rivers of the northern provinces. There is significant archaeological evidence, for example the vessels excavated at Mainz, Zwammerdam, Oberstim, and considerable epigraphic evidence, such as the many portrayals of ships on Trajan's Column. Whilst, in themselves, extremely interesting to the student of maritime archaeology, it is the writer's opinion that this type of craft is unlikely to have played a significant role on the waterways of Britain. Warships similar to the Mainz vessels were designed for the defence of a river frontier such as the Rhine or the Danube, and this role was never needed in Britannia. The cargo-carrying craft such as those at Zwammerdam (up to 35 metres in length) were designed to operate on European rivers (McGrail 2001, 201–7), and were much larger than those usable on British rivers. It is only necessary today to observe the size and quantity of barges presently operating on these rivers, and compare them with those still in use on the rivers of Britain, to note a much later parallel.

                     

It has been shown that there were periodic large-scale movements across the Channel, and that these were in addition to the routine tasks of the inter-provincial supply of food and materials, and the ferrying soldiers to and from duties in the Province. This lends support to Milne’s opinion (2000, 127-31) that the fleet may well have worked both for the Provincial governor, as was to be expected, but also for the Procuratorial office. He suggests that pre-Roman Britain had no indigenous fleet, and that it was therefore necessary for the classis Britannica to undertake this task, further suggesting that the fleet may have also been involved in the construction of the wooden quays at London, and the transport of stone for the city’s walls. Though there is no direct evidence for this hypothesis, it would result in an effective division of responsibility between the Governor and the Procurator.

       

                68

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  could be as low as 2 knots, but considers that this speed might be substantially increased if circumstances were favourable. Using Casson's average figure, a six-month grain supply for 40,000 men, could have been transported across the English Channel by thirty-five 60-ton ships, each making three round-trips over a 120-day period.

Chapter 7 Naval operations in Britannia It is not intended to discuss the naval operations of the Claudian invasion of 43, other than in passing, as this topic has been the subject of much academic debate elsewhere (e.g. Cunliffe 1968, 255–72; Peddie 1987; Hind 1989, 1–21; Black 1998, 306–7; Frere & Fulford 2001, 45–55; Manley 2002, Grainge 2005). At the time of the Claudian invasion, cross-Channel trade from Roman Gaul had taken place for almost 100 years and therefore there were pilots who knew of the southern rivers and estuaries, and who were familiar with the tidal conditions. For example, Diodorus Siculus (V, 22) comments that “A strange thing occurs around the nearby islands between Britain and Europe, for at high tide the causeways between them and the mainland are covered and they seem to be islands, but at low tide the sea recedes and leaves a large area high and dry so that they look like peninsulas”. Pliny wrote of the German Sea and the [Frisian] islands located within it (HN 4.33). Ptolemy recognised the distinction between the German Ocean (the present-day North Sea) and the Oceanus Duecaledonius to the north of Britain and the British Ocean to the South of Britain (Geog. 2.2.1, 2.10). The Straits of Dover were variously recorded as the freto Gallico (Solinus), freto Oceani (Tacitus) and the Oceanum fretalem (Ammianus Marcellinus).

Peddie (1987, 98–9) suggests that Roman military thinking would have been influenced by the fact that, when planning an advance towards the Thames, at least three rivers were navigable for considerable distances. The Lea was negotiable as far as Ware and, via the tributary Rib, as far as Braughing; the Roding was navigable to Chigwell and probably to Ongar; the Chelmer was open to Chelmsford and perhaps to Great Dunmow. Interestingly, he also comments on an opportunity to use the Mardyke, “although its appearance today belies the possibility’, pointing out that the river was used by shallow-draught vessels as late as 1870. Early naval operations in the Bristol Channel and the Irish Sea The presence of Roman naval forces in the Irish Sea is suggested by Tacitus when he states, “Its approaches and harbours are tolerably well known from our merchants who trade there” (Ag. 24). Ptolemy’s map of Ireland shares the cartographic errors of some of his other works, e.g. the 90 degree “skewing” of Scotland, but does at least indicate a Roman knowledge of the coasts, rivers and settlements of the island. Artefacts demonstrating established trading links during the Romano-British period have been recorded (e.g. Harbison 1988, 173–84).

It is generally accepted that the invasion force of 43, whether it landed at Richborough, Chichester or both, comprised four legions, some forty regiments of auxiliary infantry and cavalry, together with animal handlers and transport workers; a total of some 40,000 troops. The problems of supply for the soldiers, cavalry horses and probably some 10,000 transport animals were of great importance. It is sometimes suggested that the Roman Army “lived off the land”, but this is clearly unrealistic in this context. Firstly, the territory through which they would have passed would have been capable of only supporting the resident population, and the requirements of an army of some 40,000 soldiers were clearly unobtainable. Secondly, whilst some of the “natives were friendly”, other British tribes may well have emptied their granaries, and taken their herds of animals with them as they retreated, thus denying them to the advancing Romans.

A number of maritime features in the west of Scotland have known Roman names, i.e. Solway Firth Mull of Galloway Mull of Kintyre Isle of Bute Firth of Clyde Loch Ryan Irvine Bay Skye Mull Lewis Orkney Shetland /Iceland?

The invasion fleet comprised some 1,000 vessels and, after the initial landings, some of these would have been used to supply the army by sea, as it advanced northwards to the Catuvellaunian capital of Colchester, later to support the campaign of Vespasian along the south coast. A six-month's supply of grain for a Roman army of 40,000 soldiers weighed some 6250 tons, and could have been transported in approximately one hundred 60-ton shiploads. A supply fleet did not need to transport all the provisions for the entire campaign at one lift, as ships could make several round-trips in order to meet the demand and therefore, in logistical planning the speed of vessels is as important as their capacity. Casson (1995, 140–3) suggests that the average speed of a merchant ship

Ituna Aest. Novantarum Prom. Epidium Prom Botis Ins Clota Aest Rerigonius Sinus Vindogara Sinus Scetis Malaius Ins. Dumna Ins. Orcadis Ins Thule

suggesting that Roman sailors were also well aware of the opportunity for navigation in northern waters. The classis Britannica in the Bristol Channel? The presence of the classis Britannica in the Bristol Channel was given credence in the 1930’s, by the discovery of an inscription (RIB 2448.3) on the temple mosaic at Lydney, naming T. Flavius Senilis, as pr rel. This was expanded by Mommsen to pr(aepositus)

69

JAMES ELLIS JONES

  The only contemporary Roman accounts of naval/military combined operations are of the assaults on Anglesey, firstly probably by Suetonius Paulinus on the action in 60 (Ag. 14), secondly on Agricola’s campaign in 78 (ibid. 18). It seems that, in the first invasion, Suetonius had prepared a fleet of flat-bottomed vessels, and that the assault was executed by infantry, carried in boats, and by cavalry, swimming with their horses. It is possible that hide-covered shields, used inverted as a form of coracle, or inflated skins, were used as floats. Tacitus tells us that the plans for Agricola’s assault “had been hastily formed and as a consequence he had no ships available”. He committed a body of auxiliaries (probably the Batavians), “who knew the fords, and had that facility in swimming which belongs to their nation, and by means of which they can control simultaneously their own movements, their weapons, and their horses. He then launched them upon the enemy, so suddenly that the astonished islanders, who looked for fleets of ships upon the sea, promptly came to the conclusion that nothing was hard and nothing invincible to men who fought in this fashion”. The island was captured and a fort was built, but the Boudiccan revolt soon forced the abandonment of the conquered territory.

rel(iquationi classis) and interpreted as “officer-in-charge of a fleet supply depot” (Wheeler 1932, 103). Given the religious context of the site, the present view is that its most satisfactory rendition is pr(aepositus) rel(igionis), “director of the cult” (de la Bédoyère 1999, 144). Fulford (1996, 26–7) accepts this, but has argued that the dedication to the god Nodens by Flavius Blandinus, an armatura (weapon instructor), is an indication of a military unit at the site. Mason (2003, 187) supports this view and considers that “Senilis seems more likely to have been a senior naval officer attached to or even commanding the squadron guarding the Bristol Channel“. However, dedications at temples by military personnel are commonly found, and are not necessarily evidence of an adjacent military site. Perhaps the most compelling evidence against a naval presence at Lydney is not epigraphic, but tactical. Lydney is located well up the Severn Estuary in an area of extreme tidal range and strong currents, and is therefore in no position to have any influence over seaborne incursions. It is also poorly located to function as a supply depot (to supply what and to where?). Soldiers from the sea – amphibious operations in the west of Britannia

The problem of finding evidence for sites of amphibious landings is well demonstrated by the volume of writing on the place of landing of the invasion force of AD 43. This is not the place to discuss the arguments but, in essence there are two opposing “camps” with Cunliffe (1968, 255–72), Peddie (1987, 47–65), Frere & Fulford (2001, 45–55) and (Grainge 2002), being convinced that the major invasion force landed at Richborough, but with Hind (1989, 1–21), (Black 1998, 306–7) and Manley (2002) being equally sure that the landing took place in Chichester Harbour. In the case of the protagonists of Richborough, the archaeological evidence is limited to the remains of defensive banks and trenches and, in the case of Chichester, to some indication of a possible early supply depot. In neither case is there sufficient evidence for close dating to be established. As both these sites have been subjected to detailed archaeological investigation, it might be thought that evidence for this form of operation in the less well examined areas of the west might be impossible to uncover. However, it is suggested that some indication for amphibious operations might be revealed by a study of the distribution of Roman camps in Wales.

Jarrett (1994, 6) commented that “our land-bound experience restricts our thinking about early campaigns” and that “we easily neglect the possibility of combined operations”. Regrettably, Michael Jarrett’s premature death prevented his possible development of this theme, and the topic has not been extensively pursued by later writers. Manning (2000, 14; 28) mentions that the early fort at Cardiff probably served as a harbour for a fleet operating in the Severn Estuary, that later forts at Neath and Loughor served as minor naval bases, but does not discuss the question of amphibious operations. Davies has much detail on the land-based campaigns (some thirty pages), has only a paragraph on maritime communications, and makes no mention of operations in co-operation with the army (Arnold & Davies 2000, 35). However, this lack of coverage on possible seaborne operations in Roman Britain is not confined to campaigns in Wales for, aside from a mass of publications associated with the Caesarian and Claudian invasions, little of note has been published. A notable exception is the work of David Shotter (2004, 20; 2008, 105 and Figure 1),whose study of the incidence of pre-Flavian aes-coinage (especially Claudian copies) around the north-west’s river estuaries and coastline, led him to suggest that, in combined operations, sea-borne troops were landed at strategic points during campaigning in the early 60s. He suggests they probably sailed from a pre-fortress “fort” at or near Chester, via the Dee estuary, in order to link up with troops marching overland from their bases in the north Midlands and across the Cheshire Plain. He also considered that the Dee may previously have been used as a base from which to launch the invasion of Anglesey in 60 (Ann. 14, 29–30), as well as for embarking troops to serve in support of Cartimandua in the 50s and 60s. (2002, 25–38).

Roman Camps Fortified camps were built by the Roman army on the march, at each overnight stopping point. Vegetius writes “it was as if the army carried with it everywhere with it a walled town” (Mil. 1, 21). These structures, of which over 400 are known in Britain, have been traditionally known as “marching camps” (Jarrett 1969, 123–6, Jones & Mattingly 1990, 77–87; Frere & St Joseph 1983, 19–31; Peddie 1996, 59–79). The term “temporary camps” is sometimes used (e.g. Jones & Mattingly 1990, 77–88; Arnold & Davies 2000, 5–7), but Welfare and Swann

70

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  have been identified to the east of this line. To the west, permanent forts were established on or near to tidal estuaries, at Trawscoed, Pen-llwn, Erglodd, Pennal and Brithdir, and the apparent absence of camps leading to them is surprising as, during the Agricolan campaigns in Scotland, the average distance between camps was 17 km (Peddie 1987, 74). Because of the known line of a Roman road running westwards from Carmarthen, forts are also probable near Milford Haven and St David’s. It is unlikely that there was no military activity in the area bounded by the 200 km of coastline between Porthmadoc and Swansea.

(1998, 1) question how “temporary” any one defensive structure may have been. They point out that, without the most extensive and painstaking excavation, archaeology is unlikely to be able to distinguish between earthworks constructed for an overnight stop, those used for a longer period during a season’s campaign, or even from those sites which were regularly reoccupied, perhaps on an annual basis, in the course of prolonged operations covering several years. Some of the smaller camps may have functioned as construction camps for associated forts nearby. The defences consisted of a bank and ditch, usually in the familiar playing card shape, with protected entrances, usually on all four sides. Polybius (6, 27–34), Josephus (BJud. 3, 76) and Vegetius (Mil. 2.7) mention tents in describing the camps, saying that the interior of the camps mirrored that of permanent forts, in order that the soldiers were familiar with a standard type of layout. As it is not possible to discriminate between types of usage, the most widely used term, “marching camps”, will here be used. The size of the camps reflects the units on campaign; the two largest in Wales (Brampton Bryan and Blaen-cwmbach) covered c.25 ha. and were probably forming-up points, where the army assembled at the commencement of a campaign. Camps greater than 10 ha. indicate the separation of the assembled army into brigade group formations; sites ranging from 8.5 ha. down to under 2 ha. at the lower end, probably indicate a single unit of auxiliary troops and, at the higher end, the combination of legionaries, auxiliary cavalry and infantry, into battle group size. These camps are the primary indicators of Roman offensive operations and Jones and Mattingly (1990, 77–88) have used their locations to indicate the possible lines of advance during the campaigns leading to the conquest of Wales.

Figure 7.1 Marching camps in Wales

The problem of associating camps with specific campaigns has been mentioned above; camps in mid and south Wales could have been the result of military operations under any of the governors from the time of Ostorius Scapula (47–52) to Julius Frontinus (73–7). The situation in the area to the west and north is less confused, as Tacitus (Ag. 17) indicates that, aside from the assault on Anglesey by Agricola described above, Frontinus was responsible for the final conquest of Wales. When compared with northern England and Scotland, only a small number of camps have been found in Wales and the Marches (Arnold & Davies 2000, 5–7). It may be that this paucity of sites, particularly in west and north Wales, is the result of the conquest of the area having been achieved by amphibious operations, rather than by cross-country penetration over difficult terrain. A further probability is supply and replenishment by sea after the initial landings.

The writer is well aware that absence of evidence does not necessarily indicate evidence of absence, and that many authorities agree that one should not argue from absence. However, the “blank space on the map” (above Figure 7.1) suggests the absence of marching camps on the routes to the coast, and that the most obvious explanation is that amphibious operations were employed to establish the majority of the coastal forts. As might be expected, the mountainous areas of Wales were not densely populated and have few hillforts; therefore any military advance through these areas is unlikely to have met with serious resistance. A map showing the distribution of hillforts larger than 1.2 ha. (Jones and Mattingly 1990. Map 3:18), shows only two coastal hillforts between St David’s Head in the south and the Lleyn peninsula in the north, suggesting a lack of tribal organisation, indicating that the possibility of an opposed landing was slight. A decision to strike for the coast by sea, rather than by land, would therefore have been a choice based on military expediency, rather than the avoidance of conflict.

To the west of a line from Conway on the north coast to Neath on the south coast of Wales, only isolated camps occur in the north at Tomen-y-Mur, Derwydd-bach and Pen-y-gwryd. By contrast, between this line and a line drawn from Chester to Caerleon, twenty marching camps

71

JAMES ELLIS JONES

  timber and iron, and suitable locations for the construction of the required number of vessels. From here, the newly constructed ships could easily reach an assembly point in Cardiff Bay in the space of two tides. Prior to the Claudian invasion, the construction of large numbers of ships for a specific task is demonstrated by Caesar’s instructions (BGall. 3, 9–13) for reinforcements prior to the engagement with the Venetic fleet in 56 BC, his orders for construction of invasion craft in 55 and 54 BC (BGall. 4, 21; 5, 1–2), and the building of 1000 vessels for Germanicus’s expedition from the Rhine to the Emms in 15 (Ann. 2, 4). A military transport vessel of the period was probably capable of carrying either some 70 infantry soldiers or 10 cavalrymen and their horses (Grainge 2002, 45–51). An amphibious task force, comprising 2000 soldiers and 500 horses, would therefore have required some 30 troop transports and 35 horse transports. Peddie (1987, 40–1) has estimated the number of ships needed to meet the logistical needs of the Claudian invasion fleet (draft animals, carts, artillery and rations, including two weeks reserve of grain), and, using his figures as a basis, a first wave would have been accompanied by four cargo vessels. Warships, probably liburnae, at a ratio of one warship per five transports (Grainge op. cit. 51), could provide an escort force of six vessels. It is unlikely that the cavalry contingent would have accompanied the main amphibious task force, whose primary task would be to secure the landing area, establish a fortified camp and expand the bridgehead, with the cavalry arriving perhaps on the second or third day. On the arrival of supporting supply vessels, probably with pre-cut timber, and possibly with pre-fabricated towers and gates, a permanent fort could be rapidly constructed.

Figure 7.2 Alternative strategy for the conquest of west Wales by amphibious operations. The dark arrows, indicating advance by sea, point to either a known, or postulated position, for a fort. The provision of transport vessels could have been solved by re-using some of the vessels previously employed in the Claudian invasion. There can be little doubt that a considerable number of these were used in support of the Vespasianic advance in to the West Country, and that the subsequent conquest of the south-west peninsula would have drawn these to the west, at least as far as the great natural harbour at Falmouth. From there, with a possible overnight stop in Mounts Bay, close to either Newlyn or Penzance, and waiting for conditions of favourable wind and tide, a passage round Land’s End, to the shelter of St Ives Bay would have presented few problems. A number of writers (e.g. Allen & Fulford 1996, 257; McGrail (1985, 16); Peddie 1987, 155–7) have drawn attention to the supposed difficulties of rounding Land’s End but, having done this, in a small boat on a number of occasions, the writer is convinced that this so-called problem has been greatly exaggerated, and further discussion, together with supporting evidence, is later introduced. Whilst the north coasts of Cornwall and Devon present undoubted hazards to mariners, passage in easy stages, from haven to haven, presents few problems. After a coastal passage and a possible stay in Porlock Bay, Hurlestone Point would have provided a convenient departure point for the island of Steep Holm, with its probable Roman station, leading to a short run to the major pre-Flavian fort in Cardiff Bay.

The most likely setting for the postulated Welsh amphibious operations is during the governorship of Julius Frontinus (74–7) and, although a cavalry unit was almost destroyed in Ordivician territory in c.77/8 (Ag. 18), resistance was in its final throes. By contrast, the northern campaigns were into hostile territory, opposed by a well organised and, so far, undefeated enemy and therefore the likelihood of an opposed landing was much greater. The suggestion of the use of amphibious operations as a major factor in securing the west coast of Wales can only be considered as the first tentative steps in the development of a sustainable hypothesis. However, there is some encouragement to be derived from the investigation of similar type operations in the north of Britain, and the work on this topic of Colin Martin (1992, 1–34) is now briefly discussed. Naval/Military operations in the north of Britain Naval power fulfilled an important role during the initial period of military conquest and consolidation. Tacitus (Ag. 25) describes how, during the course of combined operations, the soldiers of the army and the sailors of the fleet often “matched the perilous depths of woods and ravines against the hazards of storms and waves, victories on land against the conquest of the ocean”. Referring to

If the re-use of the invasion ships was not the preferred option, the Forest of Dean provided a ready source of 72

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  been to set up advanced replenishment depots. These would be held for as long as it was necessary to support a mobile task force, operating without the encumbrance of heavy impedimenta. Because the size of baggage trains, and in particular, the need to provide fodder for the draught and pack animals is critical, it will be obvious that this was also a very desirable objective. Martin also considers that the punitive campaigns of Septimius Severus, and later Caracalla, between 208 and 211, were supported by the employment of water transport on a massive scale, enabling the deployment of overwhelming military force at any chosen point with despatch and economy.

Agricola’s campaigns in Scotland, he comments that “The war was pushed forward simultaneously by land and sea, the infantry, cavalry and marines often meeting in the same camps”, with the fleet being sent ahead “to plunder at various points and thus spread uncertainty and terror”. In this passage the term “marines”, translated from nauticus miles, should not be considered to have the same meaning as that of today’s Royal Marines; it is probably closer to the role of ship-borne infantry (Wintjes 2009, pers. comm.). The much later use of infantry in this role is evidenced by the Regimental Colour of the Worcestershire Regiment that bears the Battle Honour “1st June 1794” awarded to its predecessor, the 29th Foot, for its service under Lord Howe at the naval battle of the “Glorious First of June”, some 740 km west of Ushant. However, the conquest of Britain was achieved by military, not naval victories, whilst operating over extended lines of communication, particularly in the Highland Zones. The role of the fleet, in enabling the safe provision of supplies and equipment to the armies in the field was therefore of major importance; the logistics of military supply have been previously discussed in Chapter 4.

Establishing and provisioning this type of operation presents significant logistical problems, but similar events from later periods demonstrate that, provided replenishment by sea is available, an isolated outpost may be successfully implanted and maintained. For example, Edward I was content to hold and strengthen a relatively small number of strategically placed coastal castles inherited from his father Henry III, during the early period of his reign. However, North Wales was seen to present the greatest danger and, after finally crushing the uprisings of Llewellyn ap Gruffydd in 1284, he constructed a ring of massive coastal fortifications at Harlech, Caernarfon, Conway and Beaumaris. The garrisons of these castles were depleted by Edward’s expedition to Gascony and, in 1294, a new revolt broke out. Several castles were besieged, but, as long as provisions could be brought by ship, the castles were able to hold out, and the English control of the sea restricted the spread of the revolt. Further afield, during the American War of Independence, the Penobscot Expedition of 1779 took place when a force of British “redcoats” were sent to the unfinished Fort George at Majabigwaduce on the Penobscot River, with the harbour beneath the fort protected by three small sloops-of-war. The successful defence against a force of several regiments from the army of the State of Massachusetts and a fleet of 42 ships, half of which were warships, came to an end with the arrival of a squadron of the Royal Navy, resulting in the United States’ worst naval defeat prior to Pearl Harbour.

It is probable that Agricola’s experience of combined naval military operations was gained in north Wales. Tacitus records combined naval/military operations against Anglesey by Paulinus in 60 and by Agricola in 78 (Ag. 14. 18). The role of the fleet in reconnaissance is also indicated when Tacitus states that Agricola “used his fleet to reconnoitre the harbours” and that “The coast of the remotest sea was first rounded at this time by a Roman fleet which first established the fact that Britain was an island. At the same time it discovered and subdued the Orkney Islands, hitherto unknown” (Ag. 10). Appian also comments on reconnaissance when he states that it “was carried out on land by the cavalry and at sea by liburnians” (Hist. 5. 103).There are, however, clear differences between the campaigns in Wales and those in the north, in duration, scale and strategic/tactical considerations; these need to be given due consideration. It has been previously argued that the apparent absence of marching camps from the west of Wales supports the suggestion of amphibious landings, followed by the establishment of forts, and the conquest of territory by sea-power, rather than by cross-country penetration. The maritime operations in Scotland, were in support of a linear advance parallel to the coast, as opposed to landings intended to join up with existing seaward thrusts. The difference in scale is indicated by the size of marching camps, those in the Scotland indicating the massive force of a complete army, major elements of two or three legions together with a similar force of auxiliaries, whilst those in Wales usually indicate a legionary contingent, with accompanying auxiliaries.

In returning to the Roman period, I am grateful to Dr Birgitta Hoffmann for drawing my attention to the Periplus Ponti Euxini (Circumnavigation of the Fortunate Sea) written in 131, and addressed to the emperor Hadrian. The periplus was written by Lucius Flavius Arrianus (Arrian) soon after taking up his appointment as legatus Augusti of Cappodocia, a frontier province bordering the Black Sea in Asia Minor. The province had a standing garrison of two legions (legio XII Fulminatta and legio XV Appollinaris) and, when in 135, it was necessary to counter a threat posed by an invading force of the Alans, Arrian was able to assemble an expeditionary force comprising legio XV, a detachment from legio XII and the whole, or part, of twelve regiments of auxiliary cavalry and five regiments of auxiliary infantry. Apart from a much higher proportion of cavalry,

Colin Martin (1992, 1–34) has suggested that Agricola's ability to operate successfully far beyond his northernmost supply bases indicates a synchronised supply service by sea and that, when the fleet was sent ahead of the main army, one of the purposes may have

73

JAMES ELLIS JONES

  this is not dissimilar to some campaigning armies in the north of Britain. The classis Pontica was responsible for the whole of the Black Sea until the reign of Hadrian, when the operational role of the classis Moesica was expanded to cover the mouth of the Danube and the northern sector. Despite the reduction in the size of the area of operations, the status of the prefect of the Pontican fleet was raised to procuratores centenani (an annual salary of 100,000 sesterces) and, as this status was only shared with the fleets of Germany and Britain, is an indication of its importance. The fleet, numbering 40 warships and 3000 men (Josephus, BJud.II. 16.4), was originally based at Byzantium but, in 170, the headquarters were transferred to Cyzicus (Starr 1989, 77).

example, during the Severan operations in Scotland, but for the majority of the period the major function was the operation of seaborne replenishment of the Roman military, both men and materials and the conveyance of the cursus publicus. However, by the middle of the third century, sea raiding and piracy were endemic, particularly in the North Sea, and the primary naval role became patrolling to combat seaborne coastal raiders, and to protect merchant shipping from piracy. It would seem that Carausius, who was an experienced naval officer, had some success in intercepting sea raiders, for he was eventually accused of waiting for them to accomplish their mission, and then relieving them of their spoils for his personal benefit.

Arrian recorded a neighbouring Colchian tribe, the Sauni, writing, “To this day they are most warlike, and live at deadly feud with the people of Trapezus. They dwell in strongholds and have no king, and though long tributary to the Romans they are not practicable in paying their tribute, being engaged in robbery and pillage”. However, Arrian followed this with a statement that “Now, however, they shall pay regularly, or, with the help of God, we will root them out” and, once again, these comments are reminiscent of those concerning events in northern Britain. Two forts mentioned by Arrian in the periplus reappear in the Notitia Dignitatum of the early fifth century; the cohors Apuleia Civium Romanorum at Hyssi Portus, and the cohors I Claudia equitata at Sebastopolis. Sebastopolis was evacuated in 542 by the Romans and captured by the Sassanid; reoccupied and restored by Justinian I in 565; it remained a Byzantine stronghold until being sacked by the Arabs in 736. The Roman ability to establish and maintain long-term garrisons on the shores of a potentially hostile hinterland is therefore demonstrated not only in Britain, on the northwestern “edge of empire”, but similarly on its northeastern extremity.

Later naval and military operations on the English Channel and North Sea coasts. The ability of Rome to respond to affairs in the North Sea was diminished because civil war was endemic during the third and fourth centuries. For example, 60 emperors (of whom all but one died by violence) were recognised in Rome during the period between the murder of Severus in 235 and the accession of Diocletian in 284. During his nine-year reign, Gallienus (259–68) had to suppress the activities of no less than 18 usurpers, before being finally assassinated by his own officers (Goldsworthy 2007, 25– 32). Despite this, the shores of Britain seem to have been relatively free from the sea raiders during this period. There is evidence of the strengthening coastal defences, for example at Portchester in the first half of the fourth century (Cunliffe 1975, 223) and the construction of a new fort at Pevensey, in order to plug the long gap in the defences between the Dover Straits and Portchester (Johnson 1977, 141–3). However, the period from 275 to 350 demonstrates the villa system reaching its peak of prosperity and they were not fortified, as was the case on the Continent. It is possible that this was, in part, caused by perceived safety within the shores of Britain leading to an influx of wealthy landowners fleeing from Gaul (Salway, op. cit. 329).

Changing role of naval operations in Britain The demolition of the naval squadron headquarters at Dover took place c.215, and brought an end to a major establishment for the classis Britannica (but not to naval operations in Britain – see below). The last large-scale activity of a northern provincial fleet was probably that of the classis Germanica, in the campaigns of Alexander Severus and Maximinus in 235, and naval forts were abandoned by the middle of third century. There is evidence for warships (lusoriae), perhaps built by Aurelian or Probus, patrolling the Rhine c 280 and Starr considers that naval activity continued during the fourth century, not as organised fleets, but as smaller flotillas, each based on a single port and patrolling a small area (1960, 151;197).

Naval warfare in the modern sense was very rare. The interception of an enemy ship on the open sea was an unusual occurrence as and, because of the limited seakeeping ability of the vessels of the period, a system of regular offshore patrolling was, to all intents and purposes, not a practical proposition. It is probable that the Straits of Dover were the only point at which any significant degree of success could be achieved, the tactics being based on the interception of a raider on the way home, the coastal forts having been alerted by news of the raiders’ activities. It will be remembered that deliberately allowing raiders to penetrate the sea defences, and only intercepting them on their way home, was one of the charges that was made against Carausius. Whilst the other accusations laid against him may or may not be true, this part of the charge was probably unjust, as it was simply the way the system worked best (Haywood 1991, 38).

Starr (ibid. 153) comments that, after the time of Agricola’s governorship, “an almost unrelieved blackness hides any naval actions north of the English Channel for the next two centuries”. In fact, naval operations in support of the army had occurred periodically, for

74

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  capture 24 Roman galleys. A Roman fleet off the North Sea coast was attacked and the Canninefates won a significant victory, with most of the Roman vessels being sunk or captured. This was followed by an attack on the Rhine and further vessels, including a trireme serving as flagship of the classis, were captured. The final naval engagement of the rebellion took place when a Roman troop convoy, approaching the mouth of the Rhine from the Channel, was engaged by a rebel fleet, large in numbers but inferior to the Roman fleet in both the size of vessel and the ability to fight a disciplined naval fleet action, but there appear to have been a reluctance on both sides to force a close quarter engagement (Hist. IV, 16– 27). This indecisive battle is, however, of significance as Tacitus tells us that the Germans used cloaks to make improvised sails to improve both the appearance and performance of some of the captured Roman vessels; this is the earliest written evidence that exists for the use of sail by barbarians in Northern waters (ibid. 41).

However, the interception of an enemy fleet whilst at sea was an entirely different matter as, to his cost, Maximian found in 288 when attempting to recover the province of Britannia. Despite the face-saving propaganda of a failure attributed due to bad weather, his invasion fleet was forced to withdraw without landing, probably following a naval battle. The Manapian origin of Carausius led Haywood to suggest that “It would be pleasingly ironic if a Roman fleet approaching Britain from the Rhine was driven off by the Saxon shore defences manned, in part, by one-time Germanic pirates” (op. cit., 40). Maximian's failure to control the Channel coast of Gaul was a contributory factor to his defeat but, in 293, Constantius captured Boulogne following a short siege. The failure to hold Boulogne, and therefore control of the Channel, led to the assassination of Carausius and his replacement by Allectus. Under the direction of Constantius, the following three years were spent building up a fleet strong enough for the invasion of Britain. With the probable intention of decoying Allectus’s fleet away from the Straits of Dover, one section of the fleet, under Asclepiodotus, sailed cross-channel from the mouth of the Seine. This tactic was successful, as Allectus’s British fleet waited in the Solent, whilst Constantius set sail from Boulogne, crossed the Straits of Dover with the main invasion force and successfully recovered the capital city, as recorded on a commemorative medallion. Meanwhile, probably hampered by conditions of poor visibility, Allectus failed to intercept the invading fleet of Asclepiodotus, who landed unopposed and won a decisive battle against the usurper’s land forces (Salway 1991, 288–311).

A marine transgression (formerly called Dunkirk II) began in the early decades of the third century and brought about significant changes in the geomorphology of the lower Rhine and the North Sea coasts of Germany. A combination of eustatic and isostatic forces led to a rise in absolute sea level of between 1 and 2m above normal levels, but subsidence of the land combined with it to produce an effective sea-level rise of some 4.5 m around the Rhine estuary. This led to extensive flooding in the area now known as the Low Countries, with severe depopulation as a result of the ruination of the agricultural economy, followed by a consequent reduction in the Roman military presence from an area in which there was now little point in garrisoning. The fort at Vakkenburg covering the mouth of the old Rhine was abandoned c.240 and, by 250, the garrisons had been withdrawn from most of the forts on the Rhine between Nijmegen and the sea, with Aardensburg also being abandoned by the end of the century. Utrecht was garrisoned until the mid fourth century but, on its abandonment, the main Roman defence line in the region was withdrawn to a series of fortifications running from the English Channel at Boulogne, via Bavai and Tongaren, to the River Rhine at Cologne. The Frankish tribes quickly moved in to exploit this military vacuum and thus, for the first time, gained access to the open sea, whilst further north, the marine transgression led to the loss of reclaimed Saxon farm land and the consequent necessity to seek alternative “gainful employment”! Whilst some piracy had been endemic for several centuries, it has been suggested that is no coincidence that this period led to a dramatic increase in the scale and scope of North Sea raiding. Haywood (1991, 27–30) has drawn comparisons with the motivation for other Germans “to invade of the Roman Empire by land; the desire for loot and the tempting weakness of the Empire, coupled with the growing strength of the Germans, which together promised greater success”. Whilst Frisian raiders had ranged far and wide until the second century, there then appears to be a surprising hiatus in their activities for several centuries, but the reason for this may now be explained by the finding of hoards of military equipment,

Figure 7.3 Gold medallion depicting the guardian spirit of Londinium kneeling to welcome Constantius I as “restorer of eternal light” (reditor lucis aeternae). The sea raiders of the Germanic tribes Whilst our immediate interest is with matters of the third and fourth centuries, there is only very limited documentary evidence concerning naval activity at this time. It is therefore appropriate, once again, to refer to Tacitus as a possible source of information, albeit from a much earlier period, on the naval capabilities of the Germanic tribes. It seems likely that the Canninefates were the leading naval power amongst the barbarian allies of the revolt during the years 69 to 70 and their fleet was reinforced when a mutiny by Batavian oarsmen serving in the classis Germanica enabled the rebels to 75

JAMES ELLIS JONES

  Later naval and military operations in the west of Britain - the Irish raiders

from this period, on the east coasts of Denmark (Hoffman 2010, pers. com.). This suggests that invasion by sea from southern Sweden forced the Frisians to move from an offensive strategy centred on their western coast, to the need to adopt a defensive strategy on their eastern coastline.

The west of Britain faced barbarian raids from the Scotti (a name which seems to have meant “the plunderers”), and it is as raiders that the Scotti of Ireland are most frequently mentioned in classical sources. In 359/360, the magister equitum Lupicinus, was sent by Julian to take two units of Batavians (possibly auxilia palatina) and two numeri across the Channel during the height of winter, because “the wild tribes of Picts and Scots broke their undertaking to keep peace, laid waste to the country near the frontier, and caused alarm among the provincials, who were exhausted by the repeated disasters they had already suffered” (Amm.Marc. 20.1). The small number of troops, (possibly in the order of two thousand), cannot be seen as a response to a full-scale invasion, and were probably for the purpose of stiffening morale; fulfilling the traditional role of the field army in providing a high-quality strike force. Ammianus writes of “a concerted attack by barbarians” in 367/9 and states that the disturbances seemed to have been pre-planned and, as a consequence, very widespread. In his panegyric to Stilicho, Claudian writes that, about 398, “the Scots had raised all Ireland against me (Britain) and the sea foamed under hostile oars” (Stil. 2. 251).

The seagoing vessels of the Franks and Saxons The Bruges ship, a 15 m seagoing cargo vessel, has been carbon dated to AD180± 80 and, in contrast to the Nydam ship, was carvel (edge to edge) built and flat bottomed and therefore with many similarities in both construction and sailing capabilities to the Blackfriars I ship found in London and described in Chapter 3. The Hjortspring boat from Als in Denmark some 15 m in length and with a beam of c. 2 m demonstrates the ability of northern Europeans to build relatively sophisticated plank boats as early as the fourth century BC. Lightly built and probably paddled by a crew of 20, it would have been a fast flexible and seaworthy vessel. The mid fourth-century Nydam ship, also excavated from the River Als, was a flat keel, clinker-built oared longship, some 20 m long, perhaps having a mast and sail and capable of carrying a crew of 30 to 40 fighting men.

For the west of Britain, the results of these incursions are probably indicated by a number of significant finds in Ireland, dating from the late fourth/early fifth century. Several hoards of bullion have been found, one from a hillfort at Freestone Hill, County Kilkenny, containing coins of Constantine dating from 337, and a collection of silverware that had been deliberately cut into pieces of equal weight, appearing to be bullion silver. These finds were originally interpreted as barbarian “loot”, but the presence of silver ingots, most of them bearing an official stamp; suggest that they may be more plausibly seen as a Roman payment or subsidy, to keep the peace (Harbison 1988, 185). It is often supposed that the sole purpose of these raids were the enjoyable pastimes of “ravening or raving, pillaging and looting and, in other words, fun” (Reece 2002, 69), however, there is evidence that, on occasion, eventual settlement followed. By the end of the century, Scots from Ireland were settling in the Lleyn peninsula of north Wales, the Gower peninsula of south Wales and the peninsula of south-west England (Frere 1978, 406; Thomas 1972a 260–5; Wooding 1996, 37–8).

Figure 7.4 The Nydam No 2 oak boat exhibited at Flensburg in 1865 There has been much debate as to the date of adoption of sail in northern European waters (e.g. Hayward 17–21; 69–74). The statement by Tacitus that the ships of the Suiones from Scandinavia “used neither sails nor oars but were paddled instead” (Germania, 44), is often cited in support of the argument that sail was unknown in northern European waters until as late as the seventh century. However, the Tacitean passage quoted in an earlier paragraph concerning the use of sails by the Canninefates suggests that what may well have been the case in Baltic waters, need not necessarily be true in other areas. These had been exposed to the use of sails at a much earlier date, perhaps as early as the voyages of Pytheas in the fourth century BC, and it would be surprising if this influence did not impact on shipbuilding techniques. Wooding pointed out “ships are by definition mobile, so technical innovations may move from one region to another in the fabric of maritime architecture” (1996, 8) and McGrail (1995, 256) emphasised that the Massilliot Periplus describes voyages by hide boats from Brittany to Ireland taking two days and considers, if this were the case, the passage must have been made under sail.

Defence of the coastline of western Britain against seaborne raiders, was a near impossibility, whether attempted by land-based forces, or by ships of the local fleet. Starr, perhaps unfairly, suggested that patrolling the Welsh and Scottish coasts would be prevented “both by the traditional fear of the ocean and by the character of the warships, for the British fleet seems to have retained the Mediterranean trireme and liburna” (1960, 153). This comment on the supposed deficiencies of the vessels, and the morale of the sailors, is not consistent with his remarks on the following page, when he refers to the rounding of the north coast of Scotland by the Agricolan

76

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  These raids reached a climax in 367 with the Barbarica Conspiratio, a confederation of Picts, Scots, Attacotti, Saxons and the Franks. The onset of increasingly unstable and violent conditions in Britain is indicated by the fact that, of all the known Roman silver coin hoards that are datable between 388 and 410, 80% derive from Britain (Abdy 2002, 62; Laycock 2008, 138).

fleet (ibid. 154). The element of surprise gave a power out of all proportion to the probable numerical strength of the raiders and navigable rivers, such as the Severn, Wye and Lune, gave the opportunity for deep penetration. As a response, new forts were built at Cardiff and Lancaster, a small fortified harbour was built at Caer Gybi on Anglesey, together with a watchtower on Holyhead Mountain (Jarrett 1969, 135–37). There was limited reoccupation at both Neath and Loughor and these are probably an indication of locations for naval units, as well as the “secure base” role described above. The apparent absence of fortified naval bases on the English side of the Bristol Channel is perplexing, and it is surprising that one or other of the first/second-century fortlets at Old Burrow or Martinhoe was not re-activated.

Naval operations were restricted to the use of isolated detachments of naval vessels, based at some of the remaining coastal forts and facing a losing battle in attempting to protect a long and vulnerable coastline. However, defence of the shores of Britain against seaborne raiders, became a near impossibility, whether attempted by land-based forces, or by ships of the local fleet. There is some indication of small naval units endeavouring to protect vital centres of settlement from the threat of sea raiders, but it is clear that this activity was “too little and too late”.

Figure 7.5 Remains of the watch tower on Holyhead mountain -o-o-o-o-o-o-oRoman naval units, operationally responsible to the local military commander, played a significant part in the conquest and consolidation of Britain; Martin's (1992, 20–21; 25–29) description of the employment of water transport on a massive scale, thus delivering dominant military forces to any location “with dispatch and economy” is particularly relevant. Supply and provisioning would have been an area of major concern; the importance of water transport in the logistical systems employed by the Roman army has already been discussed. The requirement for transport by water of bulk cargoes of grain and full shiploads of amphorae, containing olive oil and wine continued well into the third century, but a decline in the military population, coupled with changing tastes, led to a reduction in the volume of shipping. It is clear that by the early 360s the coastal defences were finding increasing difficulty in coping with the coastal raiders and, according to Ammianus Marcellinus, the raids were continuous (XXVI, 4.5). The appearance on the western coasts of Scottish of pirates operating from Ireland and the south-west Scotland, were perhaps accompanied by those of Pictish raiders on the east coast. 77

JAMES ELLIS JONES

  by a U-shaped rock-cut ditch. Construction seems never to have been completed, with a wall having collapsed directly onto builders’ levels, and no evidence of any form of occupation debris. Pottery and coins suggest a date of construction in the late third century, at a time of considerable activity on the coast of south Wales, with a Saxon Shore type fort being erected at Cardiff and evidence of re-occupation of the forts at Loughor and Neath. The excavators (Evans, Dodwell & Thomas 1985, 57–125) considered that the building may have been intended as a mansio, serving a crossing point from the estuary of the River Parrett on the Somerset coast, connecting west and mid-Wales directly to Ilchester, Dorchester and southwest England and therefore best described in modern terms as a “ferry terminal”. However, there is no known Roman road leading from the site, and the fort at Cardiff, only 10 km further to the west, seems a more obvious landing place for crossChannel traffic.

Chapter 8 Harbours, ports landing places By considering the Roman road system and its association with known coastal and riverside settlements, Henry Cleere (1978, 36–40) identified the location of 45 possible harbours in Britain There undoubtedly were many smaller landing places, for example, a RomanoBritish pottery assemblage from Magor Pill (where, in a creek where a small Welsh port called Abergwaitha was once located, a 13th-century boat has recently been discovered) indicates that it functioned as a small landing place from the late first/second century. Place name evidence might suggest that, on the opposite shore of the Bristol Channel, the small port of Porlock Weir, sheltered by hills on either side, and shown in the Anglo-Saxon Chronicle as Portloca, could, during Roman times, have provided a welcome haven on the otherwise inhospitable Somerset coast. This section concludes with a discussion of the problems associated with problems of locating archaeological evidence for Roman waterside structures and is well illustrated by a report in British Archaeology (1999, Issue 50), confidently stating that “A man-made Roman harbour linked to the River Severn has been discovered at the Roman town of Wroxeter in Shropshire”. Despite efforts to recover structural remains, six years later it was necessary to state that “No evidence has yet been forthcoming about the harbour facilities within the town, which must have existed given the level of imports evidenced by artefacts” (White & Dalwood 2005, 18). Similarly, the fort of Arbeia, lying at the mouth of the river Tyne, four miles east of the terminus of Hadrian’s Wall at Wallsend, is the most extensively excavated Roman military supply-base in the Roman empire but as Neil Hodgson points out “There remains the difficulty that no Roman port at South Shields has been physically located” (pers. comm.). However, the preliminary (as yet unpublished) results from the 2009/11 excavations at Caerleon by the University of Cardiff demonstrate, there is much still to be discovered.

Figure 8.1 The pebble beach at Cold Knap at low tide The natural harbour at Barry is one of the few places on the Bristol Channel where a vessel may lie afloat at all states of the tide and, on the English shore, the lee of Brean Down off Sprat Beach, provides a similar, if only limited, advantage. It might therefore be suggested that a naval presence might account for the building at Cold Knap, enabling some form of patrolling of the 15 km stretch of water, hopefully intercepting sea raiders en route to pillage the rich villa area on the southern side of the Severn and its estuary.

Whilst it is certainly true that no large-scale excavations such as the Portus Project on the Tiber (www.portusproject.org) or at Myos Hormos on the Red Sea (Peacock & Blue 2006) have been undertaken, there is sufficient evidence to demonstrate the scale and extent of harbour, port and landing place development in Roman Britain. A number of sites are therefore discussed, having been chosen for their diversity, ranging in size and archaeological content from the substantial remains of quayside structures at Londinium and the pharos at Dover, to a small group of dedicatory coins found at a Severn Estuary ferry-crossing.

Caerleon Excavations at Caerleon (Isca, directed by Boon in 1963, revealed a well-built stone and timber quay, lying 230m from the present line of the riverbank, at a height of 6.56 m above Ordnance Datum. It is extremely unlikely that it was the only quay, since it was not constructed until the early third century, and there must been have port facilities from the time of the foundation of the fortress. Based on data obtained from the excavation, there have been a number of attempts to ascertain the level of tidal heights during the Roman period (e.g. Boon 1978, 24– 36).

Barry Located on the north shore of the Bristol Channel, close to a small natural harbour, the pebble beach at Cold Knap, near Barry, has a steeply shelving rocky bank, which then flattens out to sand towards the bottom end of the tidal range. A Mediterranean courtyard-type building with twenty-one rooms and with corridors arranged on all four sides of a central rectangular courtyard was enclosed 78

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  what were believed to have been vertical posts of the Roman bridge or landing stage; in 1962 some pieces of wood were recovered and lodged in Chepstow Museum, but they had not been scientifically tested to determine their date. However, during the production of an Extreme Archaeology TV program in 2003, samples were obtained from the piles and were later radiocarbon dated giving a late first/early second century date. Chester Most publications on Chester (Deva), and many on Roman Britain in general, have featured photographs of a section of the well preserved “quay wall”, alongside the racecourse on the Roodee. Recent research attempting to determine changes in sea level since the Roman period has included the study of probable tidal levels at Chester (Waddelove & Waddelove 1990, 253–66; Ward 1996, 4– 11). Mason (2002, 59) has therefore been led to reconsider this structure and he persuasively argues that the height of the so-called quayside would be some 5 m higher than the deck level of a vessel tied up alongside and, quite correctly, points out that this “is clearly a nonsense”. He conjectures that this structure was a wall, not a quay, and was intended to protect the western section of the canabae, the wealthiest area of the extramural settlement.

Figure 8.2 Third-century quay at Caerleon during excavation. The magnitude of the problem of cargo handling at a legionary fortress should not be under-estimated. In addition to the 80 consignments of grain (concentrated in the months from August to October), the requirement for wine for the fortress itself (90 shiploads), possible transhipment of olive oil, and the provisioning of military stores and equipment to the subsidiary forts, suggests at least one arrival and departure of a major vessel each day of the sailing season. This indicates a need for more extensive port facilities than had, so far, been discovered but, in 2010, geophysical survey and excavation by the University of Cardiff revealed a large stores/supply complex lying between the amphitheatre and the River Usk.

Figure 8.4 Section of the Roman “quay wall” at Chester. Excavations in 1885 revealed an ancient riverbed at a depth of about 6 m below present ground level, where lengths of oak timbers set in concrete were found (Shrubsole 1887, 80). The timbers were some 3 m in length, averaging 0.30 m in diameter, having a point at the embedded end, encased in an iron sheath, enabling them more easily to be driven into the riverbed. Roman material was found, including bricks, tiles, samian ware, other types of pottery and, most importantly, an ingot of lead bearing a date manufacture of 74. Iron sheathed timber was commonly used in the construction of bridges, wharves, and jetties during the Roman period and Mason (ibid. 64–72) suggests a landing stage projecting into the deepest part of the river channel, allowing ships to tie up alongside at most stages of the tide.

Figure 8.3 Quay at Caerleon adjacent to the site of the legionary fortress. Chepstow Chepstow is located on the west bank of the River Wye, some 4 km from its confluence with the Severn estuary. Lying on the coastal road from Gloucester to Caerleon, a road to Monmouth runs northwards. Early material from Chepstow may indicate a pre-Flavian fort; first-century coins, pottery, metal horse-trappings and early burials indicate a fort guarding this key crossing of the Wye. There is nothing to be seen of the (presumed) Roman bridge at Chepstow, but from time to time, in exceptionally low water, some timbers have been reported in this locality. In 1911, excavations exposed 79

JAMES ELLIS JONES

  A pharos (lighthouse) was built on each of the headlands overlooking the harbour, and the remains of that on the Eastern Heights still display some 13 m of Roman construction. A protected anchorage at the mouth of the River Dour was formed by the construction of a massive breakwater; the remains of 30 m of timber framing, infilled with shingle being found in 1855, some 230 m to the east of the fort; timber piles, mooring rings and groynes were found in 1860. A probable quay and timber jetty to the west of the inlet were found in 1956, and a structure located in 1974, closer to the mouth of the inlet, has been interpreted as a continuation of the harbourside. Gloucester Hurst (1999, 123) comments that “Despite the fact that a river-side setting is self-evidently the major factor explaining the siting of the fortress and colonia (and indeed all the habitation areas of Roman Gloucester), the study of the Roman and early medieval waterfronts must be counted as a failure of the last 30 years, reflecting the piecemeal response of the city’s archaeology to modern redevelopments”. He points out that opportunities for major investigation have not been lacking, but that only small-scale excavations have been carried out and that most information comes from observation of contractors’ trenches and boreholes.

Figure 8.5 Drawing (1885) of iron shoes of wooden piles.  Dover (Dubris) The construction of the headquarters for the British squadron of the classis Britannica was commenced in 116, but was left unfinished, probably because of the demands of the construction of Hadrian’s Wall. Work on the construction of a second fort, on the levelled remains of the previous structure, began c. 130 on a similar design to that of a normal auxiliary fort of the period. Rebuilt in 160, but again abandoned in 180, it underwent a final phase of refurbishment and re-occupation early in the third century; then became disused and probably demolished.

Figure 8.7 Artists impression of the colonia at Gloucester

Figure 8.6

The course of the River Severn (Sabrina Fluvis) changed during, and since, the Roman period and part of a quay, composed of massive masonry, has been found on a line set back from the present course of the river. In a creek off the river, there is evidence of a late first-century harbour, with timber wharves or landing stages. At Quay Street, limestone rubble had been packed on top of timber piles, probably as a foundation wall to protect the reclaimed ground. The main road leading from the west gate ran parallel, and a number of wooden drains hollowed from tree trunks, were nearby. The area was reclaimed with dumps of rubble, and a new quay wall constructed in the third century. This area presumably contained warehouses, but one building containing a

Remains of the Roman lighthouse at Dover 80

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  and possibly as far as Cambridge (Duroliponte), and connected with the navigable Rivers Welland and Trent. The extent to which the canals were used for navigation, or only drainage, is still a matter of debate. The riverbank was clearly a commercial focus of the city, with a possible dock to the east of the walled town and waterfront buildings extending for at least a kilometre. Jones (1999, 108) considers that substantial quays were not needed, as Lincoln was situated above the normal tidal level, and that wooden jetties, similar to that found at Chester, would have been adequate for the loading and unloading of sea-going vessels. The process of land reclamation included the dumping of waste material; painstaking excavation has led to the recovery of the remains of table and drinking vessels (including a high proportion of samian ware), jewellery, glass, coins, fragments of weapons, and armour and even leather sandals. A wooden writing tablet and nineteen styli, together with a copper-alloy balance might be also considered to be evidence of waterfront commercial activity and overseas trade is indicated by the dedicatory inscription of M. Aurelius Lunaris, a wine merchant from Bordeaux (Jones 2002, Figs. 25, 64, 10).

mosaic and a column portico indicates some residential occupation. During the third and fourth centuries, a number of buildings were deliberately demolished, with their sites remaining vacant (Wacher 1995, 150–167). Heronbridge The settlement site at Heronbridge stands on the west bank of the River Dee 2 km south of Chester, straddling the road running from the legionary fortress towards Whitchurch. Occupation commenced c. 90 and continued down to at least the middle of the fourth century, as demonstrated by coins ranging from a denarius of Titus (80) to an as of Constans or Constantius (330-360). There were numerous stone buildings, including some with hypocausts, as well as at least one shrine or temple, but the majority of buildings were strip houses set at right angles to the main road. There was metalworking and grain processing; there is evidence that building work and street repairs continued at Heronbridge into the early fourth century. An existing streambed was deepened in the second century to enable construction of a ramp leading down to the edge of an inlet from the Dee, this having been artificially straitened. There were several rock-cut pits and other features that were probably mooring posts to enable barges to be tied-up prior to unloading or loading.

Littlecote During the excavation of the extensive first/fourthcentury villa complex at Littlecote, near Ramsbury, two water-filled dykes, cut at right angles to the River Kennet, were identified. Bryn Waters (2001, 137–8) considers that these are “boat-channels” used by shallow-draught barges during loading or off-loading of produce.

Figure 8.8 Silted-up inlet with rock outcrop Ilchester (Lindinis) Thew (1994, 134) has found geo-archaeological evidence for the canalisation of the River Yeo at Ilchester (Lindinis) during the Roman period. Roman stone quays have been claimed at Ilchester but whilst accepting evidence for canalisation, Leach (2000, pers. comm.) considers a medieval date for the quays to be the most likely.  

Figure 8.9 Boat channels at Littlecote Roman villa Waters sees this as evidence for the utilisation of water transport directly onto the River Kennet and cites similar Avon riverside locations of the villas at Manningford Bruce, Netheravon, Amesbury, Downton, East Grimstead and West Dean. He contends, “Where a known villa or substantial building is located adjacent, or very close to, a river or stream, it is safe to assume that it was located on that site for reasons connected with the transport of produce downstream”. This may have resonance with the medieval period when both religious and lay estates often had landing places to dispatch goods to market place or receive incoming produce. At Waltham Abbey in Essex,

Lincoln (Lindum) The colonia at Lincoln (Lindum) lies on the navigable River Witham, some 60 km from its entry into The Wash and, like Gloucester, developed from a legionary fortress. During the Roman period, Lincoln lay close to the head of the estuary of the Witham and it is probable that the present course of the river below Lincoln was subject to some Roman modification. From Lincoln the Fosse Dyke ran in the direction of Littleborough (Segelocum) and the Car Dyke ran south towards Water Newton (Durobrivae) 81

JAMES ELLIS JONES

  (1982, 162) claimed “Londinium was a large and busy port, probably handing more tonnage than any other port in Roman Europe”. A more balanced view is probably that of Gustav Milne who considered that London played an important role in the distributive network of the province, acting as a market centre for its hinterland, as did the other major towns: “London was a, not the, Romano-British port.” (1985, 187).

a wharf on the River Lea provided access to barns within the grange adjoining the Abbey. The moat at East Hattersley in Yorkshire was enlarged to take boats and was linked to the River Aire and the moated site at Stretham in Sussex had a waterfront onto the River Adur (Gardiner 2007, 90). London The most extensive information of a Romano-British waterfront has come from the River Thames in London (Londinium), where a combination of extensive redevelopment, the provisions of PPG 16, and the activities of the Museum of London Archaeology Service, have led to a series of developer-funded excavations. These have revealed considerable details of the foundation of the port in the late first century, its development in the second and third centuries, and its decline in the fourth century when “after almost two centuries of energetic expansion, Roman London shrank to a small community mainly on the waterfront” (Milne 1985, 22–33).

Plymouth Sound At least eight hoards, and over 50 separate coin finds have been made at different points overlooking Plymouth Sound. Significant quantities of building materials, including Roman tiles, have been recovered from the foreshore of Sutton Pool, and other Roman material has been found at various locations on the banks of both the rivers Tavy and Tamar. Plymouth Sound provides an outstanding natural harbour and, taken in conjunction with the information of the Iron Age entrepôt on the Mount Batten Promontory and the coin evidence dating from the first to the fourth centuries, it is highly probable that a significant Roman harbour existed.

At Pudding Lane, some 75 m to the west of the present London Bridge, the remains of a massive timber framework, were excavated. This was built c. 80 out into the river, presumably supporting a planked platform, and has been interpreted as a landing stage, allowing the transfer of cargoes from ship to shore. It has been suggested that the structure was pre-fabricated, as it had to be erected on an open foreshore between tides. Within a few years, this landing stage was dismantled, as part of a planned development incorporating the construction of quays, terraces and warehouses. The replacement was a timber-faced quay of massive construction, using timbers of up to 660 x 400 mm in cross section, in-filled with dumped deposits and levelled off with a surface of brickearth and gravel. Development along the London waterfront continued until the mid-third century, with each successive quay being laid at a lower level than its predecessor, this indication of a fall in the river level was discussed in Chapter 1.

Sudbrook Metalwork and pre-Flavian pottery indicate a Roman garrison was installed within the ramparts of an Iron Age multivallate promontory fort, presumably to control a ferry crossing of the Severn Estuary, from the English side to Portskewett. A road runs from Sudbrook to the main Caerwent road at Crick and probably accords with a decision to ship some supplies across the Severn Estuary, rather than by the longer and circuitous route via Gloucester. “Coins are frequently picked up in the mud upon the rock or landing place” (Harris 1793, 2), indicating dedicatory offerings after a safe crossing .and cover 300 years of Roman occupation, ranging from the Emperor Claudius to one of Gratian struck between 367 and 375,

Figure 8.11 Dedicatory coins from ferry crossing Absence of archaeological evidence

Figure 8.10 Blackfriars type vessel beached to unload at a third-century London waterfront.

Having considered some examples of observable structures and artifacts, the opposite position – that of the absence of significant archaeological remains must now be considered. Jones and Mattingly (1990, 198) wrote, “With the notable exception of London, comparatively

The information obtained from the extensive program of excavation may have led to an over-emphasis of the importance of London as a port. For example, Morris 82

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  few traces of Roman harbours and quays have been properly examined in Britain. This is in part a reflection on the vulnerability of harbour installations to destruction as a result of coastal or river change, or as a result of continued use of natural harbour sites down to the present day. Yet an island province like Britain must have depended heavily on its sea communications with the Continent and a large number of harbours are unknown, or can be inferred to have existed”. This opinion is echoed by Hurst (1999, 123) who commented that “Despite the fact that a riverside setting is self-evidently the major factor explaining the siting of the fortress and colonia (and indeed all the habitation areas of Roman Gloucester), the study of the Roman and early medieval waterfronts must be counted as a failure”. There is epigraphic evidence of overseas trade at York (Eboracum) during the late second and early third centuries with an inscription referring to a sea-going merchant, L.Vidicius Placidus, a negotiator from Rouen (Rotomagus) in northern Gaul and a reference on an inscription in Bordeaux (Burdigala) to a wine-merchant, M. Aurelius Lunaris (see Lincoln below). However, despite considerable efforts to establish the location of the quays, their whereabouts is not yet known and it is only possible to suggest that there were “major waterfront structures somewhere on the banks of the River Ouse, where ships plying coastal and overseas trade could load and unload” (Ottaway 1993, 85).

by canalisation in more recent times. There is considerable evidence of overseas trade at York during the late second and early third centuries, but despite considerable efforts to establish the location of the quays, their whereabouts is not yet known. Ottaway (1993, 85) is able only to suggest that there were major waterfront structures “somewhere on the banks of the River Ouse”, where ships plying coastal and overseas trade could load and unload. That the lack of archaeological evidence is not confined to the Roman period is demonstrated by the recent survey of the Welsh coast, designed to record the archaeology situated within the coastal zone. The authors’ comment that “there is virtually no direct evidence of maritime trade in the archaeological record: apart from the dock incorporated into Beaumaris Castle Anglesey, there are no surviving quays, wharves or jetties of medieval date” (Davidson 2002, 61). However, quays, wharves or jetties were often unnecessary, and a simple gangplank combined with manual labour, provided a method of loading or unloading. Figure 8.12 illustrates this technique being used for loading sacks of wheat to a codicaria at Ostia for shipment up to Rome.

The apparent absence of archaeological evidence for waterside structures of the Romano-British period is also well demonstrated in the area of the Bristol Avon. The river was a busy Roman waterway, probably with quays at Bath, equipped with cranes for loading large blocks of Bath Stone; a landing place serving the villa at Keynsham and a significant port at Sea Mills. In addition, the recent discovery of a large villa complex near to Bradford on Avon (Corney 2002) and the presence of a medieval quay near to the town bridge indicate navigability of the river to that point. The remains of Roman riverside structures are unlikely to have survived the development of quays at Bath, serving the Kennet and Avon Canal, the construction of wharves for Fry’s chocolate factory at Keynsham and the development of a “wet dock” at Sea Mills. The destructive power of even a non-tidal river is well demonstrated at Corbridge where the remains of the southern part of the Roman bridge (which at one time consisted of 11 stone arches, standing to a height of perhaps as much as 9 m above the river) now lie on the bed of the river. Erosion threatened to destroy the road ramp leading to the bridge, and to prevent further destruction, over 300 blocks, some of them weighing more than a tonne, were removed and have now been reassembled a short distance from the present riverbank (Hodgson 2009, 101).

Figure 8.12 Fresco in the Vatican Museum

At York, where the fortress of Eboracum is situated at the junction of two rivers, the Ouse and the Foss, the course of the rivers in Roman times is not altogether certain (Briden 1997, 165), the Ouse probably has not changed a great deal, but the Foss was substantially altered by the creation of the King’s Fish Pool in the 11th century, and

Figure 8.13 Loading quarried limestone in the Avon Gorge (note gangplank to shore) 83

JAMES ELLIS JONES

  The use of a gang-plank to the shore leaves no trace in the archaeological record, similarly in the case of beaching which, to avoid stranding, is achieved by running a vessel aground in shallow water at half-tide.

Figure 8.16 The Old Passage Ferry at Aust in 1965

Figure 8.14 The trow “Palace” beached to unload stone to be used in construction of sea defences the Severn Estuary A few miles upstream, Allen (2000, 132–4) investigated the remains of a jetty, dating from the late 19th century, at Cone Pill, Woolaston. Protected by a previous sequence of deposition, the recent erosion of the riverbank has exposed it to the elements, and it is rapidly disintegrating, with only a few stumps remaining visible in 2007. There is no evidence of a road leading to the jetty, built for the convenience of one or two local landowners, as was the probable case for many riverside settlements in the Roman period.

Figure 8.17 The ferry stage remains in 2009 There are significant differences between the characteristics of the western and eastern coasts of Britannia, not only in geomorphology but for this examination, more importantly, the density of military installations. In the north of Roman Britain, harbours, ports, landing places are invariably associated with military sites. According to the fifth edition of the Ordnance Survey Map of Roman Britain, with the exception of the Yorkshire signal stations, Lease Rigg and the Saxon shore forts, no Roman fort is shown either on, or within, 20 km of the coast between Hadrian's Wall and Dover. It is probable that the forts also existed at Whitby, Bridlington, Skegness and Huntstanton, but have been lost to coastal erosion. On the west coast, between Hadrian's Wall and the Bristol Channel, there were forts at Carmarthen, Caerleon, Sudbrook, Cardiff, Neath, Loughor, Trawscoed, Pen Llwyn, Pennal, Brithdir, Tomen-y-Mur, Caer Gybi, Caernarfon, Caerhun,

Figure 8.15 Remains (c. 1998) of a jetty at Cone Pill, Woolaston More substantial structures, such as wooden landing stages, also have a limited life, for example, the Severn Estuary narrows dramatically between Aust and Beachley and, until the opening of the first Severn Bridge in 1966, England and Wales were linked by a vehicle and pedestrian ferry across the “Old Passage”. In a little over 40 years or so, the Aust terminal landing stage and slipway have deteriorated to the remains shown in Figure 8.16. John Allen observed, “Tide, wave and weather continued to damage and erode these structures, the outlook for which is bleak” (2002, 53–64). 84

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Aestuarium), the River Clyde gave access by seagoing ships as far as Old Kilpatrick, with possible small barge access further upriver to the isolated fort at Bothwellhaugh. The River Dee (Deva) in Galloway was navigable from Kirkcudbright Bay to Castle Douglas and the fort at Glenlochar; the River Nith gave river access to Dumfries and the line of forts on Nithdale. The Esk may have been navigable in the direction of Netherby and the Eden as far as Carlisle. Each of these waterways had its own network of landing places, but archaeological evidence for their existence is difficult, or in most cases impossible, to establish.

Kirkham, Lancaster, Ravenglass, Moresby, Burrow Walls, Maryport, Beckfoot, Kirkbride, Bowness, Drumburgh, Papcastle and Burgh-by-Sands. It is probable that forts also existed in the area of Fleetwood, in Lancashire, and St David’s, in Pembrokeshire, but have yet to be located. Because of the processes of erosion and deposition, coastal structures are less visible in the archaeological record than those inland, but sufficient remains survive to indicate the scale of achievement. Geomorphological changes not sufficient to mask the features of the maritime landscape of the Roman period and one must agree with John Blair (2007, Pref.), who suggested that “The landscape contains numerous man-made and modified watercourses which are yet to be explored by the readily available tools of excavation, geomorphology, sedimentology, and radio carbon dating. When more of this work has been done, we hope that the subject will not be seen as a blind load, but as a useful tributary to the mainstream of medieval economic history”. In certain cases, the changes have taken place in recent times, for example, the loss to the sea of Milefortlet Two of the Solway Frontier and part of the cliff-top fortlet at Martinhoe on the north Devon coast. There is, however, archaeological evidence for change and response as several places that flourished as ports during the Romano-British period have now no maritime function, but it is unwise to attribute this solely to geomorphological change.

It is quite clear that the Roman period saw dramatic peaks and troughs in the utilisation of the British rivers and coasts. For example, the Flavian constructions and the Severan reconstructions are at one extreme, and the virtual absence of the legions during the long campaigns in the north, are at the other. A combination of sea and inland waterway transport enabled goods to be moved, relatively inexpensively, over long distances. The evidence from historical periods demonstrates the extent to which the coasts and rivers of Britain could be used for the movement of large and varied volumes of traffic.

The Tacitean comment that the sea “passes deep inland and winds about, pushing in among the highlands and mountains, as if in its own domain” (Ag. 10.43) is true of the rivers of the south where, for example, the River Severn (Sabrina Fluvis) and its tributaries, served the fortress at Kingsholm, the fortress, later colonia, at Gloucester; the fortress, later civitas capital, at Wroxeter; the forts at Caersws, Forden Gaer, Abertanant, Pentrehyling and Wall Town; the small industrial towns of Worcester and Droitwich, and numerous villas and small settlements close to its banks. Other southern waterways were equally suitable for water transport; the Yorkshire Ouse was tidal to York and was accessible to sea-going ships and the Fosse Dyke linked the Trent at Torksey with Lincoln. The Great Ouse was navigable from King's Lynn to St Ives, the Cam is known to have been navigable to Cambridge and the Ouse was linked with the Nene at Wisbech. The Nene was navigable to Peterborough and, from London; the Thames was navigable by large sailing barges as far as Burcot. With the exception of the River Tay, northern Britain was less well provided with navigable rivers that penetrated far inland. The Tay (Tava) was navigable as far as the fort at Bertha and, with some improvement, might have been made navigable as far as the short-lived fortress at Inchtuthil; its tributary the Isla provided access to Cargill and possibly to Cardean. The Firth of Forth (Bodotria Aestuarium) provided a number of harbours and/or landing places at Inveresk, Cramond and Carriden, seagoing vessels were able to reach Stirling and the River Carron gave access to Camelon. From its estuary (Clota 85

JAMES ELLIS JONES

  system of inland waterways. By contrast with later periods, Roman law provided legally enforceable constraints protecting the rights of navigation, and this compares favourably with the situation facing the “improvers” of the 17th/18th centuries, who needed to spend much time and effort in pursuing legislation for removal of obstacles, before any work could be commenced.

Chapter 9 The development of the inland waterways Introduction Rivers were the major arteries of long-distance trade in Britain during late prehistory (Sherratt 1996, 211–34) and, although best known for its system of roads, the Romano-British period saw the development of a coastal and riverine transport infrastructure that was not equalled until the advent of the Canal Age of the 18th century. A combination of legislation, and the availability of a costeffective water transport system, meant that the roads were available for their primary task, i.e. the movement of military personnel and supplies.

The conquest of Egypt had given Roman engineers direct access to long established expertise in waterway management, linked to the Nile, and this knowledge was put into practice in Italy. The Fossa Augusta linked the River Po to the homeport of the Roman Adriatic fleet at Ravenna. Strabo describes how Scaurus drained the plains to the south of the River Po “running navigable canals from the Padus (Po) as far as Parma, for near Placentia the Padus is joined by the Trebia, as also before that by several other rivers, and is thus made excessively full” (Geog. 5.1.11). This expertise was transferred to the Northern provinces where, for example, the Fossa Drusiana was built by Drusus in 12 BC to link the Rhine to the Yssel. Tacitus describes the campaign of Germanicus in 17 when “he sent on his supplies and assigned the vessels for the legions and the allied troops, entered the Fossa Drusiana, as it was called, and he arrived after a prosperous voyage through the lakes and the ocean as far as the River Amisia” (Ann. 4.8). The canal was used again for the carriage of troops and supplies in 47, but was badly damaged in the Batavian revolt of 70 and never reopened.

Movement by water provided a viable alternative to road transport, initially for military activity and, from the later first century, for the establishment of urban communities. The needs of the cities, towns and remaining forts continued the requirement into the second, third, fourth and possibly fifth centuries. In addition to the transport of major cargoes of grain, wine and olive oil, the products of the extractive industries of quarrying, tin, copper and lead were particularly suited for movement by river. For example, Severn Valley pottery could be moved either up or down river, the iron blooms produced on the Severn Estuary sites were, because of the weight factor, ideally suited for movement up-river to Gloucester and beyond. A Roman sea-going merchant ship could carry a cargo of 60 tons, and this would be transferred, say at Gloucester, into 6 x 10 ton barges to take its cargo up river. If the cargo were taken inland by road, 110 ox-wagons and between 500 and 1,000 oxen, or 440 mules would be needed, depending on the terrain. Extra animals and wagons were also needed to carry fodder for the animals pulling the cargo. Some 20 men would have been needed to work the barges, compared to the 220 men (drivers plus leaders) required for the wagon train.

Strabo (Geog. 4.1) refers to portage from one navigable point to another in Gaul, when he comments that “The course of the rivers is so happily disposed in relation to each other that you may travel from one sea to the other, carrying the merchandise only a short distance, and that easily across the plains, but for the most part by the rivers, ascending some and descending others.”. In cases where a river became un-navigable, the Romans saw portage as an acceptable option. Strabo comments that “since the Rhone is swift and difficult to sail up, some of the traffic from here goes by land on wagons, that is all the traffic to the Arvernians and to the River Loire, although the Rhone in its course draws close to these also; but the fact that the road is level and only 800 stades (148 km), is an invitation not to sail upstream, since it is easier to go by land”.

The fact that a river might have been naturally navigable, or might have been made navigable, does not necessarily indicate that it was in use as a major waterway. Clearly, the cost-effectiveness of river improvement and maintenance were important factors. However, we should bear in mind the actions of Corbulo who is said by Tacitus that “to keep his soldiers free from sloth, he dug a canal of 23 miles in length between the Rhine and the Meuse” (Ann. 11.21). This might be considered a more useful occupation than that of soldiers of the British army who, in recent times, were employed in whitening stones and blacking coal.

River improvement Whilst the removal of obstacles to navigation is of great importance, perhaps the most significant engineering method employed is the use of locks, to permit boats to ascend or descend rivers and canals, and this technique is known from early times. Herodotus (2.138, 158) attributes the first canal linking the Nile to the Red Sea to Necho II (610–595 BC). The canal was over 60 miles in length and was wide and deep enough to take large seagoing vessels. There is a difference in levels between the Nile and the Red Sea, and “The Ptolemaic kings cut their

Roman competence in fluvial engineering An understanding of the level of competence of Rome in matters of port construction and inland water transport, in other parts of the Empire, is necessary for an understanding of the degree of exploitation of rivers in Britannia. It will be argued that Roman military engineering was more than capable of improving and maintaining these rivers, thereby establishing an effective

86

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  held the office of praefectus praetorio. The extracts from Book 43 clearly give support to the contention that river navigation, during the Romano-British period, was not impeded by the weirs etc., which be-devilled medieval navigation.

canal and made it so that it could be closed so that, when desired, they could sail into the sea without difficulty, and also sail back”. Referring to the differences in levels between the River Nile and Lake Moaris, Strabo comments that “Locks have been placed at both mouths of the canal, by which the engineers regulate both the inflow and outflow of water” (Geog. 17.1.25).

Edict 12 is intended to prevent anything from being done in a river, or on its bank, to hamper navigation (De fluminibus. Ne quid in flumine publico ripave eius fiat, quo peius navigatur.). It states that “You are not to do anything in the public river or on its bank by which the landing or passage of a boat is or shall be made worse”. Edict 13 states that nothing should be done in a river which might cause the water to flow otherwise than it did last summer (Ne quid in flumine publico fiat, atque uti priore aestate fluxit.). The provision is made to prevent a river drying up because of unauthorized tapping by watercourses or by changing its bed and applies to all rivers, both navigable and un-navigable. Edict 14 deals with navigation in a public river (Ut in flumine publico navigare liceat) and forbids “the use of force against such a one to prevent him from travelling in a boat or raft in a public river, or loading or unloading on its bank”. An English Act of Henry VI echoes this edict (9 Hen. VI. c. 5), declaring the River Severn to be a free river for all the King's subjects, to “carry on within the stream of the River”. Edict 15 deals with “Building up a bank” (De ripa munienda) and prohibits “the use of force to prevent such a one from doing any work in a public river or on its bank for the purpose of protecting the bank or the field which adjoins the bank, provided that navigation is not made worse by it” (Mommsen et al. 1985, 578-82).

In 112, Pliny the Younger, who was at that time governor of Bithynia, wrote (Ep. 10. 41) to the Emperor Trajan that “there is a sizeable lake in the area of Nicomedia across which marble, farm produce, wood and timber are easily and cheaply conveyed by boat right up to the main road, from which, with great effort, and even greater expense, carts take them to the sea”. Pliny proposed the construction of a canal linking the lake with the sea and Trajan approved the idea in principle, offering the services of an experienced surveyor. The lake was some 20 m above sea level, but the correspondence shows that the Romans were not greatly concerned about the problem of changes of level in their waterways. Pound locks, as used on modern rivers and canal systems, consist of two gates close together, which permit the transfer of a vessel from one level to another. When proceeding upstream, a vessel enters the area between two gates, the downstream gate is closed and “paddles” in the upstream gate permit water to fill the area between the two gates. When the area between the gates rises to the upstream water levels, the lock-gate is opened and a vessel is able to proceed; the process is reversed when proceeding downstream. It is by no means certain that pound-lock technology was in use in Roman Britain and, whilst some literary evidence indicates that some form of locks were in use, particularly in the Mediterranean area, there is no archaeological evidence to attest the use of these devices in Britain or elsewhere in the Northern provinces. The flash lock was a less sophisticated device, consisting of a gate, or a set of horizontal boards, spanning a gap in a dam across the river. Water flows from the higher into the lower level after the gate is raised and, when a balance is achieved, a vessel is moved either up or down river, and the gate is then closed. The use of flash locks was dependent on water levels and, from a later period, Willan (1964, 88) echoes Pliny’s comment when he points out that the replenishment of water above a lock on the Thames, to a level sufficient to allow further navigation up-river, “could take anywhere from two hours to a month or six weeks”.

The extent to which this legislation was enforced, or enforceable, cannot be ascertained, particularly in a distant province such as Britannia. However, it does at least indicate that, if an impediment to navigation needed to be removed, it could be done without recourse to the tedious legal processes which made the later improvement of English rivers such a drawn-out process. Inland water transport in Britannia Rivers Writing of the British rivers Tacitus comments; “I would only add one remark, that nowhere else does the sea make its power more felt: the tide causes long stretches of the rivers alternately to ebb and flow, nor does it simply rise and sink upon the shore, but it runs far inland, and winds about and makes its way into the very heart of the hills and mountain chains, as if the sea were lord of all” (Ag. 10). Most rivers may be made navigable, if an adequate labour force is available to construct dams to produce a ponding-up effect, similar in function to the later flash locks. A suitable supply of slave labour would have presumably been available, as a result of the defeat in battle of the Welsh tribes under Caractacus, or in Scotland after Mons Graupius. As carriage by road was expensive, but by water was relatively cheap, the Romans would have exploited the rivers to the maximum

Justinian’s Digest of the Law The importance placed by Rome on unimpeded use of rivers, is well illustrated by certain edicts contained in Justinian’s (482-565) “Digest of the Law” (Codex Justinianus), published in 530, and codifying all Imperial constitutions, dating from the time of Hadrian, that were still valid. Some 95% of the work is taken from authors of the period between 100 and 250 and, in particular, the edicts summarised below are the work of the jurist Domitius Ulpianus, who at the time of his death in 223

87

JAMES ELLIS JONES

  have been done, as in the case of the upper reaches of the River Tiber (Pliny HN. 3.53). However, the question of cost-effectiveness would come into play; Ilchester was a key point on a trans-isthmian trade route, and would justify the Roman canalisation of the river (Thew 1994. 134). On the other hand, water transport to Llandovery had reached the effective limit of navigation, with no opportunity for onward shipment, negligible downshipment, and therefore no consequent economic advantage.

potential. The great flat-bottomed lighters that plied the Rhine and its tributaries would have been far too large, but vessels the size of the Barland’s Farm boat would have had no difficulty in navigating the River Severn as far as Pool Quay, near Welshpool (Nayling et al. 1994, 596–603). It is also probable that vessels of similar size could reach as far as Forden Gaer, Caersws and possibly beyond. During the excavation at Caersws, the principal buildings of the fort were found to be of a type of sandstone of which the nearest outcrop is at Welshpool. It is the view of the excavator that the building stone, probably dressed at source, had been brought up river to the fort, either by boat or on rafts (Britnell, J. 1999, pers. comm.) Today the river has a reasonable depth of water, certainly as far as Llanidloes, and appears to have potential for earlier navigation to that point. Whilst it may be an unreliable preservation from the folklore, the area upstream of the present bridge is known as “The Harbour”.

Figure 9.1 River Severn at Caersws

Figure 9.2 Watersheds (thick line) and main rivers of Britain

The landscape of Britain displays significant topographical variations; it is only necessary to compare and contrast the mountainous topography of the Cambrian massif with the wetland plains of the Somerset Levels, for it to be obvious that differing problems for transport by water exist, and that equally different solutions are required. In the case of the Somerset Levels the problem is mainly that of tidal inundation, and the solution is the embankment of rivers such as the River Parrett, providing not only an increase in the depth of navigable waters, but also protection against flooding of valuable agricultural land. Navigability to the Roman town at Ilchester (Lindinis) is highly probable as, though 49 km from Bridgwater Bay; the town is situated at only 11 m above sea level (a rise of 0.22 m per km) and is only 3 m higher than the port of Bridgwater, itself 19 km from the sea. By contrast, the fort at Llandovery, situated on the Afon Twyi, is 68 km from Carmarthen Bay; by this point, a height above sea level of 73 m (a rise of 1.07 m per km) has been attained. A major program of river improvement, involving debris clearance and the probable use of flash locks, would have been needed in order to achieve navigability. This could unquestionably

The Romans developed the use of British rivers, even if transhipment and portage were required. This is illustrated, for example, by the movement of Blackburnished pottery from the Poole Harbour area (Allen & Fulford 1996, 223–281). They suggest that the pottery was transported by pack animals to Ilchester, then by barge via the Rivers Yeo and Parrett to a port on the Bristol Channel. Some was re-loaded to coastal vessels for distribution to the ports of south Wales, and shipment through the Irish Sea to the forts of north Wales, northwest England and the garrisons of Hadrian’s Wall. The Rivers Severn, Wye and Usk provided access far into the interior, and the apparent absence of roads leading to some isolated villas located on the riverbank suggests a reliance on water transport. Equally, the Roman programme of road building provided an effective method of cross-country transport, and pottery found in the fort at Caersws, in the upper reaches of the River Severn, consisted of 85% Cheshire Plain ware, as opposed to 15% Severn Valley ware (Britnell 1989, 85– 7). This suggests that the supply depot of legio XX at Chester was responsible for overland provisioning and 88

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  years, studied 57 km of the northern section of the Dyke and led him to the conclusion that the Car Dyke in the Witham peat fens is a natural watercourse, and that the land levels indicate that the original movement of water in early times was from north to south. In the silt fens, he identified a man-made channel within which, because of variations in levels, there cannot be a flow of water in a single direction. He further argued that, from artefact distribution, there is little evidence to suggest significant transport to or from Lincoln. He does, however, detect stronger evidence of pottery from the west, in the form of mortaria from Mancetter/Hartshill and suggests the possibility of reciprocal trade from the east, involving perhaps salt, oysters, reeds and leather. He questioned the need for a canal on the basis that navigable waterways already existed. For example, he considered it likely that, in Roman times, the Car Dyke was navigable from North Kyme to Washingborough, but thought it difficult to determine whether was actually used for transport, as it was probable that the River Witham would have provided a better contemporary route. Based on archaeological investigation, he argued that the waterway was not continuous, but was either broken or never constructed, where the several Roman road and tracks crossed its line.

maintenance, and that, in this case, shipment via the Severn played only a minor role. Inshore channels The forts at Reculver (Regulbium) and Richborough (Rutupiae) were strategically placed at either end of the Wantsum Channel to give protection against North Sea raiders. The Channel was still navigable when the Vikings carried out a successful raid on Canterbury in 839, but continued deposition of shingle at Stone Neach led to the gradual silting up of the channel and, by the 10th century, it was no longer possible for coastal vessels to reach Canterbury via the River Stour. The North Sea flood of 1953 (in which tides reached 2.4 m above prediction) made Isle of Thanet an island again, if only for a few days. Placing the Wantsum Channel in a European context, Hill (1981, 14) considered that it would appear to have been an important part of a sheltered waterway system that ran from Ribe in Denmark to Quentovic in Northern Gaul, and from the Alps to the Thames. Although it cannot be conclusively proved that such a waterway existed in Roman times, Speed’s 1612 map of Kent clearly shows the existence of the Yantlet creek, which at that time provided a navigable channel between the Isle of Grain and the Hoo Peninsula, and this waterway continued in use by shallow draft Thames barges until the 20th century. It is of interest that at the time of writing (January 2010), an Admiralty Notice to Mariners gives notice that the “grab hopper dredger Cherry Sand will commence work on the removal of wrecks from the Yantlet channel” and then goes on to specify six positions in the middle of the waterway (Figure 9.3).

Salway (1980, 338–9) considered it worth hastening to put on record his support for Simmons’ thesis that the Lincolnshire Car Dyke was not a canal, but that it acted as a ringvart drain. He further commented that some of his co-authors of The Fenland in Roman Times had been “uneasy as to how the natural streams and rivers flowing from the upland into the Fenland basin could have been crossed by the Dyke if it were indeed a continuous canal: we had to assume that an elaborate system of locks and sluices had once existed”. He also drew attention to Professor St Joseph's opinion that some of the sharp bends in the Dyke would have made difficult the use of substantial barges.

Figure 9.3 Admiralty chart of wrecks in Yantlet Creek Canals Long interpreted as a navigable canal, and approximately following the 3 m contour, the Lincolnshire Car Dyke runs for some 65 km from the River Witham at Washingborough, 5 km below Lincoln, cutting across the lines of natural drainage of Fenland, intersecting with the River Welland at Deeping Gate and joining the River Nene at Fengate near Peterborough. However, fieldwork by Simmons (1979, 183–196), over a period of over ten

Figure 9.4 The embanked Lincolnshire Car Dyke Rippon (2000, 70) noted that the occurrence of banks on both sides of the central channel suggest that the Car Dyke could not have functioned as a catch-water drain, and the causeways may in fact have served the same function as locks in a canal, maintaining the water levels 89

JAMES ELLIS JONES

  in individual stretches of the channel”. It might well be disputed that “banks on both sides” of a channel are an impediment to effective draining. As is made clear by Rippon's own extensive work on the Somerset Levels, embankment of both natural and artificial channels is standard practice in flood control levels. As he later points out (2007, 213), embanked canals do not drain water from the areas to which they pass, but help to prevent flooding by carrying fresh water discharge to the sea, from rivers flowing off adjacent uplands. In the case of the Car Dyke, short portages would have enabled full advantage to be taken of what, with an original navigable width of 12 m and a depth of 2 m, would have been an extremely effective waterway. The most recently published opinion is that of James Bond (2007, 167), who considered it possible that the purpose of the “breaks” was in order to enable a navigable depth of water throughout the entire length all year round, by separating the canal into pounds at different levels. He is further of the opinion that the necessity of transferring goods from one boat to another or carrying the boats themselves over the obstructions would be a fairly cumbersome operation, but might still be preferable to road transport.

wide and less than 2 m deep was dug immediately to the south of the road, with salterns along its banks. Subsequent marine inundation covered both the road and canal with silt deposits, but the road was later reestablished on the south bank of the earlier canal. As the canals made no contribution to the local drainage system “it is difficult to find any explanation for them other than as an alternative transport link between the upland margins” (Bond, 2007, 168). In 1724 Stukeley suggested that the Cambridgeshire and Lincolnshire Car Dykes were joined into a single waterway between the rivers Nene and Ouse by the use of further artificial cuts and intervening rivers, but no proof of this has ever been forthcoming; indeed, it is not even certain whether both parts are of the same date (Bond 2007, 165). Romano-British pottery has been reported from dredging of a canalised section of the River Lark between Isleham Sluice and Mile End and a canal, apparently associated with Romano British turburies to the south of the Fen Causeway and east of Stonea Island has been indicated by aerial photography. Other canals of Roman date extend from the Car Dyke, with the longest running from the settlement at Bourne for some 6 km to the head of a creek in Morton Fen. Scoured by the tidal flow and with an original width of 13 m and a depth of 3 m, this could have enabled access to and from The Wash by coastal vessels.

The Cambridgeshire Car Dyke links the River Cam with the Ouse and the Old West River and was probably constructed as part of a major program of imperial planning associated with the settlement of the Fenland, possibly under the direction of Hadrian at the time of his visit to Britain around 120. Its use as a navigable canal is less controversial than its Lincolnshire counterpart, as the absence of Roman settlements in the southern Fens may be seen to have obviated the need to protect the surrounding area from flooding and suggests water transport as the only significant reason for its construction. Function as a canal with an original width of 20 m and a depth of 2m, would have considerably reduced the distance for transport by water between Cambridgeshire and Godmanchester.

Roman canals are concentrated in the Fenlands, but there are isolated examples from other parts of Roman Britain including the Raw Dykes at Leicester that have been traditionally considered as part of the town aqueduct. However, Shepherd Frere (1978, 280, 298) suggests that the shape of the earthwork is that of a navigable canal, perhaps leading to docks on the River Soar. He points out that its level of water would have been over 3 m feet below that of the public baths, only 2 km away, and that for an aqueduct this is “an incredible error when water was available over 50 feet higher at a hardly greater distance from the town”. Joining the River Cam below Cambridge, three straight canals cut across the natural drainage, and Wicken Lode, Monks’ Lode and Burwell Old Lode and the Reach Lode which has produced quantities of Romano-British pottery, have all been suggested as being of Roman date. However, David Hall (1992, 6, 42) has found no compelling evidence for Roman construction at any of these waterways; there was no Roman site at their landward ends and the Reach Lode appears to have actually cut through a Roman villa. He considers that a late Anglo-Saxon or early medieval construction is more probable. In the North Somerset Levels, there is a possible embanked Roman canal, some 12 m wide and 1.5 m deep (Grove 2002, 65–87); Rippon (2007, 211) considers this is remarkably similar in scale to the Aylmer Hall, Fen Causeway, and Bourne-Morton canals in the Norfolk and Lincolnshire Fenland. In Yorkshire, the Don formerly flowed northwest entering the Trent just before its confluence with the Humber. At some stage, perhaps even in Roman times (Cole 2007, 65); a channel was cut to take the Don northwards to join

With a navigable length of 17 km, and possibly following the original course of the Trent in prehistoric times, the Foss Dyke connects the River Witham at Lincoln with the River Trent at Torksey, where some form of tidal sluice would have been needed. At the end of the Roman period, the navigation fell into disuse, but from the quantity of eighth/ninth-century coins and ornamental metalwork found at Torksey, it is probable that the Roman canal had been brought back into use. After a hiatus during the 10th and 11th centuries, the Foss Dyke was recut by order of Henry I and reopened for traffic in 1121. Failure to maintain the canal led to reduced traffic and the Foss Dyke fell out of use until its refurbishment at the end of the 17th century. The Fen Causeway was constructed as a road between Denver near Downham Market and Longthorpe near Peterborough, in the later part of the first century. In some sections it follows a natural watershed on slightly higher ground, but in others it crosses the peat fen and, probably during the second century, a canal some 12 m

90

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  it with the Went and then the Ouse, thus providing a waterway to York and an alternative entry to The Wash via the Humber.

their Abbey to the manor of Andredsye [Nyland] in their boats”. As early as the 10th century there is evidence of the importance placed on artificial canals for the transport of building materials and other bulky commodities to and from monastic sites (Bond 2001, 102–3). On the Somerset Levels, the River Brue was diverted away from the Bleadney gap into Meare Pool; the Pilrow Cut (Figure 9.6) was an artificial canal linking the Rivers Brue and Axe and thereby providing access to the coastal manors of Brent, Lympsham and Berrow.

Development of rivers and ports during later periods There is no significant documentary evidence for the use of British rivers during the Roman period. In order that an impression of the potential for navigation may be examined, some varied examples of river use, ranging from the medieval to the mid-18th century, have been chosen to illustrate the diversity and volume of traffic that was possible, even on rivers that are now restricted solely to leisure activities, such as boating and fishing. The potential for inland water transport is demonstrated by trade on the River Severn. From the early medieval period the river was, once again, a major trade route for traffic to and from the West Midlands and the Welsh Marches. In favourable conditions, the Severn was navigable for 250 km and Pool Quay, near Welshpool, became an important waterway settlement, remaining in active use until the early 19th century. In 1756, there were 376 vessels owned by traders living between Welshpool and Gloucester and, in the 18th century, some 4,000 watermen and their families worked on the river. When the sails could not be used, teams of men known as “bowhauliers” dragged the barges and trows,. The most common vessel was the Severn trow of some 50 tons, with fore and aft sails, and therefore a performance to windward superior to that of similar sized square-rigged vessels of the Roman period. However, in confined river waters, this would be of little advantage, as tacking was usually impracticable, and sailing was therefore only possible when the wind was abaft the beam.

Figure 9.5 The “Pilrow Cut” completed in 14th century by direction of the Abbot of Glastonbury During Edward I’s campaign against Llewellyn ap Gruffydd from 1277–80, three hundred “ditchers” from the Fenlands were engaged in digging a new cut, 3–4 km long, between Rhuddlan, the river’s highest tidal point, and the sea. This replaced the meandering course of the River Clwyd (Figure 9.7), and enabled the newly built castle to be supplied by sea-going ships (Taylor 1963, 319). A dock within the castle ward, capable of accommodating vessels of up to 40 tons, was constructed at the end of a short cut.

Early post-Roman river improvement The course of the River Thames underwent a number of alterations during the 11th and 12th centuries, when the citizens of Oxford petitioned Abbot Orderici of Abingdon for consent to convert the navigable channel from the abbey's meadows to the south, in order to allow their vessels more convenient passage upstream, thus bypassing the dangerous shallows near Barton Court. During the 11th century, a new canal, the Bykers Dyke, some 8.5 km in length, diverted the River Idle, which had formerly flowed north into the old course of the River Don on Hatfield Chase, to join the River Trent at West Stockwith. A 1.5 km canal connecting Glastonbury with the River Brue during the Abbacy of Dunstan has been discussed by Hollinrake and Hollinrake (1991, 117–8; 1992, 73– 94). It is possible that Glastonbury Abbey also constructed a new outlet of the River Brue to the sea near Highbridge (Williams 1970, 62–74). A 13th-century document (Joannis confratis et monachi Glastoniensis) states that the watercourse between Nyland and Bleadney “was adequate for the Abbot to take stone, lime and corn from his manor and from other places in those parts to his Abbey of Glastonbury and [they] were used to go from

Figure 9.6 The improved 13th-century cut of the River Clwyd at Rhuddlan Castle The Gough Map, dating between 1355 and 1366, is the oldest surviving roadmap of Great Britain, and, whilst the depiction of a river on the map cannot be relied on as an indication of navigability, the pictorial over-emphasis of 91

JAMES ELLIS JONES

  By 1500, the population of England and Wales had risen to some four million, and was therefore within the range of estimates for the population of Roman Britain (Millett 1990, 182). As in the Roman period, this rise in population was accompanied by a rise in urbanisation, with the population of London doubling from a quarter of a million inhabitants to half a million, in less than a hundred years. This concentration of population created a problem of supply, to which long distance transport by water was the most effective solution. The period from 1500–1750 is of particular importance, as it was the time of river improvement on a significant scale, yet precedes the construction of artificial watercourses during the Canal Age (see Figures 9.10 and 9.11 below). The length of navigable rivers in England grew from 685 miles in 1600–6 to 960 miles c. 1702, and to 1160 miles by 1730.

rivers may suggest their importance as part of the transportation infrastructure. For the medieval period, Langdon (2007, 110–132) has studied the accounts for purveyance in the sheriffs’ miscellaneous accounts from the Public Record Office. He considers that these accounts provide a sufficiently concentrated core material to make possible systematic analysis of the inland water transport system, in particular for the period from the late 13th century to the mid-14th century. Whilst accepting that this data does not provide comprehensive coverage of the entire inland water transport system and that lesser waterways were not considered a serious transport options by the various sheriffs, his contention that the purveyance accounts provide extremely rich information about many aspects of the inland (and coastal) water transport systems does seem well justified. His analysis of the types, dimensions, and loading capacities of the boats themselves, the logic of trans-shipment points, and the potential for satisfying regional and national carrying requirements is of particular value. However, the temptation to compare the medieval with the Roman period has its dangers and Langdon’s emphasis on downriver travel is a case in point. He emphasised that, as it is much easier it is to “go with the flow” than to haul loaded boats up-river. He illustrated this by reference to Leighton’s observation that, on the Rhône from Lyon to Avignon the downstream journey took 2 to 5 days, whereas the same journey upstream took nearly a month and, in addition, required the use of several towing animals (Leighton 1972, 126). This is particularly relevant in Roman Britain as, in the majority of cases, the direction of travel of loaded cargo was upstream, in order to supply military installations.

Figure 9.8 Water transport on medieval rivers. Grading indicates suggested volume of traffic The cost advantages of water transport (Duncan-Jones 1974, 366–9) and the legal restrictions placed on loads carried on roads (e.g. Codex Theodosianus, 8.5.30) suggest this form of transport as the preferred option. In the case of transport by road, Salzman commented that, during the Middle Ages, the cost of transport of stone exceeded that of the stone itself, for distances of over 12 miles (1967, 119). In some cases necessity, rather than cost-effectiveness or convenience, may have been a major factor. Finley (1973, 126) commented that, prior to the advent of railways, “Towns could not safely outgrow the food production of their own immediate hinterland, unless they had direct access to waterways”. Figure 9.7 Average loads on some rivers during the medieval period

92

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Major navigable rivers The difference in cost between road and river transport suggests that goods would be carried by water when that was possible and convenient, but river transport leaves very little evidence of its nature and extent. It is therefore difficult to determine how many of the rivers were navigable, and what was the limit of navigation on individual rivers. The five major rivers, the Yorkshire Ouse, the Trent, the Great Ouse, the Thames and the Severn, were all tidal, and therefore passage along them, within the tidal limits, was free and common to all. They were all navigable but, except for the Severn, very little is known about the goods they carried. Willan (1976, 15) considers it doubtful whether even intensive local research would reveal very much about the traffic on some of these rivers, as the absence of tolls removed any incentive to record the passage of goods. Later when the improvement of rivers led to the imposition of tolls to pay for the improvement, records of tolls were made, but they do not survive before the end of the 17th century. Traffic on the Severn is better documented, simply because Gloucester ranked as a seaport and so its trade was recorded in the Port Books. The Yorkshire Ouse was tidal to York and was accessible to sea-going ships and, during the early 16th century, was ranked as a seaport. Probably because of the increased size of ships during the 17th century and some deterioration in the river, sea-going vessels were prevented from reaching York and the city became purely a river port. There is some evidence for river traffic in the area, for example, in 1345–6 grain was transported from Boroughbridge on the River Ure and Wansford on the River Hull to Kingston-Upon-Hull (Gardiner 2007, 88), using small boats (batellae) and ships (naves), but neither the Wharfe nor the Aire seem to have carried a significant volume of goods. The citizens of Beverley made some improvements to the Beck, which gave access to the sea via the River Hull. Willan (1976, 16–17) comments that whilst the Trent was clearly an important channel of communication, it seems impossible to get any accurate picture of the traffic on the river. Only one of the Trent's tributaries, the Idle, was navigable but the Foss Dyke, an artificial channel which linked the Trent at Torksey with Lincoln, had been navigable but seems to have deteriorated in the 16th century, when attempts to scour it were never completed. However, at times of high water levels, it is possible that travel by waterway from York to Lincoln could have taken place. The Great Ouse was navigable from King’s Lynn to St Ives, the Cam is known to have been navigable to Cambridge and there was some traffic on the Little Ouse and the Lark. The Ouse was linked with the Nene at Wisbech; the Nene was navigable to Peterborough. To the east of London, the Lea was navigable from Ware in Hertfordshire, and a significant trade in malt and corn developed. From London, the Thames was navigable by large sailing barges as far as Burcot, where it is probable that transfer took place into smaller river craft. Using the record of cargoes carried from the sheriffs’ purveyances, Langdon (2007, 1-6) has

Figure 9.9 Areas shown in black were more than ten miles from a navigable waterway in c. 1600 (after Willan)

Figure 9.10 Reduced area shown in black demonstrates greater access to “improved” waterways by 1760 (after Willan)

93

JAMES ELLIS JONES

  dredger was used to break through the hard crust of gravel on the shoal. McDonald commented, “That, of course, is by no means the whole of the story. Down to 1818, after much time and labour had been expended in improving the waterway, it sometimes happened in seasons of neap tide that lighters, drawing only 4 ½ feet of water, took six weeks to complete the voyage from Greenock to Glasgow” (1932, 240–41).

suggested that barges carrying loads of up to perhaps 50 tons could have worked the Lower Thames. At the other extreme, boats on the upper reaches of the rivers Parrett and Tone were much smaller, carrying only some 2 to 3 tons but with coastal vessels below Bridgewater usually carrying up to 90 tons. There are very few records of the products that were carried up or down, but Willan (1976,18-19) cites examples of hay, wood, stone, slate and “the occasional shipment of salt fish and wainscot” being unloaded at High Bridge in Oxford and considers that coal was probably sent upstream.

Small craft were able to reach the city of Exeter in Roman and medieval periods but by the late 13th century, the River Exe was blocked by weirs so that sizable vessels could no longer pass between the city and the sea. In 1540, an Act of Parliament was obtained, permitting clearance of all obstructions to navigation but, despite considerable efforts in the following years, it proved impossible to make the river navigable even for small boats. The City Council eventually decided to build a canal to bypass the most difficult sections of the river and construction began early in 1564, being completed two years later. The canal had a fall of about 2 m over a length of 2.8 km but a depth of only 1 m. Seven sluice gates and three poundlocks permitted passage of vessels from 8 tons on neaps up to 16 tons on spring tides. At the end of the 17th century, the canal was deepened to allow passage of sea going vessels of up to 4 m draught and eventually ships of 200 to 300 tons were able to reach the city (Henderson 1991, 124–135).

The port books of Bristol (Exch. K.R. Port Books, 1128/13, 14) show that the port shipped upriver to Berkeley, Frampton-on-Severn, Newnham, Gloucester, Tewkesbury, Worcester, Bewdley and Bridgnorth. Wine was the most important single item but there is also a record of a wide variety of goods: soap, raisins, linen, canvas, oil, iron, dry-wares, upholstery-wares, tin, brass, pitch, tar, hops and fish. The wine had probably been imported from the Bordeaux region, and the tin came up by coaster from Cornwall. An equally wide variety of goods was sent down river. Tewkesbury and Gloucester shipped large quantities of barley, wheat and peas to Bristol, Bewdley sent skins, tallow, leather and cloth; Bridgnorth wool, honey, candles and wax; Shrewsbury sent wool. Tewkesbury and Gloucester had a coasting trade of their own, and there are records of direct shipment of large quantities of wheat and barley malt as far as Caernarfonshire, Devon and Cornwall.

The potential for a rapid rise and fall in the fortunes of a small port is well shown by Parkgate, some 33 km downstream of the Roman fortress of Chester, by Sunderland Point lying 10 km downstream of the “Saxon Shore” fort at Lancaster, and by Lilstock (see Chapter 1) on the North Somerset coast Each of these harbours demonstrate a different reason for its demise, in the case of Parkgate it was caused by increased sedimentation, Sunderland ceased to be viable because of the opening of a newer and more efficient port; at Lilstock the natural process of longshore drift eventually denied access to the port.

In 1699, over 90% of voyages from Chepstow to Bristol returned in ballast and, at the same time, 20% of all Severn trows returning to Gloucester were unladen or carried only a small quantity of dutiable goods. Trade with Europe was significant, with the following voyages being recorded in 1699 (Hussey 2000, 40); Iberia France N. Europe Mediterranean

22 8 31 18

The now silted-up old port of Parkgate (Figure 9.12) lies on the English side of the Dee Estuary and, during the 18th century, became Chester's principal outport, with a principal trade in passengers to and from Dublin. The port usually handled vessels between 80 and 100 tons; the largest ship recorded at Parkgate was 300 tons, but it ran aground and damaged its keel. The construction of the New Cut in 1737 improved the navigation of the Dee to Chester by creating an artificial channel for 7 km below the city and initially this was to the benefit of Parkgate as it diverted the main channel from the Welsh to the English side. However, the combination of an unstable channel across the estuary and engineering works on the Welsh side by the River Dee Company, led to the abandonment of the port, with the last occasion when ships were reported at Parkgate being in 1850 (Place 1996, 72–4).

In the 18th century, the Clyde was fordable on foot at Dumbuck, some two miles below Old Kilpatrick. The preamble to the first of the Clyde Navigation Acts, which dates from 1798 states that “The river Clyde from Dumbuck to the Bridge of Glasgow is so very shallow in several parts thereof that boats, lighters, barges or other vessels cannot pass to and from the City of Glasgow except it be in the time of flood or high water at spring tides.” Several abortive attempts were made to improve navigation but eventually, John Golborne of Chester, recommended what proved to be effective measures for dealing with the problem. The level of the water was raised several feet by running a series of rubble jetties out from the south bank creating a scouring effect; a powerful

94

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

 

Figure 9.11 Silted-up port at Parkgate on the River Dee

Figure 9.13 Skinners Quay, Chester 1836

By contrast, at Sunderland Point there is still a reasonable depth of water (Figure 9.13), but only small pleasure craft are still to be seen in the area.

By contrast, at Caerleon, also the site of a major Roman port, there is no evidence for any significant man-made interference with the natural course of the Usk. The greater volume of freshwater passing downstream, combined with the scouring potential of Bristol Channel tides, means that, unlike the Dee, the river maintains its earlier navigability. However, only a few small fishing boats now lie in the river, with all sea-going vessels being handled on the coast at the appropriately named city of Newport. The real reason for the demise of inland ports is perhaps not the result of natural processes. There is no insurmountable reason why, at the present time, a vessel similar to the St Peter Port ship of the Romano-British period should not navigate to either Chester or Caerleon, as was the case in the 19th century (Figure 9.14). It is the size of vessel that determines access to a port, and perhaps it is not questions of siltation or the reduction in coastal traffic and the like, that have fundamentally affected Chester or Caerleon, but simply that ships have become larger.

Figure 9.12 Abandoned port at Sunderland on the River Lune The old port lies on the north bank of the River Lune, some 8 km downstream of Lancaster and, at the beginning of the 18th century, was developed as an outport where ships could unload, or wait for the tide before moving upriver to the main docks at St George’s Quay. Once part of the “slave triangle”, a thriving Atlantic trade had cotton as the major import; several of the buildings that once stored rum and tobacco remain as private houses. The opening in 1797 of Glasson Dock, equipped with lock gates, on the opposite side of the estuary led to the abrupt demise of Sunderland as a port.

-o-o-o-o-o-o-oThe preponderance of forts and settlements on the coasts, estuaries and navigable rivers provided an excellent case for supply by water transport. Roman Britain lent itself well to this method of supply, because of its long coastline, the isolation of the forts and the lack of an effective road system during the early campaigns. For example, the case for the widespread use of water transport in Wales and the Marches is well supported by an analysis of Maps 4.33–38 in Jones and Mattingly (1990, 103–5). Of the 66 sites shown, 51 are on coasts and potentially navigable waterways; only 15 are unlikely to have been directly supplied by water transport.

In Roman times, sea-going ships on passage to Chester could take advantage of the incoming tide, sail up the wide estuary of the River Dee and unload their cargoes at the quays of the fortress of Deva. Silting of the estuary was recorded in the Middle Ages, and navigation up to Chester became increasingly difficult. In an attempt to solve the problem, the New Cut was dug in 1735–6 giving a depth of 16 feet of water at moderate spring tides enabling access for ships up to 200 tons (Ward 1996, 4– 10). However, the only vessels now to be seen in the area of the once prosperous Roman port are small pleasure craft.

There is no doubt that the Romans were capable of constructing large-scale port installations and improving and maintaining river navigation. By contrast with later periods, Roman law provided legally enforceable constraints protecting the rights of navigation, and this compares favourably with the situation facing the

95

JAMES ELLIS JONES

  “improvers” of the 17th/18th centuries. It is arguable that the rivers of Roman Britain were exploited to a greater extent than at any period up to the commencement of the Canal Age. This was made possible by a combination of engineering skills, military necessity, economic advantage, plentiful labour resources, suitable vessels and an absence of riparian rights. Inland transport in Roman Britain, whether by road or river, at least equalled, and in the majority of cases surpassed, that of the medieval period, and it is probable that it was not until the late 18th century that the transport systems of the Roman period were bettered. The term “improvers” is, stricto sensu, perhaps unjustified when referring to individuals responsible for earlier efforts and, “restorers” might be considered more appropriate when describing attempts to return rivers to their natural state, by removing man-made obstructions. The river improvements made during the 17th and 18th centuries were driven purely by the profit motive, and it is therefore clear that an economic case existed for this type of venture. However, as was the case during Roman times, the needs of the military are not subject to the same economic constraints; an example of this from the more recent past is the Huntspill River on the Somerset Levels, a completely artificial watercourse, constructed during the Second World War to service the needs of the Royal Ordinance Factory at Puriton. An interesting footnote to this is that the majority of the labour was provided by prisoners of war, echoing what has been suggested for the Romano-British period.

96

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  earth that there is a law forbidding its production beyond a certain amount”. There is little evidence in Britain for the use of this metal during the Iron Age, for example, a few lead net sinkers and spindle whorls from the Lake Villages near Glastonbury (Bulleid & Gray 1911, 241). By contrast, the Roman world made great use of lead, particularly for plumbing, and it is probable that this was the only major metal industry introduced into Britain following the conquest (Manning 1979, 113). In addition to the better-known domestic use, the scale of demand from the construction industry is demonstrated by the probable usage of some 7 or 8 tonnes of lead in the construction of the bridge across the Tyne at Chesters fort (Hodgson 2009, 107). As the ownership of all mineral resources was vested in the Emperor, the early extraction was probably organised and supervised by Roman troops, possibly with the use of slave labour.

Chapter 10 Cargoes within the Province Products of the extractive industries

The Bishops Castle/Montgomery area on the Shropshire/Clwyd border was a significant mining area. Lead ingots of Hadrianic date are known from the vicinity of the “villa” at Linley, and extensive areas of the hillside give indications of hydraulic sluicing. The nearby fort at Brompton has evidence of secondary industrial activity within the fort itself, with bowl furnaces and clay hearths with charcoal and furnace debris, together with pits containing considerable quantities of litharge from cupellation. To the east of the fort, the vicus had a dense concentration of industrial pits and gullies, yielding ironsmithing slags, hearth-bottom material and lead waste. This suggests that the fort was a processing plant for the ore obtained from Linley. Further north, It was suggested by Eric Birley that the fort at Whitley Castle controlled mines at Alston on the present Northumberland/Durham boundary (1958, 41). On the Mendips in Somerset, the fort at Charterhouse had only a limited existence; with a lead ingot bearing the stamp of Legion II indicating that the mines were initially under military control. However, a lead pig dated AD 60 is marked “from the lead-silver works of C. Nipius Ascanius”, making it clear that control soon passed to private lessees (Webster 1953, 8–9, 22– 24). The last securely datable ingots from Charterhouse were made between 164 and 169, but coin finds indicate that occupation of the settlement, and presumably the working of the mines, continued into the fourth century. Recent landscape survey suggests that, rather than being a Roman innovation, production of lead developed from an existing Iron Age production complex (Fradley 2009, 99– 122). There are known lead-mining settlements at Prestatyn (Blockley 1989b) on the north Wales coast, Pentre Flint (O’Leary et al. 1989) on the Dee estuary and Ffridd (Blockley 1989a, 135–66) to the north of Wrexham. All have produced evidence for large-scale ore processing, dating from the later first century, with each site being provided with a bathhouse. In south Wales, lead was mined at Risca on the River Ebbw, where a bathhouse is also known, whilst on the River Rhymey at Lower Machen there is extensive evidence of lead ore and slag; at Draethen first-century pottery and coins have been found at the gallery mine at Coed Cefn Pwll-Du (Nash-Williams 1939a, 108-10).

Figure 10.1 Mineral resouces of Britain The comment that “Britain yields gold, silver and other metals, the fruits of victory” (Ag. 12) is appropriate, as the study area is rich in metalliferous minerals; those mentioned by Tacitus, plus copper, lead and iron. There are major deposits of these materials, as well as widespread and plentiful locations for stone-quarrying, clays suitable for the manufacture production of pottery, and sources of high quality brine for the production of salt. The movement by water of the products of these industries presented a cost-effective alternative to carriage by land, and there can be little doubt that, during the Romano-British period, a thriving system of sea and river transport existed. It is probable that mineral resources were under state control in the Roman Empire, and the army may have been involved in the initial exploitation of newly conquered provinces but, following consolidation, the mines were not worked directly by the state, but by comparatively small-scale leaseholders. Some areas were placed under the control of a procurator metallorum, who regulated mining activities and supervised the workings of individuals and companies; this was probably the case at Caernarfon. Lead Pliny (HN 34, 17. 164) commented, “Lead is used for pipes and sheets. In Spain and throughout the whole of Gaul it is extracted with considerable effort; in Britain, however, it is so abundant within the upper layers of the 97

JAMES ELLIS JONES

  it is produced very carefully. They beat the metal into masses shaped like knucklebones (astragali) and carry it to a certain island of Britain called Ictis, for at the time of the ebb tide the space between the island and the mainland becomes dry and they can take the tin in large quantities over to the island on their wagons. On the island of Ictis the merchants purchase the tin of the natives…whence it is then taken to Gaul and overland to the Mediterranean.” As Belerion can be identified with Cornwall, either St Michael’s Mount or Mount Batten (in Plymouth Sound), is probably the site of Ictis (Frere 1978, 30). According to Pliny (HN. 4.30.1 04) “There is an island named Mictis lying inwards, six days sail from Britain, where tin is to be had and to which the Britons cross in boats of osier covered with stitched hides”.

There is some archaeological evidence for both interprovincial movement, and overseas export, of British lead. Excavations on the River Frome at Wadebridge Street in Bristol uncovered two large Roman lead pigs with inscriptions dating from the reign of the emperor Antoninus Pius (139–161). It is probable that the pigs fell from a vehicle transporting them across the Frome, probably en route to Portus Abonae (Sea Mills). A group of 20 lead ingots found in the River Mersey in 1697 bear the marks of Vespasian, Titus and Domitian and are probably the remains of a sunken cargo; a lead ingot (Figure 11.2), mined and smelted in Flintshire, was found alongside a jetty at Chester.

Whilst both Pliny and Diodorus Siculus describe British tin exports, there is only limited archaeological evidence for tin mining during the Roman period. It is probable that one of the functions for the small fort at Nantstallon was to control the local mineral resources; there is significant evidence of metalworking within the fort (Fox and Ravenhill 1972, 110). Tin bearing lodes are known from Mulberry, near to Nanstallon and there is some evidence for native tin-working. Excavation at Killegrew Round revealed an “unusual hour-glass-shaped furnace containing rich slag” and part of a 95% pure, hand-beaten plaque of tin was also found (Britannia 1998 29, 423).

Figure 10.2 Lead ingot from Chester Copper There is clear evidence for copper mining at Llanymynech, but the proximity to the so-called siege works at Abertanant and the “supply base” at Llantsantfraid, led Jones (1988, 417–9) to suggest an association with the “Last Stand” of Caractacus. An alternative view is the direct association of a military presence with mineral extraction, as is evidenced at both nearby Brompton, and further afield at Charterhouse-onMendip and Dolaucothi. There is evidence of Roman copper mining at Parys Mountain on Anglesey, and on the north coast of Wales on the Great Orme. The abandoned mining settlement at Dylife is close to the isolated fortlet at Pen-y-Crocbren, and is reputed to be associated with Roman mining; possible evidence for extraction from the Plymlimon foothills has been found at Cwm Ystwyth and Trefeglwys. Copper deposits also occur at Alderley Edge and the Peckforton hills in Cheshire, in Cumbria at Alston Moor, Coniston and Keswick (Tylecote 1976). There are undoubtedly many other sites of Roman period extraction for which the evidence has been obscured or eradicated by multi-period activity ranging from the medieval to the early 20th century.

Precious metals There is evidence of the extraction and processing of gold (notably at Dolaucothi), and the de-silvering of lead on Mendip and Cornwall, but these commodities are unlikely to have been locally transported by water, as the weights were small and the value considerable. Iron mining and smelting Some indication of the demand for iron, during Roman times, is provided by Shirley’s estimate that 12 tonnes of iron (requiring 72 tonnes of ore) would have been needed to produce the Inchtuthil hoard of almost one million nails. The fortress also needed iron furniture for its c. 3,000 doors (2001, 107–8). In a civilian context, the extent of iron working at the small town of Worcester is indicated by the fact that many of the streets were surfaced with iron slag (Burnham and Wacher 1990, 234). There is evidence of Roman mining of the deposits of iron that are widespread throughout Britain, ranging from Cornwall in the south to the valleys of Cumbria in the north. As an illustration of the Roman expansion of the iron industry, Tylecote (1976, 53) has observed that whereas the amount of slag found on pre-Roman iron smelting sites is measured in kilograms or hundreds of kilograms, the slag heaps on the Roman period are measurable in hundreds of tonnes. Near to the small town of Ariconium (Weston-under-Penyard), the slag tips covered 200 acres and were so large that, in the 16th and 17th centuries, re-mining them became a commercial

Tin Writing in the first century BC and probably basing his observations on information from the voyage of Pytheas, Diodorus Siculus (5.22) commented that “In Britain the inhabitants of a promontory called Belerion are particularly friendly to strangers and have become civilised through contacts with merchants from foreign parts. They prepare the tin, working the ground in which 98

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  undertaking. Manning (1979, 117) has suggested that the increase in the Roman period was achieved, not by new mining techniques, but by greatly increasing the number of mines, of types so simple “that a Neolithic miner would have regarded them as amateurish”.

and both Willan (1938) and Hussey (2000) have studied he data for the 17th and 18th centuries, providing much valuable information for this period. Whilst recognising the danger of transferring data over such a long period of time, it is suggested that information from these sources may, at least, provide some very general indication of the cargoes that could have been involved.

A number of iron-making sites have been located along both banks of the Severn Estuary; the spreads of slag are limited and, with the exception of the Chesters (below), there is little or no evidence for any masonry buildings at these locations. Sites of this kind extend from Oldburyon-Severn on the south bank, to Rumney Great Wharf on the north bank. The proximity of the river suggests that the iron was probably moved by water, perhaps upstream along the River Severn, to sites such as Worcester (Burnham & Wacher 1990, 232). The Roman villa at Woolaston in the Forest of Dean, near to the banks of the River Severn, was excavated by Scott-Garrett in 1938. Fulford and Allen investigated the site of an enclosure close to the villa from 1988–90; these excavations revealed two furnace sites, with geophysical survey indicating the probability of other furnaces in the immediate area. There is evidence of a highly organised enterprise for the production of iron blooms, covering an area of about 7250 square metres. Based on the accumulation of waste in the area, the excavators estimate an annual output of between 0.6 and 1.8 tonnes. The working of the blooms into billets probably took place near to Ley Pill; the iron was shipped in billet form, rather than as finished articles (1992, 159–215).

In 1721, Chester shipped 2,486 tons of lead of which 1,191 tons were sent to London and 920 tons to Bristol. The Mendips are said to have been in their prime in the first half of the 17th century (Willan 1938, 72) and in 1685, shipments of lead from Bristol amounted to 74 tons. Surprisingly, in 1735, Bristol shipped 484 tons of lead, of which 405 tons went to London, but some of this was probably re-shipment of lead from Chester. The average of annual shipments of lead from Bristol was 187 tons, for the period from 1695–1701, with the average consignment being c. 25 tons. Although a small quantity of copper was smelted at St Ives, the bulk of copper ore mined in Devon and Cornwall was shipped to Bristol, Monmouthshire and South Wales for refining. The total shipments from Padstow, Truro, St Ives and Plymouth totalled some 2,400 tons in 1717, with forty ships said to have been employed in the trade. Chester shipped 43 tons of copper ore to Warrington, where Cornish ore was also imported. The re-emergence of the Forest of Dean as an ironworking centre is demonstrated by the fact that in 1632, Chepstow shipped only 32 tons of iron, but 50 years later, its production is the tonnage had risen to 1372 tons. Of this total Bristol received 628 tons, Newport 488 tons, Gloucester 157 tons and Carmarthen 71 tons. Aside from that needed for local consumption, considerable quantities were re-shipped from Bristol to a diversity of twenty-three destinations as far apart, as London, and Ayr on the west coast of Scotland (Willan 1967, 55-77).

A proportion of the iron production of the Forest of Dean served Imperial needs but, unlike the Wealden industry where the classis Britannica played a significant role, the re-interpretation (Hassall 1980, 82) of the inscription on the Lydney temple mosaic, means that there is no direct evidence of any military or imperial involvement in the organisation of the Forest of Dean industry (Fulford & Allen 1992, 199). Roman iron working on the Blackdown Hills, the Bredon Hills and on Exmoor has long been suspected; a site at Burcombe is known as the “Roman Lode”, but the problem of identification is that later mining of the same ore body often destroys the evidence. Recent investigation of associated charcoal from a slag heap at Sindercome, south of Twitchen, has given a radio carbon date of 170 BC – AD 75, and there is sufficient evidence from coins and pottery to indicate significant Roman exploitation of the iron bearing lodes (Riley & WilsonNorth 2001, 79–81). Distribution of Exmoor iron may have been achieved through ports or anchorages on the Taw estuary or Combe Martin Bay.

Salt Salt, being soluble, leaves no trace in the archaeological record and recognition of the patterns of production is dependent on the evidence from briquetage (the coarse pottery used in the production process). Transport of salt took place in special ceramic vessels, for example those made in the Malvern potteries, as containers for the salt produced at Droitwich, but distribution patterns have been, as yet, subject to only limited investigation (Woodiwiss 1992, 183–6, Morris 1994, 371–393).  The long held presumption that salt was an Imperial monopoly has been challenged by Hilary Cool (2006, 57– 8), who comments that there is no evidence of how the industry was organised in Britain and that the pattern of widespread, dispersed production is at odds with a model of a commodity held under monopolistic control. Using information from the Vindolanda tablets, she calculates that bacon was approximately 15 times the price of salt, arguing against it being a significant source of revenue for the state, and wonders whether the view of state

Data from 17th/18th-century Port Books We have no way of knowing the quantities of metalliferous minerals extracted, processed and shipped during the Romano-British period. However, Exchequer Port Books provide some details of shipment of cargoes,

99

JAMES ELLIS JONES

  control that has arisen was unconsciously adopted from the experience of British rule in India.

of a number of sites of small-scale production during the first/second centuries.

There is evidence for a significant salt production industry in the Central Somerset Levels from the late first to mid-fourth centuries. Briquetage mounds were first recorded in the early 19th century, and 169 sites have been identified, mostly in the area of Gold Corner and the River Gripps. The modern Huntspill Cut provides a section through the upper part of the alluvial sequence and briquetage mounds, buried at depths of 0.3-0.6 m below the present ground surface, have been recorded. The mounds extend westward for some 2 km beyond the area where they survive as earthworks; their burial may reflect the extent of the late/post-Roman inundation and indicate that significant remains of the industry are present below the level of the Clay Belt. Concentrations of Roman material in Highbridge probably indicate salt production on the banks of the proto-Brue river (Rippon 1997, 66–71). In the Bristol Channel, salterns appear to have been restricted to the Somerset Levels and only at Badgworth and Banwell is there evidence for salt extraction from the Late Iron Age. This is in contrast to many other areas where there is evidence for widespread salt production from the Early Iron Age and, in some cases from the Middle Bronze Age. It seems probable that the marked expansion of production in the Somerset Levels was a response to the demands of the fortress at Caerleon and the forts of south Wales and this would have been met by shipment across the Bristol Channel.

In addition to those noted above, a number of sites of limited salt production have been identified around the coasts of Britain at, for example, Trebarveth and Carngoon Bank in Cornwall, Ingolmells and Skegness on the Lincolnshire coast, on both sides of the Humber Estuary, Port Einon on the Gower peninsula of South Wales and St Monans in East Fife, Scotland. Droitwich lies on the River Salwarpe, not far from its junction with the Severn, and Roman roads from Metchley, Gloucester, Alcester and possibly Wroxeter met here. The brine springs were used in the Iron Age, and it was probably this industrial basis that led to the foundation and growth of the Roman town. Droitwich brine is saturated at about 25% sodium chloride and will therefore produce 0.294 kg of salt from 1 litre of brine. By comparison with the Somerset Levels, the same amount of seawater will produce only about 0.031 kg, and the economic advantages of the Droitwich area are immediately apparent (Woodiwiss 1992,4). Seven brine settling tanks dated to the early first century have been found, the largest being capable of the production of 4.17 tonnes of salt at a time and the smallest 1.56. Land routes for packhorses, known as “saltways” formed an extensive network, but during the Roman period, it is probable that the River Salwarpe was improved to enable access to the Severn. The Salwarpe was navigable to Droitwich in the Middle Ages, as is shown by the fact that in 1378 Richard III granted the right to levy tolls on the river to the bailiffs of the town (Rolt 1950, 32). The trade in salt continued to the end of the 19th century and salt was shipped from Droitwich both upriver, and to many ports in the Bristol Channel, in specialist trows that were known as “Wich” barges.

The marshes of Kent and Essex have produced significant evidence of Iron Age and Romano-British salt production; mounds of thick layers of burnt clay and sand have led to the salterns being known as the “Red Hills”". Major centres of production were located around the Thames and Medway estuaries with, further to the north, areas of activity to the south of Colchester cantered around the Blackwater and Colne estuaries. Although it is probable that large-scale production had ceased by the late second century, some fragments of troughs and firebars have been recovered from Chigborough on the Blackwater estuary, indicating that production continued into the third and possibly fourth centuries, albeit on a limited scale (Wallis & Waughman 1998, 164). There were large areas of salt marsh surrounding the estuaries of the Welland and Nene and, whilst not on in a semiindustrial scale, these were exploited at small centres, declining by the late second century with only handful of sites continuing in operation during the third and fourth centuries (Hallam 1970, 70).

Central Cheshire is endowed with many salt-rich brine springs, for example at Northwich, located at a strategic position on the west bank of the navigable River Weaver near to its confluence with the River Dane. Whilst there is evidence for industrial activity in the form of a pottery kiln and ironworking furnaces, there is no doubt that the major activity was the exploitation of the local brine springs. This was demonstrated by the discovery, in the 19th century, of five lead pans for use in salt working together with a nearby first-century brine kiln. Middlewich is another of a number of places in Cheshire where, during the Roman period, salt was produced on a semi-industrial scale, and finds include numerous pieces of briquetage, brine pits, evaporating pans and other items of salt-making equipment.

On the south coast, a significant area of production was concentrated around Poole Harbour and Purbeck, where production appears to have commenced in the late Iron Age and continued until the fourth century. Salterns are also known from the Solent and around Portsmouth and Chichester harbours, but in these areas production appears to have ended by the mid second century. Further east, in several areas of Romney Marsh there is evidence

Stone The enormous volumes of material needed for the construction, and periodic reconstructions, of forts and fortresses led to fluctuation in the volume of shipping. In the west of Britain construction was aided by the fact that

100

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Kenchester, on the River Wye on the opposite side of the Severn Estuary.

building stone was readily available in most areas and, because of estuarine or riverine siting, sand, pebbles and cobbles were often locally available. This is in marked contrast to the North Sea coasts where the construction of the Saxon Shore forts in the later second and third centuries required the transport of stone over considerable distances (Allen & Fulford 1999, 163–84). This was also the case with the building of the walls of Londinium in the early third century, where Marsden (1994, 126–7) has estimated that 35,000 cubic metres of stone were used in the construction of the wall. This would have needed 1,750 voyages, by ships of similar size to the Blackfriars ship, to bring about 45,000 tonnes of ragstone from the source, some 112 km away at Allington on the west bank of the Medway. A number of pieces of architectural stonework, carved from oolitic limestone, have been found, dating mainly to the early third century and the stone has been provenanced to Roman quarries close to the rivers Welland and Nene. In view of the considerable quantity, size and weight (from 0.122 up to 0.869 tonnes), transport by road is unlikely and it is probable that the stone blocks were taken downstream by barge for loading onto sea-going vessels for shipment to London (Marsden 1994, 23).

Stone from local sources was used when available, for example, the Corinthian capitals and columns of the forum-basilica at Caerwent were made from Sudbrook sandstone. At Loughor, limestone, Millstone Grit and Pennant Sandstone were locally available and were used throughout the fort; sand for mortar and cobbles and gravel for the roads and foundations were obtainable from the nearby beaches.

The oolitic limestone quarried on the Downs to the south of Bath was in great demand, because of its suitability for fine carving and architectural detail; at Caerleon the columns and column bases of the basilica were executed in this material. It was widely distributed throughout southern Britain, being used at London, Colchester and Silchester; in the Cotswolds at Cirencester and numerous villas, in The Marches at Kenchester and Wroxeter, in the north at Caernarfon and Chester, as well as at the forts and villas of south Wales. Because of the very considerable quantities and weight involved (a cubic metre block of limestone weighs 2.4 tonnes [Marsden 1994, 23]), it is probable that the stone was shipped on barges down the Avon to Sea Mills (Portus Abonae) and then loaded into sea-going vessels for dispatch to its final destination.

Figure 10.3 Artist’s impression of the legionary fortress, canabae, bridge and River Dee at Chester Mason (2002, 91–7) has produced data estimating and summarising the quantities required for the fortress at Chester. The building of the Flavian fortress during the 70s presented the greatest challenge, as the construction took place on a greenfield site, without the benefit of an existing transport infrastructure. It is probable that as part of the original construction, only the baths (requiring 25,000 tonnes of stone) were built of masonry and concrete, and all other buildings were constructed of timber on stone foundations. At Caerleon, the situation would have been similar, the Old Red Sandstone outcropping beside the tidal River Usk, just downstream of the site; being the predominant stone in structures of all phases of the fortress. The nearest source of freestone was the Sudbrook Sandstone, with a coastal outcrop at Sudbrook near Caldicot, Gwent, with thick beds of the stone forming a low cliff (about 5 m high) around the site of the Iron Age promontory fort (Allen 2005, 21–44). It would have been convenient for flat-bottomed boats of the period to “take the ground” before high water, be loaded with stone, floated on the next incoming tide, and then taken by sea and river the 30 km or so to Caerleon. Large blocks of Sudbrook sandstone was used in the late first-century amphitheatre and the legionary bathhouse, where Purbeck Marble was also used as facing panels. The construction of the legionary baths required large quantities of lime, necessary for the original concrete construction, and the Liassic Limestone, outcropping on a limited scale at Goldcliff (8 km south of the fortress), and more extensively in the sea-cliffs between Penarth and Southerndown, could well have provided the quantities

At Chester, shipping of stone both in and out of the port is demonstrated by the amount of Welsh slate (identified as originating at quarries near Caernarfon) in stratified Roman contexts. Traffic in the opposite direction has been confirmed at Caernarfon, where the rebuilding in stone of the fort took place in stages from about 140 onwards. The construction took place using Lower Keuper sandstone imported from quarries in the area of Chester. However, from the third century onwards, local resources of Carboniferous sandstone and limestone were used for construction of the praetorium and other buildings. Building materials were brought to the fort at Caerhun by sea and river; for example, a sandstone column from Bodysgallen, near Llandudno, a sandstone base originating from Runcorn in Cheshire, and a plinth from the angle of the West Gate was identified as having been quarried at Tattenhall, also in Cheshire. Oolitic limestone was shipped from the Cotswolds and used for columns and architectural detail in the civitas at

101

JAMES ELLIS JONES

  required, being loaded onto vessels for transport along the south Wales coast and up the River Usk.

from the fort. It was concluded that most stone appears to have been obtained from within the 30 km radius.

At nearby Caerwent the coastal outcrop of sandstone lies only 2 km from the civitas; the forum-basilica of the early second century used large quantities of this stone for column bases, steps, gutters and box drains and also in the main temple of the town of the early fourth century. The outcrop lies close to the Nedern Brook and Pearson (2006, 120) suggests that this small waterway may have been used for the transport of stone to Caerwent. There has been a persistent local tradition that Roman ships were able to reach Caerwent via the Nedern Brook, and that mooring rings, enabling vessels to tie up to the walls, were in existence until modern times. However, Waite has convincingly argued (1990, 13–22) that the brook could not have been made navigable and that the socalled “mooring rings” were probably used for tethering animals. Nevertheless, it may be of interest that the Llandaff Charters note that c.895 Bishop Cyfeiliog was granted “free landing rights for ships at the mouth of the Troggy (Nedern)” (quoted in Blair 2007, 40).

Over 75% of all raw materials for the Saxon Shore Forts building program exploited the coast as a source of stone (ibid. 91). Much of the stone, for example flint and septarian cement stones were, without any need for quarrying, collected from beaches and loaded into vessels drawn up on the foreshore. If not available directly on the coast, other stones were present in outcrops near to navigable rivers. The Kentish rag stone quarries that supplied stone to Bradwell, Reculver and Richborough all lay within easy reach of the River Medway. The volume of raw materials required for each fort varied considerably, ranging from 13,420 m³ at Burgh Castle, to that at Pevensey (with the longest and widest superstructure of any fort) being estimated at 31,620 m³. In order to present some order of guidance as to the comparative size of task involved, Pearson estimated that the total volume of raw material needed for the construction of the perimeter defences of the eleven Saxon Shore forts was in the order of 200,000 m³ (roughly equal to 500,000 tons of stone). By comparison, calculations for the curtain wall, mile castles and turrets of Hadrian's Wall required some 1,178,000 m³ of stone (ibid. 87).

Silchester is not on a navigable river and lacked suitable building stone in its immediate vicinity. Most structures were built primarily of flint; Bath Stone, from 90 km away, was used as the principal material for fine architecture in the columns and capitals of public buildings. The town wall constructed in the third century consumed approximately 40,000 m³ of stone and mortar, probable sources included Farnham in Surrey and Kingsclere in Hampshire, both some 20 km distant and Boon estimated that this equalled 45,000 cartloads of bonding stone and 105,000 loads for the core (1974, 1012). It has been suggested that, as Silchester was not situated on a navigable waterway, this led in the medieval period to the subsequent growth of Reading, some 15 km away on the River Thames, and the demise of Silchester as a significant settlement (Pearson 2006, 100).

Buckland’s (1988) survey of the building stones of Roman York represents the most complete examination of quarrying and supply in the North of England. No readily accessible building stone of any quality was available within 20 km of York, with the only locally available stone being Pleistocene boulders, only of use in rubble foundations and probably obtained during the digging of foundations and the fortress ditches. The main building stone used was Lower Magnesiam limestone with, for example, the fortress defences consuming approximately 14,000 m³ and it is probable that the defensive walls of the colonia required a similar amount. Millstone Grit was another significant building stone, and massive blocks of this material, some weighing more than 3 tons have been located, 250 such blocks being used for the sewers for the draining of the legionary baths. Various other buildings stones have been identified but not in large quantities. It is probable that the Lower Magnesiam limestone was quarried to the south west of York in the area of Tadcaster, where the Roman name of Calcaria may suggest a connection with limestone extraction. Tadcaster is located on the River Wharfe, requiring cartage of several km from the known outcrops, from where it could be shipped directly to the fortress. It is possible that outcrops adjacent to the Wharfe some 10 km west of Tadcaster, between Collingham and Boston Spa, could have been loaded directly from the quarries onto barges. The distance by river from Tadcaster to York is approximately 32 km but by road, over relatively level countryside, is only 15 km, so it is not clear whether either or both methods of transport were used during the Roman period. However, during the medieval period, stone from the quarry at Thevesdale, 2 km to the west of

A central theme of his study of The Construction of the Saxon Shore Forts was stated by Andrew Pearson to be the identification of the lithologies and, where possible, the associated provenances, of the stone used in their construction. In addition he set out, in the opinion of this writer with great success, to establish the scale of the shore forts as a building project and to determine the limits of what might have been possible in terms of speed of construction” (2003, 28). He concluded that the Romans relied on local, coastal resources for the bulk of the items needed for the construction of the forts and used the term “local” to denote an area within a radius of 30 km of the building site. In some cases, notably the case at Pevensey, outcrops of suitable building stone occurring within 1 km of the fort were quarried, providing 93% of the requirement. By contrast, at both Burgh Castle and Caister-on-Sea all suitable building stone was obtained from locations between 20 and 50 km from the site. The only fort receiving significant quantities of building stone from a distance of over 50 km was Reculver, where 18% needed to be brought from quarries at least this distance

102

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  water transport of small quantities of coal across the Bristol Channel is illustrated by the use of the Watchet “flatner”, a 5.5 metre flat-bottomed boat in common use on the north Somerset coast (Banks 1999). Within living memory, this craft was used to “pop across to Barry [on the south Wales coast] for a free load of coal” (John Nash, pers. comm.), showing not only the sea-going capability of this small craft, but also the effective crosschannel use of local water transport.

Tadcaster, was carted to the river and then shipped onwards down the River Wharf. Millstone Grit occurs to the north and west of York, for example in the areas around Knaresborough and Plumpton Rocks. The tributary rivers Swale (100 km), Ure (66 km) and Nidd to all pass through the Millstone Grit outcrops before they join the navigable River Ouse. Millstone grit also outcrops around Collingham and in the area of Leeds, from where stone might have been transported downstream along the River Aire to the Ouse and then upstream to York. The distance by road was only approximately 30 km and might have presented an acceptable alternative.

Products of the agricultural landscape Caesar commented that “the population [of Britain] is very large, their homesteads are thick on the ground and very much like those in Gaul, and the cattle are numerous” (BGall. 5. 12); Strabo referred to the production of grain and cattle, some of which was exported (Geog. 4. 5.2). Tacitus wrote, “except for olives, vines and other products of warmer countries, the soil will produce good crops. They are slow to ripen, although they shoot up quickly, both facts being due to the same cause, the excessive moisture of the soil and atmosphere” (Ag. 11). At the time of the Roman invasion, much of the landscape had already been extensively cleared of woodland and converted to agriculture, but there seems to have been some further clearance in the Roman period, presumably to increase the area available for pasture and for cultivation of crops.

The stone for the bridge on Hadrian’s Wall at Chesters was locally available with outcrops of sandstone suitable for the large rocks used in construction being available from outcrops 2.5 km to the east. Collingwood Bruce thought that the source of the stone was the Black Pastor quarries which were still being worked in his time (1867, 142). Limestone and ironstone are also exposed at Brunton (ibid. 445). Outcrops of limestone on the Isle of Purbeck have, because of its visual similarity to a true marble, become known as Purbeck Marble. It is particularly useful as a suitable stone for use in inscriptions and these have been found at, for example, Verulamium, Silchester, Colchester, Cirencester, Caerleon and Chester, perhaps the best-known being of the “Togidumnus” inscription found at Chichester. Oolitic limestone from the southwest was used for the monument to Julius Classicanus, first-century Procurator of Britain.

The first two hundred years of Roman rule saw the continuation of the agricultural pattern established in the later Iron Age, but a winter, as well as a summer wheat crop, was introduced, having the effect of allowing an increase in the number of livestock kept year round. The introduction of a plough with coulter, share and mouldboard provided a better seedbed, obviated cross ploughing and could be used to break up the heavy clay soils that had resisted earlier attempts at cultivation. A number of new or improved plant varieties were introduced, including rye, oats, vetch and flax, as well as cabbage, parsnip, turnip, carrots and fruit trees such as the vine, plum, apple and mulberry (Millet 1990, 91–9).

Coal The use of coal became widespread in Britain during the Roman period, and has been reported from excavations at over 200 Roman sites, in association with, for example, metal-working activities, hypocausts and a cremation pyre. In order to establish provenance, samples from 70 localities have been analysed by Smith, using chemical and/or microscopic methods (1996, 373–89). Coal deposits are widespread; there are coalfields in north and south Wales, Shropshire and Worcestershire, the Forest of Dean, the Coalpit Heath Basin in south Gloucestershire, the Radstock Basin in Somerset, and in Lancashire and west Cumbria.

Grain In his detailed study of the grain market in the Roman Empire, Erdkamp pointed out that “transport over land depended on the energy provided by humans and animals, transportation by water largely on wind and current” (2005, 10) and this meant that “long distance transportation of grain overland was impossible, because the input required by way of human and animal energyand, worse, consumption would make it impossibly ineffective”. He further considered that the high cost of transportation overland virtually ruled out the conveyance of large amounts of corn across any meaningful distance inland (ibid.199–201).  Manning (1975, 112–6) challenged the traditional view of a largely pastoral highland zone, arguing in favour of important pockets of arable production in Wales (this was the case until the 19th century, when reduced transport costs, enabled by the

Coal has been recovered from many military sites, including Carlisle, Lancaster, Manchester, Caerwent, Caerleon, and Usk and from villa sites at Llantwit Major, Great Witcombe, Chedworth, Chew Valley, Star, Gatcombe and Marshfield. Coal from the nearest coalfield (at Wigan it was underneath the town itself) could have been transported to many sites using the known Roman road system, but some coal samples from Caerleon and Llantwit Major, have been provenanced to the Forest of Dean, clearly indicating the use of water transport (Smith 1996, 373–89; 1997, 297–324). Recent

103

JAMES ELLIS JONES

  advent of the railways provided a source of cheap grain). Heslop (1987, 117–20) showed that seed assemblages from the northeast of England were indicators of a mixed economy, with significant arable components. Manning’s view, that the basic demands of an auxiliary unit could be met by the local population being stimulated to move towards a system of surplus grain production, was not shared by Higham and Jones (1975, 16). They point out that evidence from the Solway region indicated that the Roman influence did not involve an accumulation of native settlements around forts, and that the distribution pattern of native sites was primarily dictated by natural and geographical conditions, such as flood-free levels, or altitude.

consequent reduction in the possible shipping period. However, it would be possible to minimise the effect by completing shipments to tidal and coastal locations in the late summer/early autumn months, leaving consignments to the distribution points until the late autumn, or even early winter. In this way, out-of-season voyages would be avoided and, equally importantly, river levels would be high enough to permit navigation to the upper reaches. For example, in the early Flavian period, the west coast forts at Tomen-y-Mur, Brithdir, Pennal, Pen Llwn, Trawscoed might be supplied first, the north coast forts at Prestatyn, Caerhun and Caernarfon and the south coast forts at Carmarthen, Loughor, Neath and Cardiff next, then the legionary fortresses at Chester and Caerleon. Over the three-month period from August to the end of October, and assuming two voyages per vessel per month, this would give a requirement for 40 vessels of the Blackfriars type. Unlike later periods, there would be no need to wait for availability of cargo or return loads, and the number of voyages might well be increased. If 50% of the remaining grain consignments were loaded onto barges, near to the source of supply, and delivered direct to the destination, there would be a remainder of 160 consignments. This could be met by 10 vessels operating in a similar manner to the Severn trows of later periods, plus the redeployment, from the beginning of November, of the ships previously used for coastal supply.

The areas surrounding Hardknott fort or the isolated fortlet at Pen-y-Cochbren would have been incapable of the supply of even the most meagre amount of cereals, but the valleys and plains of Wales, The Marches, Cheshire, Lancashire and the Solway may have been capable of satisfying some of the demand for grain for the nearby auxiliary forts. However, the relative ease by which large quantities of cereals from the fertile English lowlands could be transported by water, even to remote garrisons such as Pennal on the Welsh coast, Maryport on the Cumbrian coast, the forts of Hadrian’s and the Antonine Wall or far into the interior, is of considerable importance. Supply of grain from a distance may have been the normal case at Loughor, and Bidwell has pointed out that granaries of the supply base at South Shields (Arbeia) continued in use throughout most of the last two centuries of Roman rule. It is probable that Carlisle fulfilled the same function for the garrisons of the west. Bidwell suggests that although forts obtained some grain from the immediate localities, a supplementary system of long-distance supply was still necessary (1997, 85–7).

Grain for beer at Vindolanda In addition to wine, the Vindolanda tablets show that beer was consumed in some quantity in the praetorium of Flavius Cerialis, the garrison commander. For example, on one day four modii (a little over thirty litres) was consumed. That the taste for beer was shared by the “other ranks” is epitomized in a letter from the decurion Masclus (93/1544), who states that “the comrades have no beer” (cervasam comilitiones non habunt) and asks Cerialis “that you order it to be sent” (quam rogo iubeas mitti). Birley has calculated that an as, the smallest coin in use, would have bought the equivalent of over 11 pints of beer, so the purchase of similar large quantities would have been within the reach of the soldiers of the regiment.

Recent, and well-published, excavations at Caernarfon (Casey & Davies 1993) and Loughor (Marvell & OwenJohn 1997) have provided some factual information on the types of grain consumed. The cereal assemblage at Segontium, in north Wales, was dominated by spelt, but with a significant quantity of barley (c. 30%) being present. Bread wheat was present from the mid to late second century, but never seems to have been particularly important. The cereal assemblage from the fort at Leucarum (Loughor), on the coast of south Wales, is of a different character, with a high incidence of bread wheat (of probable non-local provenance) occurring in late first/early second century contexts. In this case, the fort’s estuarine location and ready access to a short sea-route for obtaining agricultural supplies from the fertile West Country were major factors, if grain was obtained from this source.

Barley for animal feed “Hard” fodder, usually either barley or oats, was provisioned and distributed in a similar manner to wheat for the soldiers. It is probable that Roman cavalry horses were slightly smaller than their modern equivalent, and a reasonable estimate of their requirements is approximately 2.5 kg of hard fodder and around 7 kg of green fodder per day. Whilst a significant part of a horse’s dietary requirement could be obtained by grazing, a Roman warhorse needed to be fed some grain in order to remain in good condition (Roth 1999, 62–5). During the Peninsula War, Wellington's mules were given 2.3 kg of barley and 4.5 kg of hay per day and they performed very well on this diet (White 1984, 128–9). Because of their large size, oxen require large quantities of food, and

The requirement for grain for an auxiliary fort was some 250 tonnes (8 shiploads) and for each legionary fortress 2500 tonnes (80 shiploads). Obviously, grain is not available for shipping until the harvest has been gathered in (traditionally by Lammas Day, 1st August), with a

104

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  iron working; 2,800 for lime burning and 31,000 for tile kilns) would be needed to produce basic building materials for a timber-phase fortress (2001, 106–7).

Roth (1999, 66–7) has estimated that this would amount to 7 kg of hard fodder, 11 kg of green fodder or, because oxen can obtain a large percentage of their nutritional requirement through grazing, approximately 22 kg of pasturage.

Hanson’s comment that, in a newly subjugated area, timber would not be hauled any further than was absolutely necessary (1978a, 297–8) invites the question as to whether local felling and planking was always employed. As all permanent British fortresses were located on either an estuary or a navigable river, and as this was also the case for many of the forts of the study area, it is probable that water transport was employed. Timber could then be felled and processed in a secure rearward area (for example, the Forest of Dean), and then shipped direct to the fort, ensuring a much earlier operational readiness, and enabling the garrisons to undertake a military, rather than a constructional role.

The ink-written tablets from the fort at Carlisle (Luguvalium) were written during the period between 72/73 and 125, when the fort was probably occupied by the ala Gallorum Sebosiana, a 500-strong auxiliary cavalry unit. Of particular interest is a series of documents detailing the issues of barley and wheat to the sixteen troops (turmae) of 30 cavalrymen and their horses, plus the mounts of headquarters staff and remounts. From the documents, Tomlin determined the daily consumption of barley per horse per day as 1.96 kg, an annual total of 715 kg, and therefore a unit consumption of some 392 tonnes per annum (1988, 49).

The manufacturing industries In a military context, barley is normally considered for use either as animal feed (Pliny HN 18.15.94), in brewing or as an emergency/punishment ration, but Davies (1997, 269) suggests that, particularly in the fourth century, at Segontium, it could represent locally grown grain consumed by a garrison long-accustomed to a readily available staple crop. Stallibras and Thomas (2008, 164) support this point of view and go further in suggesting that classical tests making recommendations in the core of the Empire might be less relevant to peripheral areas. However, the diet of the Roman soldier, even when belonging to a barbarian tribe has been considered by van Waateringe to be “composed of all the elements that a true Roman would have expected” (1997, 264).

Brick and tile manufacture. Tileries were established in many areas in order to meet the requirements of nearby forts. In the north, for example, a tilery at Scalesceugh supplied Carlisle, Muncaster served Papcastle and Maryport, and Quernmore supplied Lancaster. Finds from the brick and tile depot at Holt, near to the fortress at Chester, indicate that it was not established until the rebuilding of the fortress in stone at the beginning of the second century, and therefore would not have been available for the initial production of the 70,000 bricks and 73,000 roof tiles (Mason 2002, 91). The options for this task were therefore the construction of a tilery on site, land transport from established tileries by some 2100 wagonloads, from perhaps Wroxeter or further south, or by sea in vessels of the Blackfriars or St Peter Port type, with probable cargoes of 50 tonnes, and therefore some 35 voyages. Although we have no evidence as to which alternative was employed, it does seem probable that the most likely was production by a tilery specially constructed on site, as the necessary raw materials were available on nearby. A small civilian tilery at Tarbock, 13 km east of Liverpool, was used as a stopgap to manufacture roof-tiles for the fortress at Chester at a time (c. 167) when the legionary tilery at Holt was not in use. It has been suggested that the tiles were transported from Tarbock to Chester by water, using the estuaries of the Mersey and Dee. The distance of some 7 km from Tarbock to the Mersey shore could have been traversed by using the tidal inlet of the Ditton Brook (Swan and Philpott 2000, 57/8).

Other foodstuffs Significant quantities of oyster shells are found on sites well inland, and speedy transport, packed in casks filled with salt-water and seaweed, would have been necessary for them to arrive in fresh condition. Freshwater fish would also have provided a return cargo for riverboats transporting inland goods to ports. Timber On palynological evidence, Hanson (1978a, 293–305; 1996, 354–60) and Dumayne (1994, 217–224) disagree as to the extent of forest clearance, in northern Britain, during the Roman period. However, they agree that because of the high cost of transport, timber would, wherever possible, be obtained locally and also agree that local sources would often be sufficient to meet the demand. Hanson estimates the total requirement for a timber fort (palisade, gates, towers and internal buildings) as 22,000 cubic feet (662 cubic metres) and that to obtain this volume of structural timber would require felling of an area of between 6.5 ha. and 12 ha. Elizabeth Shirley has suggested that about 70,000 trees needed to be felled to provide the variously sawn timber lengths used in the construction of the legionary fortress at Inchtuthil. She also estimates that 36,000 tonnes of timber (2,000 for

Pottery Black-burnished pottery became a dominant kitchen ware in Britain and, from the mid third century to the late fourth century was used for almost all bowls, dishes and cooking pots, “being commoner than all other wares put together” (Gillam 1973, 54). Black-burnished ware 1 (BB1) was predominantly made in southeast Dorset and

105

JAMES ELLIS JONES

  distributed in the west of Britannia, in many cases, by the rivers and coasts of the west. It has been suggested that the production of BB1 was driven by the demand for Dorset salt (e.g. Allen & Fulford 1996, 268) and, more recently, Gerrard has argued the links between salt and pottery production in Poole Harbour would appear to be remarkably close. Both potters and salt workers needed large quantities of fuel, the former to fire pots and the latter to boil brine (2008, 121).

Gloucester to the northern garrisons took place via the Western Seaways (Webster 1976, 18-46). Fulford & Hodder (1975, 26-33) used regression analysis to examine the relationship between the distribution of New Forest and Oxford wares that are contemporary and similar in quality and utility. They concluded that one of the factors leading to a wider distribution of Oxford ware was due to the availability of water transport via the Thames, thus enabling access to the large cities of Cirencester and London. The New Forest kilns produced a wide variety of pottery including mortaria, coarsewares, colour-coated beakers and jugs, white flagons and bowls with dark red painted decoration known as “parchment ware” and colour coated ware based on terra sigillata forms. Fulford has demonstrated that each variety had different distribution characteristics, reflecting their fineness and technical complexity. Coursewares were restricted to an area within 40 km of the kilns, whereas elegant pieces of colour-coated tableware have been found up to 100 km distant from the production centres. Fulford convincingly argues that this distribution reflects the fact that the higher quality pottery had a higher market value, and could therefore absorb a higher transport cost from the than the more bulky and less expensive coarse ware (1973, 161-178).

Black-burnished ware 1 (BB1), produced near the Poole Harbour on the south coast, may have been successful on Hadrian's Wall because a merchant was shipping grain and textiles from the south-west to the same area, and pottery was piggybacked onto these cargoes (Green 1979, 103). By contrast, Black-burnished ware 2 (BB2) was made and distributed (mainly via the Thames estuary) in the east, with very little overlap other than in the northern frontier region (Allen & Fulford 1996, 223–81). Gillam considered direct shipment from the potteries to the Antonine Wall as the most likely option, commenting, “It is as though ships from the Thames estuary unloaded none of their cargo in the Tyne on their way to the Forth” (1973, 57). Gillam studied the quantities of Blackburnished ware reaching forts in the areas of the Hadrianic and Antonine Walls and clearly demonstrates that, whilst both personal preferences and cost play a part in creating demand, BB1 was able to achieve market penetration of sites on the eastern side of Hadrian’s Wall. For example, in the eastern turrets, an average of 15 km from the point of unloading of BB2 at Newcastle, it accounts for 41% of the recovered ware. By contrast, on the Antonine Wall, Cramond showed 73% BB2 against 27% BB1 and Mumrills displayed a similar trend with 62% BB2, as opposed to 28% BB1 and at Balmuildy, less than 15 km from the western end of the Wall, BB2 still managed to account for 40% of the ware (1974, 9–24). At Bar Hill, 25 km from the west end of the Antonine Wall and about 34 km from its east end, a similar pattern, e.g. 220 bowls of BB2 and 97 of BB1, was observed (Robertson 1975, 162–6).

Dales ware was produced in kilns near Dere Street on Humberside and Gillam has argued that in the fourth century this product effectively eliminated Midland and Southern competition, such as BB2 and Colchester mortaria, in the northern military market (1973, 61).  Movement in the opposite direction, from the Humber south to Kent, is demonstrated by the coastal distribution of Dales Ware that was being produced by the later second century, as well as, in more general terms, by the distribution of various types of coal from the North down the East Coast. There was large-scale production of mortaria at Wilderspool, and a high proportion of Wilderspool mortaria has been discovered at forts along Hadrian’s Wall. During the second century, there was also a wide distribution across south-west England, north Wales and southern Scotland. The Wilderspool potters Austinus, Docci. and DIS/LDB moved subsequently to the Carlisle (Eden Valley) region, where they produced very similar mortaria. It is not always possible to distinguish Wilderspool and Carlisle products on fabric grounds alone, but different dies were used in the two workshops (Tyers 1996, 134-5). Heavy mortaria were transported by river over wide inland areas, as well as by coastal shipping (Fulford 1982, 412, fig. 4).

Severn Valley ware occurs as a range of forms including jars, tankards and bowls, and is one of the most widespread classes of Roman coarse pottery found on sites in the west of England. Only a few kilns producing this type have been identified, these are not confined to the immediate area of the Severn, with possible sites being identified at Perry Barr, Shepton Mallet and Wroxeter, in addition to Gloucester itself. No major production site has, as yet, been identified and the industry would seem to be one of small-scale production of similar forms, in a similar fabric, but from a variety of centres. The main distribution area was concentrated in the catchment area of the river Severn, its estuary and main tributaries and the presence of Severn Valley Ware as far up-river as Caersws indicates that the river was used for transport. An exception is the supply, mainly of tankards, to the western sector of Hadrian's Wall during the second century, and it is probable that shipment from

The Nene Valley was well established as a major supplier of mortaria by 270 (Gillam 1973, 59), and Mancetter (Hartshill) ware was also important in the northeast. Fulford suggests that this may have been because they were the furthest removed from the Continent and the competition of Central Gaulish and Rhenish potters (1977a, 309). Service roads leading from the Mancetter

106

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  including clay-lined pits suggesting evidence for tanning. The high number of silver coins, samian pottery and amphorae recovered suggest, particularly in the second century, an industrial centre under direct military control with army personnel in permanent residence.

kilns to Watling Street and from the Hardman complex to the main Harlow-Braughing road (Swann 1984, 19/48) suggest overland transport, and Hartley argues that the distribution of mortaria from Hartshill and Mancetter is clear evidence of the use, basically, of road transport but that the distribution to the Antonine Wall and Corbridge is clearly indicative of transport by sea (1973, 39). Gillam (1973, 57) commented that not only mortaria but many other identical types of pottery are recorded at both Corbridge and the Antonine Wall and that this kind of similarity, between two sites are so far apart, is probably without parallel elsewhere in the Province, at this period. However, he considered it unlikely that Corbridge was a ceramic depot for Scotland, as direct shipment to the north by sea transport was so much more efficient and economical that it was almost certainly used.

-o-o-o-o-o-o-o-oIt was shown in Chapter 9 that the rivers of the region were well able to sustain a high level of waterborne transport. For example, during the Romano-British period it is probable that the River Severn was navigable as far as the long-lived fort and vicus at Caersws, some 300 km from the Bristol Channel. Changes in the coastal and riverine landscape were shown, in Chapter 1, to have had little negative effect on the river systems and it is probable that sea levels during the later Romano-British period were beneficial to navigation. Chapter 3 demonstrated the wide range of coastal vessels, boats and local craft that were available to handle all types of cargo both within the region and from imports to the Province. A vessel similar to the Blackfriars ship would have been capable of reaching the legionary fortress at Wroxeter and beyond; the Barland’s Farm boat could have carried cargo to the fort at Caersws.

There is evidence that, in contrast to Hadrian’s Wall, the east coast was the major route of pottery distribution for the Antonine Wall. Gillam commented that the pattern of distribution of pottery changed as troop distribution changed and suggests that “The pottery ships no longer ran mainly from the Bristol Channel to the Solway Firth but from the Thames Estuary to the Firth of Forth” (ibid.).

Clearly, the volume of traffic in particular commodities considerably varied through time. For example, massive quantities of stone were needed for the episodes of fortress-building, followed by a period of reduced demand. Similarly, as a result of an extensive period of villa construction and the building of defensive walls, there was a floruit in the area of the River Severn and its estuary during the late third to mid fourth centuries. The movement of large quantities of grain to meet the demands of the Roman army and its dependents provided the greatest potential for water transport and many statements within the classical literature (e.g. Libanius. Ora. 18, 82–3; Appian. BCiv. 4.12, 100) suggest that this was the preferred option. Although some northern areas may have been capable of satisfying the demand for grain for the civilian population, in the early stages of conquest and consolidation of the study area, most of the grain for the army would have been supplied from southern Britain. It is probable that much of the widespread distribution of pottery to the north was the result of piggy-backing onto this type of cargo.

Ironworking Each Roman fortress had its own manufacturing facility, the fabrica. At Inchtuthil, the main workshop measured 58 m x 59 m with aisled halls on three sides of a courtyard and a range of large rooms on the fourth side. When the fortress was abandoned, a pit was dug in one of the rooms and almost 10 tonnes of unused nails were placed in it. A workshop of this size and capacity was, to all intents and purposes a factory, a form of building not present in Europe until the late 17th and early 18th centuries. Each auxiliary fort had its own fabrica, obviously on a much smaller scale. Most small towns and many vici have produced evidence of ironworking. Lying on the banks of the River Severn, the small town of Worcester was engaged in the metal industry, on a substantial scale, making use of processed iron from the Forest of Dean. A foundry with at least six smelting hearths was in use and the large quantity of slag used for road surfacing indicates that the processing of iron was a major activity in the town. A large number of hearths have been found at in the vicus at Manchester, either within buildings or in the open space to the rear. Many of these were used to forge weapons, tools or domestic equipment; there is also some evidence for the production of bloomeries. In a different category is the extensive area of industrial settlement at Walton le Dale, located near the junction of the former course of the River Darwen with the River Ribble (Belisama fluvis). At this point, the river is tidal and sea-going ships would have been able to dry out on the tidal mud flats. Excavation has revealed a number of rectilinear enclosures, together with hearth/furnace bases indicating iron smelting and other industrial processes, 107

JAMES ELLIS JONES

  value in economic analysis, if there is some indication of their place of origin, their likely contents and their date range. In addition to wine, amphorae were used for transporting commodities such as olive oil and fish sauces, and therefore provide an indication of the movement of certain foodstuffs of considerable economic importance, and a significant demonstration of part of the Roman way of life. The shape of amphorae may well have evolved in response to the need for effective packing, as they can be conveniently stacked in the hold of a ship for transportation, with the spikes of an upper layer fitting securely into the spaces between lower layers. Millet (1990, 158) commented that amphorae used for the transport of perishable products, such as wine and olive oil and fish products, provide direct evidence for the overall pattern of trade. He did however point out that even the distribution of amphorae cannot be taken to provide an unbiased measure of trade because of the widespread use of other containers like barrels, and the probability that changes occurred in the packaging used. He suggested that a reliance on amphorae tended to bias the archaeological evidence in favour of Mediterranean products against those from Northern Europe, which were more likely to have been carried in barrels.

Chapter 11 Imports and exports It is clear that there was an adequate transport infrastructure for large-scale movement of cargoes, both for the inter-provincial and the import/export trade. It is, however, unlikely that exports from Britain were of sufficient quantity to provide return cargoes for the imports of, say, olive oil and wine, and it is therefore probable that many vessels returned to the Continent in ballast. Wine in Britain We may be surprised at Roman wine consumption but, in The Ancient Economy, Finley (1973, 206) described Rome as a “fabulous” consumer of wine. We have some examples of consumption from the Roman period, for example, Horace suggests that a sextarius (0.54 litres) is a reasonable accompaniment to a meal (Satirae, I, 1, 74). Suetonius (Augustus, 77) describes the Emperor as a “habitually abstemious drinker”, never taking more than three drinks of wine-and-water at dinner; in later life his limit was a pint, and if he exceeded this, he would deliberately vomit. By contrast, Martial comments, “one is drunk on ten quincunces” (2.27 litres)! Tchernia reached the conclusion that in Rome, the annual consumption per head of the population ranged from 146 litres per annum (if the adult male drank 0.80 litre of wine per day), to 182 litres if he drank 1 litre per day (1986, 21–7). Wines to suit all tastes were available, ranging from very sweet wine from grapes of Cos and Rhodes to a particular variety of Falernian that was very dry. Variation also existed in quality; Pliny dismissing the wines produced in Pompeii, because they gave him “a headache that lasted till noon of the following day” (XIV.viii. 62).

A further problem is that amphorae are often underrepresented in published excavation records, for example, the monograph on the excavations taking place from 1875 to 1975 at South Shields contained details of 375 pots, pans, flagons and dishes, 235 items of samian, but only six amphorae were recorded (Dore & Gillam 1979, 98–163). Even in more recent excavations, such as those at Caerleon, the under-reporting of amphorae continues. The report on the excavation of the canabae (Evans 2000, 283) commented, “Given the very high proportion of Baetican olive oil sherds in this collection, it was considered neither useful nor economical of time and money to quantify the number of vessels by weight, or to attempt any statistical analysis of the quantities involved”. Other finds remain completely unreported, for example, the writer is aware of some 30 kg of Dressel 20 amphorae from an unpublished and ill-defined site, at Crandon Bridge on the south Somerset Levels, and there are undoubtedly many similar situations.

Diodorus Siculus commented that the Gauls are “exceedingly addicted to the use of wine” and that the merchants, who “transport the wine on navigable rivers by means of boats, and through the level plains on wagons”, receive a slave in exchange for a jar of wine, “thereby getting a servant in return for a drink” (5. 26. 2– 3). There is ample evidence for the appetite for wine in pre-Roman Britain but, from the finds of Dressel 1 amphora, Sealey has argued that the import of Italian wine peaked at the end of the first century BC and then declined by some 70% in the 50 years before the Roman invasion. He considers this to be the result of a growing demand for wine from within Italy itself and a consequent reduction in the amount available for export (2009, 1– 40). It might, somewhat facetiously, be argued that the wine-drinking habits of the native elite were enhanced, or at least maintained, by the influx of wine generated by the Claudian invasion!

After pointing out that properly quantified amphora data are not common, Hilary Cool (2006,61) produced a table (7.2) showing the presence or absence of amphorae in 15 sites including, for example, Longthorpe, Castleford, York, Canterbury, Dorchester, Cirencester and Greta Bridge. The site dates range from the mid first to late third centuries and demonstrate the presence of olive oil at every location but, surprisingly, suggest the unlikely absence of wine amphorae at Nantstallon, Chelmsford, Towcester and Cramond. Cool rightly considers that, in the quantification of amphorae finds, only recording by weight allows valid comparison across sites. She points out that quantification by minimum numbers is very dependent on the size of the assemblage, and sherd count is conditioned by site formation processes (ibid. 18).

Amphorae as evidence for trade Amphorae survive well in the archaeological record, but as Parker (1973, 363) has commented, they are only of

108

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  from the eastern Mediterranean and are most common in the third and fourth centuries, with some examples from 475–550 being found in western Britain.

Dressel 1 and Dressel 2–4 amphorae (30 litres) usually contained wine, produced in Italy, Iberia, central and southern France, and amphorae of these forms are found in pre-Conquest contexts. Some Haltern 70 amphorae (30 litres), found in pre-Flavian contexts, have similar stamps to those on Dressel 20 and suggest manufacture in the same Baetican workshops, but in this case containing wine, rather than olive oil. There is epigraphic evidence from Bordeaux of L. Solimarius Secundinus, a negotiator Britannicianus (CIL, xiii, 634), a probable supplier of the local claret to Britain. Marcus Aurelius Lunaris built an altar, dedicated in AD 237, in recognition of his safe return from York (JRS 11, 1921 101–7). Cunliffe (2001, 101) suggests that these two merchants were involved in exporting the local wine to Britain “along the route that had been used to carry wine to the island three centuries earlier”.

From the distribution of amphorae in the first century BC, there is evidence for the extensive export of wine from Italy. Cunliffe (2001, 100) considered that some of that product, taken by river and overland portage to the Gironde will have been loaded onto ships for onward passage to Amorica and Britain. For the Garonne route, Mediterranean cargo was loaded at the port of Narbo Martius then taken via the River Aude as far as Carcassone (Carcaso) where it was discharged and taken by road transport some 90 km to Toulouse (Tolosa) for transfer to barges on the navigable River Garonne (Garumna). At its confluence with the River Dordogne the Garonne entered the estuary of the Gironde; reaching the port of Bordeaux (Burdigala) after 22 km, where cargoes were again transferred, this time to sea-going ships, the journey from the Mediterranean to the Atlantic having taken between fifteen and twenty days. Cargo from the Mediterranean destined for the Loire was loaded at the mouth of the Rhône near Marseille (Massilia). Strabo (Geog. 4.1.14) describes the Rhône as fast flowing and difficult to sail up, and that wagons were used for traffic up the valley of the Gier and then across to the Loire, near to Burges (a distance of some 40 km), from where river transport was used to reach Nantes (Portus Namnentum), 55 km from the Atlantic coast. The total journey time from coast to coast was about forty days, but the reverse journey from Atlantic to Mediterranean was reduced by some ten days as, from Vienne, the fastflowing Rhone could then be used to advantage.

Figure 11.1. Distribution of Dressel 1 amphorae in Western Europe Rhodian amphorae (25 litres) are found on first-century sites, with the earliest examples from pre-Flavian military sites in southern Britain, where they are a common wine container and continue to the mid second century. The island of Rhodes was incorporated into the Empire as a punishment for the deaths of some Roman citizens, and the presence of significant quantities of Rhodian wine on British sites was probably due to consignments exacted as tribute, acquired and distributed as part of the annona militaris (Peacock & Williams 1986, 62). By the end of the second century, the major source of Continental wine importation to the east of Britannia was the Rhineland, and a lesser quantity was received from Atlantic Gaul. However, for the west of Britain, the more cost-effective Western Seaways should be considered as a viable alternative. Flat based Gauloise 4 wine amphorae (37 litres) are widely distributed in Britain, from the Flavian period and throughout the second and third centuries AD. British B4 amphorae (6 litres) probably contained wine

Figure 11.2 Major rivers and roads of Gaul Peacock’s map of shipment routes to Britain, based on Strabo’s first-century writings (N.H. 4.5, 2), includes continental departure points from the mouths of the Rivers Loire and Garonne. Peacock’s routes show entry to the English Channel, but the same arguments that have previously been advanced for direct shipment of olive oil 109

JAMES ELLIS JONES

  mentions of wine, beer and a whole variety of spices. However, finds of Dressel 20 amphorae show that olive oil was part of the soldiers’ ration (Birley 2002, 92).

via the Bristol Channel and Irish Sea hold good for wine. For example, this was certainly the case in the Middle Ages when, from these areas, Bristol handled 24% of England’s imported wine (Horton 1997, 12–13). Barrels

The province of Baetica, on the coast of Iberia, was the principal source of olive oil, with some 130,000 globular amphorae being shipped each year. Referring to Cadiz (Roman Gades), Strabo wrote, “Here live men who fit out the most and largest merchant vessels, both from Our Sea and the outer sea” (Geog. 3.5.7). Most production sites were located on the Guadalquivir River that followed a meandering, but navigable, course from deep inside the province, to the Atlantic near the Straits of Gibraltar. This provided an outlet for oil that could be shipped northwards either to Gaul, Britain or southwards into the Mediterranean (Greene 1986, 110). The Guadalquivir carried vast quantities of olive oil - the so-called Monte Testaccio in Rome is estimated to have contained the sherds of 53 million amphorae. At the peak period of production, towards the end of the second century, some 130,000 amphorae were exported each year to meet the annual demand in Rome for at least 7.5 million litres of olive oil. However, in the mid third century Spanish olive oil production declined, or at least ceased to be exported on a large scale.

Finds of barrels occur in favourable conditions, but are normally found in secondary usage, for example, as brine tanks at Droitwich (Woodiwiss 1992, 192–3); this provides little information about their original function. Wine barrels were often reused as well linings and a number of complete barrels of the first to second centuries have been found in London, and it is therefore possible to calculate their weight when full. The empty barrels weighed about 200 kg and the calculated volume would hold 966 litres of wine, weighing 966 kg, giving a total weight when full of about 1200 kg. Another barrel had a capacity of 550 litres and two barrels found at Silchester had capacities of 800 and 930 litres (Marsden 1994, 22). Remains of barrels have also been recovered from Newstead, Colchester, Bar Hill, Caernarfon, Silchester, London and Carlisle; the evidence suggests some were of continental (larch, silver fir) or Mediterranean (cedar) origin, as well as from local (oak) sources. Sealey (2009, 25) points out that, by comparison with amphorae, barrels are not well suited to the ageing and preservation of wines. He suggests that barrels would have been equally appropriate for the carriage of Celtic beer (ceresin) and, drawing attention to the fact that the Roman garrison at Vindolanda was issued with more beer than wine (Bowman & Thomas 1983, 83–93), suggests that this may well have been the case.

The Dressel 20 is one of the most common, and widely distributed, of all amphorae, especially in the western Roman provinces, with a major axis of distribution to the north, via the Rhone and the Rhine. However, it will be later argued that the Western Seaways would have provided a more cost-effective distribution route to the west of Britain. The Dressel 20 is a large globular form, with two handles and a distinctive plug of clay sealing the base of the vessel. Capacities range from 40 to 80 litres, with an average of 60 to 65 litres. The rim shape developed from more rounded forms in the first century to forms that are more angular in the third century; it is therefore possible to date the time of production with some degree of accuracy. The vessel is often stamped in relief with a name (often that of the estate owner), set in a rectangular frame on the summit of the handle, painted inscriptions (tituli picti) may sometimes be found on the shoulder, between the handles. On average, a Dressel 20 amphora weighed 28.42 kg and contained 62.83 litres of oil. Since a litre of olive oil weighs 0.91 kg, the average weight of oil in an amphora would be 57.17 kg. As the total weight of an amphora filled with oil would therefore be 84.59 kg, it would be difficult to lift and carry, even by two men each holding a porter’s bar (Peacock & Williams 1986, 52). It is therefore apparent that, whilst small quantities might be unloaded by hand, some form of lifting mechanism was desirable to load and unload a cargo. This could either be a shore-based crane, or by swaying out with a ship’s boom.

Figure 11.3 Stowing barrels on the Danube (Trajan’s Column) Olive oil Analysis of the remains of Dressel 20 amphorae from a number of late Iron Age sites in Britain, indicate that significant imports of olive oil, from Iberia, took place prior to the Roman conquest (Williams & Peacock, 1983). It has been identified throughout the study, in sites as far apart as Nantstallon in Cornwall, the island of Steep Holm in the Bristol Channel, Caernarfon and Rhoson-Sea on the coast of north Wales and the Antonine and Hadrian’s Walls. At Vindolanda, only one writing tablet (203) mentions olive oil, by contrast with frequent

A sixth-century Egyptian papyrus contained a ration scale of an eighth of a pint of oil per day This suggests that a soldier was entitled to 40 pints (23 litres) of olive oil per year, “although there is no certainty that this ration was 110

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  require half a shipload (13,000 litres) per annum, it is likely that cargoes from Iberia would be shipped to ports, such as Chester, Caerleon, Carlisle or Gloucester, and the oil re-distributed to the forts, as part of a mixed cargo. It has been mentioned above that decorated hemispherical cups and ovoid beakers, manufactured in the valley of Guadalquivir, have been found in Britain (Greene 1979, 63–73), probably piggybacked on olive oil shipments. Unfortunately, to the writer’s knowledge, there is no evidence for return cargo to Iberia, although it is probable that this occurred.

always issued” (Jones 1964, 191–2). However, there is some evidence from the Cumbrian coast that olive oil was in plentiful supply. Twenty lookout towers, out of a probable forty-five, have been excavated on the so-called Solway Frontier. They were stone-built, generally square, with internal dimensions of some 3.8 m. sufficient for two sentries on daily (?) detachment from the nearby fortlets or forts. Olive oil (Dressel 20) amphora fragments have been found at the towers at Herd Hill North, Brownrigg, Low Mire, Bitterlees West, Mawbray, Mawbray Sandpit and Rise How (Breeze 2004, 85–91). This represents finds of amphorae at 44% of the excavated sites and clearly demonstrates that the “common soldiery” were able to consume olive oil, even when serving at small outpost locations.

In the Mediterranean, the disappearance of Spanish olive oil was mirrored by an increase in a rise of imports from North Africa. In Britain, the remains of amphorae from Tunisia (examples of “North African Cylindrical Amphorae” have been found at Caerwent, Gloucester and Cheddar) are insufficient to compensate for the absence of the Spanish product, and indicate that the use of olive oil declined during the later Roman period (Tyers 1996, 72). The reduction in the quantities of olive oil amphorae around the middle of the second century (Cool 2006, 124) may reflect the growing economic independence of the province, with British equivalents being substituted for Mediterranean products (Mattingly 2006, 594).

Assuming that the full ration was taken, the requirement for an auxiliary regiment of 500 soldiers was 11,500 litres, and 115,000 litres for a full strength legion. Converting these figures to the average capacity of a Dressel 20 amphora gives a demand of 183 and 1830 per annum respectively. Putting these figures into a historical context, the fortress at Caerleon was in use from c. 75 to 275, with some periods of complete and other periods of only partial occupation. On the assumption of a 50% occupation over the 200-year period, this suggests the supply of some 180,000 Dressel 20 amphorae.

Fish Sauce In Roman Britain, the fish sauce, known variously as garum, liquamen or muria, was usually obtained from the Iberian Peninsula, being shipped in distinctive shaped amphorae. A very pungent seasoning made from fermented oily finish, herbs, seasonings and salt, it was often used in military and Romanised environments on special occasions such as festivals. However, this particular type of amphora is absent from sites after the middle of the second century and, whilst other possible overseas sources of supply e.g. Brittany, were available, it seems probable that this decline is accounted for a change in Romano-British dietary preferences (Cool, op. cit. 58–62).

It has been argued that the distribution of Dressel 30 amphorae, used in the transport of Massaliot wine from southern Gaul, is a valid example of preference for the Rhône–Rhine route over the Western Seaways (Peacock 1978, 49–51). Cunliffe agrees, considering it is abundantly clear that the Rhône/Rhine route was used to the total exclusion of the Atlantic route (2001, 115). As the map shows, this is undoubtedly true for Massiolot wine, but it might be argued that the choice of any other route for a product from the Mediterranean coast of southern Gaul, would have been totally illogical, and the question of the use of an alternative route via the Western Seaways did not therefore arise. It is suggested that the situation concerning a product such as olive oil, produced in Baetica on the south west coast of the Iberian Peninsula, and shipped from the Atlantic coast via the estuary of the Guadalquivir, leads to a complete reversal of the argument. As was the case for the Germanic provinces, it is of course probable that, at certain times, olive oil destined for Britain was shipped via the Straits of Gibraltar for transport via the Continental river systems. However, whilst the Dressel 20 is a particularly robust amphora, each trans-shipment or cross-decking, requires significant handling and re-stowage effort, and the option of direct shipment to ports on the west coast of Britain has many advantages.

Grain and other foodstuffs Whilst there is clear evidence that Britain exported grain to the continent, imports from the Continent also occurred. For example, charred grain recovered from Rocester and South Shields has been identified as a bread-wheat, Triticum aestivum s., probably imported from north Gaul. Occasional importation from much further overseas is indicated by a large quantity of charred grain, probably comprising a single cargo, found inside a London shop, probably burnt down in 60–1; seeds of plants mixed with the grain indicate that the crop had been grown in the eastern Mediterranean (Straker 1987, 151-5). Salway considers that this reflects an exceptional bad harvest in Britain (and probably the Continent) that year, as it is otherwise difficult to understand why this consignment had been shipped from such a distant source (1981, 618–20).

Assuming an average distance covered of 70 nautical miles per day, the 1500-mile passages to and from southern Spain would take approximately 40 days. Although there are 210 days in the sailing season, it is improbable that more than three round trips, and possibly only two, would be made. As a single fort would only

111

JAMES ELLIS JONES

  The excavated remains of many imported fruits and vegetables, ranging from grapes and dates to cucumbers and coriander, have been found in Romano-British towns and villas; a wide variety of commodities were consumed at Vindolanda, and described in the tablets. However, the quantities of these commodities were relatively insignificant, when compared to the imports of wine and olive oil, and were probably only carried as “part” cargoes.

southern Spain, probably in the valley of the Guadalquivir (Greene 1979, 67–73). This ware has been found in France, Britain and Germany, and in large quantities around the coast of the Mediterranean. The Port Vendres II shipwreck yielded a quantity of this type of pottery, shipped together with several varieties of Baetican amphorae. It seems clear that a few cases of fine ware were included with the major cargo, thus incurring no transport costs and permitting distribution in an area that would normally have been serviced from the major kilns at Lyon. It is probable that the finds in Britain may have been the result of a similar type of piggy-back cargo loading. Fulford (1977, 45) has suggested that ceramique a l'eponge, manufactured in the Bordeaux region, arrived in Britain on the back of the wine trade; Exeter has produced the largest quantity found from any site in Britain, together with significant quantities of Spanish and North African amphorae (Holbrook & Bidwell 1991, 21, 217–8). Dickinson and Hartley considered that Bordeaux received La Graufesenque samian from southern Gaul, destined for the north-east from the conquest to c.110, but that it was shipped first to London or Richborough, before delivery to a north-east port (1971, 131). However, in the case of East Gaulish samian (accounting for 16.5% of the total stamps found along Hadrian's Wall) arrived on the Dutch coast by way of the River Rhine and then directly to South Shields as merchant shipping from the Rhineland to Britain would naturally tend to seek harbours on the East Coast and clearly the Tyne and Humber both received direct shipment from the Rhine (ibid., 130).

Pottery Pottery is considered of particular importance in the study of patterns of trade, as it survives well in the archaeological record. It has been argued (Peacock 1982, 1954; Fulford 1977, 1978), that even though the survival rate is only a fraction of 1% of that produced, the study of distribution may help to evaluate the direction and quantity of trade. It has been stated that it can be considered as a fossilized marker, representative of the total pattern of exchange, because most pottery was not transported in its own right, but was carried as a part of a mixed cargo, probably together with perishable commodities (Fulford 1978a, 59–69). However, Hopkins asked the question “would the distribution maps for the ancient wool trade look like recent maps for the distribution of ancient pots?” (1983b, xxii). Tyers has drawn attention to the pottery recovered from the wreck of the St Peter Port Ship and points out that the small pottery assemblage (probably the personal possessions of a three-man crew) is extremely diverse. He suggests that, if the remains of this cargo had not been preserved by accidental burning, it would have disappeared from the archaeological record and that, if the pottery assemblage had been recovered and taken ashore, it might have “dissolved into a few enigmatic points on our distribution maps” (1996. 73–4).

Stone The Richborough arch is estimated to have required some 200 m³ of Carrera marble (400 tons). The Fishbourne palace of Togidumnus displays a variety of imported stone including Pouilleny Rosé from the Côte d’Or, Porfido Vere Antico green porphory from Greece, Carrera Marble from Italy and a variety of other marbles from the eastern Mediterranean. Pearson suggested, “The stone from Gaul and the Mediterranean arrived directly to the site from the sea” (2003, 135–40). A small quantity of exotic stone, dominated by marbles from Greek, Turkish and Egyptian quarries has been found in Roman London and this might possibly suggest western seaway transport, however the quantities are too small to draw any firm conclusions.

Erim & Reynolds (1973, 108-9) have pointed out that the prices of earthenware vessels given in Diocletian's Price Edict of 301, show that they were cheaper than building bricks, and suggest that it is difficult to argue that it had the same level economic importance as it is given in archaeological publications. De la Bédoyère (2003, 130133) draws attention to Fulford’s suggestion that the movements of legio II Augusta are traceable by “bulges” in the frequency of south-east Dorset Black-burnished ware (1996, 17-19), but he points out that this ignores the incidence of “bulges” prior to the legion being testified in the area. He suggests that this is “one of those examples of archaeological optimism that believes history can be reconstructed from artefacts”. He agrees that it may be possible to link pottery industries with the army, but suggests that attaching them to specific units “is taking things beyond the limits of archaeological evidence”.

Exports from the Province Writing at the time of Augustus, Strabo (Geog. 4. 5. 2–3) enumerates exports comprising grain, cattle, gold, silver, iron and hides, slaves and hunting dogs, and two hundred and fifty years later Eumenius wrote that “so plentiful are its harvests, so numerous are the pasturelands in which it rejoices, so many are the metals of which seams run through it” (Constantius 11.1). The Edict of Diocletian of 301 stipulated prices for a byrrus Britannicus, an early type of duffle coat. Although there is no evidence for the quantities of exports involved, it is unlikely that they were of sufficient quantity to provide

It seems probable that pottery was only a minor shipping trade, tacked-on to bulkier cargoes of perishables. For example, during the first century, decorated hemispherical cups and ovoid beakers in colour-coated fabrics were manufactured in the province of Baetica in 112

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  then either to the Continent or, in particular, to Wales, Chester and Maryport. In either case, direct shipment would obviate the need for unloading and reloading at Dover, but the iron production of the Forest of Dean would have provided a more viable alternative for transport to the west coasts of Britain. (ibid. 83)

return cargoes for the imports of olive oil and wine, and it is therefore probable that many vessels returned to the Continent in ballast. Grain The province often produced a significant surplus of grain, as indicated by the account of the interruption of shipments of British grain to the mouth of the Rhine in the mid fourth century (Libanius Ora. 18, 82-3). Ammianus Marcellinus records (18. 2, 3) that Julian “constructed granaries in place of those burnt, in which could be stored the supply of grain usually brought over from Britain”.

Slaves The text on a wooden writing tablet found during 1996 in the City of London, and translated by Roger Tomlin, reads “Vegetus, assistant slave of Montanus the slave of the August Emperor, has brought the girl Fontunata, by nationality a Diablintian (from near Jublains in France), for 600 denarii. She is warranted healthy and not liable to run away”. The tablet is dated around 80–120 and, as the price paid for the slave was twice the annual salary of 300 denarii of a comparatively well-paid legionary soldier, the export of British slaves could have been a highly profitable enterprise.

Oysters The presence of shellfish remains on inland sites is clear evidence for long distance movement of perishable foods, for example, the oysters from Benwell fort on Hadrian's Wall are believed to have been harvested from southern England (Davies 1971, 129). Furthermore, an interesting, even if low volume, export from Britain was oysters, said to have been introduced to Rome by Agricola and praised by Juvenal. Macrobius described Rutupiae (Richborough) oysters as being ever-present on the tables of Roman pontiffs in the fourth century, “no doubt emulating from Constantine the Great who returned from Britain with a decided liking for them”. Oysters sent by sea to Rome might have been laid loose in the boat (presumably in the bilges) and continually doused with fresh seawater, a relatively cheap and effective method of transportation.

Lead and tin The speed of Roman exploitation of British resources is illustrated by a sample from a lead cistern, buried in the eruption of Vesuvius in 79 that, according to isotope analysis, was produced in the Mendips, where lead had been mined by the Romans for at least 30 years (Brill & Wampler 1967, 63–77). Most British lead was utilised within the Province and Pliny (HN 34. 49) recorded that it was so readily mined in Britain that a law was passed limiting the amount that could be produced, presumably to prevent exports undercutting the Spanish lead industry. There is, however, some further evidence of exports. A Neronian lead ingot of Mendip origin was found at Valéry-sur-Somme and the cargo of a third or fourthcentury shipwreck at Ploumanach off the coast of Brittany consisted of 271 lead ingots marked variously with the stampings CIVTBR, CBRIGAN, ICIN P and CIVTICENORP, indicating an origin from within the tribal areas of the Brigantes or Iceni. Ten Roman tin-alloy ingots recovered from the Thames in the 19th century suggest the possibility of export in unprocessed form and there was probably a small-scale export trade in pewter to the Continent, where over 100 vessels have been identified at Late-Roman sites in France, Belgium and the Netherlands (Beagrie 1989).

Pearls and jet Small items for export are indicated by Pliny’s reference (HN IX, 116) to “pearls of a discoloured appearance” being produced in Britain; Solimarius (CRM 22, 19) refers to “a very large supply of that excellent rock, jet”. Because of the comparatively high value and small bulk, such items were probably in the personal care of the ship's master. Iron Cleere (1985, 61–6) has suggested that products, in the shape of iron blooms, were shipped from a port installation near to the head of the Brede estuary but, if such existed, it has not yet been located. It is probable that Bodiam, at the mouth of the Rother estuary superceded the hypothetical Brede port at some time in the mid second century. The iron production capability from within Britain significantly exceeded the military and civilian requirement of the Province. In view of this surplus, from about 120 until 240, some 400 tonnes of iron were exported annually to the army on the Rhine frontier, which hitherto had been dependent for its supplies on the industry of distant Noricum (mostly modern Austria) or on local small-scale operations. This may well be the case, but it is difficult to support the proposed methods of shipment, i.e. initially to Dover and

Pottery During the later Roman period, there is evidence for export of Black Burnished pottery to the Continent with over 1000 BB1 pots identified in Normandy. That there was common use of this product is demonstrated by fourth-century assemblages from Lillebone displaying 40% of this ware and from Bayeux where BB1 accounted for 20% of the assemblage. Military personnel For the whole of the Romano-British period, a major

113

JAMES ELLIS JONES

  of cargo, and is therefore probable that the personnel were used to supplement the activities of the seamen of the classis Britannica, involved in not only the movement of significant numbers of military personnel, but also the movement of large quantities of supplies to the operating areas. The main theatre of operations lay to the north of the Forth/Clyde isthmus in the territory of the Maeatae, presumably relying on naval transport and avoiding the overland routes (Shotter 1975, 88). This method for supply and provisioning resulted in the reactivation of the Antonine fort at Cramond on the Firth of Forth, and the construction of a coastal fortress Carpow on the Firth of Tay.

requirement was the transport of military and state personnel, both across the Channel, and around the coasts of Britain. The scale of movement ranged through individual soldiers on posting, sometimes not without mishap (see Figure 6.4 for an illustration of the memorial from Chester commemorating an optio who lost his life in a shipwreck); to the transfer of complete legions. The range of activity undertaken is illustrated by tasks such as the carriage of members of the cursus publicus, the movement of vexillations, transport of units on a routine change of garrison, and major operations, as in the case of the Sarmatians, when 5000 heavy cavalry, plus dependants, were transported from the Continent, some to the fort at Ribchester. However, despite the evidence from Chester, the vast majority of personnel were undoubtedly transported via the short Dover/Boulogne crossing.

Cross-channel troop movements continued after the demise of the classis Britannica, with legionary vexillations from Britain, probably accompanied by auxiliary units, being deployed against the Alamanni and the Franks in the campaigns of the mid-250s. Troops were sent to join the Danubian expeditions under Gallienus, but did not return, as the establishment of the so-called Gallic Empire removed Britain from the influence of the central government in Rome. The integration of Germany, Gaul Britain and parts of Spain to form the Imperium Galliarum, resulted in further movement of troops to the continent, but we have no records of the quantities involved; the defeat of the Tetricus by Aurelian in 274, ending the breakaway empire.

Legio XIV was withdrawn from Britain for Nero’s campaign in the Caucasus in c. 66/7, returned in 69 following the battle of Bedriacum, and was transferred to Germany in 70. Legio II Adiutrix was posted to Britain in 71 and transferred back to the Danube in 85/92 and some time before 130, the 9th Legion was withdrawn from Britain. Revolt in the north c. 117-20, in which a centurion from Vindolanda was killed, resulted in the reinforcement of the British garrison by the despatch of 3,000 troops from Germany and Spain to the Tyne. In the mid-140s, troops were sent to a campaign against the Moors and by contrast, reinforcements were sent from the continent to Britain in the same year. The peace settlement of c. 175 following the conquest of Sarmatia, resulted in the transfer to the Roman army of 8,000 heavily armoured cataphract cavalry, of whom 5,500 (plus horses and dependants) were shipped across the Channel to Britain). Vexillations of two British legions were transferred to Gaul to aid in the suppression of a revolt in Brittany. The last decade of the second century saw major troop movements, back and forth across the English Channel. Following the murder of Commodus in 192, three separate Emperors were proclaimed, Septimius Severus in Pannonia, Pescennius Niger in Syria and the Governor, Decimus Clodius Albinus in Britain. Civil war lasted for four years during which time Albinus removed part of the British army (the number is not known) to support his campaigns on the Continent, but after his defeat by Severus at Lyon, Britain was recovered in 197. Periods of comparative calm were punctuated by episodes of frenzied activity such as the arrival of Septimius Severus, accompanied by his family and the Imperial court. Prior to the event, the supply base at Corbridge was subject to a major refurbishment, together with the establishment/re-establishment of a military port and supply base at South Shields. Elements of the classes Germanica, Pannonica and Moesica were included in the expeditionary forces brought to Britain by the Emperor. However, the vessels of these fleets were designed for use on the inland European waterways and would have been unsuitable for operations on the open waters of the British coasts. In addition, their design was intended for the armed patrolling of the frontier rivers, not the carriage

The Carauasian power base was his fleet, he continued to hold Boulogne and significant parts of northern Gaul, and was probably the victor in a naval battle with Maximian in 289. Boulogne was re-captured by Constantius in 297, Carausius was murdered and replaced by Allectus, his finance minister who, after a full-scale invasion across the Channel was killed in battle and Britain was returned to the Empire. Constantius again came to Britain in 306 to campaign in Scotland against the Caledonians and Picts and, after achieving victory, he died at York in 310. His son, Constantine, was proclaimed Emperor and withdrew troops from Britain to support his successful claim to the Purple. Friction between Constantine and Constans, who held the central sector of the Empire, escalated into war in 340 and the consequent removal of troops from Britain to fight in the abortive invasion of Italy. In 359/360, Lupicinus, the magister equitorum, was sent by Julian to take two bodies of auxilia palatina and two numerii across the Channel during the height of winter. The so-called Great Barbarian Uprising of 367 brought a response in the person of Count Theodosius, who crossed the Channel with four regiments of the field army and restored a semblance of order to Britain. Yet another usurper, Magnus Maximus, probably holding the office of either Dux Britanniarum or Comes Maritimi Tractus, led a considerable number of British troops on a successful invasion of the continent in 383, capturing large tracts of territory. In 398, the Vandal general Stilicho brought/sent? troops to Britain to counter an incursion from Caledonia, but in 401 troops were withdrawn to the continent in order to counter a

114

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  of the fourth century Britain was utterly Roman. Legally and culturally there was little to distinguish her inhabitants from any other province, except for some Celtic rather than Hellenistic or Italic undertones”. This may well be true, but there were some significant economic differences, as the English Channel had ensured that Britain had remained insulated from the major effects of the barbarian invasions of the Northern Province. However, in terms of trade, this insulation led to a form of isolation. There is some evidence of shipment of British grain to the Rhine at this time (Ammianus Marcellinus 18. 2, 3; Libanius 18, 82–3) and there may well have been return cargoes, for example wine, but these imports were largely confined to the south-east.

Visigothic threat to Italy. Some of these troops probably returned to Britain because the last British usurper, Constantine III, was able to muster sufficient force to launch a successful assault on the continent in 407, controlling all Gaul and Spain until his eventual defeat by the forces of Honorius in AD 411. Wrecks Wreck sites may serve as indicators of trade and concentrations of artefacts and finds of Roman material on a coastal site may also be considered as potential indicators of a shipwreck. On the Isles of Scilly, Fulford (1989, 245–9) suggests that portable items and personal equipment deposited as votive offerings at temples on Nornour are distinctive because of the narrow range of artefacts present and points out that this is a characteristic of Roman ships' cargoes, assemblages from Roman waterfronts and of Roman “ship” deposits. He suggests that the most satisfactory explanation for the Nornour finds is that they derived from a second/third-century shipwreck from a merchantman blown off course as it rounded Land's End on its way to serve markets in Wales and the northern frontier. A group of 20 lead ingots, found in the River Mersey in 1697, bear the marks of Vespasian, Titus and Domitian and are probably the remains of a sunken cargo (Britannia 22, 319). In the Thames Estuary, at least four mortaria with the stamps of Q. VALERIVS SE... were brought up by fishermen in Herne Bay, but the wreck has not been located. Also in the Thames estuary, six mortaria stamped CAVARIUS were recovered from the Ouze Deep but the precise location is not known (Dean 1984, 78/9). A cargo of mixed form samian from Lezoux, with the stamps of more than forty potters, was recovered from the Pudding Pan shoal, 10 km northeast of Whitstable. The remains of the vessel were not located, but Roman tiles and an intact amphora containing 6000 olive pits found in the area may have come from the same wreck (Smith 1909, 395–414; Dean 1984, 78/9). At Little Russel, east of St. Peter Port in Guernsey, several Beltran 2B amphorae were found protruding from the seabed; fragments of Baetican Dr 711 amphorae were brought up by a trawler (Keen 1979, 3). Several Haltern 70 form amphorae of first century date were netted by French fishermen off the Isle of Wight and a number of Dressel 1A amphorae from Yarmouth Roads and the Ryde Middle Bank (Lyne, 2005, 4). On the estuary of the River Tyne, the wreck of a Roman troopship on the Herd Sand is indicated by a shield-boss of Legio VIII Augusta, a helmet cheek-piece, a patera, 67 coins and other items (Hodgson, 2009, 71). Six second-century amphorae from southern Spain in a cave close to Fife Ness may perhaps best be explained by a shipwreck on the adjacent Carr Rocks, a major hazard to vessels plying between the Forth and Tay (Martin 1992, 22–3).

Whilst the degree of barbarian attacks in 367 have perhaps been exaggerated, they caused widespread disruption and a loss of confidence, and the effect on trade and commerce cannot be overestimated. Whilst there was still considerable use of the inland waterways (for example, Wroxeter remained a prosperous settlement until the latter part of the fifth century), it has previously been considered that imports from the Mediterranean fell to insignificant levels, perhaps to a level even less than that before the Roman invasion. However, opinions that are more recent suggest that this may not have been the case. Roger White (2007, 149–176) writes that, because of the Germanic incursions into Gaul, the Roman elites of the province were isolated from further contact with the core of the Roman world in the Mediterranean, while paradoxically the Britonnic elites on the western coast of Britain, who had largely been excluded, or had excluded themselves, from the trappings of civilisation in the Roman period found themselves able to receive, and control, access to the prestige items brought from both the Mediterranean and Western Gaul. White’s map (Figure 12.4) showing trade routes from the Mediterranean to Britain, with density of Mediterranean imports of the fifth to seventh centuries, concentrates on river transport via the Garonne and Loire. However, the case has already been made that this distribution could equally well have been via the Straits of Gibraltar and the Iberian coast. There is considerable archaeological evidence for Mediterranean exploitation of the Irish Sea littoral in the late/post-Roman period, with Byzantine ceramic imports being found at several dozen sites across Western Britain and Ireland. In Somerset, significant amounts of imported pottery and glass, including some Byzantine amphorae and tableware, have been found at Cadbury Congresbury (Ratz et al. 1992, 214–18), South Cadbury (Alcock 1995, 14–30; 128–32) and Cannington (Ratz et al. 2000). In north Wales, the hillfort at Deganwy, overlooking the estuary of the River Conwy, provided evidence of Byzantine amphorae and late Roman pottery (Edwards & Lane 1988, 50–3). Its proximity to the Roman period copper mining and smelting site on the Great Orme

Decline and fall Millett (1990, 212) commented, “During the third quarter

115

JAMES ELLIS JONES

  suggests possible return cargoes. Further north, it is clear that Meols continued to trade with the Mediterranean in the post-Roman period as evidenced by a pottery flask from the early Christian shrine of St Menas in Egypt, found in the 1950s, and more recently the discovery of three sixth/seventh-century Byzantine coins.

Figure 11.5 Post-Roman/Dark Age trading site at Tintagel

Figure 11.4 Trade routes from the Mediterranean Ewan Campbell (2007, 26) has suggested that these imports may be dated to between 475 and 550, indicating a phase of importation lasting no more than 75 years. It is unclear whether the widespread distribution across the Irish Sea Province is the result of direct imports from Byzantine ships or secondary distribution from a number of high status and/or emporia in Western Britain (Harris 2003, 147). The presence of discrete packages of products, undiluted by goods from either the Western Mediterranean or the Atlantic coasts, suggests direct shipment as opposed to cabotage; and slaves could have been a welcome additional part of a return cargo (ibid. 137). The best-known and most recently excavated site is at Tintagel, on the north Cornish coast, where North African Red-Slipped Ware has been found in abundance, together with amphorae, pottery from Carthage, the Greek Islands, Turkey and fragments of a sixth-century Spanish glass flagon (Thomas 1993, 67–81; Morris et al. 1999, 206– 15).

116

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  AD 120 AD 210 AD 300 AD 400

Chapter 12 Patterns of demand Introduction An exhaustive search of published material to recover information on the pattern of cargoes (whether imports, exports or locally produced) on with the coasts and rivers of Roman Britain was unsuccessful. Notable exceptions are Marsden’s (1994, 80–84) assessment of the number of shiploads of Kentish rag stone needed to construct the city walls of London, the studies of the supply by sea of building stone to the Saxon Shore forts (Allen and Fulford (1999, 163–84; Pearson 2003, 2006), the sources of building materials in Roman York (Gaunt & Buckland 2003, 135–43), and Fulford’s (2000, 41–50) examination of the organisation of support for the Claudian invasion of Britain. These studies are, however, based on specific events and comparatively confined geographical areas, and it was decided to consider this topic in the context of a demand based approach.

45,000 to 53,000 47,000 to 55,000 21,000 to 25,000 28,000 to 33,000

moreover, if the figures are accepted as accurate, there is an increase of some 20% from lowest to highest figure. Similarly, Simon James (1984, 164) has pointed out that the available evidence can be interpreted to suggest that the Province was either “swarming with troops” through most of the fourth century or, using a minimalist approach, that the army may be numbered from 20,000 to as low as 15,000, from the late third century through to the fourth century. In terms of the transport requirement for supply and provision, the location and purchasing power of the military also had an effect. Davies (1984, 93) suggested that, at that time, the much reduced legionary fortresses at Caerleon and Chester remained “as monuments of military inertia”. In any one period, some forts were occupied in full strength, others completely abandoned and some maintained only on a care and maintenance basis. For example, Peter Carrington (2008, 9) suggests that the fortress at Chester was held in strength for less than three generations, c. 90–120 and 210–250 and that, even during these periods it is prudent to assume that the absence of detachments of the legion would have been the norm. Later the size of units was much reduced and, by the early fourth century, the strength of a legion was only about 400 soldiers and that of a cavalry unit was much less than 100 (Tomlin 2000, 172). At Caernarfon, a coin of the Emperor Theodosius I (379–95) suggests that the fort remained in use until the end of the fourth century. From that time, no gold or silver coins were being shipped in to Britain, the state was no longer paying its servants, the most important of these were the remnant of the Army but, “Cease paying the soldiers and the soldiers will, in due course, cease to be soldiers” (Cleary 1989, 140).

Clearly, size of population is the dominant factor, but estimates for the population of Roman Britain have ranged from under 1 million (Collingwood & Myres 1937, 180) to 4–6 million (Salway 1981, 544). Millett endeavours to provide better figures for England and Wales, but in the end, he is forced to resort to a midrange estimate of 3,655,000. He has assessed the rural, urban and military populations for the later Roman period and has arrived at mid-range figures of 3,300,000 rural (90% of total population), 240,000 urban (6.6%) and 125,000 soldiers + dependents (3.4%), indicating “a fundamentally agrarian society”186). He comments that “The uncertainties make the range very large, although we can be reasonably certain that the real value lies somewhere within it” (1990 181–6). Military demand

Civilian settlements were established outside the forts and fortresses of Britain; the settlement outside a fortress was known as a canabae, the term vicus was used for a settlement outside a fort. As their prime source of revenue, the activities of the inhabitants were directed to the demands of the soldiers, either as suppliers of goods and services, or as common-law wives of the soldiers. As such, they acquired (or brought with them) Romanised tastes and patterns of consumption, and therefore, created a significant demand for a wide range of products.

The army played a predominant role in the British economy and therefore the pattern of cargoes, either within the Province or as imports from abroad. Army units on the frontier zones, and at strategic positions along the lines of communication, created not only a series of micro-economies in the immediate environs, but also created an economic pull-through effect. It has been argued that long distance trade was dependent, indeed parasitic, upon official supply lines, and was almost certainly directed to the needs of the military market; there is no evidence that the civilian market generated long distance trade (Middleton 1979, 91).The establishment of a large standing garrison on Hadrian’s Wall required the regular long distance shipment of food supplies from the grain growing areas of the south. Middleton considers that it is no coincidence that this is also the period at which the earliest instances of largescale, long-distance, distribution of British pottery fabrics to the north occur.

There has been considerable discussion on the role of civilian contractors in supplying the army. Hilary Cool (2006, 133) suggested, “One thing that most writers on the Roman army are united is that the mechanisms of army supply are not understood”. Whatever the mechanism, it is highly probable that, in Britain, the Procurator was ultimately responsible for the control of this task, and had the necessary authority to direct the suppliers to operate in the most effective manner. It is probable that this included direct delivery to the point of consumption, wherever this was a practical proposition,

David Breeze (1984, 264–8) suggested the following wide-ranging estimates for the Roman army in Britain

117

JAMES ELLIS JONES

  reaching the settlements would have originally formed part of a piggy-backed cargo and, at any one time, are unlikely to have produced a significant increase in the volume of vessels involved in sea-borne traffic.

thus minimising the volume of transhipment and cargo handling. Regulation, rather than the market forces, governed “official” supplies to the military, but market forces covered the provision of “unofficial” supplies, though these were parasitic on the officially stimulated trade routes. Jones (1974, 845) considered that longdistance trade of bulky and relatively inexpensive commodities, such as grain, to a civilian area was only profitable, when both the production area and the market for the grain lay close to a port or navigable river. However, the provision of grain to forts was a military necessity, with the costs being borne by the state and, even if the supply had been left to private enterprise rather than to an official system; economic factors need not have been the over-riding consideration.

In Rome, the average male ate about one kilogram of bread a day; the corn dole gave an average monthly issue of five modii (a modius equals 6.67 kg), probably sufficient for about two persons (Erdkamp 2005, 240–4). When considering provincial non-military demand, the amount of grain, wine and olive oil consumed by women and children would be much less than that of a soldier. Erdkamp (2002, 49), Garnsey (1988, 229–35) and Carrington (2008, 45) have produced estimates of civilian cereal consumption for males, females and children. It is estimated that 220 kg per annum wheat-equivalent represents subsistence level and, as the Roman soldier was well-to-do when compared to the civilian population, this amount should certainly be increased. A suggestion, and it is no more than that, of an increase by c. 10% to 240 kg per annum seems reasonable.

Non-military demand Towns were an essential part of the Roman way of life, not only fulfilling a central place function, but also simplifying the tasks of administration and tax collection. By contrast, Iron Age society in Britain had revolved almost entirely around individual farms, hill forts and villages, with comparatively few larger collective settlements (oppida). Wacher (1978, 36–7) has suggested that one of the main achievements, by which the success of the Roman administration in Britain may be judged, was its ability to persuade a population, totally committed to a rural way of life, to accept the alien concept of urban settlements in its midst. Jones and Mattingly’s invaluable Atlas of Roman Britain (1990, Map 5:12 94) illustrates the distribution of towns and small towns, not including military vici, and shows 94 unfortified small towns, 40 fortified small towns, 14 civitas capitals and 4 coloniae. Urbanisation led to changes in taste, and the desire for Roman “luxuries”, such as Italian wine and Spanish olive oil, but these show variability of demand, dramatically rising in the early period of occupation, but falling by the fourth century to a fraction of the earlier volumes.

Changes in styles and standards of living Changes in demand based on population size are obvious, but changes in the standard of living also have an effect and some of these are detectable in the archaeological record. For example, the inhabitants of the stone houses of the mid-second to mid-third century military vicus at Chesterholm (Vindolanda) “possessed all the material comforts of Romanisation”, including much imported pottery. On the other hand, in the wooden houses of the later vicus (c. 270–350), coinage was scarce; pottery was markedly inferior and usually of local manufacture (Birley 1977, 71–2). Geographical and social diversity was observed by Timby (1999, 40-1), who examined whether Gloucester, a colonia with an established population familiar with Roman culinary habits and vessel forms, had a different pottery assemblage from contemporary rural sites in the area. Frocester, Brockworth, Birdlip, Cowley, Kingscote and Uley were considered; the later first and early second centuries show a greater diversity of traded wares and continental imports in the colonia. For example, samian ware accounts for 6% (by weight) of the Berkeley St assemblage, but is less than 1% at Frocester Court. In the later second to early fourth centuries, the assemblage at Gloucester became less distinctive, having similar wares as at the nearby sites. In the later fourth century, Gloucester alone seems to have received imports, presumably by seaborne coastal trade.

Many villas (of which over 600 have, to some level, been excavated) in the countryside, had access to supply by water transport but were usually self-sufficient in agricultural produce, with many producing a surplus; therefore, there was no requirement for the supply of grain. We are therefore left with the requirement for “Roman luxuries” as demand creation factors. Clearly, these items would usually be the preserve of the villa owner and his family but, in the case of villa estates, some senior employees such as bailiffs, farm managers and their immediate family should also be considered. Native settlements also produced some evidence of “Roman luxuries”, for example, the enclosed hut group at Din Lligwy, on Anglesey, had a quantity of decorated samian and later fine wares, including Oxford and Nene Valley products (Arnold & Davies 2000, 113); the courtyard house at Carn Euny, in Cornwall, produced a Dressel 20 amphora fragment, samian and Blackburnished ware, glass beads and rotary querns (Christie 1993, 27). Demand in rural areas indicates supply at the level of travelling tinkers; any luxury goods eventually

Variations in taste Occasionally, changes in taste may be detected in the archaeological record. In the report on the fourth-century town houses at Beeches Rd, Cirencester, the excavators comment on the virtual absence of Oxfordshire Ware drinking vessels. They point out that this is in marked contrast to periods when beakers and flagons were much in evidence and suggest that, by comparison with earlier

118

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  and plateaux to suggest that, as suggested above, variations in source of supply, taste and patterns of living had a measurable effect. These are the major bulk commodities affecting the volume of shipping, and most other items, such as pottery, were probably piggy-backed onto these cargoes. However, whatever the level of demand, it is only relevant if a suitable source of supply is available. For example, Iberia apparently ceased to export olive oil in the later third century and this was mirrored by an increase in a rise of imports to Britain from North Africa. However, the remains of amphorae from Tunisia found, for example, at Caerwent, Gloucester and Cheddar, are insufficient to compensate for the absence of the Spanish product, indicating that the use of olive oil in Britain declined during the later Roman period (Tyers 1996, 72).

periods, this indicates a fall in wine consumption, (McWhirr et al. 1986, 175). Ethnicity and social status is perhaps reflected by the fact that legionary sites often have a much higher proportion of pig bones and fewer sheep/goat bones than auxiliary sites (King 1984, 189). Legionary soldiers were Roman citizens, but “it would be a mistake to think that they should be automatically equated with an Italian ethnicity” (op cit. 179). Whilst the first six governors of Britannia were Italian, as were the known legionary commanders and many of the lower ranks, there had been extensive recruitment from outside Italy, especially from Spain and southeastern Gaul. As three of the Conquest legions had been serving on the Rhineland and the fourth on the Danube, an element of men recruited from those areas is also to be expected. With the possible exception of the Batavians, who continued to recruit from their homeland in the Netherlands, auxiliary regiments in Britain drew replacement recruits from the nearest convenient source. Within a generation of a regiment moving out of the area in which it was raised, its ethnic/tribal identity became less clearly defined. By the third century the composition of the frontier troops (limitanei) would have become almost entirely hereditary, since sons of soldiers were regularly recruited, and from 313 were obliged to serve (Jones 1973, 615), so the taste for Mediterranean styles of food, including wine and olive oil, may have gradually declined; at Vindolanda, the Batavian troops were complaining that they had run out of beer, not wine. The case for direct shipment of olive oil from Iberia via the Western Seaways was made in Chapter 2, as was the effect of the cessation of the supply of Spanish products in the mid third century. There is evidence for some importation of African olive oil, but it is probable that the use of this was confined to the more affluent sectors of the population, and that the bulk of the population, both military and civilian, reverted to the use of animal fats for cooking, lighting and hygiene. As regards wine, a steady decline in military consumption from the mid third century was not only the result of the departure of the more Romanised legionaries, but also because of a reversion to traditional Celtic tastes for drinks such as mead and ale. There was a reasonably consistent civilian demand for wine from the early second to mid fourth centuries. There is some evidence for British viticulture, but only on a limited scale; villa owners and well-to-do urban citizens undoubtedly continued their enjoyment of imported Gaulish wine.

-o-o-o-o-o-oWe should not be surprised at the variations this chapter has highlighted – it is certainly happening today. In this writer’s lifetime, the consumption of olive oil has moved from an occasional teaspoon for medical purposes, to a “must have” on every TV cookery programme and, aided by extensive advertising, to most domestic kitchens. Wine, that only a few decades ago, was a drink for “special occasions” is, we are now warned, consumed in such quantity that it poses a serious threat to the nation’s health! Chicken tikka masala, described by (at that time) Foreign Secretary Robin Cook, as “a truly British national dish”, is stated to be today’s most popular choice in British restaurants. In recent years, the influence on demand created by the army has been much reduced, the numbers having fallen from some 500,000 in 1957, to less than 100,000; the ethnic mix of the population has changed, and is still changing, at a dramatic rate. Tempora mutantur, nos et mutamur in illis

Changes in patterns of demand A number of references have already been made to significant changes in demand over time, and they are indeed significant. For example, at the beginning of the second century, consumption of wine was some 20,000,000 litres, but a hundred years later, it had fallen to 6,000,000 litres, similarly, olive oil had fallen from 1,200,000 to 400,000 litres. Graphs have been produces indicating the variations in the demand for grain, wine and olive oil, over the Romano-British period (Jones 2009, Appendix 9) and display sufficient peaks, valleys 119

JAMES ELLIS JONES

  As the result of a previous publication, I now have a better understanding of the modern military expression “blue-on-blue”, having experienced coming under socalled “friendly fire”, even if only in the literary sense! However, Michael Fulford must be singled out, who, in response to my criticism of some of his opinions, kindly wrote to me that this had not caused any of his “bloodvessels to boil”. I must acknowledge the inspiration of Sheppard Frere who, to re-cycle the title of Sir Mortimer Wheeler’s book, was “Still Digging” at Colwyn Castle in his nineties, publishing with commendable speed and, when occasionally necessary, proving well capable of an imperious response to his critics (e.g. Britannia 42, 263274). I am grateful to Sir Barry Cunliffe for encouraging my efforts and, in particular, for his suggestion “surely an eightieth birthday is no obstacle” to my completion of this book!

Conclusion As the length of the bibliography demonstrates, this is a work of synthesis and the reader will have observed that it is liberally besprinkled with quotations. In certain instances, this is because the writer has “stood on the shoulders of giants” but, however, does not claim to have seen further, but perhaps gazed in somewhat different directions. The maritime course pursued has, at some times, ventured into uncharted waters, and has therefore been subject to the inherent dangers of this form of voyage. Metaphorical reefs, sandbanks, adverse tides and weather conditions have been encountered; sometimes in the doldrums or entangled in the Sea of Sargossa and progress has been somewhat erratic. It may be suggested, with some justification, that the chart on which the voyage has been based should, particularly in the area of statistics, be liberally decorated with mythical seacreatures. Nevertheless, it is arguable that, by arriving at a reasoned assessment of the role and extent of water transport in Roman Britain, the voyage has been accomplished and the cargo safely landed. Whether this particular cargo has been worth carrying is for others to decide! In a previous volume, dealing with only the west coast of Roman Britain, the conclusion was largely devoted to a recapitulation of the previous chapters, with the primary objective of “hammering home” the concepts and arguments contained therein. It is now realised that this approach was a hangover from the originating doctoral thesis, intended to convince examiners of its inherent value, and is of little value to the audience this work is intended to address. The present readers, having reached this point, are well aware of these concepts and arguments, and have undoubtedly formed their own judgements on their validity and will therefore not be subjected to further reiteration of the preceding text. David Breeze pointed out that “Archaeologists are used to working with a jigsaw of which many pieces may not have survived” and also that “the jigsaw may never have been completed by its maker” (2009, 4). Having been assembled from so many diverse parts, this volume might well be considered in the context of a complicated jigsaw; undoubtedly much valuable information is missing and it is certainly far from complete. It may well be that some pieces may have been forced into spaces for which they were never intended. However, if on some occasions the temptation to “make things fit” has not been successfully resisted and some conclusions may best be considered as highly speculative, comfort may perhaps be drawn from the advice of Richard Reece (2002, 69) that “It probably does not matter that your house of interpretation is built upon sand, because if the foundations are undermined by proper research you will simply shift your castle into the air, where it will be safe from all attack”.

120

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

 

Ancient Sources Ancient sources quoted, and abbreviations used Ammianus Marcellinus (Amm. Marc.) Appian (App.) Arrian (Arr.) Aurelius Victor (Aur. Vic.) Caesar (Caes.) ,, Cato (Cato) Dio Cassius (Dio. Cass.) Claudius Claudianus (Claudian) Eugippius (Eug.) Eutropius (Eutr.) Frontinus (Frontin.) Herodian (Hdn.) Josephus (Jos.) ,, Libanius (Lib.) Livy (Livy) Lucan (Luc.) Pliny (the Elder) Pliny (the Younger) Plutarch (Plut.) Polybius (Polyb.) Sallust (Sall.) Scriptores Historiae Augustae (Var.) Solinus (Sol.) Statius (Stat.) Strabo (Strab.) Suetonius (Suet.) ,, ,, Tacitus (Tac.) ,, ,, ,, Ulpian (Ulpian) Vegetius (Veg.) Velleius Paterculus (Vell. Pat.) Vitruvius (Vit.)

Historia (Hist.) Historia (Hist.) Tactica (Tact). Caesares Bellum Civile (BCiv.) Bellum Gallicum (BGall.) De Agri Cultura (Agri.) Epitome (Ep.) Panygerics Stilicho (Stil.) Vita Sancti Severini (Vit.Sev.) Breviarium ab urbe condita (Brev.) Strategemata (Str.) Historia (Hist.) Bellum Judaicum (BJud.) Antiquitates Judaicae (AJ) Oration (Ora.) Epitomae (Epit.) Naturalis Historia (HN) Epistulae (Ep.) Vitae Parallelae (Vit.) Bellum Iugurthinum (Iug.) Historia Augusta (SHA) Collectanea Rerum Memorabilium RM) Silvae (Silv.) Geographica (Geog.) Divus Claudius (Claud.) Divus Iulius (Iul.) Divus Vespasianus (Vesp.) Agricola (Ag.) Annales (Ann.) Germania (Germ.) Historiae (Hist.) Digesta (Digest) De re militari (Mil.) Historiae Romanae (Vell. Pat.) De Architectura (De Arch.)

121

JAMES ELLIS JONES

  version. However, whilst producing this volume, it was often necessary to refer back to the previous publication and the writer’s own experience has led, somewhat reluctantly, to present the gazetteer in simple alphabetical order. It is hoped that the National Grid references and the maps provided within the main text will go some way to aid the problems of location and that the series of tables (by Location, River and Coast) that follow the gazetteer, will aid those seeking further information.

Appendix 1 Gazetteer of locations on the coasts, estuaries and rivers considered accessible by sea-going ships, coastal vessels, barges and/or local craft Introduction This gazetteer is not intended as a work of general reference, but rather as an aid to considering the problems of supply and replenishment. The most significant factors relating to supply and provisioning are location in relation to alternative routes of supply, size of garrison, population of civilian settlements, significant periods of construction, and periods of occupation and abandonment However, in some instances it not been possible to resist the temptation to include certain items not directly relevant to the maritime context, such as the 12th century description of existing Roman remains at Caerleon by Gerald the Welshman (Giraldus Cambrensis) or those of Samuel Lysons, writing about Bryn-y-Gefeiliau in 1807.

The entries for individual sites have also presented a variety of problems, not least in selecting the level of information to be provided. For example, no plans of sites are included for, whilst there are plans available for the majority of the military sites or civilian settlements, many are based on early excavations and, in many cases owe as much to the inspired use of the dotted line, as to archaeological evidence. There is also considerable variation in the size of the entries for individual locations, with extra space being allocated to well-published excavations, for example Caernarfon (Casey & Davies 1993), South Shields (Bidwell & Speak 1994), Loughor (Marvell & Owen-John 1997). By contrast, Hanson’s excellent publication on Elginhaugh (Hanson 2007) is given only a brief paragraph as the forts relevance to water transport is limited in both time and utility. On a smaller scale, a single sentence covers Whitehouse Farm, where the only evidence is a series of aerial photographs showing crop marks of a pair of ditches with rounded corners, whereas the nearby campaign base at Clyro, of similar size and period, has sufficient information to justify a full paragraph. In order to reflect military activity on the west coast of Scotland, some marching camps from this area (e.g. Irvine) have been included. In some cases, it has been decided to concentrate on specific aspects of a location, and this is typified by the information presented for the permanent legionary fortresses. Because of the information available from recently published work by Carrington and Mason, the entry for Chester concentrates on the most significant periods of occupation and disuse, whereas the entry for Caerleon is able to focus on the role of the canabae, because of the excellent excavation report on the civilian settlement by Evans in 2000. More attention is paid to the early fortresses at Kingsholm and Usk than is often the case, as it is arguable that, in terms of active military operations, their significance was as great as that of the later permanent locations at Caerleon, York and Chester.

The overall structuring of the data, and the detail of the individual items, is the result of several attempts at finding a cohesive and readily understandable method of presentation. The first, and perhaps most obvious, attempt grouped the individual sites on a geographical basis. This had the advantage that all locations, whether military or civilian, were presented in a manner illustrating the traffic on a particular stretch of coastline or river. However, this method had the disadvantage of obscuring patterns of size and therefore demand, changes over time resulting from the ebb and flow of military occupation, or the growth of civilian settlements. On the River Usk, for example, the legionary fortress of Caerleon and its associated canabae supported a population of perhaps as much as 10,000, or as little as 1,000, at different times between the late 1st century and the late 3rd century, but by the late-4th century the fortress and the canabae had been virtually abandoned. Clearly, this variation had a massive effect on the amount of transport required to supply the needs of the garrison. Some 35 k upstream, the fort at Abergavenny was occupied from the second half of the 1st century until the early 2nd century but, when the fort was abandoned, the site found a new role in the iron industry and, in effect, changed from the function of consumer to that of supplier. An attempt to structure solely by time was equally unsuccessful, as this resulted in not only a grouping together of locations of dis-similar function and size, with widely varying patterns of supply and demand, but also created considerable confusion because of the varying periods of occupation. An earlier version, covering only the west coast of Britannia, was based on a grouping by type, using categories of fortress, fort, city, town and settlement, and to place these alphabetically in a series of time-frames. This in itself has several disadvantages, notably the appearance of a single site under more than one heading, (for example, consider the case of Abergavenny mentioned above). Peter Carrington kindly “test-drove” this gazetteer and pointed out that there was difficulty in moving between a text reference and its entry in the gazetteer and, in order to retain the structure and to, hopefully, minimise this problem a ‘Gazetteer Index’ was included in an earlier

Any evidence for maritime activity is of major importance, whether direct or indirect, and is referred to in the site entry. Indirect evidence is demonstrated, for example, by the use of Bath-stone at Kenchester and of Cheshire sandstone at Caernarfon, direct evidence includes the 3rd-century quay at Caerleon and the tombstone from Chester commemorating a soldier who was lost in a shipwreck. When considering the relative merits of road or water transport the relationship of a site to the extensive system of Roman roads is important, and each site entry contains this data.

122

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  In a previous volume, it was stated “In order to accurately assess the nature of supply and replenishment of the military and civilian settlements in the study area, all sites that may have been supplied by water transport were visited. From the considered limit of navigation, bridging points down-river were visited, and, based on experience of small boat handling on the coasts and rivers of Wales, the Irish Sea and the Bristol Channel, a judgement was made as to the probability of access by water transport” (Jones 2009, 6). Regrettably, the constraints of time and reduced mobility mean that this does not hold good for this volume. Whilst the writer is able to claim an adequate knowledge of the south coast, its rivers and the Thames Estuary, a gap exists to the north as far as the Humber and again northward for the eastern seaboard of Scotland. It has therefore been necessary to rely on charts, maps and the work of others; undoubtedly, a number of misjudgements have been made and the writer would be most grateful for advice ([email protected]) on where these exist. -o-o-o-o-o-o-

123

JAMES ELLIS JONES

  the River Ure, 2 km downstream from Roecliffe, has yet to be located, but items of military equipment and 1st century masonry structures from the presumed site of the vicus have been reported and confirm its undoubted presence. By the early 2nd century, Aldborough had become the civitas capital of the Brigantes with a wellappointed town based on two main cross streets and a regular grid system. Modern excavations have concentrated on the defences but, from 19th-century excavations, the site of the forum is known, as are several houses with impressive mosaics and a probable amphitheatre. It is probable that Aldborough was not so very different to other civitas capitals further south, most of which had an earlier start date, but numismatic evidence suggests a gradual decline in prosperity during the 4th century.

Abergavenny (SO299140) – Gobannium The mid-1st century fort lies on the brow of a low spur, 500 m to the north of the River Usk, blocking the valley where the river passes between the Black Mountains and the Brecon Beacons. The fort lies at the junction of the road from Brecon Gaer to Caerleon, with a branch road leading to Kenchester. As only very limited excavation has taken place, with a small granary and a short length of the defences being located, it is not possible to determine the extent of the fort. The majority of datable finds from the site are from the late-1st to mid-2nd centuries, but a small quantity of Claudian pottery suggests that Abergavenny was the site of an early fort, established during the course of campaigning against the Silures. The postulated stores depot at Usk would have provided an incentive for the improvement of the river up to Abergavenny, and perhaps beyond to Pen-y-Gaer.

Bank and ditch defences enclosing an area of 22.3 ha. were built late in the 2nd century and reinforced by the addition of a masonry wall, a heightened rampart and internal towers before the middle of the 3rd century. In the early 4th century, massive external corner towers and interval towers were added. Later 2nd- and 3rd-century military equipment found within the town, together with a high proportion of Valentinian coins dated to the house of Valentine suggests the presence of a unit of the field army.

The vicus developed into a small settlement after the closure of the fort; the coin series extends to the close of the 4th century. It is probable that native iron smelting took place before the establishment of the early fort and, as smithing has been identified at nearby Usk, there is a strong case that the Roman settlement developed to exploit this activity. The major consumers of iron products in the area would have been the fortress and canabae at Caerleon, and, as no iron smelting activity has been detected at the large fort at Y Gaer near Brecon, it is possible that iron products were shipped upriver to that point. Abergavenny would also have value as a market centre between the rich arable lands to the east and south and the pastoral areas to the west and north.

Ambleside (NY372034) - Glannoventa Located at the head of Lake Windermere, a fort of 2.2 ha. was established c.90. A mountain road (High Street) runs north-east to join the Manchester/Carlisle main route at Brougham. A westward road, via Wrynose Pass and the fort at Hardknott joins the coast at Ravenglass. The site of the early fort was converted into a platform, raised above flood level, and replaced by a stone fort, and occupied until well into the 4th century. A 3rd century tombstone records the name of a soldier killed in the fort by the enemy (‘in cast(ris) inte(rfectus) ab hosti(bus)’).

Abertanant (SJ 248214), Llansantfraid-y-Mechain (SJ 229207) Situated on the River Vyrnwy, a tributary of the Severn, Abertanant was recognised from aerial photographs in 1976, and two separate sites were identified. There is no evidence for a Roman road in the area, and supply and replenishment along the Afon Vyrnwy, from its confluence with the River Severn some 12 km to the southeast, is probable. The site occupies a low-lying and tactically disadvantageous location, usable only in the drier months and dominated by Llanymynech Mountain. The first fort covered an area of 0.6 ha., with a wide defensive ditch, timber gate, box rampart and titulum, but no interior structures were located. A ditch surrounded the second fort, covering c 0.9 ha., and the remains of a timber-revetted rampart were identified. Within the defences, six timber buildings were revealed, four of which might have been barracks. An excavation report has not been published, and it is understood that no dateable artefacts have yet been identified. The nearby 1.1 ha. fort at Llantsantfraid-y-Mechain, Meliniog was subject to limited excavation in 1987; within a polygonal double-ditched enclosure a probable barrack building and granary with loading bay were located; the excavator identified the site as a military supply depot.

A depression in the ground marks the old course of the River Brathay which, in wet weather, still contains standing water; it is probable that the Romans brought boats up to this gate, though there are also traces, now visible in summer when the water is low, of what may have been a Roman quay in the lake south of the fort. In the 19th century, sea-going luggers operated on Lake Windermere, transporting Cumberland roofing slate down the west coast, and in some cases as far as London. The River Leven is, at present, tidal to Hatherthwaite, and it is possible that Roman improvement of the 5 km stretch of the river, followed by a short portage near the point where it leaves Lake Windermere, would have enabled waterborne supply and replenishment of the fort. Annan Hill (NY192654) Annanfoot (NY180653) This Annan Hill marching camp (1.3 ha.) commands a good view to the south over the Solway and is located on top of a natural ridge running north-east to south-west,

Aldborough (SE406664) - Isurium Brigantum Structural evidence for a Flavian fort on the east bank of

124

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  However, there is no known Roman road leading from  the  site  and  the  fort at  Cardiff, only  10  km  further to  the  west,  seems  a  more  obvious  landing  place  for  cross‐Channel traffic. The natural harbour at Barry is  one of the few places on the Bristol Channel where  a  vessel may lie afloat at all states of the tide and, on the  English shore, the lee of Brean Down off Sprat Beach,  provides a similar, if only limited, advantage. It might  therefore  be  suggested  that  a  naval  presence  might  account  for the  building  at Cold  Knap,  enabling  some  form  of  patrolling  of  the  15  km  stretch  of  water,  hopefully  intercepting  sea  raiders  en  route  to  pillage  the  rich  villa  area  on the  southern  side of  the  Severn  (Sabrina Fluvis) and its estuary.  

with the ground falling away quite steeply to the northwest and south-east. Overlooking the Solway Firth, the Annanfoot 5.2 ha. marching camp, lying on the right bank of the River Annan (tidal for 5km northwards), near its mouth, has been associated with early coastal or marine operations conducted by Agricola. Axminster (SY298974) - Moridunum? Axminster is situated on the Fosse Way with Exeter to the west and Ilchester to the north-east, a branch road leads eastwards towards the coast and Dorchester is 35 km distant. Although lying on the on the River Axe, a road leads due south for 8 km to Seaton, and it is improbable that direct water transport to Axminster was involved. The 1st century fort covered 1.9 ha. and contained a significant number of stone buildings, indicating continuing occupation well beyond the norm for the area. A mansio has been identified within the fort and the vicus developed into civil settlement of 12 ha., mainly on the road to the west of the fort, with evidence of occupation continuing into the 4th century..

Bath (ST755653) - Aquae Sulis Bath is located on the navigable River Avon and on the Fosse Way between Cirencester and Ilchester, with a road also running to Sea Mills to the west and Mildenhall to the east, with also a probable road to Old Sarum Claudio-Neronian pottery and coins indicate a fort at this strategic location. Bath lay at the centre of an extensive economic hinterland with a concentration of at least thirty villas within a 25 km radius. A temple was constructed in the late 1st century and by the 3rd century had developed into an extensive religious complex. The ‘probable’ Roman walls enclose an area of some 10 ha. and an extensive ribbon development has been located to the east and west.   The oolitic limestone quarried on the Downs to the south of Bath was in great demand, because of its suitability for fine carving and architectural detail, and was widely distributed throughout southern Britain. It was used, for example, at London, Colchester, and Silchester in the south and east, Cirencester and numerous villas in the Cotswolds, at Kenchester and Wroxeter in the Welsh marches, as well as at the forts and villas of south Wales. Because of the very considerable quantities and weight involved (a cubic metre block of limestone weighs 2.4 tonnes, it is probable that some of the stone was shipped on barges down the Avon to Sea Mills (Portus Abonae) and then loaded into sea-going vessels for dispatch to its final destination.

Barochan Hill (NS413690) Overlooking the Firth of Clyde, the 1.3 ha. fort is rectangular in plan, with an annexe of c.0.4 ha. attached to the eastern side of the fort. A fragment of a samian bowl and a number of bronze asses dated to the year 86 confirm a Flavian date for foundation. The fort is situated on the summit of Barochan Hill and commands excellent views of the surrounding area. To the east, the forts at Old Kilpatrick, Duntocher, and Castlehill are visible, whilst to the west the north bank of the River Clyde is visible as far as the Ardmore peninsula. The fort, however, had only limited views westward along the south bank of the Clyde, and it has been suggested that communication with the mouth of the estuary was maintained by a chain of watch-towers. Barry (ST099664) Located the shore of the Bristol Channel at Cold Knap, and close to a small natural harbour, a Mediterranean courtyard-type building was constructed of stone to roof height and with a tiled roof. Comprising twenty-one rooms and with corridors arranged on all four sides of a central rectangular courtyard, the building was enclosed by a U-shaped rock-cut ditch. Construction seems never to have been completed, with a wall having collapsed directly onto builders’ levels, and there is no evidence of any form of occupation debris. Pottery and coins suggest a date of construction in the late 3rd century, at a time of considerable activity on the coast of south Wales, with a Saxon Shore type fort being erected at Cardiff and evidence of re-occupation of the forts at Loughor and Neath.

Bawtry (SK626956) The fortlet lies near the medieval river-port of Bawtry, on the River Idle, 15 km from its confluence with the Don, some 8 km to the south of the vexillation base at Rossington. Evidence of later development as a settlement is suggested by a hoard of coins ranging in date from Valerian to Diocletian (235–305).

It has been suggested that the  building may have been  intended  as  a  mansio,  serving  a  crossing  point  from  the estuary of the River Parrett on the Somerset coast,  connecting  west  and  mid‐Wales  directly  to  Ilchester,  Dorchester and southwest England and therefore best  described  in  modern  terms  as  a  “ferry  terminal”. 

125

Baylham House (TM112529) - Combretovium Situated on the River Gipping, a tributary of the Orwell, a major road runs north to Caister St Edmund and south to Colchester with apparently minor roads running northeast, east and west in the direction of Cambridge and the estate centre at Stonea. A 2.1 ha. fort lies within the defences of another covering 5.8 ha. but the sequence of occupation is not known. Numerous finds over an area of

JAMES ELLIS JONES

  situated at about 95 m above the River Wear and the remains of the stone bridge used by Dere Street can still be seen when the river is low. Two altar-stones dedicated by holders of the office of Beneficiarius Consularis are known and underline its prominence in the Roman system in the north of England. A Flavian fort of between 5 and 7 ha. in area was constructed during the governorship of Petillius Cerialis (71–74) and at some time in the first half of the 2nd century the original fort was levelled and a new fort of some 3.9 ha was built by legio VI Valeria Victrix, with Dere Street forming the via principalis. The clay rampart was cut for the installation of a stone wall later in the 2nd century . There is some evidence that the fort was occupied into the 5th century, when some of the buildings were reused for non-military purposes. Substantial stone strip buildings lined Dere Street on the north-west and south-east sides of the fort and, at some time, those to the south-east, extending to at least 225 m from the fort were surrounded by a doubleditched defensive enclosure.

60 ha. have led to the suggestion that the un-walled settlement was a cantonal capital of the Trinovantes. Beckfoot (NY216647) Bibra lies on the coastal road, close to the shore of the Solway Firth, with the Dumfries coast 12 km to the north-west and the forts at Kirkbride to the north and Maryport to the south. The1.3 ha. turf and timber fort was replaced in stone, probably in the mid-2nd century and held to the 4th century (coin of Constantius II). The site of the vicus, seen as cropmarks and mapped from air photographs, is visible on either side of the road which runs out from the north gate of the fort, for a distance of 236 m. Bertha (NO097268) The 3.7 ha. fort at lies on the banks of the Almond, just north of its confluence with the Tay and it has been suggested that it may be Ptolemy’s horrea classis (storehouse of the fleet). A road runs for 20 km to Strageath via the Gask Ridge but, despite many efforts, the course of a possible road to Inchtuthil, some 15 km to the north, still awaits precise location. In the 18th century, the remains of a Roman bridge across the River Tay were mentioned by Maitland, and Roy marked six bridge-piers on his plan of the area. The old Statistical Account (SA 1795, 15,528) said that the remains of the bridge consisted of “large oak planks, from six to eight inches in diameter, fastened together by long skairs (splice or scarf joints), but coarsely jointed, and surrounded with clasps of iron, frequently twisted”. Recent investigations have failed to find any trace of this structure.

The potential navigability of the Wear in Roman times is indicated by its naming as Vedra Fluvis. At the present time, the river at Binchester is obviously capable of handling reasonably large craft and further downstream there are 14 boathouses and 20 boat clubs based on the Wear at Durham. The possibility of navigability during the Roman period is obscured by two weirs, originally created for industrial activity, which impede the flow of the river at Durham. The first, at the Old Fulling Mill, is now an archaeological museum and the second, beneath Milburngate Bridge, includes a salmon leap. Downstream from Durham, the river shallows at Finchale Priory and Roman riverine engineering would have been required to enable water transport to Binchester.

Despite the failure to establish the course of a road to Inchtuthil, it is extremely unlikely that a “penetration road”, built quickly, probably without heavy-duty metalling, was not constructed. It is probable that, had the fortress been completed, it would have been linked to the rest of the road network by a more permanent form of construction. If this were the case, the opportunity for Bertha to act as a trans-shipment port, where sea-going vessels could transfer their cargoes to river barges, for shipment to Inchtuthil and beyond, is unlikely to have been ignored. The Severan construction of a 12 ha. fortress base at Carpow, on the south side of the Tay Estuary reduced the significance of Bertha and this is perhaps the reason why there is no evidence of a 2nd century road.

Bishopton (NS418720) The 1.3 ha. Antonine fort at Whitemoss Farm, Bishopton, lies on the south shore of the Clyde Estuary with a magnificent vista across and along the Clyde. On the opposite bank of the Clyde, about 5 km to the west, lies the fort at Old Kilpatrick at the western end of the Antonine Wall. An earlier Agricolan fort is situated 3 km to the south at Barochan, on a low hill overlooking the Clyde Estuary. Excavation of the principia revealed four different phases of construction, all during the Antonine period. The third period had ended in destruction by fire. On aerial photographs, a ditch was located running from the north-west angle of the fort for some distance down the slope to the north and may indicate an annexe.

Berwick-on-Tweed (NT997560) The Devil’s Wades Causeway runs northwards for 40 km from Learchild directly towards present day BerwickUpon-Tweed. It would be surprising if a Roman fort and harbour, as yet undiscovered, did not lie on the south bank of the Tweed. There is little evidence of Roman finds in the area and it is possible that, prior to erosion, the fort/harbour lay to the east of the present coastline.

Bitterne (SU434133) - Clausentum Situated on a bend in the River Itchen above Southampton Water, the small town served as a seaport for Winchester which lies 18 km to the north;; a road leads eastwards in the direction of Chichester, some 40 km distant. During the Flavian period, a ditch was cut across the Bitterne promontory enclosing an area of 11 ha. About 150 a wooden palisade was erected inside the ditch, reducing the fortified area to 3.2 ha., within which wharves were built and stone buildings, including a

Binchester (NZ210313) – Vinovia/Vinovium One of the largest and longest permanently occupied Roman forts in the north of England, Binchester is

126

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  bathhouse, were constructed. Towards the end of the 3rd century, there was renewed constructional activity, including the replacement of the wooden stockade by a stone wall and the rebuilding of the bathhouse, leading to an unresolved debate as to whether Bitterne replaced Portchester as a Saxon Shore fort

Bradwell (TM031082) - Othona The late-3rd century Saxon Shore fort was constructed on a 30 m high promontory, based on a trapezoidal shape with external bastions, but the eastern wall has been eroded, making it impossible to determine the full extent of the fort. It is probable that most of the structures inside the fort were of timber construction and there is evidence for extramural settlement in the areas to the south and west of the defences including finds of roof tiles that suggest the existence of some substantial buildings. Coins dating to Arcadius (383-408) have been recovered.

Blackwardine (SO534567) The small town lies at a height of between 100 m and 120 m on the line of the Roman road from Weston-underPenyard, but the destination of the route to the north has not been established. The town is situated 2 km to the west of the River Lugg (a tributary of the River Severn) and 3 km upstream from Leominster, to where later navigation is known. Artefactual evidence indicates that the Romano-British settlement was established in the later 1st century, and it is probable that the town was provided with defensive walls in the 2nd century. The quality of the archaeological evidence is not sufficient to determine the development or extent of the town, but there is evidence for timber and stone buildings with stone and tile roofs. Evidence for hypocausts, painted wall plaster and mosaic floors, indicate some higher status buildings within the town. Coin and pottery evidence suggests that occupation continued into the 5th century, the latest coin being one of Honorius (393-423).

Brancaster (TF782440) - Branodunum Lying on a tidal inlet on the Norfolk coast during the Roman period, deposition means that the remains of the early Saxon Shore fort are now almost 1km inland. Probably preceded by a 2nd century fort, the walls of the fort enclosed 2.56 ha and the absence of projecting bastions suggests an early foundation. On the eastern side of fort, a large (23 ha.) civilian settlement, with an apparent regular grid system of roads, has been dated to the 2nd century, suggesting development from the vicus of an earlier fort. Brecon Gaer (SO002297) – Cicucium ? The 3.4 ha. fort at Y Gaer is located 150m to the north of the River Usk, close to its subsidiary the Yscir, 5km west of Brecon. The late-1st century fort lies close to the eastwest road from Caerleon to Llandovery and thence to Carmarthen and west Wales. A further road leads to Neath to the south-west, and northeast to Clyro, Clifford and Kenchester. The falls on the river near Llangynidr prevent direct access by river, and a portage would have been needed. The defences and principle buildings were rebuilt in stone at some time during the middle of the 2nd century, with the barracks remaining in wood. A small internal bath building suggests probable reoccupation by a small holding-force in the 3rd century and, in the final phase of occupation probably in the late-4th century, this building was converted to living quarters. There is evidence of a substantial civilian settlement on both sides of the road, leading from the north gate for a distance of some 300 m. One stone building may have been a mansio, a second was probably a workshop; most of the other buildings were of timber construction.

Blennerhasset (NY190413) The 3.4 ha. fort is located on a bluff to the south of the River Ellen, on the road from Old Carlisle to Papcastle. Late-Neronian pottery suggests that the fort was founded during the earliest Roman military campaigns in the north-west during the governorships of Cerialis or Frontinus. There are crop marks of a rectangular building adjacent to the fort’s rampart on the north-west side. The fort is larger than any other known Cumbrian fort, suggesting that it was garrisoned by a unit larger than the standard quingenary garrison of about 500 troops. Bowness-on-Solway (NY230627/NY223627) - Maia A marching camp is situated at The Old Police House on the south side of Solway Estuary and has been partly been eroded by the sea. The camp is located to the north of Hadrian’s Wall, which it clearly predates and might be interpreted as a bridgehead for crossing to the north coast of the Solway to link up with the marching camps on the River Annan. The Admiralty Pilot for 1922 notes a ‘cartford to Dornock on the opposite shore, a distance of 1¾ miles’.

Brithdir (SH773188) This late-1st/mid-2nd-century 0.4 ha. fortlet lies south of the Afon Wnion and its confluence with the Afon Mawddach; the tidal limit is at Dolgellau, 4 km to the west, where sea-going vessels berthed in the Tudor period. The fort is situated at an altitude of 150 m commanding good views of the valley of the Wnion, and lies on the road leading to Caer Gai, some 17 km to the north-east. Extramural buildings, situated within a large polygonal enclosure to the south of the fort, included tanning and metalworking sheds and a bathhouse. The site was abandoned about 120; the buildings were dismantled, the ditches backfilled, and a road built over the remains. The location of Brithdir is not suitable for the control of the Mawddach estuary, and it is probable

An early timber fort was replaced by a 2.38 ha. stone fort in the late Hadrianic period and underwent major rebuilding in the 3rd century. Apart from the Antonine advance into Scotland, the fort was held until the 4th century (coins of Gratian). The remains of the vicus, an alleged bath house, and a possible road, have been observed on the three landward sides of the fort. In 1599, Camden noted walls, traces of streets and a silted-up harbour.

127

JAMES ELLIS JONES

  that a fort existed to the west, but this has not, as yet, been located.

Roman road, running through mountainous country to the north.

Brough on Fosse (SK837585) - Crococalana Straddling the Fosse Way, to the south-west of Lincoln, and 3 km to the east of the River Trent, a bank-and-ditch enclosure of 3.25 ha. defended the small town of Crococalana. Occupation continued until the late-4th century and a lead fragment, with a symbol similar to Chi-ro, has been interpreted as part of a Christian baptismal font.

An annexe to the west of the fort, covering an area of 1.3 ha., was surrounded by a rampart, and contains a complex of substantial stone buildings, still standing to a height of about 1 m above ground level. In 1807, Samuel Lysons recorded “an abundance of building materials have been taken from these remains, for several years past” and that “a room containing hypocaust pillars was visible”. The recovery of tiles and short stone pillars indicates a hypocaust system and suggest use as the fort baths, later expanded to include accommodation to be used as a mansio.

Brough-on-Humber (SE938268) - Petuaria Brough lies on the coast of the Humber Estuary, providing the northern terminal of the 4 km ferrycrossing from Wintertingham. Roads lead north-west to York and north to Malton and, from the other side of the Humber, Lincoln lies 50 km to the south. A fort defended by a rampart and ditch enclosing 1.8 ha. was established c. 70, with a possible break in occupation during the later-1st century, and was later expanded to include a military supply-base function,. It is probable that a harbour lay to the south-west of the fort, but there is evidence that a rise in sea-level resulted in flooding and silting, making it unusable by the late-3rd century. A defensive circuit enclosing 5 ha. was constructed in the late-2nd century and a stone wall was added in the late-3rd century with, in the 4th bastions and projecting gatetowers being added to the circuit. At this stage, Brough had many features in common with the Saxon Shore forts to the south.

Burgh-by-Sands (NY3258) - Aballava Lying on the south side of the lowest fords across the Solway, a late-1st century 1.3 ha. fort is aligned with a Roman road that runs from the direction of Carlisle, some 4 miles to the east, along the humpback ridge of Fingland Rigg, to the fort at Kirkbride. The road passes directly in front of the fort, and probably represents a westward extension of the Stanegate frontier system, sometime during the Trajanic period, although it is possible that the fort may have been founded during the campaigns of Agricola around 79/80. A two-period fort initially of seven acres but later reduced in size to 3.7 acres is situated on top of a commanding hill, straddling the road some 1000 yards to the south of Burgh village and was built as part of the initial scheme for Hadrian's Wall. A 19m wide circular crop-mark lying within the defences of the fort, but of earlier construction, has been identified as a four-post watch-tower with a circular ditch. Pottery from the site lies within the period from 90 to 130 and suggests that the fort was briefly occupied prior to the construction of the Wall itself. Recent excavations at the village of Amberfield, approximately 400 m to the south of the fort suggest the presence of a vicus. There is clear evidence of rubble surfaces, possibly roads or yards, a timber building, hearth, and possible well. The evidence suggests that occupation of this area was fairly intensive, and probably covered the whole of the field.

On the basis of a 2nd dedication of a new proscaenium of a theatre by a junior magistrate of the vicus Petu[ariensis], that had been incorporated in the masonry of an early-4th century building, Brough is usually assumed to be he civitas capital of the Parisi, but no evidence of the range of massive public buildings usually associated with such status have been discovered. However, it is probable that the civil town (including the theatre mentioned on the dedicatory inscription) lay outside the defended circuit in the direction of the area now occupied by North Ferriby. Bryn-y-Gefeiliau (SH746572) Lying on the road running between Caerhun to the north and Tomen-y-Mur to the south, the late-1st century 2.4 ha. fort (later reduced to1.57 ha.) was clearly intended to control the valley of the Afon Llugwy on whose banks it stands. Pottery indicates a late-Flavian foundation, with occupation lasting into the second half of the 2nd century. The fort occupies a level meadow within a broad loop of the river, and stands only a couple of metres above flood level. The river falls rapidly over 5 km to its confluence with the Conwy at Betws-y-Coed; a number of waterfalls, the most notable being Swallow Falls, would have prevented navigation along the river. However, the 15 km stretch of the River Conwy between Caerhun and Betwsy-Coed could have been navigable, and presented a favourable alternative to the presumed route of the

Burgh Castle (TG475046) - Gariannum At present lying 5 km inland on the River Waveney. overlooking the entrance to the River Yare, (Gariannus Fluvis - “babbling river”), the Saxon Shore fort, probably with a 1st century predecessor, lay on the southern side of the 7 km wide “Great Estuary”, only a short distance from Caister-on-Sea on the opposite side. The early-4th century fort was built on a 30 m high spur and of a trapezoidal quadrilateral shape with projecting bastions, enclosing 2.8 ha., but there has been insufficient excavation to determine the interior of the fort. An extensive vicus became established on three sides of the fort. Burrow Walls (NY004300) - Axelodunum On the coastal road from Beckfoot, with a presumed road

128

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  to Moresby to the south, the c. 2 ha. fort is situated on an old clifftop, slightly higher than the marshy area to the west and a little to the north of the mouth of the River Derwent. The western third of the fort has been lost to coastal erosion. Excavation revealed it to be a typical 2nd century fort, with evidence for a possible smaller fort within the original ramparts. The site has produced evidence for Hadrianic occupation but little pottery other than 4th century. It may therefore be that the smaller fort is a new foundation, built in response to a threat from across the Irish Sea, possibly at the same time as some of the Solway fortlets and towers were re-occupied.

origin of both the dock and the road leading to it”. Building materials were probably brought in by sea, as demonstrated by a sandstone column from Bodysgallen, near Llandudno, a sandstone base originating from Runcorn in Cheshire, and a plinth from the angle of the West Gate identified as having been quarried at Tattenhall, also in Cheshire. Caerleon (ST340906) - Isca Trans-shipment to barges was needed to supply the existing legionary fortress at Usk; the move to Caerleon was intended to provide supplies to the army through a base accessible to sea-going ships. Constructed by legio II Augusta, the new fortress was located on the River Usk, the Roman Isca Fluvis, 14 km from the sea, with the tidal limit a further 11 km upstream at Newbridge-onUsk. The Roman bridge was located up-river of the early 19th century stone bridge from where, in medieval and later times, merchant ships sailed from quays on the riverside to Bristol, Ireland and many Atlantic coast ports. A well-built stone and timber quay, lay 230m from the present line of the riverbank, at a height of 6.56 m above Ordnance Datum. It is extremely unlikely that it was the only quay, since it was not constructed until the early 3rd century, and port facilities were needed from the time of foundation of the fortress. On-going excavations by Cardiff University are revealing extensive riverside structures adjacent to the fortress walls. An inscribed stone from Gold Cliff on the Caldicot Level records the completion of 33 pedes of work by the century of Statorius Maximus of the First Cohort of the legion. This may have been part of the reclamation of land to enlarge the territorium of the legion, without encroaching on the lands of the civitas at Caerwent.

Caer Gybi (SH247826) The fortified harbour is located on a 5 m cliff in the modern Inner Harbour at Holyhead on Anglesey. The north and south walls of the fort run down to the lowwater mark, allowing boats to be beached on the shore. At the points where the walls cross the cliffs, and at both inland corners, there were round towers some 5 m in diameter. The positioning and style of building of Caer Gybi is similar to the group of coastal fortifications of the 3rd and 4th centuries, the Saxon Shore Forts and, whilst no similar example of a three-sided landing place is known in Britain, there are others on the Rhine frontier, such as Engers and Zullestein, dated to the reign of Valentinian. However, Symonds (2009, pers. comm.) is of the opinion that the fort was constructed with four sides and that the seaward structure has been destroyed by coastal erosion. A mile to the north is a watch-tower intended to give warning of approaching hostile vessels. Coins recovered from the interior date to the 390s, indicating that it was contemporaneous with the late occupation of Segontium. Caerhun (SH775704) – Canovium Canovium is situated 5 km from the mouth of the Afon Conwy at a point that is tidal and accessible to vessels of up to an estimated 100 tons burden. The 1.97 ha. fort is situated at the crossing of the Conwy by the main eastwest road linking Chester with the major fort at Caernarfon, 32 km to the west. There is no known fort between Caerhun and Chester, a distance of 63 km, but a possible fort at Prestatyn on the coast may have controlled the valley of the River Clwyd and the junction with the road leading south to Ruthin. A mountainous road to the south leads to Bryn-y-Gefeiliau and Tomen-yMur, and probably continued via the coastal forts as far south as Carmarthen. Coins date the occupation from the Flavian to the Theodosian periods, with a peak occurring in the late 3rd century, suggesting an intensification of occupation, or perhaps a reoccupation, possibly as a reaction to the threat posed by Irish raiders. The vicus lay 200 m to the north of the fort and suggests that the open space may have been used as a market place, for stalls or the penning of animals.

The fortress covered an area of 20.5 ha. and, when first built about 75, the defences were of turf, clay and timber, with the bank being fronted with a wall of mortared masonry, the gate towers being replaced in stone by110. The baths were constructed around 80 and refurbished thirty years later, with a major reconstruction carried out c.170. An amphitheatre with seating for 6000 was built around 90, the lower part built in stone, with an upper timber grandstand. When Hadrian’s Wall was constructed, a large part of the Roman army in Britain was drafted to the north and inscriptions from the Wall record the work of seven out of the ten cohorts of legio II Augusta. When the frontier was advanced into southern Scotland under Antoninus Pius, a further series of inscriptions show that much of the legion was still on the northern frontier. There is archaeological evidence for the continued occupation of perhaps one barrack block per cohort during the later 2nd/mid-3rd centuries, but the fortress became a regimental base depot rather than an active military garrison. The fortress baths were still in use until about 235, perhaps under the supervision of a care and maintenance unit, when the 7th Cohort returned in 253–8 their barracks needed to be totally rebuilt; a recently discovered

A possible Roman dock was excavated by Reynolds in 1930, but without conclusive proof of dating, however the excavator considers that “on the whole, the balance of probability may be said to be in favour of the Roman

129

JAMES ELLIS JONES

  viable alternative route. This river is tidal to the fort at Cardiff, and improvements over a further 11 km would have reached a point within 4 km of Caerphilly.

inscription records further building activity under Aurelian in 275. There is evidence of dismantlement of the legionary fortress in the last years of the 3rd century’, but some parts of the site continued in occupation through most of the 4th century. The generally squalid traces of habitation were of civilian rather than military origin, and occurred after the legion had left Caerleon.

Caersws I (SO041926) Caersws II (SO029920) Caersws I is situated 1.1 km due east of the modern village and is almost certainly a pre-Flavian foundation. Located on a spur overlooking the Severn, the 3.8 ha. fort does not occupy a commanding position, and is liable to flooding, with better tactical positions nearby on higher ground. The later Caersws II fort formed the pivot of the road system controlling central Wales; roads to Pennal, Forden Gaer, Caer Gai and Castell Collen converged on the site, whose possible Roman name was Mediolanum. The site is a natural communication centre for the upper Severn Valley and its importance is reflected by its size, which places it among the largest of the Welsh forts. There appears to have been several phases of occupation, the first fort being defended by a clay rampart with all buildings of timber construction, with possible preFlavian pottery. At some time later in the 1st century, the buildings were demolished and the rampart slighted; it is probable that a break in occupation took place at this time. The buildings of the second fort were also of timber, within a raised rampart and with a more spacious layout, perhaps implying a change in the composition of the garrison. In the Hadrianatic/Antonine period, the fort was completely rebuilt within a stone-faced rampart, and while the barrack blocks were again of timber, the central buildings were built from sandstone that outcrops 20 km down river, and transport by water is therefore a strong probability. The vicus covered some 7 ha. with the possible temenos of a Romano-British temple, and a flourishing commercial centre with a continued existence into the 4th century.

Caernarfon (SH 485625) – Segontium The 2.27 ha. fort is located on the tidal mouth of the Afon Seiont, guarding access to the Menai Straits and the Isle of Anglesey, and though no evidence has been located, it is probable that a road ran due south to the otherwise isolated fort at Pen Lystyn, and from there to Tomen-yMur and the forts of the west coast. Established in 77 on the final conquest of the Ordovices by Agricola; a coin of Theodosius I (379-95) demonstrates that the fort remained in use until the end of the 4th century. The continued occupation over such a long period is attributable not only to the strategically important position, but also to the mineral wealth of the area, important silver-lead ores being found in Flintshire, and copper ore being mined on Anglesey and the Ormes. A courtyard house of 2nd century origin and a small bathhouse situated within an enclosed area to the west of the via praetoria has been interpreted as the residence of an official in charge of mineral extraction (procurator metallorum) for the surrounding region. As the fortress at Chester is some 100 km distant by road, it is probable that civilian administration of the area was also centred on Segontium. The praetorium, granaries and barrack blocks were constructed from Cheshire sandstone, and bricks and tiles from the legionary works-depot at Holt on the River Dee were used in quantity. A rectangular enclosure measuring 70 by 15 m with walls 1.8 m thick, and in places still standing to a height of 5.8 m, lies about 140 m to the west of the fort, at the mouth of the Afon Seiont. Known as Hen Waliau (“Old Walls”) it was originally interpreted as a late Roman fort, but details in construction of the wall resemble those of the fort proper, having the same round scaffold holes. It is therefore likely to be of the same date as the fort wall, and is best regarded as a large military stores compound..

Caerwent (ST489905) - Venta Silurum This present small village is 3.2 km from the Bristol Channel, lying on the main road from running from Gloucester), to Caerleon and beyond. Known as Venta Silurum (Market of the Silures) it became the administrative centre and tribal capital of the Silures covering an area of some 18 ha., with a planned grid system of streets dividing the town into 20 insulae containing the public buildings, private houses and shops. In addition to the forum-basilica, there were two temples, a public bathhouse and a mansio. It is probable that Caerwent did not acquire defences until late in the 2nd century, comprising ditch, earthen rampart, and wooden palisade and timber gatehouses. Probably around 330 a stone wall was built in front of the rampart, stone gatetowers were added and a number of bastions were added to the south and north walls.

Caerphilly (SO154873) The late 1st century fort is situated on a mound of glacial gravel, commanding good views eastward down the valley of the Rhymney, towards the mining settlement at Machen and the fortress at Caerleon. The fort is adjacent to the road leading south to Cardiff and north to Gelligaer. As the crow flies, Caerphilly lies only 11 km from the sea, and at first sight appears to be a prime candidate for supply and replenishment by water transport along the River Rhymney, but the tortuous course of the river increases the distance by water to 25 km. However, if the movement of ore from the mining settlement at Machen had merited improvement to the river, further improvements over the remaining 6 km might have been justified. If Caerphilly was, in fact, supplied by water, the River Taff may have provided a

Caerwent clearly enjoyed a significant period of prosperity during the 4th century, with the building of elaborate and well-constructed new houses with lavish decorations, and alterations, including the laying of mosaic floors, carried out to existing dwellings. This period coincided with the menace of seaborne raiders

130

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  from Ireland, using the Bristol Channel and Severn Estuary to plunder the rich hinterland of the Cotswolds. Late military equipment has been found at Caerwent, including two plumbatae (lead weighted darts characteristic of the 4th century), military belts and buckles, and crossbow brooches usually associated with military uniforms and, given that the legionary fortress at Caerleon had been virtually abandoned by that time, it is probable that there was some form of military presence in the town at this time.

Cambridge (TL444592) - Duroliponte Cambridge is situated on the River Cam, at the highest navigable point for seagoing vessels, on the site of an earlier Belgic settlement. The Via Devana ran north-west for 25 km to Godmanchester and south-east for 20 km to Great Chesterford, whilst Akerman Street joined Ermine Street 15 km to the south-west and then north-east towards The Fens. Duroliponte is an obvious site for a fort, and whilst its site has not been precisely identified, there is sufficient evidence of military occupation to confirm its existence during the Claudian period. Most of the subsequent settlement buildings were built of timber, with either thatched or tiled roofs; only one stone building has been located. The 4th century defences, comprising a bank, ditch and stone wall enclosed an area of 10 ha. and occupation continued into the 5th century with the coin series extending down to the reign of Honorius.

There has been a persistent local tradition that Roman ships were able to reach Caerwent via the Nedern Brook, and that mooring rings, enabling vessels to tie up to the walls, were in existence until modern times. However, it is more likely that the landing place was at Black Rock, near to Sudbrook, where a Roman garrison was installed within the ramparts of the Iron Age fort, presumably to control a ferry crossing of the Severn Estuary.

Camelon (NS863809) The Antonine fort at Camelon occupies a small eminence in a meander of the River Carron that had previously been the site of a fortified native enclosure. Its military significance is emphasised by the size of the fort (3.2 ha) and the presence of ten marching camps in the immediate area. A Flavian fort or forts originally occupied the south part of the site, but whilst the Flavian period is relatively well represented in terms of pottery and coins, its structural remains have been almost completely destroyed. The Antonine fort was constructed c.139, when the Wall was built some 1070 m to the south. A road leads south to Watling Lodge fortlet on the Wall and another leads north towards Ardoch and it has been suggested that an Agricolan road ran to Mumrills and may have been still usable. It is clear that imports reached Camelon in large quantities, for example, the whole surface of at least one path had been made up from amphorae sherds.

Caister-on-Sea (TG517123) During the Roman period Caister was on a large island at the mouth of the “Great Estuary”, a major tidal inlet, some 7 km wide, leading to the rivers Bure, Yare and Waveney. On the opposite side of the estuary to the Saxon shore fort at Burgh Castle; there are no known Roman roads serving the area. The site was originally interpreted as a harbour and settlement, defended by a wooden palisade, with stone walls enclosing an area of 3.5 ha. being added in the early-3rd century. The unusually early date for the addition of stone walls to a civilian settlement and similarities of morphology to both Reculver and Brancaster, have led to reinterpretation as early Saxon Shore fort. The interior of the fort covered 2.8 ha with a regular street grid and numerous masonry buildings, including a high-status courtyard house with window glass, painted wall plaster and a hypocaust. Caistor St Edmund (TG230035) - Venta Icenorum Lying 8 km to the south of Norwich, on River Tas, a subsidiary of the Yare, Caistor St Edmund was founded about 70, following the Boudiccan rebellion. The only known road leads southwards for 50 km to Baylam House, then for a further 25 km to Colchester; supply by sea via the estuary of the Yare (the location of both Caister-on-Sea and Burgh Castle) must be a distinct possibility. With no known military predecessor, the town was planned and organised as the civitas capital of the Iceni, with a forum, basilica and public baths being built about 130. At the beginning of the 3rd century, the city was enclosed by stone walls, defending an area of 14 ha. and at the same time, two temples were constructed to the north of the walls; a possible amphitheatre lies to the south of the town’s defences.

During the campaigns in Scotland, the River Carron may have been navigable to its tidal limit near to the fort and Arthur’s O’on, usually interpreted as a temple to the goddess Victory, situated on a 30 m raised beach above the river, and clearly visible to ships coming up the Firth of Forth may have been a leading mark, or even a lighthouse, indicating a channel leading into a harbour. In 1695 Camden referred to the discovery of an anchor “within these hundred years”, and to a former “road for ships” on the river at Dorrator. In the 18th century, Macfarlane referred to Carron as “A good harbour for small boats and barks, yea sometimes at spring tides there comes ships here of 60 tun burden”. A fortified annexe on the north side of the fort contained an industrial area with a stone-built building, five timberframed buildings and a number of iron-smelting furnaces and smithing-hearths. An area of ditched enclosures, some 500 m west of the fort, indicated activity of a semiindustrial nature. In 1848, a lead pig weighing 38 kg was found in association with Roman pottery to the northwest of the fort and, a year later, a second pig was found,

Calstock (SX436691) A probable 1st century 1.82 ha. fort was located in 2008, and the results await publication. Located on a spur above the former river-side port of Calstock on the River Tamar 1.82 ha., the site is some 4.8 km downstream from the historic port of Morwellham Quay.

131

JAMES ELLIS JONES

  weighing 102 kg. The pigs were probably from one of the outer annexes and may have been landed at the nearby harbour as raw material for reworking elsewhere, possibly at Rough Castle, where a stone mould for pewter has been found.

Cardiff (SO180765) Finds of pre-Flavian pottery and copies of Claudian asses, indicate an early foundation, followed by a succession of forts on approximately the same site; the first was probably a vexillation base (c.55), followed by a conventional fort of the Flavian period, with a further construction early in the 2nd century. A Saxon Shore type fort was built in the late 3rd century and was occupied until the time of Valentinian (364-78). It is likely that the fort served as headquarters for naval units operating in the Bristol Channel, with detachments being based at reoccupied forts at Neath and Loughor, and probably at the civitas capital of Carmarthen. Excavations within the walls have not revealed any substantial buildings and it is probable that the garrison lived in wooden hutments - a similar situation has been observed at the Saxon Shore forts at Richborough and Portchester.

Canterbury (TR148576) - Durovernum On Watling Street, with Rochester to the south-east and London to the north-west and Dover to the south, with roads also leading to the Saxon Shore forts at Reculver and Lympne, Canterbury was situated on the (in Roman times) navigable Great Stour, 23 km distant from the Wantsum Channel at Richborough. As Canterbury lay on the direct route westwards from Richborough some form of military installation would be expected, but has not, so far been located. Canterbury became the cantonal capital of the Cantiaci, possibly as early as the Neronian period, and the usual complement of public buildings, forum, basilica, bath houses, temples and undoubtedly existed but, of these, only the theatre, built during the Flavian period and completely rebuilt in the early-3rd century, has been subject to extensive excavation. There is no evidence for fortifications being erected around Canterbury until the later-3rd century when walls built of coursed flints and mortar, together with internal towers and external bastions, enclosed an irregular octagonal area of 48 ha.

A reconstruction of part of the fort at Cardiff Castle was carried out at the expense of the Marquis of Bute at the end of the 19th century, and is similar to the Saxon Shore forts of south-east England. The reconstruction illustrates the semi-octagonal towers and two single-arched gates, each protected by guard towers. Although the gate towers are probably constructed to approximately the correct height, the adjacent curtain wall is certainly too high and its crenulations are more likely to have been level with the first floor windows in the gate towers.

There is evidence of occupation at Canterbury extending into the 5th century, for example, a worn coin of Arcadius (395–408) found within the collapsed floor of a fired hypocaust in one of the town houses. There is some evidence for the presence of foederati in the mid-5th century but in contrast to, for example, Cirencester, there is no evidence for sub-Roman occupation. The River Stour was still navigable when the Vikings carried out a successful raid on Canterbury in 839, but the continued deposition of shingle at Stone Neach led to the gradual silting up of the channel and, by the 10th century, it was no longer possible for coastal vessels to reach Canterbury.

Cargill (NO166379) A timber-and-turf Flavian Fort of some 1.94 ha. and a 0.5 ha. fortlet lie on the southern side of the River Isla just above its confluence with the Tay, 13 km upstream from Bertha and 4.8 km downstream from Inchtuthil. A road runs south-west towards Bertha and probably extended north-eastwards to Cardean. Occupation of the Flavian fort probably ended c. 80 but recent excavations have produced well-worn coins of the Trajanic and Hadrianic suggesting Antonine and/or possible Severan reoccupation. The Isla is navigable at this point and an annexe running down to the river may have been a defended landing area.

Cardean (NO289460) The Flavian fort is situated on a promontory close to the confluence of the River Isla and the Dean Water and, at 3.7 ha. is one of the largest forts in Scotland, with annexes on both the south-west and south-east sides. The south-west annexe may have extended down to the Dean Water, acting as a fortified landing area, as was the case some 15 km downstream at Cargill.

Carlisle (NY399559) – Lugovalium Lying on the navigable River Eden, 12 km from its entry to the Solway Firth, the 3.2 ha. fort was located on the main road to the north, with roads leading east to Castlefields and eventually South Shields, west to the Solway Frontier forts and south-west to Old Carlisle. A turf and timber fort was constructed during the autumn/winter of 73, known from dendrochronological dating of timbers from the south rampart. A major rebuild within the existing defensive perimeter occurred in 83, and a minor refurbishment took place in c. 93-4, with work on the defences and some new buildings were erected but the fort was completely demolished around 103-5. After a short break in occupation, the fort was entirely rebuilt, again in timber, probably similar in size and layout to its predecessor. During the Hadrianic period, the fort may have evolved from a conventional auxiliary fort into some form of works depot. In the fabrica, an important cache of articulated armour

In addition to the usual range of buildings of an auxiliary fort, granaries, barracks, etc., one large rectangular building has been interpreted as a fabrica and a second, which produced an unusually rich finds assemblage, has been suggested as a high status, residential building, possibly an extended praetorium.

132

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  There is only limited evidence of pre-Flavian military action in south-west Wales and it has therefore been argued that, in contrast to the Silures, the Demetae were rather more peaceable, or actively pro-Roman, than their neighbours. It is possible that this led to the granting of self-government and the establishment of Moridunum Demetarum as the civitas capital of the tribe. By 150, the town had a planned and regular street layout, even if not strictly conforming to the rectangular insulae layout seen at Caerwent, and this may coincide with the designation as a civitas capital. A bathhouse and a possible mansio have been identified on the southern side of the town; a Romano-British temple, similar to that found near the forum at Caerwent, was built during the late-1st century but demolished before the end of the 2nd century. Cutting back into the hillside and using the excavated soil and rock to form the outer seating bank enabled the construction of a theatre/amphitheatre, lying to the north of the defences.

fragments (lorica segmentata) was deposited towards the end of this period. The change in function may have been related to the need to provide the equipment required for the construction of Hadrian’s Wall and for the fort at Stanwix, less than 1 km distant. The fort was again demolished in the early 140s, perhaps as a consequence of the Antonine occupation of southern Scotland. Intermittent occupation occurred during the second half of the century but the site does not seem to have been used as a conventional auxiliary fort. Extensive rebuilding in stone took place during the early 3rd century, however the functions of the new buildings are uncertain. Minor repairs were undertaken to many buildings and a hypocaust was constructed within the principia. There was continued occupation of the site into the 5th century; heavy coin loss outside the principia suggests that a market place existed in the late 4th century. Carlisle (NY399559) - Civitas Carvetiorum The civilian settlement at Carlisle initially developed in the form of a vicus, already extensive by the early 70s, on both sides the road leading southwards, and later by occupation and extensive modification within the walls of the fort. By the later-2nd century, it had expanded to a settlement of 33 ha. and performed an important regional function as, by 105, a centurio regionis had been appointed. At some time between 222 and 230, the town became the Civitas Carvetiorum, probably so constituted by Septimius Severus. This change was swiftly followed by the construction of civic buildings including a forum, temples and bathhouses. New roads were laid out, creating rectilinear insulae with yards and metalled lanes overlying the remains of the earlier fort

Defensive walls were built to enclose an area of some 13 ha., during a single period of construction during the late 2nd century. Carmarthen was provided with new stone defences at some time in the late-3rd/early-4th century, with stone gate-towers along the circuit walls. The importance of water transport to Carmarthen cannot be overestimated. During the late-3rd/early-4th century coastal defence work involving the reoccupation of the forts at Loughor and Neath was carried out, and at this time the position of Carmarthen would have made it ideal as an advanced base for a naval flotilla. Carnforth (SD 489712) In the late-18th century a Roman anchor and a quantity of timber ‘evidently’ (?) prepared for shipbuilding purposes, were found at Galley Hall on a stream leading to nearby Morecambe Bay.

There is good evidence, both archaeological and documentary, that some parts of Carlisle continued to be inhabited well beyond the fourth century. For example, in a house with a hypocaust; a gold solidus of Valentinian II (375 to 392) was sealed below several subsequent floor surfaces. The ‘Lives’ of St Cuthbert of Lindisfarne (c. 634-87) refer to Roman walls and a fountain in ‘Lugubalia’ still functioning in the late 7th century.

Carpow (NO207179) The 12 ha. Severan vexillation base at Carpow lies on the south side of the Tay Estuary, just below its confluence with the River Earn, and is located within a 28 ha. polygonal enclosure, with ditch and rampart, which may have been a defended beachhead for an initial landing.. The main buildings were solidly built of stone, there was a building inscription dedicated to Caracalla by legio II Augusta and a tile stamp of legio VI Victrix, and an aqueduct had been constructed, indicating that permanent occupation was intended, but the fortress was abandoned by 215.

Carmarthen (SN416214) - Moridunum Demetarum An early fort was succeeded by the civitas capital of the Demetae. The fort lay on the tidal estuary of the Afon Tywi (Tuvius Fluvis), 15 km from the sea, and beneath the modern town. Roads leading east along the valley of the Twyi to Llandovery, and north-east across hilly terrain to Llanio, have been identified. A road leading westwards towards the coast has been traced over a distance of 42 km, and probably led to a coastal fort or forts, whose sites have yet to be discovered. The fort was probably founded following the campaigns of Sextus Julius Frontinus in 75, and it has been estimated that it originally occupied an area of 1.75 ha. Occupation until 110 is probable, when there is evidence of slighting of the defences, a systematic programme of demolition and salvage, and a final clearance of the site by burning.

The fortlet at St Madoes on the opposite side of the estuary may have acted as a fortified bridgehead, supporting the crossing of the Tay. A coin of Caracalla dated 209 shows an army crossing a bridge of boats, and it has been suggested that such a crossing may have been made at Carpow. Carriden (NT025808) - Velunia The Antonine fort at Carriden) lies 1.5 km to the east of the Wall terminus at Bridgeness and is 8 km distant from

133

JAMES ELLIS JONES

  navigable and that Catterick acted as a trans-shipment point between river, road and depot for military supplies.

the fort at Inveravon and 17 km to the west of the fort at Camelon, At 1.6 ha it is similar in size to Old Kilpatrick at the western terminus of the Wall. The fort lies within 100 m of the present shoreline, with an annexe on its eastern side extending to a tidal inlet with a further annexe to the south, surrounded by a ditch and a stone rampart, and containing a bath-house. An altar was discovered during ploughing in the field to the east; this was dedicated by the vicani, clearly demonstrating the presence of an extensive extramural settlement..

Carzield (NX968818) It was considered that Carzield was garrisoned only during the first occupation of the Antonine Wall, probably in support of its western supply-route but a range of Roman material was recently recovered from the vicinity of the fort, including coins from Trajan to Antoninus Pius, so an earlier foundation is possible. The 3 ha. fort lies 4 km to the north of Dumfries on the River Nith. The large concentration in Nithdale of Roman forts and marching camps (e.g. Dalswinton) would have made supply by sea of great value; improvements could have been made to make the river navigable above Dumfries itself.

Castle Hill (SK700416) - Margidunum Situated on the Fosse Way to the south-east of East Stoke and north-northwest of Leicester, significant finds of military equipment indicated army occupation but no structural evidence for a fort has yet been recovered. Following abandonment by the army, the vicus grew into a civil settlement on both sides of the Fosse Way and, in the late-2nd century, was enclosed by an earthwork defence, and with an area of 2.2 ha., and a stone wall was added in the 3rd century.

The Nith flows into the Solway Firth, and Dumfries developed as a port for sea-going ships, becoming a major port after improvements were made in 1815. The river is tidal to the north of Dumfries and, despite its decline as a port, is still navigable for small craft.

Castleford (SE426258) - Lagentium Located at the point at which the road from the Doncaster to York crosses the River Aire downstream of its confluence with the Calder, the timber and turf fort of some 3.2 ha. was built by legio IX Hispana in c.70, during the governorship of Cerialis, with another fort being constructed shortly after 86. The fort was abandoned in c.95 and military occupation came to an end, but the vicus developed into an extensive settlement with a mansio, defended in the late-Roman period by the construction of a stone wall, fronted by a system of ditches enclosing an area of 3.75 ha.

Charterhouse (ST500565) - Veb? A 1 ha. Claudian fort of c.45-50 was succeeded by fortlet, measuring 55 m by 50 m, with pottery ranging from the mid-1st to the early-2nd century. A military presence in this location would have little tactical significance, and its function was most probably for the administration and control of mining activities. It is probable that lead mining took place on Mendip before 43, and the Romans rapidly exploited the mines around Charterhouse and a small town grew up in the area, with the settlement lasting into the 4th century. The earliest ingot from the mines is dated to 49; Mendip ingots of the same period have been recorded in northern France, indicating that lead was being exported to the continent, only six years after the Claudian invasion. It has been suggested that a cistern at Pompeii, buried by the eruption of Vesuvius in 79, was made of British lead, probably from Charterhouse. The last securely datable ingot was made between 164 and 169, but coin finds indicate that occupation of the settlement, and presumably the working of the mines, continued into the 4th century.

Catterick (SE225991) – Cataractonium A 2 ha. Flavian fort lies on the south side the River Swale, immediately to the west of Dere Street and was succeeded in the mid-2nd century by a stone fort of similar size. A bank-and-ditch defended vicus of some 10 ha., including a high quality mansio supplied by an aqueduct, enclosed both the river crossing and the settlement on both sides of the Swale. Outside the defended area, ribbon development was extensive on both sides of Dere Street. By the beginning of the 4th century, the settlement was equipped with a stone wall on the south side of the river enclosing a reduced area of 6.27 ha. with a rudimentary grid pattern and at least one temple, some attempt was made to refurbish the derelict mansio baths. There is evidence of a wide range of industry and craft activity, including pewter working and ceramic production. It is probable that a late-Roman fort of some 2 ha. lay to the west of the civilian settlement and late-4th or early-5th century belt fittings indicate that Catterick may have been a base for comitatensian troops..

There are three possible routes for the transport of Mendip lead. The road from Charterhouse to the Fosse Way was probably constructed for this reason, as was the road to Chew Park, where cupellation took place, and from there to the River Avon giving access to Sea Mills and the Severn Estuary; or via Cheddar Gorge to the River Yeo and then to the Bristol Channel and the Western Seaways. Chelmsford (TL709062) - Caesaromagus Situated at the confluence of the tributary River Can with the navigable River Chelmer that, 15 km downstream, joins the River Blackwater near its estuary. London lies 45 km to the south-west and Colchester is 35 km to the north-east with a further road leading towards Bury St Edmunds, the Pedlars Way, and eventually The Wash.

Two altars dedicated by beneficarii consularis suggest some form of official status; one of the Vindolanda tablets refers to the delivery of a hides currently stored at Catterick and another to expenses occurred whilst travelling in the area. It is probable that the Swale was

134

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Caesaromagus (the origin of the name is contentious) had its origins in a short-lived Claudian fort and in the aftermath of the Boudiccan revolt it seems probable that an unsuccessful attempt was made to found a civitas capital for the Trinovantes. Stone buildings are rare, as nearest source of suitable material is 50 km distant but, in the mid-2nd century, an earlier timber mansio was replaced in stone. Ditch and rampart defences, enclosing some 8 ha., were constructed during the late-2nd century, but almost immediately back-filled. A late Roman-Celtic temple dated to 320–25, was destroyed early in 5th century.

place in the Dee Estuary and these, together with the effect of probable changes in tidal levels since the Roman period, are discussed in some detail in Chapter 1. The fortress was constructed and initially occupied by legio II Adiutrix Pia Fidelis, but this legion was moved to the Danube frontier by 88 and was replaced by legio XX Valeria Victrix, earlier commanded by Vespasian in his campaigns in the south-west of England. As was the case with other British legions, a significant part of the garrison was on detachment to Hadrian's Wall and, in the reoccupation of southern Scotland, Legio XX took part in the building and garrisoning of the Antonine Wall. The absence of troops serving in the north, led to construction being suspended with some buildings still incomplete, and some of the existing barrack blocks appear to have fallen into a semi-derelict condition. There is evidence for extensive dumping of refuse in a number of barrack blocks, with metal-working waste and debris also being found. The mansio was used as a rubbish dump, clearly indicating a period of virtual abandonment. Some elements of the legion remained at Chester, but it seems unlikely to have functioned as a conventional legionary depot. It may have been retained as a rear works depot, in which equipment was repaired and manufactured for the army operating in the north. This hiatus continued into the early 3rd century, at which time Chester once again resumed its role as a legionary fortress. At this time almost every building within the fortress was reconstructed, many from ground level. The western defences of the fortress were destroyed shortly after the beginning of the 4th century, suggesting a terminus post quem for the fortress as a defensive structure.

Chepstow (ST531942) Chepstow is located on the west bank of the River Wye, some 4 km from its confluence with the Severn estuary. Lying on the coastal road from Gloucester to Caerleon, a road to Monmouth runs northwards. Early material including 1st-century coins, pottery, metal horse-trappings and early burials indicate a fort guarding this key crossing of the Wye. There is nothing to be seen of the (presumed) Roman bridge at Chepstow, but from time to time, in exceptionally low water, some timbers have been reported in this locality. In 1911, excavations exposed what were believed to have been vertical posts of the Roman bridge; in 1962 some pieces of wood were recovered and lodged in Chepstow Museum, but they have not been scientifically tested to determine their date. However, during the production of an Extreme Archaeology TV program in 2003, samples were obtained from the piles and were later radiocarbon dated giving a late-1st to early-2nd century date. Chester (SJ404665) - Deva By the mid-70s, a major base was needed to the northwest of the existing fortress at Wroxeter to control the newly occupied areas of north Wales and the western Pennines. A 24 ha. fortress was therefore built on the River Dee at Chester (Deva) and, like most Roman fortresses in Britain, was placed near the lowest convenient crossing point and/or highest navigation point of a large river. It has been suggested that there is evidence for pre-Flavian activity, albeit on a much smaller scale and this might indicate Chester as a base for naval action in support of campaigns in the north-west, during the governorship of Quintus Petillius Cerialis (71– 3). In any event, it is probable that, from its conception, maritime use was a strong consideration as there was a deep-water channel at Chester, and there is a possibility that there was a naval base supporting Agricola's campaigns in the 70s.

Chester-le-Street (NZ275513) – Conganis Situated on a high bluff between the River Wear and the Cong Burn, during the Roman period the Wear was navigable to at least this point. A road runs northwards to Newcastle, with a branch road approximately half way leading to north-eastwards to South Shields. An early timber and turf fort is datable by samian pottery and a coin of Vespasian (69–79), indicating a Flavian origin. A stone-built fort, covering an area of 2.5 ha., was constructed at some date in the second half of the 2nd century and an inscription of 216 indicating construction work by on a bathhouse and aqueduct. Settlement has been traced for some 250 m to the south of the fort, with three phases of building producing pottery dated from the mid-2nd to the early-3rd century. Chichester (SU854044) - Noviomagus Reginorum Lying to the west of the River Lavant but probably supplied by sea via Fishbourne Creek, Chichester is at the commencement of Stane Street which runs for 80 km to London, other roads run north to Silchester, west and then north-west to Winchester and a further road leads south in the direction of Selsey Bill. Whilst its precise location has not yet been located, there can be little doubt that a fort was constructed by legio II Augusta in the friendly territory of the Atrebates, immediately following

Wooden jetties extended well out into the river, in association with 2nd-century Roman material and there is the probability that these were later rebuilt in stone. Shipping by coastal trading in and out of the port is demonstrated by the amount of Welsh slate, identified as originating at quarries near Caernarfon, in stratified Roman contexts. Traffic in the opposite direction has been shown by Cheshire sandstone found at Caernarfon and Caerhun. Significant channel changes that taken

135

JAMES ELLIS JONES

  invasion. The 4000 ha. oppidum was divided into two principal areas of occupation, Sheepen in the north being primarily the industrial and commercial area and Gosbecks in the south as the tribal and religious centre. Occupation continued into the Roman period; the site was destroyed by Boudicca in 61 but re-established over the following 10 years.

the landings in 43. Following abandonment by the army c. 45 the “New market of the Regenenses” developed in a somewhat haphazard fashion, but in the later-1st century a formal street grid was laid out and a forum, basilica, public bathhouse and amphitheatre were constructed, together with the temple to Neptune and Minerva recorded in the famous Togidumnus dedication plaque, whose palatial residence at Fishbourne is nearby. It appears that Chichester was an open town, with many indications of prosperity, including statuary, industry and masonry houses with mosaics and hypocausts until the mid-3rd century, when an area of 52 ha. was enclosed within ditch, stone wall and earthen rampart defences, with stone gates; projecting bastions were added at the end of the century, but by the 380s the city seems to have been largely abandoned.

A fortress for legio XX Valeria Victrix was constructed 4 km to north-east of the royal enclosure and a timber and turf fort was built within the limits of the royal enclosure at Gosbecks. Following the transfer of this legion to Kingsholm (Gloucester) in 49, the Colonia Victresensis was on the site of the now redundant fortress; the legionary defences were levelled and a range of formal buildings, forum, basilica, etc. were constructed within a standard insula grid pattern with the massive Temple of Claudius dominating the area. The city was sacked by Boudicca in 61, burnt to the ground and then rebuilt on the same plan. The construction of stone-walled defences enclosing 44 ha. began in the aftermath of the revolt and the ceremonial Balkerne Gate was built (with subsequent modifications) during the late-1st century.

Clifford (SO249467) The campaign base at Clifford lies close to the east bank of the Wye; above normal flood level, but the northern area of the fort is liable to be submerged in exceptional conditions. A further 7 ha. fort lies close by and is immediately adjacent to the Norman motte and bailey castle. Clifford is only 2.5 km away from the fort at Clyro and is therefore unlikely to be contemporary. It is probable that the forts at Clifford were the earlier, and were replaced by the fort at Clyro which is a better site and not liable to flooding. The fact that the Roman road lies on the opposite side of the river suggests that the forts were established before the line of the road was determined, and is a further factor indicating these as the earlier locations.

Also in the late-1st century, a double porticoed temple and a wooden-walled theatre were constructed at Gosbeck’s Farm. A group of four Romano-Celtic type temples lay north-west of the colonia, at least one of which survived into the late-4th century. A Roman circus (race-track) with a length of 450 m was located outside the walls to the south of the city - in the north-west provinces of the Empire, only six have been securely identified. Within the colonia itself there is little evidence for industry, but in the surrounding area 30 pottery kilns have already been located, probably only reflecting a fraction of those originally in use. The city was abandoned c. 440 and Saxon occupation began some ten years later.

Clyro (SO227435) This 10.4 ha. campaign base is located 1.7 km away from the village of Clyro, on a hill overlooking the west bank of the Wye. The river is fordable at this point, and there are extensive fields of view in all directions. It is probable that an early marching camp of 8 ha. preceded the fort, but this has not been tested by excavation. The few finds from the site include samian dating to 60, but a flagon neck in buff fabric is probably of Claudian date. A Roman road has been traced from Kenchester to Clyro, and probably continues to Brecon Gaer. This would undoubtedly have been the route taken by marching troops, but the supply and replenishment of the garrison of some 2000 troops would have been more effectively achieved by water transport. Hay-on-Wye with its Norman castle and a history of river transport is close by, and the defences lie within the fields of a farm known as ‘Boatside’.

Corbridge (NY982648) - Coria The main Roman site at Corbridge lies 4 km to the south of Hadrian’s Wall on the north bank of the Tyne, on a slight promontory some 45 m above sea level and sitting astride the intersection of Dere Street and the Stanegate. Its predecessor, the Agricolan supply-base at Beaufront Red House lies 1 km to the west and was deliberately demolished when the new fort was constructed c.87. This fort was reduced in size before being abandoned around 103 and then rebuilt around 105 and later reconstructed in stone with the Antonine advance into Scotland. Corbridge became a garrison town covering 12 ha., with two walled military compounds and a predominantly legionary population. The military presence is well illustrated by numerous dedicatory inscriptions, fountains and seven pedimented temples in the classical style. An inscriptions dating to 139 records the erection of the building by legio II Augusta, indicating the re-modelling of Corbridge as a supply base, prior to the Antonine invasion of Scotland.

Colchester (TL997253) - Camulodunum A road to the south-east leads for 85 km to London via Kelvedon and Chelmsford a road to the north reaches Caistor St Edmund after a similar distance and a further road leads to an unidentified location on the estuary of the River Stour. Situated on the southern bank of the navigable River Colne, some 20 km from the sea, Colchester was the tribal capital of the Trinovantian king Cunobelin and was the principal target of the Claudian

Supply by water from the estuary of the Tyne would have highly desirable, but the apparent evidence of a medieval quay near the Roman bridge has now been shown to be

136

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  an Anglo-Saxon water-mill. However in 1530, by Act of Parliament, the Mayor and Burgesses of Newcastle, were empowered to “pluck down all wears (sic), gores, and engines, that should be made in the river, to the great obstruction of the navigation thereof” and from time when the coal trade first began to be of consequence, the banks of the river became encumbered with stone discarded by vessels returning "in ballast". These are just two of many examples of the impediments to water transport occurring, after the Roman period, to rivers which may, at that time, had been navigable to barges for much of their length and this topic is more fully discussed in Chapter 9. The potential for navigability is illustrated by the fact that, in 1711, there was a proposal for rendering the Tyne navigable as far as Hexham.

and was identified as a sandstone statue of a lioness devouring a hapless male figure, probably one of a pair at the tomb of a military commander. Dalswinton - Bankfoot (NX933841) Dalswinton - Bankhead (NX 933848) On the major penetration road to the north through Nithsdale, the large Flavian complex, with two periods of occupation, lies on the level floodplain of the River Nith and predates the nearby Bankhead fort. The earlier (8.7 ha.) is a probable vexillation base that, after a brief period of occupation, was reduced in size to that of a large (3.7 ha.) fort. This in turn was abandoned after a short time, probably after severe flooding one winter, in favour of the site at Bankhead, situated on higher ground some 150 m to the north.

Cowbridge (SS995745) – Bovium Cowbridge grew in the late-1st century as a ‘ribbon development’, continuing well into the 4th century, with excavations in parts of the modern town revealing workshops and houses fronting the road to west Wales. Finds of a ballista bolt, and several other pieces of military equipment, indicate an early fort, but no structural remains have been found to confirm this suggestion. A small bathhouse using tiles with the legio II Augusta stamp indicates military involvement, and it is possible that the settlement owed its origins to the establishment of a mansio. There is evidence for ironworking from the 2nd to the 4th centuries, with large quantities of iron slag found, but none of the material was found in situ and no furnaces or hearths have been identified. The town probably acted as a market centre for the developing villa economy of south Wales; the importance of agriculture is indicated by several kilns for drying corn.

At Bankhead, there are two almost exactly superimposed forts, each with an annexe attached to its north rampart. Situated in a commanding position on high ground 150 m to the north of the forts at Bankfoot, excavations here have shown that both periods of occupation are Flavian in date. The earlier, and smaller (3.3 ha.) of the two, was replaced without any apparent break in the occupation of the site by a larger 4.3 ha. fort, with either of sufficient size to have held cavalry or two infantry units. Doncaster (SE578032) - Danum Situated where the road from Lincoln to York crosses the River Don, the fort was probably founded under Petillius Cerialis c. 70. The timber and turf fort covered some 3.75 ha. and was reconstructed in c.90 and again in c.105. The fort was later evacuated and systematically destroyed but, as a result of the withdrawal from the Antonine Wall, a smaller fort of 2.3 ha. was constructed c. 160 on the same site and the fort continued in occupation to the end of the 4th century. An extensive vicus surrounded the fort with the settlement on the south-west side being defended by a series of ditches, one of the 2nd century and two others dating from the mid-4th century. Within the defended area there was a significant pottery industry producing jars, dishes and bowls, with mortaria by the potter SARRIUS reaching as far north as the Antonine Wall.

Cramond (NT189768) Well situated to perform the function of a maritime supply base, the 1.92 ha fort at Cramond is an eastern outlier of the Antonine Wall, lying on the eastern bank of the tidal River Almond, at the river’s entry into the Firth of Forth and is protected from the north and east by Cramond Island. A road from the via principalis leads to the river and, although no supporting evidence for a Roman date has been found, in 1707 Sibbald reported a concrete mole on a rock to the east of the river mouth. Two phases of Antonine occupation, followed by a period of abandonment and subsequent reoccupation during the Severan campaigns have been suggested, but there is a lack of evidence for an intervening phase of demolition, and the Antonine buildings appear to have survived and been repaired and reused during the Severan campaigns. There is evidence of pottery of the late-3rd to early-4th century date, therefore post-dating the Severan occupation and it has been suggested that this may be accounted for by either Roman patrols or by trading by local native inhabitants.

Dorchester (SY692907) - Durnovaria Dorchester lies on the River Frome, 9 km inland from Weymouth Bay and 3 km north of the Iron Age hillfort of Maiden Castle, which was successfully assaulted by legio II Augusta under the command of Vespasian. Although not yet located, it is probable that a fort was established in the area and, by the end of the 1st century, Dorchester had become the civitas capital for at least part of the Durotrigian tribe. In the 2nd century, the town was defended by an earthen bank and ditch system enclosing some 30 ha. and reinforced in the 3rd century by the addition of a stone wall. Only very limited excavation has been possible but the discovery of a number of large masonry-built town houses with impressive mosaics suggests the presence of a school of mosaicists working in the area. A henge monument at Maumbury Rings, 800 m outside the town was originally constructed in the

In 1997, the “Cramond” one of the most ambitious pieces of Roman sculpture to have survived in Scotland was uncovered in the harbour mud by a local boatman,

137

JAMES ELLIS JONES

  Neolithic period and was provided with a timber superstructure in order to form an amphitheatre. The town was served by an aqueduct carrying water in a roofed conduit, capable of supplying some 50,000,000 litres of water per day, for a distance of over 19 km, making it the longest in Roman Britain.

The earliest evidence for Droitwich as a centre of salt production comes from the 5th/6th centuries BC, and briquetage, petrologically identified as coming from Droitwich, has been found at several sites distant from the source. A number of clay-lined pits up to 2.3 m across were located near brine springs and were probably used as storage or settlement tanks for brine. It appears that a considerable length of the Salwarpe Valley, with evidence occurring as far east as the Hanbury road, was involved in the industrial process and it has been suggested that special ceramic vessels were made in the Malvern potteries, as containers for the salt produced at Droitwich. The production of salt is dealt with in some detail in Chapter 10.

Dorchester served an important economic function as a distribution centre for Kimmeridge shale, Purbeck “marble” and pottery from the New Forest and Poole Harbour production areas. A road runs north-west for 30 km to Ilchester, Old Sarum is 50 km to the north-east and a road runs westward eventually reaching Exeter after some 80 km. The River Frome was navigable at least as far as Wareham, but it is probable that the road leading to a harbour in the Weymouth area was the route by which waterborne supplies reached the city.

Dover (TR319414) - Portus Dubris Dover succeeded Richborough as the major crossChannel port to Boulogne. The construction of the headquarters for the British squadron of the classis Britannica was commenced in 116, but was left unfinished, probably because of the demands of the construction of Hadrian’s Wall. Work on the construction of a second fort (1.05 ha), on the levelled remains of the previous structure, began c. 130 on a similar design to that of a normal auxiliary fort of the period. Rebuilt in 160, but again abandoned in the early-3rd century, which, in view of the requirement to supply the Severan campaigns in the north, is somewhat surprising. Dover underwent a final phase of re-occupation around 280, when it was rebuilt as a Saxon Shore fort with external bastions and internal timber structures. The fort was in use until at least the mid-5th century and there is some evidence of occupation into the 6th century.

Dorchester on Thames (SU577942) The only Roman military site on the River Thames to the west of London, the Flavian fort, and later walled town, at Dorchester on Thames lies not on the river whose name it bears, but a kilometre north on the River Thame. Oxford lies 13 km distant by road, but the sinuous course of the Thames means that the distance by river is doubled to some 26 km. Situated on the major Roman road between Silchester to the south and Alcester to the north, the fort controlled the crossing of both rivers, but was probably abandoned in the late-1st century. A late 2nd century earth rampart was reinforced by a ditch and stone wall in the late-3rd century, and it is possible that bastions were added in the mid-4th century. Occupying an area of only about 5.5 ha, and with a probable population of 600, Dorchester was one of the smallest walled towns in Roman Britain. A late-2nd or early-3rd century altar to Jupiter Optimus Maximus and the deity of the Emperor (Septimus Severus) was excavated in 1731. The altar was dedicated by Marcus Varius Severus, a beneficiarius consularis, a Roman army official responsible for the collection of tolls, taxes and supplies, possibly in connection with the Severan campaigns in Scotland. Public buildings such as a forum or temple might be expected in a town of administrative importance but, as yet, none have been discovered.

A pharos (lighthouse) was built on each of the headlands overlooking the harbour, and the remains of that on the Eastern Heights still display some 13 m of Roman construction. A protected anchorage at the mouth of the River Dour was formed by the construction of a massive breakwater constructed of timber framing in-filled with shingle, some 230 m to the east of the fort. Timber piles, mooring rings and groynes have been found together with a probable quay and timber jetty to the west of the inlet and a structure closer to the mouth of the inlet, has been interpreted as a continuation of the harbourside. Drumburgh (NY265598) - Congavata Located on a small drumlin on the edge of Burgh Marsh overlooking the wide mud-flats of the Solway Firth, the fort at Drumburgh lies about 4.5 km to the east of Bowness-on-Solway, the western terminal fort of Hadrian's Wall and 2.5 km to the north of line the Stanegate. An earlier turf and timber fort was converted to stone c.160, with the ramparts enclosing an area of 0.73 ha., making it the smallest fort on Hadrian’s Wall.. Squared blocks of red sandstone found in a foreshore creek could indicate the presence of a Roman harbour (Simpson & Richmond 1952, 14). Pottery of Huntcliff type, current after 367–9, indicates late occupation.

Droitwich (SO389263) - Salinae Drotwich lies on the navigable River Salwarpe, not far from its junction with the Severn, and Roman roads from Metchley, Gloucester, Alcester and Wroxeter met here. The River Salwarpe was navigable to Droitwich in the Middle Ages; in 1378, Richard III granted to the bailiffs of the town the right to levy tolls on the river. In the mid1st century, a fort was built to guard the river crossing and later a new fort was constructed in a commanding position at Dodderhill, overlooking the river valley and close to the main road junction. There is evidence of two periods of internal timber building construction, the first indicating a cessation of occupation by 70, but a probable revival of Imperial interest in the production of salt led to a second occupation from 120-50.

Ebchester (NZ103555) - Vindomora The fort at Vindomora (Bright Waters?) lies on the River 138

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  east, a road heading in the direction of Exeter has been identified. A local road leads to Topsham on the estuary of the Exe where trans-shipment of cargoes from seagoing vessels was necessary. Small craft were able to reach the city but, by the late 13th century, the River Exe was blocked by weirs so that even these vessels could no longer pass between the city and the sea until an Act of Parliament was obtained permitting clearance of all obstructions to navigation. Despite considerable efforts, it proved impossible to make the river navigable even for small craft until the City Council eventually decided to build a 2.8 km canal to bypass the most difficult sections of the river and eventually ships of 200 to 300 tons were able to reach the city.

Derwent, 10 km due south of the River Tyne, at the point where it is crossed by Dere Street. Corbridge lies 15 km to the north-west and Lanchester is 11 km to the southeast. The Agricolan fort covered 1.65 ha. and was occupied until c. 140, when it was abandoned following the Antonine advance into Scotland. On the withdrawal from Scotland in about 163 the fort was reoccupied and rebuilt in stone and, with at least two episodes of 3rd century rebuilding, remained in use until the mid-4th century. Elginhaugh (NT321673) The Agricolan 1.3 ha. turf and timber fort, with an extensive military annexe, was built c.80 on the line of Dere Street and was intended to protect the crossing the River Esk. Occupation came to an end in 86/7 when the fort was demolished on the withdrawal from Scotland and there is no evidence of subsequent reoccupation. Although the Ordnance Survey map of Roman Britain shows a road extending for some 8 km to the coast of the Forth estuary, it is probable that this construction did not take place until the foundation of the Antonine fort at Inveresk.

Filey (TA125817) Yorkshire coast 4th century “signal station”. Fingringhoe (TM047193) Eight km to the south-east of Colchester, the Claudian port and supply base lies on the Roman River close to its confluence with the Colne. A ditched enclosure covering 3 ha. includes Beacon Hard, an old landing place with embedded Roman tile and pottery, and coin evidence indicates occupation indicates ended c. 120..

Exeter (SX920925) - Isca Dumnoniorm The Neronian fortress of legio II Augusta was constructed on a commanding position overlooking the River Exe and naturally defended on two sides by steep valleys. Legionary fortresses in Britain usually covered an area of 20 ha. and, as those at Exeter enclosed only 15 ha. it is probable the full complement of the legion was never expected to be in garrison at the same time. Fortlets at Ide to the south-west and at Stoke Hill to the north may have been intended as lookout/signal stations. Excavations in 2010 have located a hitherto unknown fort during development on the site of the former St Loyes campus.

Fleetwood (SD485335) The archaeological evidence does not justify an entry of this length, but as a possible location of a “lost” port it is worthy of inclusion. There has long been a rumour of a lost Roman settlement at Fleetwood, near the mouth of the River Wyre, either under the sands or having been eroded by the sea. Roman occupation is indicated by significant numismatic evidence, folklore memories of Roman roads and buildings, and the course of the known road from Ribchester literally pointing to a Roman presence in the area. The road has been identified to a point some 5 km beyond the fort at Kirkham where, for no apparent reason, it makes a 60° turn to the north and heads in the direction of Fleetwood. Between 1832 and 1883, there was considerable mention of a road near ‘The Abbot's Walk’ in Fleetwood. It was found, with its pavement complete, during the excavation of foundations for the sea wall. The road was visible in 1836 at the shore margin of a rabbit warren and could be traced along the Naze, below Bourne Hill towards Poulton-le-Fylde.

The legion was transferred to a new base at Gloucester c. 68 and, during the next decade, the layout of the fortress was adapted, in a similar manner to the colonia at Gloucester, to create the civitas capital of the Dumnoni. The fortress defences were retained until the late-2nd century when they were replaced by a bank and ditch, 2.5 km in circumference, enclosing the much larger 93 ha. area needed to protect the urban areas outside the boundaries of the original fortress. It is not clear whether the wall fronting the rampart was constructed at the same time, or whether it was a later addition; equally, it is not certain whether the bastions and turrets of the visible walls of the present-day have a Roman or medieval origin. Partial rebuilding of the basilica and forum took place in the mid-4th century but virtually no coins dated after 380 have been found, suggesting a rapid decline.

In 1840, a hoard of 400 clipped siliquae originating in the 4th century (originally mis-identified as 4th-century denarii) was found near to Rossall Point, some 2 km to the west of Fleetwood, where the River Wyre enters Morecambe Bay. Hackensall Hall is located on the west bank of the River Wyre, approximately 1 km south of Knott End and in 1926 a hoard of 325 coins, the majority of which were silver washed Antoinani, was discovered.. A large paved platform, thought to have been a landing stage, located near to the Fleetwood lighthouse, was reported as being destroyed c. 1832. Local tradition has it that a rude thick-walled circular building was also noted. The proximity to the coin hoard discovered in 1840

Exeter is the terminus of the Fosse Way, with the nearest major settlement lying at Ilchester some 70 km to the north-west. A road leading south-east as far as Newton Abbot has been identified and probably continues much further to the west, possibly as far as Plymouth. It is probable that a road led north-east to the forts at Killerton and Cullompton and at North Tawton, some 25 km to the

139

JAMES ELLIS JONES

  coastal road has been located to the west of the fortlet. In the 18th century, the Water of Fleet was navigable for 5 km up to the Gatehouse of Fleet, for vessels of 160 tons burden.

suggests that both structures were from the Roman period. Fleetwood is sometimes equated with the site of the unlocated Portus Setantiorum of Ptolemy. Perhaps the strongest argument against a port in the area, either Roman or Iron Age, is that it would appear to serve no significant logistical purpose. It is difficult to see it as being having any real value for other either imports or exports, both functions being served more effectively by either Walton le Dale or Lancaster. The nearby fort at Kirkham was likely to have been on the earlier course of the River Ribble, and therefore capable of obtaining direct supply from sea-going vessels.

Gateshead (NZ255635) Excavation on the Bottle Bank site of the Gateshead Hilton has revealed a settlement, with evidence of occupation spanning the mid/late 2nd to the 4th/5th centuries, lining the road leading down towards the Roman bridge across the Tyne and this may possibly be associated with an undiscovered bridgehead fort. A hoard found in Church Street, Gateshead in 1790 is said to have contained Hadrianic coins and in 1927 a further hoard containing “radiates” was found.

Perhaps there was a need for the coastal fort at Fleetwood during the Flavian period, as Kirkham was some 20 km from the open sea. There are, however, no finds from the area that relate to this period, the only dateable evidence being the coin hoards, dating from the 3rd and 4th centuries. In a highly speculative attempt to reconcile the conflicting evidence of the road and the coins, it might be suggested that an earlier Flavian fort or fortlet, since lost to the sea, was reactivated at the same time that the Saxon Shore fort at Lancaster was established, in order to meet the threat of Irish sea-raiders.

Girvan Mains (NX192992 and NX188991) The two Agricolan camps (14.9 ha. and 6 ha.) are located on the coast, 0.8 km north of the estuary of the Water of Girvan. Girvan has a natural harbour and is 80 km marching-distance from nearest known forts at Glenlochar or Loudon Hill. Girvan is the southernmost point on the west coast from which interrupted penetration to the north and east is feasible, indicating support for the concept of a seaborne invasion of southwest Scotland. Glenlochar (NX738651) The location of seven late 1st-century marching camps (largest 12.7 ha.), the site extends for 1.25 km along the valley floor of the River Dee, 3.5 km north of Castle Douglas. The largest camp could have accommodated a legion and a large force of auxiliaries, and the presence of rubbish pits suggests that the camp had some degree of permanence. The 3.36 ha. fort faces west towards the River Dee; an annexe houses a possible mansio. Three successive timber and turf forts were built on the same site. The first Flavian fort was destroyed by fire, the second Antonine; the third, also Antonine, was a later modified version. It is possible that there was an earlier Flavian fort to the north.

Forden Gaer (SO208989) - Lavobrinta The fort is situated on the line of the road connecting Caersws and Wroxeter and lies on a slight rise on the east bank of the Severn, at the point at which it is joined by the River Camlad. On the evidence of samian pottery, the occupation at Forden Gaer began in the mid-Flavian, and the defences enclose an area of 3.1 ha. but although the site was occupied into the 4th century, there is nothing to suggest that Forden Gaer was ever provided with stone defences. Only limited excavation has been undertaken in the interior of the fort and none of the principal buildings have been examined in detail, although a single block of dressed sandstone suggests that the internal buildings were of stone. At some time in the 3rd century, dismantling took place in the interior, and the fort may have been briefly abandoned. A refurbishing of the rampart, and the relaying of the major interior roads seems to have taken place during the second half of the 4th century, dated by a coin of Valentinian 1 of the period 364–78.

I am grateful to Andrew Nicholson for informing me that he has reached Castle Douglas in a replica Viking longship. Glenluce (NX 198566) A 20 ha. marching camp, capable of accommodating a legion plus auxiliaries, lies on the left bank of the Water of Luce (Abravannia Ostia?) facing north-west towards the crossing of the Luce. The camp is 20 km to the west of the fortlet at Gatehouse of Fleet, and lies on the known road heading towards Stranraer on Loch Ryan.

The combination of a long period of occupation, the absence of stone walls (common on most forts in the later period) and an extensive extra-mural area, might point to something more than just a military function. There is, after all, no civilian administrative centre closer than Wroxeter; proximity to mineral extractive areas might suggest some form of pseudo-military form of control of resources and taxation.

Gloucester (ST830180) - Glevum Which legion built the 17.4 ha. fortress and subsequently occupied it is uncertain , the most likely suggestion is that part of legio II Augusta moved from Exeter for the construction and initial garrisoning of the fortress, and that the remainder of the legion moved up at a later date, as part of the movement into South Wales. In one of the

Gatehouse of Fleet (NX595573) - Lucopibia? This small 0.33 ha. fortlet could have provided quarters for one century; the few pieces of pottery recovered all point to a single Flavian occupation. A section of the

140

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  some 50 ha. and indicates that a previously held view that Gloucester was a ‘failed town’, with its growth stunted by nearby Cirencester (Corinium), is no longer tenable.

partition walls of a barrack block, a coin dated to 64–66 was found, showing that it was not built before that date, and pottery and coins from within the barrack blocks indicate that occupation lasted until 87–8. The fortress was re-furbished in the late 80s, with some rebuilding in stone and the defences being re-faced with a dry-stone wall. Military occupation seems to have ceased during the 90s, with the conversion of the site to the Colonia Nervia Glevensium under the Emperor Nerva (96–8); there is evidence of civilian occupation of some of the military buildings before the site was cleared and rebuilt.

Godmanchester (TL238703) - Duruvigutum Godmanchester was at an important road junction, about 25 km south of Water Newton, where Ermine Street crosses the navigable Great Ouse, with branch roads leading to Cambridge (25 km to the south-east), Sandy and Wimpole Lodge. A Claudian fort of indeterminate size was succeeded, in the Neronian period, by a larger, short-lived, fort probably capable of accommodating half a legion.

Gloucester (SO 830180) - Colonia Nervia Glevensium Situated at the head of the Severn Estuary, the importance of Gloucester as a major port may be judged by the evidence of maritime activity, but the course of the River Severn changed during, and since the Roman period. In about 98, a colonia was established on the site of the former fortress, and therefore on land that for some time had been Imperial territory. There was riverine land nearby that may previously have been of little value, but could be reclaimed by Roman drainage methods. It is therefore probable that the bulk of land redistributed to the colonists had not previously been settled by the local population. This tactic would have avoided the resentment caused half a century earlier, by the appropriation of the best tribal land by the colonists of Colchester.

The vicus to the north expanded along both sides of Ermine Street and, following the departure of the military and the levelling of the forts, a planned town centre developed at the crossroads. Early in the 2nd century, a mansio (with a possible administrative component) and associated bathhouse were constructed and, in the early 3rd century, a basilica was built and a colonnaded marketplace developed. Stone walls were constructed c. 270 enclosing an area of 11 ha. and external bastions were added in the mid-4th century. Goldsborough (NZ835151) Yorkshire coast 4th century “signal station”. Hardknott (NY219015) - Mediobogdum A Trajanic foundation, the 1.3 ha. Fort at Hardknott lies on the road between Ambleside and the coast at Ravenglass; a 50 m length of road lies between the fort and the main Roman road. The timber barrack blocks of the first construction were replaced in stone in the early 2nd century. The parade ground, with a probable tribunal, is a levelled area of c.140 m. x 18 m. to the north-east of the fort. The artificial levelling of the parade ground is estimated to have required the removal of 5,000 cubic metres of soil and rock. The fort was probably given up on the Antonine advance into Scotland; was never reoccupied and, unusually, appears not to have had a civilian settlement, its small bathhouse servicing only a military community. Small quantities of later pottery and two 4th-century coins have been suggested as evidence for a later military occupation but are more likely to indicate occasional use as temporary shelter for travellers to and from the coast. The extreme difficulty of the road from Ambleside, via Wrynose Pass, would make supply from the coast highly desirable. The Esk is tidal for 7 km above Ravenglass and therefore could have been used bring supplies closer to the fort.

The fortress was converted into a colonia by the simple expedient of adapting the barrack blocks into domestic dwellings. Some other buildings in the fortress were retained for civilian use, but others were demolished without immediate replacement. The town was defended by re-using the fortifications of wall, rampart and ditch belonging to the legionary fortress; the north, east and west gates have been identified and are nominally retained in the modern city. During the late 3rd century, a programme of refurbishment and replacement of the walls was carried out; projecting towers were added, possibly indicating a quasi-military role at this time. A pottery kiln and iron-working surface have been located, three other 1st century kilns are known from outside the east gate, probably belonging to the legionary period. A brick and tile works, supplying the needs of the developing town, developed due north of the walls and close to the contemporary course of the Severn. Iron working production seems to have grown in the 3rd and 4th centuries, when slag was used for resurfacing the streets and yards, and a large lime-burning area was located along the London Road, about half a mile from the north gate. The most important suburb was outside the north gate, and its origin probably lay in the civilian settlement attached to the fortress. Alongside the roads leading to Kingsholm and Cirencester were timber, and then masonry buildings fronted by colonnades, some of which were associated with commercial use. The extensive and prosperous built-up area lying outside the walls which, when added to that within the walls, virtually doubled the total inhabited area of the town to

Healam Bridge SE323836 The early/mid-2nd second century 2 ha. fort was situated astride Dere Street at the point where it crosses the River Swale, with York lying some 50 km to the south-east via the Swale and then the Ouse. Lying roughly between Aldborough and Catterick it is possible that, after abandonment, part of the fort was converted into a mansio. The vicus, including a water-powered grain mill, extended as a ribbon development along both sides of

141

JAMES ELLIS JONES

  Dere Street to both north and south, and appears to survived as a roadside settlement until the 4th century. The Swale is known to have been navigable to nearby Morton, from where timber was regularly shipped during the medieval period.

Huntcliff (NZ687219) Yorkshire coast 4th century “signal station”. Ilchester (ST525225) - Lindinis Settlement commenced at the junction of the Dorchester road with the Fosse Way, extending on either side of the crossing of the River Yeo. By the end of the 1st century, the settlement had developed into a town of around 20 ha. Excavations have revealed the regular plan of a Roman town with paved streets, and the foundation walls of houses, some of which had mosaic floors and painted wall-plaster. Construction of the defences, consisting of an earthen bank and ditch has been dated to the late 2nd or early 3rd century. Sometime later, the earthen rampart was fronted by a masonry wall, though no dating evidence has yet come to light. Nearby sites of more than 20 villas have been found and it has been suggested that Ilchester was the civitas capital for the northern half of the area previously controlled by the Durotriges. However, the presence of the municipal buildings, which might be expected within a civitas capital, has yet to be proved.

Heronbridge (SJ410635) The settlement site at Heronbridge stands on the west bank of the River Dee 2 km south of Chester, straddling the road running from the legionary fortress towards Whitchurch. Occupation commenced c. 90 and continued down to at least the middle of the 4th century, This is demonstrated by coins ranging from a denarius of Titus (80) to an as of Constans or Constantius (330-360). There were numerous stone buildings, including some with hypocausts, as well as at least one shrine or temple but the majority of buildings were strip houses set at right angles to the main road. There was metalworking and grain processing; there is evidence that building work and street repairs continued at Heronbridge into the early 4th century. An existing streambed was deepened in the 2nd century to enable construction of a ramp leading down to the edge of an inlet from the Dee, this having been artificially straitened. There were several rock-cut pits and other features that were probably mooring posts in order for barges to be tied-up prior to unloading or loading.

Navigability to Ilchester is highly probable as, though 49 km from Bristol Channel, the town is situated at only 11 m above sea level (a rise of 0.22 m per km) and is only 3 m higher than the port of Bridgwater, itself 19 km from the sea. Roman stone quays have been claimed and geoarchaeological evidence for the canalisation of the River Yeo during the Roman period has been found. It has been suggested that black-burnished pottery from the Poole Harbour area was transported by pack animals to Ilchester, then by barge via the Rivers Yeo and Parrett to a port on the Bristol Channel.

Holt (SJ405545) The legionary tilery at Holt, situated on the banks of the River Dee some 9.5 km upstream from Chester, was in operation from c. 90 to 240, but with reduced production from the middle of the 2nd century. In 1930 Grimes considered transport by river as the most likely option, as he described transport by road as ‘a much more difficult and troublesome business’. A battery of 6 double-flued kilns producing a wide range of oxidised coarse wares, as well as ‘legionary’ fine wares and building materials. A number of distinctive African forms were produced on a small scale in the 150s, possibly by troops posted to Britain after the Mauretanian War. A bath-house and barrack blocks for some 60 soldiers indicate military control and probable involvement in the production process.

Inchtuthil (NO117396) Founded under Agricola in c.83 the legionary fortress of Inchtuthil occupies 22 ha. of a large, steep sided plateau standing some 30m above the north bank of the River Tay. Geomorphology indicates that the Tay has changed its course in the past and, during the Roman period, may have flowed to the north of the fortress. Construction was undertaken by Legio XX Valeria Victrix and some three years later the original turf and timber defences had been modified by the addition of a stone wall. A 1.65 ha. enclosure, with no internal buildings, lies near the easiest access up to the plateau and could have served as the access point for heavy supplies brought up from the river. Extending to the edge of the plateau a further enclosure, usually known as the “officers’ compound” was defended by a ditch and rampart and contained not only standard timber barracks but also a somewhat palatial structure with hypocaust under floor heating which has been interpreted as an interim proprietoria, to accommodate the legate whilst his permanent quarters were under construction.

Horncastle (TF258695) - Bannovalum? In Roman times, Horncastle may have stood at the head of the navigable inlet of the River Bain and was clearly much closer to the sea than it was today; Stukeley noted that “some don’t scruple to affirm that it was a seaport, that is navigable” and the massive stone walls with external bastions have led to suggestions that (in conjunction with Caistor) it functioned as a Saxon Shore fort, intended to protect Lincoln. A Roman road running from Lincoln to an as yet undiscovered location on the coast passes 7 km to the north. Horncastle comprises two main elements: a walled enclosure of about 2 ha. at the junction of the rivers Bain and Waring, and an un-walled settlement covering up to 5 ha. on a slightly higher gravel terrace, south of the Waring and east of the Bain.

By late 86 or the beginning of the 87 the interior was almost complete, with bath-house, sixty-four barrack blocks, four houses for tribunes, six granaries, a large hospital, a drill hall, a workshop and a temporary

142

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  headquarters building, having all been completed. The transfer of legio II Adiutrix to deal with problems on the Danube frontier led to a decision to withdraw from Scotland abandon Scotland and, in a process of thorough destruction, the fortress was deliberately and systematically dismantled in order to prevent anything of value falling into enemy hands.

constructed timber buildings, several stone structures and a possible timber-framed amphitheatre. A substantial kiln/furnace and a number of wasters, including an almost complete mortaria waster, indicate both pottery and mortaria production on the site and a wide range of pottery forms (Inveresk ware) has been identified as being unique to the site.

It is perhaps relevant that a road leading directly from Bertha to Inchtuthil has yet to be discovered, and this might indicate a dependence on water transport to the fortress. Navigation of the Tay beyond Perth is a matter of some dispute and certainly may have presented difficulties because of the rapids below Stanley, and, in particular, of a barrier of hard rock, the cause of the Falls of Campsie Linn, but a short portage would have enabled supply and replenishment by water. When the fortress was abandoned, large quantities of iron objects, including 900,000 nails and a number of unused wheel-tyres, were buried in a pit within the fabrica. This study seeks to demonstrate the use of rivers during the Roman period, but it has to be accepted that this abandonment of valuable material casts some doubt on the navigability of the Tay. The apparent absence of a road would have made difficult the recovery of over 10 tonnes of ironwork by wheeled transport and, if available, the opportunity of loading to shallow draft barges for transport downstream to Perth would have been an attractive option to discarding such valuable material. Recent magnetometer survey within the fabrica revealed the existence of a second pit (Woolliscroft 2010 pers. com.) and this must add a further question mark to the navigability of the river during the Roman period.

Two altars dedicated to Apollo Grannus by Quintus Lusius Sabinianus, Imperial Procurator in the province of suggests that Inveresk may have been an important centre, especially as only four other instances of procuratorial inscriptions have been found in Britain. The size and importance of the civilian settlement Inveresk is probably to be explained by its location away from the physical frontier and on the junction of four main communication routes. Irvine (NS380202) Irvine Bay was known to Ptolemy as Vindogara Sinus, suggesting a significant Roman presence in the area. Ixworth (TL931695) Ixworth lies on the Pedlars Way (which runs north to Holme-next-the-Sea on the Wash), a road ran north-west to Caistor St Edmund, another ran southerly to Chelmsford and another south-easterly to Colchester. The short-lived 2.9 ha. fort was a probable post-Boudiccan foundation and the vicus developed into a civil settlement at the crossing of the River Lark, a tributary of the Great Ouse, which was once navigable to nearby Bury St Edmunds.   Kelvedon (TL864185) - Canonium Fifteen km to the south-west of Colchester and 20 km to the north-east of Chelmsford, Kelvedon is situated on the navigable River Blackwater. Preceded by a prosperous Late Iron Age settlement there is some evidence of a post-Boudiccan fort. In the late-2nd century, a linear development of timber-framed buildings was enclosed by a hastily-constructed defensive earthwork covering 5 ha. that was backfilled during the early-3rd century

Inveresk (NT342720) The 2.7 ha fort at Inveresk, 45 km distant from the eastern terminus of the Antonine Wall, stands on a 20m high promontory in a meander bend of the River Esk, some 1.5 km to the south of its entry into the Firth of Forth at present-day Musselburgh. There is evidence for deliberate demolition of the fort followed by reconstruction on a slightly different internal plan, together with further indications of sub-phase repair work or modifications. The discovery of a Flavian fort at Elginhaugh, less than 10 km away makes the presence of a contemporaneous fort at Inveresk unlikely. It is probable that the port for Inveresk lay on the coast at Fisherrow, where the remains of Roman buildings and a road leading towards the fort were reported in the 19th century.

Kenchester (SO440428) – Magnis New Weir (SO436418) Kenchester is strategically located in the valley of the River Wye, on an important east-west road heading towards the early military bases at Clifford and Clyro. Other roads lead northeast towards Leintwardine and Wroxeter, south across the Wye to Abergavenny and east to Sutton Grandison. This has led to speculation that the town owed its origins to an early military presence, supporting the Roman advances into the Welsh Marches. However, no trace of a fort has been located at or near the later town, and only three isolated pieces of military equipment are known. Evidence for the maritime use of the River Wye includes tile and pottery from Gloucester and Cirencester, Forest of Dean iron ore and quern stones and, most significantly, stone from the Cotswolds used for architectural stonework, including columns of oolitic limestone.

Inveresk displays Scotland’s largest known civil settlement outside a fort, where a combination of stray finds, aerial photography and excavation has revealed timber and stone buildings, based on a rudimentary streetgrid, extending for approximately a kilometre from the fort, and surrounded by extensive field systems. A hypocausted building, with an aqueduct leading to it, lies 300 m from the fort. The final phase of civilian occupation suggests a well-established, prosperous community, with evidence for cobbled roads, massively

143

JAMES ELLIS JONES

  The town defences enclosed an area of 10 ha. and, when erected c.150, comprised a rampart and ditch with wooden gateways. In the middle 4th century, a stone wall was added to the rampart, and stone gateways and bastions were erected. Prior to the construction of the town defences, the settlement was a linear development stretching some 700 m to the east and 400 m to the west of the crossroads. The economic base was predominantly agricultural, but there is some evidence for iron and copper alloy working. The town underwent a number of internal re-organisations and rebuilding, with a major remodelling of the town centre occurring in the mid to late 2nd century. It has been suggested that Kenchester acted as a pagus centre for the western Dobunnic area, which was isolated from the civitas capital at Cirencester by the territorium surrounding the colonia at Gloucester. It is also possible that, following the re-modelling of the area enclosed by the new defences in the later 2nd century, the status of the town was raised - possibly by elevation to that of civitas capital.

Kirkbride (NY323557) Kirkbride is the western terminal fort of the Stanegate frontier, overlooking the River Wampool at the head of an inlet on Moricambe Bay. The 3.3 ha. fort was defended by a turf rampart and double ditches and was probably abandoned when Hadrian’s Wall became operational. A road runs north towards the Wampool, passing a recently identified compound and, on the line of the Roman road running east to Carlisle, there are signs of an extensive vicus, with an industrial site showing evidence of lead and ironworking. The silting of the Moricambe Estuary is arguably of post-Roman date, as the fort would not be intelligible other than as a port at the end of the western Stanegate. A double denarius of Tetricus (270-74) might suggest continued civilian occupation. Kirkham - Dowbridge (SD430320) The fort is located some 12 km from the Irish Sea, at the highest point (30 m OD) on a glacial moraine running from Preston to the coast at Blackpool. Three temporary marching camps dating from the 70s were succeeded by a small (60 m square) defended enclosure, similar in morphology to the fortlet at Martinhoe. There is no evidence of a full-size timber and turf fort, but a stonebuilt fort (incorporating a timber tower) was constructed late in the 1st or at the beginning of the 2nd century. The fort was built of sandstone, which does not occur closer than Preston, and would therefore have been transported down river by barge. The fort, lying on the road from Preston to the coast, had little tactical or strategic value other than as an entrepôt or protected landing place. Following the development of a secure road system from Manchester to Carlisle, the fort was only occupied for some 30 years, before being abandoned around the middle of the 2nd century, coincident with a decline in the activity at Ribchester.

New Weir (SO436418) lies 1 km south of Kenchester, and 0.5 km west of the line of a Roman road, leading from the east gate towards a postulated crossing over the River Wye. A pair of buttresses, still standing to a height of 4 m, rise from the river to a terrace on which there was a complex of rooms, including at least one with a mosaic pavement. 50 m south-east is a further complex of rooms that were probably joined by a corridor. Downstream is a six-level octagonal stepped cistern which, when found in 1891, contained numerous tessarae and is a possible water-shrine or nymphaeum. It has been suggested that, because of the significance of water transport on the River Wye, the site may be the villa of a prosperous merchant engaged in the supply of goods to nearby Kenchester. Kingsholm (ST835195) Kingsholm lies on a gravel terrace on the east side of a former major course of the River Severn, now a partly culverted stream (the Twver) and it is probable that in Roman times the Severn was divided into two channels of more or less equal size, spaced about 1.5 km apart, with a flood plain of alluvium silts between them. At some time from 49 onwards, a fortress was built on the east bank of the river to defend the crossing. Kingsholm controls a major north-south communication line on the east bank of the Severn, extending from Sea Mills up to the west midlands and Watling Street. It has been supposed that the move of 1 km to a new fortress, at what is now the centre of modern Gloucester, was probably occasioned by localized flooding at Kingsholm. However, Roman levels exist at c. 9 m A.O.D or below to the south of the main Kingsholm site, and to the west of the later colonia lie between 8.2 and 9.2 m A.O.D. Therefore if Kingsholm flooded, so did a substantial part of the colonia at Gloucester and the move is more likely to have been as the result of re-location to a more favourable natural river crossing.

The course of the River Ribble has changed and it is probable that, during the Romano-British period, the fort was accessible to sea-going ships. A wooden tower could have served as a beacon or lighthouse indicating a landing place on the Ribble estuary. Incoming vessels would have ‘sat out the tide’ and waited for the ‘first of the flood’ before proceeding upriver to Walton le Dale, and possibly to Ribchester. When proceeding to sea, the reverse operation would have taken place. Alternatively, if Ribchester was not accessible by sea-going vessels, cargoes could have been transferred to/from barges at either Walton le Dale or at Kirkham itself. In common with many other sites of this period, there is evidence for the import of olive oil in Dressel 20 amphorae. Lake Farm (SY999992) Developed on the site of a marching camp, a vexillation base covered 12 ha. and was almost certainly constructed as a result of the campaigns of legio II Augusta against the Duritroges, whilst under the command of Vespasian. Some 9 km to the north of the sea at Poole Harbour, the probable semi- permanent nature of the base is suggested by the recovery of tiles which may have come from an, as yet, undiscovered bathhouse. 144

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Leicester (SK582044) - Ratae Coritanorum Leicester lies on the Fosse Way, with Lincoln 90 km to the north-east and Cirencester about 115 km to the southsouth-west. A road runs westward to Mancetter and another leads east-south-east to Godmanchester. The River Soar runs through Leicester and meanders northwards for 46 km until it reaches the River Trent, being joined at Syston by its major tributary, the once navigable River Wreake.

Lancaster (SD474610) A turf and timber fort, probably Ceralian, was located on the navigable River Lune, on the road from Preston and then north-east towards Burrow in Lonsdale and a wooden bridge across the River Lune has been suggested. On the evidence of a Trajanic inscription, the fort was rebuilt in stone, in the early 2nd century. The site was levelled c.160 and a new enlarged fort constructed with some stone and some timber buildings. About 330–40, a new fort with massive stone walls and bastions, was built on a different alignment to that of the previous forts. This appears to have coincided with the abandonment of the extramural settlement, and the possible relocation of the inhabitants into the new fort itself. The fort has similar characteristics to the so-called ‘Saxon Shore’ forts of southern England and was clearly designed, like the fort at Cardiff, to strengthen the late Roman coastal defences. Of particular interest considering the riverine location, is the inscription of a numerus barcariorum, a unit of bargemen.

A small fortlet, possibly post-Boudiccan, was established and it is probable that this was the stimulus for the organic growth of a civil settlement on the east bank of the river. In the late-1st century Leicester became the civitas capital of the Corieltauvi, with a regular street grid and, eventually, a range of public buildings. The forum/basilica, public baths and mithraic temple were built during the mid-2nd century, possibly as a result of Hadrianic stimulation, and a macellum was constructed in the early-3rd century. Two further temples and a range of luxurious masonry residences were concentrated in the area to the west of the Fosse Way, with timber-built houses and industrial premises lying to the east. It seems probable that Leicester was provided with major defences on three sides, with the river providing adequate protection to most of the western approaches. A rampart, ditch and palisade are dated to the late-2nd century, with stonewalled fronting being added to the rampart in the 3rd century. There is evidence for fire damage to public buildings during the late-3rd century, but there is no reason to suppose anything other than the normal (if such exists) rate of decline for Roman Leicester.

A large courtyard building with a small bathhouse may have been a mansio, although it is equally possible that it was the official residence of the beneficiarius consularis recorded as having been present in Lancaster. The site of the vicus has been excavated, revealing substantial masonry and structural remains, with many coins and much pottery. The main occupation appears to span the period c.80 to c.330, at which time the final fort was constructed. Lantonside (NY010662) This Antonine fortlet, sited 70 m. from the mouth of the Nith on its eastern bank, replaced the nearby Flavian camp at Ward Law. The fortlet measures 42 x 37 m, enclosing an area just over 0.15 ha. There is a rhomboidal annexe of about 0.5 ha. attached to the side of the fortlet.

Whilst local granite was used in the construction of rubble masonry, millstone grit quarried at Melbourne in Derbyshire, 48 km to the north-west, was employed for most better-class work and architectural decoration. However, high-quality oolitic limestone occurs only some 25 km to the east and south-east, and the preferred use of millstone grit, from a more distant source, may suggest that water transport, via the Soar and Trent, was the preferred option. The Roman settlement at Barrow upon Soar, where there is later evidence for limekilns, may have been the source of lime for the mortar used in building construction. In the late 18th century, after straightening and the removal of shallows, the Soar became the most profitable commercial navigation in Britain; it is certain that Roman fluvial engineering skills would have been more than capable of a similar achievement.   Lincoln (SK975716) – Lindum At Lincoln a 15 ha. fortress was founded in c.60 for legio IX Hispana, with this legion subsequently moving its new fortress to York, and its place being taken by legio II Adiutrix but in c.78 this legion moved to Chester. The majority of the buildings were demolished but a caretaker garrison probably remained within the walls that were left standing. Lindum Colonia was founded in the late-1st century; the turf and timber defences of the fortress were replaced in stone and a “Lower City”, surrounded by

Learchild (NU101113) - Alauna On the River Aln, some 16 km from the sea at Alnmouth, Alauna is situated on the Devil’s Causeway, midway between Corbridge and Berwick-upon-Tweed. Known only from aerial photography, and with no surviving earthworks, it appears that a 1st century 4.4 ha. fort was succeeded, presumably during the 2nd century, by a smaller fort within the same area. Ptolemy’s identification of fluvius Alaunus indicates that the river had some importance during the Roman period, and may suggest that the fort was supplied by sea. Lease Rigg (NZ815041) The 1.1 ha. Flavian fort, occupied until c. 120, lies at the apparent end of Wade’s Causeway, some 15 km north of Cawthorn, which appears to be the southern terminus of the road. At Wheeldale Moor, this seemingly disconnected road contains one of the best-preserved stretches of Roman road in Britain. Lease Rigg lies on the River Esk, some 15 km to the south-west of Whitby, the site of a probable fort and harbour and, because of the apparent isolation of Wade’s Causeway, supply by sea must be a probability.

145

JAMES ELLIS JONES

  stone walls enclosing an area almost twice the size of the original fortress was constructed southwards to the river, probably covering the same area as the earlier canaba.

Llandeilo (SN620225) At Llandeilo on the Afon Tywi there is a ‘Quay Street’ and an oral tradition of ships (barges?) unloading cargo there. Two superimposed forts (earliest c. 3.9 ha./later 1.54 ha.) have been recorded, during a geophysical survey at Dinefwr Park. The larger first fort was possibly constructed in the pre-Flavian period, with a smaller fort being built at some later date. Four 1st century silver denarii have been recorded, one of Augustus and possibly two of Vespasian. There is evidence of a vicus lining the road leading to the fort at Llandovery.

Ermine Street runs south to Ancaster and due north for 55 km to the Humber ferry-crossing; York may also be reached by a branch road running to the north-west via Doncaster. The Fosse Way runs south-west, reaching Leicester after 80 km and another road leads eastwards in the direction of an unknown location on the coast. During the Roman period, Lincoln lay close to the head of the estuary of the Witham and it is probable that the present course of the river below Lincoln was subject to some Roman modification. From Lincoln the Fosse Dyke ran in the direction of Littleborough and the Car Dyke ran south towards Water Newton and Cambridge, connecting with the navigable Rivers Welland and Trent. The extent to which the canals were used for navigation, or only drainage, is still a matter of debate. The riverbank was clearly a commercial focus of the city, with a possible dock to the east of the walled town and waterfront building extending for at least a kilometre. Substantial quays were not needed, as Lincoln was situated above the normal tidal level, and wooden jetties, similar to that found at Chester, would have been adequate for the loading and unloading of coastal vessels. A wooden writing tablet and nineteen styli, together with a copper-alloy balance found near the waterfront might be connected with maritime commercial activity and overseas trade is indicated by the dedicatory inscription of M. Aurelius Lunaris, a wine merchant from Bordeaux.

London – Londinium Dealt with in some detail in Chapter 8. Long Sandall (SE 604067) Aerial photography in 2006 revealed a double ditched enclosure indicating a probable 1.65 ha. fort on the east bank of the river Don, 4.5 km downstream of Doncaster. Longthorpe (TL158977) Logthorpe is situated 4.5 km to the east of the junction of Ermine Street and King Street but itself lies on no known Roman road. Lying 4 km distant from Peterborough on the banks of the River Nene, a vexillation base covering 10 ha., constructed c. 44–5, together with a significant industrial complex, was located to the south-east containing evidence of pottery production and military metal-working. The vexillation base was succeeded by a 4.5 ha. fort, using part of the defences and some of the existing buildings, and remained in occupation until around 60. A probable reason for this reduction in size is that the troops under the command of Petillius Cerialis who were defeated by Boudicca near Colchester, were from a vexillation of legio IX Hispana, stationed then at Longthorpe, thus leading to the reduction of the base until the founding at about this time of the fortress at Lincoln, some 70 km to the north.

Littleborough (SK822829) - Segelocum The settlement, possibly walled, with evidence of 1st (possibly as early as the 50’s) to 4th century occupation, lies on the west side of the River Trent where it is crossed by the road from Lincoln to Doncaster, and is 8 km north of the point where the Fosse Dyke joins the Trent. Earthworks on the opposite side of the river at Marton are the site of a small (0.7 ha.) early fortlet.

Lostwithiel (SX102611) Situated on a hillside overlooking the navigable River Fowey, the recently discovered 0.42 ha. fort is closely similar in size and shape to the fort at Nantstallon, only 9 km to the north on the River Camel. Located within 300 m of a prominent lode, finds of iron slag indicate metal processing during the Roman period and the finds assemblage suggests continuous occupation from the 1st to the 4th century.

Llandovery (SN770352) The pre-Flavian fort lies on the River Twyi at an important crossroads, with a south-west/north-east road leading from Carmarthen to Caerau, Castle Collen and eventually to Caersws, and a south-west/north-west road leading from Brecon via Llandovery to Pumpsaint and then Llanio. The early timber buildings were destroyed by fire at some time, and later rebuilt in the same materials. A third phase involved the addition of a stone revetment and the building of some internal stone buildings, with a final phase, probably c.120-5, being the construction of an internal ditch and rampart, resulting in a reduction in size. There is no firm indication of occupation beyond about 160 but, in common with Bryny-Gefeiliau and Gelligaer, there is some slight evidence for re-occupation in the 3rd/4th centuries.

Lurg Moor (NS295737) Looking out over the Clyde estuary this 0.20 ha. fortlet is of the Antonine period. Loughor (SS565979) - Leucarum Loughor is located towards the west end of a gravel spur on the east bank of the tidal Afon Llwchwr (River Loughor), where the river enters the Burry Inlet to the Bristol Channel, at a point where the width of the river is less than 200 m. Excavation took place between 1985 and 1988, and was one of the most comprehensive investigations of an auxiliary fort in Wales in recent years, being well published in 1997 by Marvell & Owen146

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  John. In particular, the demonstration of frequent changes of garrison, and consequent re-building, is of major interest. The fort possessed a sheltered anchorage and controlled the lowest crossing of the river and, as the river widens out significantly to the north of Loughor, the next feasible crossing-point is some 8 km upstream at Hendy. The only known road from the fort runs towards Neath (Nidum), following the line of the modern A 470 to Swansea, but it is probable that a road ran north-east along the eastern bank of the River Loughor to the next crossing point, and perhaps continued northwards to Llandeilo to link with the main Brecon to Carmarthen road. It would be surprising if a road did not lead to the west from the river crossing, and a road running in the direction of Loughor can be traced for some distance from Carmarthen. Based on the size of the fort and the fact that horse accoutrements were found, the excavators suggest an initial garrison of a cohors equitata quingenaria, but that the garrison of the reduced fort of the 2nd century was an infantry unit of similar size (cohors peditata quingenaria).

century, the pre-existing Iron Age mines were exploited and this activity continued into the 4th century. It is possible that the site complex was involved in the management of the mining and the distribution of ore from the Forest. One excavated shaft extends 15 m underground, reaches a depth of 4.5 m and still bears the marks of Roman picks. Towards the end of the 3rd century, the prehistoric rampart was given a minor rebuild. Shortly after 364, a large temple dedicated to the native British cult of Nodens, was erected in the southern half of the hillfort.

The fort was established during the early years (73–4) of the governorship of Sextus Julius Frontinus, and formed part of a network of forts and roads controlling the main routes through the South Wales uplands and coastal plain. The fort was fully garrisoned until c.85, but that this was followed by an apparent downgrading with a significant reduction in the numbers of troops, and even a possible brief period of abandonment, coincident with the withdrawal from Britain of legio II Adiutrix in 86/7 to deal with problems on the Danube frontier. The return of a full garrison in the earliest years of the 2nd century was marked by a major refurbishment of the fort, including the construction of a bath-house and a new praetorium. At some time between c.105–110, and probably indicating a change of garrison, further construction took place including the building of a new praetorium. Between c.115 and 130, the fort was reduced to approximately two-thirds of its original size, by the construction of new western defences comprising a ditch with clay and turf rampart, accompanied by filling in the original fort ditch. However, the fort appears to have fallen into disuse before construction of the interior buildings had been completed. There is the probability of a brief period of reoccupation in the mid-2nd century, as was also the case at nearby Neath. A further reoccupation, with the construction of a stone wall on the crest of the rampart, accompanied by limited rebuilding within the reduced fort, took place in the late 3rd century; the latest evidence for occupation being a coin of Constantine II (324-330). It is the probable that similar activity took place at Neath in the same period and that both forts, controlled from the newly-built Saxon Shore type fort at Cardiff, formed part of a system of coastal defence along the Bristol Channel.

Lympne (TR117342) - Portus Lemnanis In contrast to Reculver and Walton Castle, both of which have been subject to erosion, deposition means that the late-3rd century Saxon Shore fort at Lympne is now some 4 km distant from the sea. A road runs north to Canterbury and a road running to the north-west eventually reaches Rochester. An irregular pentagon with perhaps as many as 14 external bastions and an interior area of 3.4 ha. the fort was built at an altitude of some 90–100 m above a major tidal inlet. Probably preceded by a naval base for the classis Britannica (a re-used altar, dedicated by Aufidius Pantera, prefect of the fleet and dated to c. 140, CLBR stamped roof tiles and a coin of Antonius Pius (138–61) have been recovered), the Saxon Shore fort was constructed in the late 270s and abandoned c. 340/50.

In Tudor times, men-of-war were built at a shipyard adjacent to the church at Lydney that, at the present time, is 2.5 km from the sea. Access to the tidal pill became impossible as the coast had retreated seawards because of the accretion of sand and silt. In 1664, Samuel Pepys was informed by the local master shipwright, Daniel Furzer, that “Lydney is not so fit a place now for building a ship as formerly, on account of the growing of the sands, not known in man's memory before”.

Manchester (SJ833976) - Mamucium The first fort at Manchester was founded c. 78 and became the hub of a regional transport network. Situated between Northwich and Ribchester on the major north/south highway, other roads led to Ilkley, York, Buxton and Wigan. Located near the navigable River Irwell, at its confluence with the Rivers Medlock and Irk, the Irwell joins the Mersey 32 km downstream. There is evidence for the fort being destroyed and abandoned in the mid-2nd century but by c.160 it had been rebuilt, again in turf and timber, but on a larger scale covering 2 ha. in order to accommodate extra granaries. The final phase of construction came c.200, when the rampart was faced with a stone wall and the gatehouses were rebuilt in stone. The coin series runs down to the Theodosian period, indicating occupation throughout the 4th century. The vicus appears to have been a linear development alongside the road extending westwards from the fort’s north gate, with a cemetery and religious complex on the east and south-east sides. A large number of hearths have been found, either within buildings or in the open to the rear; many of these were used to forge weapons, tools or

Lydney (ST 618025) The Romano-Celtic temple complex lies at the southern end of an ore-bearing stratum, 4.5 km north of ironworking site at The Chesters. Commencing in the 2nd

147

JAMES ELLIS JONES

  other equipment. The vicus declined considerably during the 3rd and 4th centuries, in contrast to numismatic and construction evidence indicating continued occupation of the fort during this period.

Coriosolites, a tribe of northern Brittany, three Carthaginian drachmas, and one early Celtic gold coin of uncertain origin have been recovered. The coins of the Coriosolites date from the early 1st century BC, prior to any direct Roman interest in Britain or Gaul; the Carthaginians coins are similar to a type known to have been minted in the 2nd/3rd centuries BC and, despite difficulties in identification, the gold coin probably dated to the last half of the 1st century BC. Carthaginian coins are not a rarity in Britain, and coins of the Coriosolites have been found as far north as Hexham in Northumberland and Lesmahago in Lanarkshire. The Roman finds suggests military activity at Meols in the pre-Flavian period, before the foundation of the fortress at Chester and perhaps a market function afterwards and the port grew to be the largest settlement in Merseyside. Over 70 Roman brooches and 120 coins have been found, showing that Meols was a busy trading community with ships sailing up the west coast of Britain stopping to pick up goods or raw materials and trade pottery and other items.

Martinhoe (SS663493) Old Barrow (SS788493) It has been suggested that, with a garrison of perhaps thirty men, these mid/late-1st century fortlets were intended to observe hostile naval action in the Bristol Channel when the war in South Wales was at its height. The excavators believed that one fortlet succeeded the other, with Old Burrow being established c.48, being replaced by Martinhoe about 60 and then evacuated in c.78 on the final subjugation of the Silures. However, the identical and unusual layout of the earthworks, suggest that that they are contemporary, intended to operate as a pair, each being responsible for a different sector of the coastline. Maryport (NY 038373) - Alauna A probable Flavian fort at Maryport was located on the coastal road, near to the mouth of the River Ellen, on a sandstone ridge near the cliff edge, with views across the Solway Firth to south-west Scotland. Two possible sections of Roman roads to Papcastle have been identified from air photographs. The adjacent Hadrianic fort, built c.124, was occupied with periodic interruptions, until the late 4th /early 5th century. The 2.5 ha. fort is larger than needed for a 500 strong auxiliary unit and a role as a supply point for the western part of Hadrian’s Wall has been considered. The strong-room is some three times the size of that in most other auxiliary forts; this suggests that, in a similar manner to South Shields, it was used for the storage of bullion in transit.

Middlewich (SJ702669) - Salinae Near to the River Dane, between Northwich and Whitchurch on the north/south road, the 1.4ha. timber and turf fort at Middlewich, was probably founded during the late 70s and may have been relatively short lived, although a bronze diploma, issued to an auxiliary soldier on his retirement in 105, indicates that there was still a military presence in the area at this time. A settlement to the south of the Flavian fort was probably occupied from around the last quarter of the 1st century to c. 400 and contained a number of timber buildings dating to the 2nd and early 3rd centuries together with evidence for pottery making and tanning leather. The area was one of a number of places in Cheshire where, during the Roman period, salt was produced on a semi-industrial scale; finds from Middlewich include numerous pieces of briquetage, brine pits, evaporating pans and other items of saltmaking equipment.

An extensive vicus, covering an area of some 16 ha. lies along a 420 m length of road leading from the north gate, with large strip buildings, two temples, a possible mithraeum, a mansio and large buttressed buildings, which were probably granaries. Evidence for industrial activity is indicated by deposits of coal and iron debris.

Monmouth (ST510130) - Blestium? Situated on the banks of the River Wye, near its junction with the River Monnow; major roads run west to Gloucester and south-west to Usk and Caerleon, with a branch road leading to the major iron-working settlement at Weston under Penyard (Ariconium). Finds of early samian pottery and bronze brooches, probably attributable to the military, meant that a large Neronian fort has long been suspected. Excavation in 1992 revealed a section of military ditch, with pre-Flavian samian pottery in the ditch slot. Massive deposits of waste lie beneath the present town and on the riverbed of the Wye, and although much of the slag was produced during the medieval period, excavation revealed sealed Roman contexts lying beneath 4 m of later iron slag. There is evidence of iron working on the site from c.75 to the end of the 3rd century, and the quantity of smelting and primary smithing slags suggest that iron working was extensive throughout this period. A single coin of Gratian (367–385) is the latest Roman coin found in Monmouth.

A Roman road running towards the lowest crossing point of Barney Gill has been suggested as a possible site for a harbour. However, the River Ellen, to the south of the fort, is both a more substantial river and also considerably closer than is the mouth of Barney Gill (650 m as opposed to 1400 m); and the mouth of the Ellen is the most likely location for the harbour. Meols (SJ275935) The maritime trading site (emporium) at Meols, on the northern tip of the Wirral peninsula, was situated on the coast of the Dee estuary, some 30 km from the Roman fortress at Chester. Discovered in the early 19th century, the site has now been eroded by the sea, but sufficient Iron Age material was recovered to indicate clear evidence of long distance trade prior to the Roman occupation of Wales, together with artefacts of the Romano-British period. Two silver coins of the

148

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  evidence of Roman activity during this period, and a military presence is the most likely explanation.

Moresby (NX982210) - Gabrosentum The 1.5 ha. fort is located on a coastal headland, with commanding views on all sides, and a small natural harbour at the foot of the cliff. No significant excavation or survey has been carried out. Founded during the Hadrianatic period, 3rd and 4th-century occupation is attested by coins and inscriptions. It is probable that both the vicus and the military bathhouse lay to the south of the fort. A small natural harbour once existed at the foot of the cliff west of the fort and may have been used during the Roman period.

Netherby (NY 339571) – Castra Exploratorum Located on the banks of the River Esk, 12 km from the Solway coast, on the road from Carlisle, the fort was the site of ‘Castra Exploratorum’ (Camp of the Scouts) of the Antonine Itinerary but there are no longer any visible remains, as the mansion, Netherby Hall, occupies the presumed site, so the precise size and orientation of the fort remain uncertain. In 1539, Leland reported “ruined walls of marvellous buildings” and also talks of rings and staples in the wall for mooring ships. Stukeley states that foundations and streets were visible and mentions finds of a cemetery and altars in gardens. An alleged bathhouse was discovered in the 1730s with an inscribed altar and an aqueduct. The fort was probably constructed in the late 1st century but there is no definite evidence of this. A coin from the reign of Gordian gives evidence of occupation to 230–40, but fort may still have been occupied into the 4th century.

Morwenstow (SX210139) High Cliff, St Genny’s (SX130945) These possible fortlets, on cliff-tops 20 km apart, were located by aerial photography, with only the latter being visible on the ground as a distinct earthwork. Nantstallon (SX034670) The mid/late-1st century fort is situated in mid-Cornwall on the south bank of the River Camel and, at 1.2 ha. is small in comparison to other forts of the period. The Camel is tidal and navigable and so could provide a means to bring in sea-borne supplies to within 6 km of the fort. Nantstallon is situated in a district rich in metals, with deposits of silver-lead, associated with copper (mined commercially in the 19th century), being present at Lanivet, some 3 km from the fort. The discovery of silver-rich slag, fragments of crucibles and an 8 dram weight indicates that the source had been located, and that cupellation was being practiced in the fort. An ironbearing vein of haematite with banded quartz lies a mile to the east, and evidence of smelting and smithing was found inside the fort.

Newcastle upon Tyne (NZ251639) - Pons Aelius Probably constructed by legio VI Victrix in the late2nd/early-3rd century, the small (0.5 ha.) bridgehead fort was situated on a bluff overlooking the Lort Burn to the east, and the bridge over the River Tyne to the south. The fort was detached from Hadrian's Wall which ran to its north and was intended to protect the crossing of the Tyne (Tinea Fluvis) by the 235 m long Pons Aelius (from Publius Aelius Hadrianus) of the road from Chester-leStreet, some 15 km to the south. The bridge is estimated to have been over 200 m in length, probably constructed on 10 stone piers supporting a timber superstructure. The bridge stood at the original terminus of Hadrian's Wall and it has been suggested that this may have been the reason for its eminent name and monumental design. Despite its obvious potential, the relatively modest size of the fort suggests that it is unlikely to have functioned as a supply-base for other forts on the Wall.

Neath (SS748977) - Nidum The 2.4 ha. fort lies on a low plateau of the western bank of the tidal River Neath, controlling access to the Neath Valley, which penetrates deeply into the hills. EarlyFlavian samian and coarse ware recovered from the fort is consistent with foundation during the governorship of Sextus Julius Frontinus. Overlooking the crossing of the road from Caerleon to west Wales, it was easily accessible by land and sea, and is therefore one of the most strategically important forts in South Wales. A road ran north-west to Coelbren and then Brecon Gaer; a milestone of Diocletian (284-305), was found some 2.5 km south of the fort, and probably marks the line of a major road running east to Cardiff and Caerleon, and west to Loughor and Carmarthen.

Newton Kyme (SE456454) - Calcaria ? The fort is situated on elevated ground, south of the River Wharfe, on the Rudgate, some 20 km south of Castleford and a similar distance to the south of Aldborough. A small fort of 2.5 ha. was later overlain by a larger stonebuilt fort of 4.7 ha. The vicus covered an area of some 15 ha. flanking both sides of a broad road running for some 550 m to the south of the fort. A rectangular late-Roman enclosure with a road running along its central axis, lined with buildings and side streets, encloses an area of 4.7 ha. within its stone walls, and has been suggested as a defended circuit added to an urban area. An inhumation cemetery and several possible mausolea outside the west rampart strengthen the suggestion that Newton Kyme might be considered as a “small town”.

The history of this site is complicated, with initial construction and occupation dating from 80/90. At some time, rebuilding took place, and the site was occupied until c.110/115, followed by a short period of abandonment and then reoccupation from about 117 to 125. The fort was again abandoned from about 125 to 140, with a second re-occupation lasting from 140 to 170. There is evidence for late-3rd/early 4th-century reoccupation, with all excavated sites producing some

Newton on Trent (SK823736) A vexillation base of some 10 ha. lay on a 30 m high ridge alongside the River Trent, 25 km downstream from Thorpe. Newton is 25 km west of Lincoln and 15 km

149

JAMES ELLIS JONES

  north of Brough, there are no known Roman roads in the immediate area so proximity to both the Trent and the Foss Dyke (5 km) suggests probable provisioning by water transport.

suggested that a port or landing place was probable at Old Kilpatrick. It might be argued that the siting of the fort at Old Kilpatrick, and the absence of an extension of the Wall to the west, suggests that there was no perceived danger from outflanking via a ford and that no such crossing place existed during Roman times.

Northwich (SJ656736) - Condate The timber and turf Flavian fort was located at a strategic position on the west bank of the navigable River Weaver near to its confluence (Condate) with the River Dane, with major Roman roads leading to Chester, Manchester, Wroxeter and Wigan. There appear to have been two periods of military occupation, the first beginning around 70 followed by abandonment until the early Antonine period, with reconstruction on a smaller scale, when a stone wall was revetted onto the rampart of the early fort. Military occupation of the fort ceased c.140, but there was probably a continued small-scale military presence connected with manufacturing activity. Whilst there is evidence for industry in the form of a pottery kiln and ironworking furnaces, there is no doubt that the major activity was the exploitation of the local brine springs. This was demonstrated by the discovery, in the 19th century, of five lead pans for use in salt-working, together with a nearby 1st century brine kiln.

Osmanthorpe (SK678564) Situated on the west bank of the River Greet, and probably established c. 60, the 8.8 ha. vexillation base is 13 km to the north west of Thorpe (Ad Pontem) on the Fosse Way. Although the nearest point on the River Trent is only 10 km distant, a number of small rivers and marshy ground suggest that access to the River Trent would be via a (probable) road leading in the direction of Thorpe. Outerwards (NS231666) Situated on the highest point of a broad, moorland ridge at approximately 272 m OD, the 0.18 ha. fortlet commands extensive views to the Firth of Clyde to the west and north. The site had two separate periods of occupation during the Antonine period, the second following the complete destruction of the earlier fortlet.

The discovery, to the north and west of the fort, of large quantities of 1st century pottery, coins, spearheads, melon beads and a possible hearth, indicates the presence of a vicus. This may well have developed into an independent manufacturing settlement after the fort was abandoned, as its location on the road network and sited on a navigable river may have ensured continued activity, albeit at a reduced level. The River Weaver continued in use as a major route into the Cheshire plain, and by 1732 the river had been improved from Frodsham Bridge to Winsford Bridge and eventually allowed vessels up to 160 tonnes to travel up to Northwich Bridge.

Papcastle (NY110315) – Derventio The first fort at Papcastle, lying on the road from Carlisle at its junction with the road to Maryport,, was established during the Flavian period and was constructed of timber and turf. This fort was levelled and a 2.8 ha. stone fort was built, probably during the early part of the 2nd century. At some time between 293 and 306, the area was again levelled and the fort rebuilt for a third time. The pottery indicates occupation from the Flavian period to the end of the 4th century with, however, a possible break between 120 and 160. The fort and vicus, with a bathhouse and a probable temple to Apollo, are more substantial than others in Cumbria; the location at this important road junction appears pivotal and suggests that the fort exercised a command and control function over the Lake District

Old Kilpatrick (NS460732) The west terminal fort of the Antonine Wall, on the banks of the Clyde; it is probable that the 1.7 ha. fort was constructed prior to the Wall itself. A little pottery of the late 1st century has been found and this may indicate an earlier Agricolan occupation at this location. Excavations have revealed a stone-built principia and horrea, and timber-framed barrack blocks. During cutting of the Forth-Clyde canal in 1790 a bath-house within an annexe was discovered outside the south defences of the fort.

The fort lies about 1.5 km from the present town, at a meeting place of Roman roads from Maryport, Carlisle and Penrith. Papcastle is near the confluence of the River Derwent and the River Cocker; downriver the sea is reached at Workington, the site of the fort at Burrow Walls (Magis ?). An annexe on the east side of the fort reached the river, and barges could have been used to bring in supplies from the coast, 16 km downstream. Alternatively the major road from Maryport (9 km away) might have been used for this purpose.

A Roman port or landing place might be expected; at the present time we are familiar with the passage of large ocean-going vessels to the port of Glasgow. However, in the 18th century the Clyde was fordable on foot at Dumbuck, some two miles below Old Kilpatrick. There is, however, no evidence that a ford existed in Roman times, indeed the incidence of higher sea levels and increased wave action on shores makes it less likely. Even if such a ford had existed, the deepening and stabilisation of the river would have been well within the capability of Roman engineering and this topic is extensively discussed in Chapter 9, and it is therefore

Pen Llwyn (SN650806) The 2.8 ha. fort is situated close to the village of Capel Bangor, on a spur overlooking the Afon Rheidol, at a point 8.5 km from the sea. The fort is 6 km from the present tidal level, but extensive engineering works in Aberystwyth have clearly reduced the point to which the tide flowed in Roman times. The course of the road linking Trawscoed to the south, and Pennal to the north

150

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  has not been established, but such a road probably existed; a road to the west leading to the fort at Cae Gaer is also a possibility. Only limited excavation has taken place, but evidence from pottery suggests a brief occupation during the Flavian-Hadrianic period.

Pen-y-Gaer (SO169219) The fort stands on a detached knoll in the valley of the Rhiangoll, where the road from Abergavenny to Brecon, previously following the valley of the Usk, was diverted to the north of Myarth in order to avoid the confines of the steep-sided valley in the area of Llangynidr. It is probable that the fort was constructed during the campaigns of Sextus Julius Frontinus c.75 and was clearly intended to control the most westerly route through the Black Mountains, thereby dividing contact between Silurian territory to the north and south of the Usk. There appear to have been three phases of construction, firstly a rampart and ditch followed by the raising of the rampart and construction of a stone wall, with a final rebuilding involving dismantling the wall and reconstruction of a replacement to further increase the height of the rampart. The lack of stratified pottery means that there is inadequate evidence to date the two rebuildings in stone, or the final abandonment of the fort, that is, however, unlikely to be later than 130.

Pennal (SH705001) Until recently, Pennal was known as the site of a 1.7 ha. late-1st century fort but geophysical survey has revealed the existence of an earlier fort of some 2.3 ha, Finds of early or pre-Flavian pottery from the site could indicate a foundation during the Frontinian campaign, but a preFlavian fort remains a possibility. A civil settlement extended for over 200 m along two of the roads from the fort and a bathhouse, mansio, parade-ground, and a possible circular tomb have been identified, together with the usual range of buildings associated with a vicus. Pennal effectively illustrates the problems of land transport, and the advantages of sea communications. Located on low spur 415 m to the north of the Afon Dyfi, at this point tidal, the fort commands the first good rivercrossing above the estuary some 15 km to the west. The mouth of the estuary is now less than 1 km wide, having been reduced from an earlier width of 7 km by the growth of the Borth Sands, a bar deposited since Roman times. This view is supported by the positioning of a small fortlet at Erglodd (ST653905), on a hillside now separated from the river by 4 km of marshland, but the topography suggests that in Roman times it was located on a sheltered bay on the south side of the estuary.

Pevensey (TQ644048) - Anderitum Constructed during the reign of Allectus (293–6) as an irregular oval in order to fit within the peninsula on which it was built, the bastioned walls enclosed some 4 ha., Pevensey is now approximately 1 km inland. There is some evidence for timber-framed structures with mortared floors, possibly barrack blocks. The classis Anderetianorum is recorded as being the stationed in Gaul in 395, indicating that Pevensey had been given up by that date. In common with the other Saxon Shore forts of the south coast, no evidence for a vicus has been found.

The mountainous road running westerly to Caersws is indicated by the short stretch of road running through the small fortlet at Pen-y-Crocbren. The course of the road running south to Erglodd, Trawscoed and Llanio is still conjectural. The routing to Brithdir where the road continued north to Tomen-y-Mur and branched northwest to Caer Gai and Llanfor, presents even more problems. Pennal is on the south side of the Cader Idris massif that reaches a height of 893 m, and both the modern A470 via the Dyfi Valley, and the more direct, but even more precipitous A487, would have presented very severe difficulties for any form of Roman wheeled transport. It is probable that there was no alternative to a much longer coastal road, as followed by the modern A493.

Piercebridge (NZ452115) - Morbium/Vinovium It is highly probable that a Flavian fort preceded the exceptionally large 4.4 ha. 3rd-century fort that is situated on Dere Street, on the north side of the River Tees to the north of Catterick and south of Binchester. Four inscriptions of early 3rd-century date attest the presence of legionary detachments, probably from legio VI Victrix suggest a Severan foundation but a later date of c.260–75 is favoured by some authorities. An elaborately decorated courtyard building is on a scale resembling that of a legate’s residence of a legionary fortress and may have been constructed during the Severan period, or possibly converted from a former praetoria in the late-3rd//early-4th century. The fort was abandoned around 320 but was later re-occupied and remained in use until the early-5th century.

Pentrehyling (SO245931) The 2.5 ha. fort at Pentrehyling, near Brompton, is thought to be a Flavian foundation, with a long period of military, then civilian, occupation, as the ditch was re-cut in the early 4th century. Bowl furnaces and clay hearths with charcoal and furnace debris, together with pits containing considerable quantities of litharge from cupellation indicate industrial activity within the fort itself. To the east of the fort, the vicus contained a dense concentration of industrial pits and gullies, yielding ironsmithing slags, hearth bottom material and lead waste, suggesting a processing plant for the ore obtained from nearby Linley.

An early timber bridge carried Dere Street across the Tees and many Roman coins, indicating votive offerings, have been recovered from the riverbed. At some time in the late-2nd/early-3rd-century this bridge was replaced/supplemented by a new bridge based on stone foundations and a timber superstructure, some 200 m to the east; the south abutment and four piers are visible at the present time.

151

JAMES ELLIS JONES

  An extensive vicus developed in Toft’s field adjacent to the fort containing a range of 30 stone-built structures, including an underground temple to Jupiter Dolichenus. The find of a North African amphora datable to c. 400 demonstrates not only the import of olive oil, but also the relative affluence of Piercebridge at this late date.

fort was occupied from c. 130 until c.400, with periods of abandonment in both the 2nd and 3rd centuries, there is evidence of destruction by fire on three occasions with the final rebuilding taking place c.350–70. An extensive vicus to N and E of the fort, still has an upstanding bath house known as Walls Castle and a probable mansio, dated by coins to the 2nd to 4th centuries.

Plymouth (Mount Batten) - SX487532 At least eight hoards and over 50 individual coin finds have been made at different points overlooking Plymouth Sound. Significant quantities of building materials, including Roman tiles, have been recovered from the foreshore of Sutton Pool, and other Roman material has been found at various locations on the banks of both the rivers Tavy and Tamar. Plymouth Sound provides an outstanding natural harbour and, taken in conjunction with the existence of the Iron Age entrepôt on Mountbatten Promontory (possibly the Ictis of Diodorus Siculus), the Roman settlement at Plymstock Quarry and the coin evidence dating from the 1st to 4th centuries, it is highly probable that a Roman port, as yet unlocated, lies in the area of Plymouth Sound.

A fragmentary military diploma dated to 158 belonging to a veteran of the Cohors I Aelia Classica (1st Cohort of the Aelian Fleet) was found on the seashore near to the fort. The name suggests that the Romans made use of the natural safe-harbour formed by the triple confluence of the Rivers Irt, Mite and Esk (tidal for 7 km in the direction of Hardknott). Reculver (TR227693) - Regulbium Preceded by a Claudian fortlet, the late-2nd/early-3rd century Saxon Shore fort, originally 3.3 ha., has now been partially eroded by the sea with only the southern half remaining. A building inscription dated to c.260 and the absence of towers confirm an early foundation. An underground strong-room lay beneath the macellum of the large (33m x 42m) principia, a bathhouse and backhouse and the barrack blocks were similarly constructed of stone. Successive re-metalling of the roadway through the south gate raised the level of the road by over 1 m, indicative not only of significant wheeled transport but also an extensive period of occupation. An antiquarian report of the 18th century describes "foundations of great bulk" and “cisterns” to the north of the fort and there is more recent evidence for settlement to the west, along the road to Canterbury.

Portchester (SU625046) - Portus Adurni Built in the later-3rd century, the fort was a regular square enclosing 3.43 ha. with 20 bastions. Situated on a still navigable tidal creek, it is probable that, during Roman times, the shoreline may well have extended several hundred metres to the south-west. Extensive excavation revealed only limited evidence of a limited number of timber buildings. However, the interior has been much disturbed during Saxon, Norman and medieval constructions and further Roman remains may have been destroyed and finds of hypocaust and flue tiles, imbrices and tegulae suggest the existence of, at least, a tile-roofed bathhouse. The absence of evidence for a vicus and finds of "civilian" occupation inside the fort suggest that both military and civil personnel lived within the fort’s defences.

Ribchester (SD650350) - Bremetennacum Veteranorum The fort is located at a major north/south and east/west junction, with roads leading to York, Manchester, Carlisle and the Fylde coast. Spring tides sometimes reach Ribchester and it is highly probable that the river was navigable in Roman times. Stukeley mentions Anchor Hill, just to west of the fort, where there had been discoveries of “anchors and great quantities of iron pins of all sizes for ships or barges”. The 2.7 ha. fort was built on the bank of the River Ribble and some thirty per cent has been destroyed by river erosion. A turf and timber fort constructed c.72 was modified c.82–86, and later demolished to be replaced by a new fort with stone defences and central buildings c.117–25.The principia was rebuilt in stone c.161–9, perhaps as a result of reactivation following the withdrawal from the Antonine Wall. There was continued occupation at Ribchester until the 4th century, the latest coins include a copper issue of Valentinian I (364–75) and a gold coin of Gratian (367– 383); final evacuation may have been as late as 383 or even under Stilicho in 396. It has also been suggested that occupation of parts of the fort may have continued into the post-Roman period.

Prestatyn (SJ 062817) Located on the North Wales coast, probably with a harbour on a now vanished inlet, a Roman fort has been suggested on this site, but recent investigations cast doubt on the identification of the general area as the site of a fort. The site is probably best interpreted as a settlement serving the lead-mines at nearby Meliden. Two workshops with crucibles, moulds, hearth linings and bronze scrap, including brooches and horse trappings have been located, together the remains of a bath-house. Both the bath-house and the workshops were abandoned by 160, probably as a result of reduced demand from the fortress at Chester. Ravenscar (NZ981018) Yorkshire coast 4th century “signal station”. Ravenglass (SD 088958) - Itunocelum Located on the shoreline with an excellent natural harbour, serious coastal erosion of the 1.8 ha. fort is taking place. Preceded by an early-Hadrianic fortlet, the

Ribchester’s importance not only as a fort, but also as a regional centre is demonstrated by an inscription, dated between 225–35, which describes Titus Floridus Natalis,

152

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  a legionary centurion, not only as commander of the garrison but also as praepositus regionis in charge of the surrounding area Some 350 m to the north of the fort, large timber buildings of military pattern were constructed c.125 and these have been suggested as part of an extramural settlement for veteran soldiers; implied by Veteranorum as part of Ribchester's Roman name.

provisioning via the Humber Estuary is a strong possibility. Rochester (TQ738695) - Durobrivae A short-lived Claudian fort at Rochester controlled the crossing of the navigable River Medway by Watling Street, with London 40 km to the west and Canterbury the same distance to the east. A road to the south headed in the direction of The Weald and the English Channel. The earthen rampart and ditch of the late-2nd century enclosed an area of 9.5 ha. and was fronted by a stone wall during the first half of the 3rd century. There is no clear evidence of industrial activity in the immediate area of Rochester but, pottery was produced at Upchurch on the eastern bank of the estuary and in the same area, there are indications of large-scale extraction of salt from seawater. If, as is often suggested, salt production was under Imperial control, then Rochester may have served as an administrative centre. The valley of the Medway is home to a significant concentration of villas to beyond Maidstone; Rochester may have served as a shipping and service area for their products.

Richborough (TR324601) - Rutupiae Richborough was the landing place of, at least, part of the Claudian invasion force of 43 and, somewhat ironically, served as a protected embarkation port for departure of the army at the end of the Roman period. It is a further irony that legio II Augusta, part of the invading army of Aulus Plautus, is recorded in the Notitia Dignitatum as being garrisoned in Richborough at the end of the 4th century. Now inland, during the Roman period the site was an island on the south side of the Wantsum Channel, which at that time separated the Isle of Thanet from the mainland of eastern Kent. There is substantial evidence for Late Iron Age imports to the area, and it is probable that, prior to the Roman arrival, Richborough was already developing as a port.

Roecliffe (SE 387665) The Flavian fort lies to the north-west of York, south of Catterick, on the east bank of the River Ure, 2km upstream of Aldborough (Isurium Brigantum) and 2 km to the west of Dere Street. Occupation of the 3 ha. fort was confined to the period c.70–80; it is possible that the site was prone to flooding and was superseded by a fort (yet to be discovered) at Alborough. An annexe with a ditch and probable rampart contained evidence for metalworking and there is evidence for some form of settlement on the north side of the Ure.

The bridgehead fort covered 4.5 ha., initially forming a defensive perimeter, and then developing into a secure landing place for supplies and provisions, as the army broke out into the surrounding countryside during its advance towards Colchester. A range of large granaries, similar to those of (later) South Shields, indicate that continental wheat was imported in large quantities prior to the Agricolan governorship, at which time Tacitus indicates that local supplies of wheat were available. It is, however, unlikely that Richborough acted as the sole source of import, as direct supply to the advancing army in the north and west by coastal shipping is a more effective option. In the later-1st century many of the existing buildings were demolished and replaced, new roads were constructed (Watling Street runs for 100 km to London, via Canterbury and Rochester) others relayed and, beginning in c. 85, a huge monumental arch, 25 m high, clad in white Carrera marble and decorated with sculpture and bronze inscriptions, was constructed. There is evidence of a decline in prosperity in the late-2nd century with buildings and roads falling out of use, probably as a result of transfer of the majority of the import/export trade to the port at Dover. The monumental arch was converted into a watchtower defended by ditches, and finally demolished in the late-3rd century to enable the construction of a Saxon Shore fort defended by bastions and interval towers. Dating from the late-3rd century, an amphitheatre lying 500 m to the south of the fort, appears to have been occupied until the early-5th century.

Rossington (SK629991) Lying on the River Torne, 7 km southwest of Doncaster on the road to Lincoln, the 9.3 ha, vexillation base was possibly founded c.60 at the time of the activities of Queen Cartimandua of the Brigantes. A number of kilns produced a wide range of black burnished ware and mortaria produced by the potter SARRIVS has been found at Corbridge, Catterick and as far north as the Antonine Wall. Ruthwell (NY102677) This 2.2 ha. marching camp, of probable Flavian date, lies close to the shores of the Solway Firth. St Albans (TL134073 ) - Verulamium On the banks of the River Ver, St Albans lies on Watling Street, 33 km to the north-west of London with other roads leading 60 km west to Alchester, the same distance north to Godmanchester and 90 km east to Colchester. The distances to the nearest military establishment are clear evidence of the ease with which this part of the country was first conquered and subsequently policed with the significant exception of the Icenian rebellion. Part of the River Ver was artificially banked by a revetment of wood and clay; the existence of a wharf at

Roall (SE564252) Known only from aerial photography and geophysical survey, the 1.3 ha. Flavian timber and turf fort is located on a low outcrop overlooking the River Aire, 14 km to the east of Castleford and 18 km north of Long Sandall. No Roman road is known in the area and supply and

153

JAMES ELLIS JONES

  Park Street villa, 3.6 km downstream, indicates significant use of water transport.

garrison had departed. Unlike Caerwent, and despite its proximity to the Bristol Channel, Sea Mills does not seem to have possessed defences, but the coin series, terminating with an issue of Arcadiius of 408, indicates that the settlement was still occupied at that late date.

Prior to the Roman invasion, an early tribal capital of the pro-Roman Catuvellauni was situated at Verlamion (Roman Verlamium) where a short-lived Claudian fort was a replaced c. 49 by a civil settlement with a planned street system, surrounded on three sides by defensive ditch and other fourth side by the River Ver. Destroyed during the Boudiccan revolt a new forum and basilica were built in 79 within a 48 ha. ditch and earth bank enclosure that, at the beginning of the 2nd century, was filled in and new defences enclosing an area of over 200 ha. erected. St Albans was given the self-governing status of a municipium, probably by Vespasian, in the early 70s and is the only town in Roman Britain where there is specific evidence for such distinction, in the form of a building inscription on Hadrian’s Wall.

Bath stone was used extensively in the fortress at Caerleon, and throughout southern Britain, and it is probable that this was transported down the River Avon in barges and, together with smaller quantities of limestone from nearby Dundry Hill (also used in the late Roman fort at Cardiff), was shipped from Sea Mills. Shadwell (TQ352807) Recent excavations some 2.5 km to the east of Londinium have revealed a high-status building with a bathhouse, and at least ten rooms with hypocausted marble-slabbed or mosaic floors. Built at some time in the 2nd century the building was in use until demolition took place around the late-4th/early 5th century. An extensive complex of timber buildings, interpreted as warehouses or granaries, suggest that as the tidal head of the Thames moved downstream (see Chapter 1) Shadwell became the port for Roman London. Built at some time in the later 3rd century, an 8m square structure originally thought to be a Caruasian/Allectian watch tower/signal station covering the approaches to London has now been reinterpreted as a mausoleum.

A fire destroyed at least 21 ha. of the central area in about 155, followed by rapid reconstruction in stone of a new forum, basilica, Gallo-Roman type theatre, several temples, marketplace and building of a number of large town houses, some with as many as 30 rooms, took place. In the late-3rd century, monumental arches were erected on Watling Street and stone defences with internal towers were built, enclosing a smaller area than the previous ditch and bank. As in the case of Canterbury, there is a scarcity of villas in the immediate surrounding countryside, suggesting that the majority of landowners lived in the town For the next 200 years, substantial town houses continued to be built or altered, and many mosaics were laid; it is only from the early-5th century that there are signs of decline, but it is probable that some level of urban life was maintained within the city until at least 450.

South Shields (NZ365679) - Arbeia The Antonine fort at South Shields, lies on the Lawe, a 20m OD headland at the mouth of the River Tyne (Tinea Fluvis), 6.5 km east of the terminus of Hadrian's Wall at Wallsend. Finds of South Gaulish samian from the site, suggest the presence of a yet to be discovered earlier fort, presumably Flavian. The topography of the river has undoubtedly changed significantly since the Roman period with sandbanks, bars, reefs and the channels appearing and disappearing on numerous occasions. For example, before infilling during 1816, the Mill Stream flowed through a channel to the south of the fort and a map from the time of Henry VIII shows that, at high tide this afforded an alternative passage from the estuary to the River Tyne. Despite a number of developer-funded excavations, the location of any port facilities remain to be discovered. The notorious Herd Sand, lies on the south side of the river estuary and objects from a probable shipwreck indicate the use of the Tyne by troop transports in the Roman period. These include a shield-boss of legio VIII Augusta (at that time based in Strasbourg), a helmet cheek-piece and 67 coins, mainly silver denarii, which cluster in the second half of the 2nd century with the latest coin being from 176–80.

St Madoes (NO209196) Situated on the north bank of the Tay Estuary this 0.5 ha. enclosure lies opposite the Severan fortress at Carpow and has been suggested as a fortified bridgehead, supporting the crossing of the estuary. Scarborough (TA052892) Yorkshire coast 4th century “signal station”. Sea Mills (ST760551) - Portus Abonae Sea Mills is located on the tidal River Avon, some 6 km from the Severn Estuary, and at the confluence with the smaller River Trym. To the east was a major road leading from Bath to the Fosse Way, and a northerly route led to Gloucester and the Severn Valley. A ferry service operated across the Severn to the fortress at Sudbrook in south Wales, probably from Sea Mills, although Aust is an alternative terminal. The presence of ClaudioNeronian coins and samian ware, together with military equipment, indicate an early military presence. Tiles and bricks, stamped “LEG II AUG” suggest a continued military presence into the 2nd century, at which date these tiles first occur in the fortress at Caerleon. A vicus developed and continued to exist after the military

The stone built fort, constructed around 160, covered an area of 1.67 ha., similar in layout to the earlier fort at Wallsend. In order to provide a supply base for the Severan campaigns in Scotland the fort was extended in area to 2.1 ha with modifications to the barrack blocks taking place around 225–235. Possibly as the result of assault by seaborne raiders, parts of the fort were

154

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  devastated by fire in the late-3rd century. Redevelopment and refurbishment took place during the period c. 286 – 318 and the fort continued to operate as a supply base, but on a much reduced scale. The number of granaries was reduced to nine with some others being converted to provide accommodation for the incoming numerus barcariorum Tigrensium (the Tigris bargemen). Despite the unit title, it is likely that quayside structures still existed from the floruit of the supply base during the 3rd century, thus rendering lighterage unnecessary and it is probable that, in addition to their land-based duties, the bacarii functioned as a light naval unit, charged with the defence of adjacent areas of the coast.

the needs of worshippers and passing travellers, a roadside settlement developed along Watling Street including a bathhouse, smithies, bakeries and shrines. Staines (TQ036716) – Pontibus The settlement at Pontibus (Bridges), probably preceded by a Claudian fort, occupied a gravel island raised a little above the floodplain at the confluence of the Thames and the Colne, on the road from London to Silchester. Early development of the town began c. 65-70 with several timber buildings being known and, as the Latin name indicates, more than one bridge. The town was pillaged during the Boudiccan uprising but was immediately rebuilt and developed into a prosperous town and, from the early-2nd second century, there is evidence of buildings with tessalated floors and hypocausts. A significant decline in prosperity commenced towards the end of the 2nd century and continued throughout the 3rd century, possibly exacerbated by episodes of flooding and the once prosperous commercial centre declined to a small settlement. At some time in during the late-4th century, surrounding ditches and embankments were constructed, but whether these to prevent flooding or were military defences has not yet been determined.

Excavation has revealed at least 16 granaries and it has been suggested that, were the area to be completely uncovered a total of 24 would be found. This is one of the largest numbers of military granaries in the whole of the Roman Empire, providing a capacity of some 2800 tonnes of grain, enabling supply and provisioning of the Severan campaigning army for up to 2 months. The 4th century praetorium was a high status courtyard house of Mediterranean style with two triclinia (one for – and one for winter use), a baths suite, porter’s lodge, kitchen and stables, with the evidence of Theodosian coins indicating continued occupation, even if in a somewhat less luxurious manner.

Stanwix (NY 402571) – Uxellodunum/Petriana? A small Hadrianic fort was built on a plateau contained by the 30 m contour, overlooking the River Eden at Stanwix (now a suburb of Carlisle), but as it was later replaced on the same site, little is understood of its size or exact location. It is not known whether the putative early small fort was built in stone or turf and timber, although the latter is the most likely interpretation. Presumably to house the incoming garrison of the ala Petriana, the fort was rebuilt in stone c.60 and was enlarged to 3.9 ha. In several instances, the internal buildings of the Antonine fort overlie a gravel surface, presumably contemporary with the smaller and earlier Hadrianic fort.

There was, undoubtedly, an extensive vicus but, as the fort is surrounded by buildings, its extent and composition is not precisely defined. However, there is evidence for extramural activity on all sides, with a road and of timber buildings, including an unfinished granary, having been found 125 m to the south-west of the fort. Cremations and inhumations took place in an area 230m south of the fort, close to the location of the well-known later-2nd century tombstones of Victor and Regina. There is evidence, in the form of a table altar, for the construction of a Christian church within the principia but early-5th century burials, from the same location, show evidence of violent death.

The most intensive use of the fort appears to be in the later 2nd century after the arrival of the ala Petriana in the 160s. There is a noticeable falling off in pottery datable to the 3rd century, before an increase in activity in the 4th century. The coin list from Stanwix shows an almost complete absence of 3rd-century radiates; it has been suggested the fort may not have been fully garrisoned at some time during this period. Recent excavations have identified timber-built phases postdating the stone buildings of the fort, possibly suggesting civilian occupation. It is probable that military occupation of the fort ceased around 330; activity of a non-military nature continued into the last quarter of the century or later.

Shapwick (ST948023) Only 7 km to the south-west of the vexillation base at Lake Farm, it might be argued that the Shapwick fort was the earlier site, intended to give protection during the construction of the base being was vacated on the arrival of the full vexillation garrison. Springhead (TQ617724) - Vagniacis Lying on Watling Street, with London to the west and Rochester to the east, Springhead was the site of an early undefended supply base on the shore of the River Ebbsfleet. A pool fed by eight natural springs at the head of the Ebbsfleet, had long been a Celtic sacred site with a 600 m ceremonial way and ritual pits filled with animal remains and pots, as well as numerous coins. During the early Roman period, this was developed into a religious complex covering 15 ha., with at least 12 temples of various designs, many of which remained in use until the advent of Christianity in the 4th century. In response to

The evidence of civil occupation outside the fort is limited to a 1st century civilian tombstone, slight traces of 2nd century ironworking and a bronze-smith’s workshop, dated to the mid-2nd century with evidence of manufacture of both military and civilian products. The apparent absence of 3rd/4th century material, at a time when the fort itself was still occupied, might suggest that the town of Carlisle came to serve in place of a vicus for

155

JAMES ELLIS JONES

  Claudius to one of Gratian struck between 367 and 375, and indicate dedicatory offerings from passengers in return for a safe crossing.

the Ala Petriana, limiting the growth of any civil development across the river. Stracathro (NO618658) On the banks of the West Water, close to its confluence with the Esk, the 2.6 ha. Flavian fort is 10 km inland from the sea. The fortlet at Dun (NR689585) on the Montrose Basin probably enabled supply by sea. Just to the west of the fort, a 15.6 ha. Agricolan marching camp has given rise to the type site for “Stracathro” (claviculae and straight ditch) gateways.

Tarbock (SJ463890) The small civilian tilery, 13 km east of Liverpool, was used as a stopgap to manufacture roof-tiles for the fortress at Chester at a time (c.167) when the legionary tilery at Holt was not in use. It has been suggested that the tiles were transported from Tarbock to Chester by water, using the estuaries of the Mersey and Dee. The distance of some 7 km from Tarbock to the Mersey shore could have been traversed by using the tidal inlet of the Ditton Brook.

Stranraer (in the area of) It is unlikely that the small fortlet at Gatehouse of Fleet was the most westerly Roman military installation in Galloway and the evidence of a well-built Roman road leading to a crossing of the Water of Fleet suggests the presence of a Roman fort (and probable harbour) on Loch Ryan (Ptolemy’s Rerigonius Sinus), probably in the vicinity of Stranraer.

Thorpe (SK759504) - Ad Pontem On the banks of the River Trent, at a point where it was crossed by the Fosse Way, with Lincoln lying 35 km to the north-east, a 0.5 ha. fortlet was built and abandoned during the Flavian period. To the north-east of the fortlet, a ditched-and-ramparted 2.5 ha enclosure has been interpreted as a supply depot. In the 2nd century, a small town developed on the site of the fortlet with, in the late3rd/-early 4th century, the turf defences being replaced by stone walls enclosing an area of c. 2.1 ha.

Stretton Grandison (SO636431) A 1st century 1.8 ha. fort was located to the south of the River Frome, a tributary of the River Wye. Aerial photography has shown an oblong enclosure some 200 m to the west, indicating either a second fort or a walled settlement. Lying to the east of the military road from Gloucester where it crosses the River Frome, close to an east-west road linking Worcester and Kenchester, a possible local market and service centre is indicated by the find of a steelyard and a quantity of iron slag.

Tomen-y-Mur (SH707388) The late-1st century fort at is located on the southern flank of the Vale of Ffestiniog, 4.5 km distant from the tidal Afon Dwryd, at a height of 275 m above sea level. Clearly, any approach by road would involve the negotiation of steep gradients, but the site commands extensive views to the south, west and north, and this is the probable reason for the siting on such an exposed position. A road runs north to Bryn-y-Gefeiliau and Caerhun, and south to Brithdir and Pennal. A road from the supply base at Caer Gai joins the north/south road; leads west to Pen Llwyn and then north to Caernarfon. The size of the fort was originally 1.7 ha. but at some stage, probably Hadrianatic, the fort was reduced in size to 1.34 ha., indicating a conversion to a smaller garrison and the fort was abandoned c.140.

Stirling (NTNS800943) – Velunia Stirling lay at the lowest bridging point on the Forth, just below its confluence with the Teith and, as “The Gateway to the Highlands”, it would be surprising if an Antonine fort had not been established in the vicinity, in order to fill in the gap between Camelon and Ardoch. Whilst there has been no positive identification, the “King’s Knot”, a set of formal gardens to the west of Stirling Castle and on the line of a Roman road from Camelon (12 km to the south), has been suggested as its possible location. Until the mid-20th century, the port of Stirling was engaged in overseas trade, including tea from India and timber from the Baltic. During the Flavian period, it would have been well situated to act as a supply port for Ardoch and the southern sector of the Gask Ridge, and to perform the same function during the Antonine and Severan campaigns.

Topsham (SX 962884) As sea-going vessels were unable to reach Exeter until the development of the Ship Canal in the 18th century, it is clear that, during the Roman period, a trans-shipment point on the River Exe was required. Stray finds of preFlavian pottery at Topsham on the east bank of the estuary suggested the possibility of a fort/supply base and excavation revealed part of a double ditched military enclosure containing 1st century pottery.

Sudbrook (ST504873) A mid-1st century garrison was installed within the compact (1.2 ha.) ramparts of the Iron Age multivallate promontory fort, presumably to control a ferry crossing of the Severn Estuary, from the English side to south Wales. A road runs from Sudbrook to the main Caerwent road at Crick and probably accords with a decision to ship supplies across the Severn Estuary, rather than by the longer and circuitous route via Gloucester. Coins recovered from the foreshore at Black Rock span 300 years of Roman occupation, ranging from the Emperor

Trawscoed (SN671727) The fort at Trawscoed lies on the northern bank of the Afon Ystwyth (Estuctia Fluvis), 12 km from the sea at Aberystwyth. A sandbar now obstructs the Ystwyth estuary and an artificial breakwater protects the harbour at Aberystwyth; it is therefore not possible to determine the tidal extent of the river during Roman times. To the north-east of the fort, a 2.5 km stretch of the river appears

156

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  The vicus lying outside the fortress developed into a small settlement after the closure of the fort and later expanded into the site of the fortress, serving as a market centre for agriculture and a source of tools and equipment, with iron working becoming an important element in the economy. Several small hearths were probably used for lead reprocessing, as lead sheeting (some with nails still attached), lead splash and two lampholder wasters were found.

to have been artificially straightened and, as there is no suggestion of medieval or later navigation, it is possible that this took place in Roman times. At 3.8 ha. the fort is large enough for a regiment of 1000 infantry, or of 500 cavalry. The course of a road running south to Llanio has been established, but no traces have yet been found of a road running north to Pen-llyn, Erglodd and Pennal. However, these forts are accessible by sea and river, and it must be a possibility that the main method of communication was by water transport.

Wallsend (SZ300660) - Segedunum Constructed c. 125, the 1.7 ha Hadrianic fort at Wallsend was the easternmost terminal fort of Hadrian's Wall. The praetorium, principia, fabrica, horrea and valetudinarium were all constructed in stone from the outset, as was the bathhouse to the south-west of the fort, but the barrack-blocks were constructed in timber. Remodelling took place between 180–230, with a crosshall constructed in front of the principia, a bath-suite inserted in the praetorium and the barrack-blocks rebuilt in stone. Situated in the northern part of the fort were six standard barrack-blocks for the accommodation of infantry and, in the southern part of the fort, was a group of four specially adapted cavalry barrack-blocks, each of which provided accommodation for a troop (turmae) of cavalrymen living alongside their horses. This is now recognised as a standard form of accommodation for cavalrymen and their mounts and indicates that the Hadrianic garrison at Wallsend was a part-mounted cohort (cohors quingenaria equitata) of about 500 men. During the mid-3rd century this form of accommodation was replaced by a series of barrack-blocks similar to those at South Shields; the hospital was demolished and a number of industrial buildings and strip houses were constructed.

A substantial vicus developed to the north-east of the fort. A street extending the via principalis was 11 m wide and another parallel to it was 9.5 m in width, both being flanked by timber buildings, some with long axes parallel with the rampart, and extending to within 1 m of the ditch. At least two phases of construction were identified, with the first commencing c.75, at some time the buildings were demolished and a deposit of metalling laid over the area as far as the ditch of the fort. The second phase of occupation extended further north-east from the fort, but it appears that the site was levelled with clay c.120 and there is nothing to indicate occupation after that date. Usk (SO379007) - Burrium During the Governorship of Aulus Didius Gallus (52–7) a 20 ha. fortress was built for legio XX Valeria Victrix, on the banks of the river at Usk, on the site of a pre-Flavian fort. The site is 34 km from the Bristol Channel and 9 km north of the present tidal limit, at a point on the banks of the river where the main road from England, running to the north of the Forest of Dean, entered the Usk valley. This position controlled the road south to the coastal lowlands, and north to Abergavenny and the hills of Breconshire, and the siting of the fortress at this point was clearly for strategic reasons. However, the site was liable to flooding, and was too far inland to be directly supplied by sea. Shallow draught vessels might well have been able to navigate the Usk as far as the town in Roman times, with trans-shipment at some point nearer to the Bristol Channel. The suggestion of limited navigability to Usk is perhaps supported by the present-day existence of a ‘Barge Field’, where, within living memory, this type of craft was built.

The so-called "Branch Wall" ran for 180 m from the south-east corner of the fort, continuing across the foreshore and into the River Tyne, extending to the lowwater mark. The furthest point perhaps formed the base of a monument to mark the end of Hadrian's Wall, and a gated-tower gave access to the vicus and the riverside area to the south of the fort. The vicus extended for almost a kilometre along the Military Way to the west of the fort and, during the early-3rd century, was enclosed by a system of defensive ditches and banks extending down to and along the riverside, probably including a quay or landing place. A series of postholes containing with sharpened stakes (lilia) provided an extra line of defence in front of the Wall, but the civilian settlement was largely abandoned by the late-3rd century.

The later presence of legio II Augusta is attested by the recovery of two tiles bearing a LEG II AVG stamp. The construction of the new fortress at Caerleon (73–4), some 12.8 km further down the River Usk, made the base redundant, and led to abandonment and demolition of some buildings taking place in c.75. Although the Usk was not navigable for large vessels, reusable material from the site was probably transported down river to the new fortress. A 1.5 ha. walled area within the abandoned fortress may be a works depot and/or a centre under military control for the administration of industrial exploitation of the region, with final abandonment taking place c.120-5.

Wall Town, Neen Savage (SO693783) With ramparts on the south side still standing to a height of 3m, the 1st century fort dominates a small valley to the east of the crossing of the River Rea at Cleobury Mortimer (a tributary of the River Teme). The fort, now bisected by the B4363, appears remote from the main Roman road systems. In 1975, Webster commented that ‘one might also speculate on the need for a road south from Uriconium to communicate with the fort at Wall

157

JAMES ELLIS JONES

  Town, which at present stands in total isolation’. Despite its proximity to the River Rea, it is probable that the River Severn, possibly at a crossing near Cressage, was used for supply and provisioning by water transport.

Washing Well, Whickham (NZ219603) Washing Well farm is situated on the plateau of DunstonHill, about 5 km south of Newcastle, the location of Pons Aelius, the Roman bridge crossing the River Tyne. Aerial photography appears to indicate two different periods of construction, one fort of 1.8 ha, another is somewhat larger but neither the size of this fort, or the order in which the forts were constructed has not yet been determined. T probable Trajanic foundation, it is likely that the fort was abandoned when the fort at Wallsend was constructed.

Whilst the evidence is not conclusive, it is probable that a Flavian-Trajanic timber and turf fort, with three defensive ditches, was followed by a Hadrianic-Antonine successor with stone buildings and stone-revetted walls, the phases being separated by a build-up of cultivation soil. From pottery evidence, it is probable that, after military abandonment during the late 2nd century, occupation of the vicus continued to the 4th century. On a recent visit, two large blocks of stone were examined which are very similar in shape and size to the excavated colonnade bases at Wroxeter.

Water Newton (TQ741686) - Durobrivae Three km downstream of the vexillation base at Longthorpe, where the River Nene was crossed by Ermine Street, a 2 ha. fort was constructed around 47 and probably abandoned within 20 years. The vicus flourished and extended for 3 km along Ermine Street and on both sides of the Nene until it covered area of over 60 ha. Lying within a rich villa landscape, the area became one of the most extensive industrial complexes in Roman Britain. The late-2nd century defences enclosed an area of 18 ha. and were fronted by a stone wall, with external bastions, in the 3rd century. A mansio, forum/basilica complex, temples and shrines (the “Water Newton Hoard” is the earliest group of Christian religious silver from the whole of the Roman Empire) have been located and, outside the defended area, there are many industrial buildings involved in metal-working, leather-working and agricultural processing. From the mid-2nd to the late-4th century a highly successful pottery industry produced a wide range of colour-coated vessels (known as Nene Valley or Castor ware) traded over much of Roman Britain. It is probable that, by the 270s. Durobrivae (the fort at the bridges) had achieved civitas status and, whilst there is some evidence of later reduction in size, a hoard of gold coins dated to c.350 indicates continuing prosperity.

Walton Castle (TM322358) On the outskirts of Felixstowe, the Saxon Shore fort, probably built between 265 and 285 and standing 30 m above sea level was destroyed by erosion in the 18th century. However, an early 18th-century copy of a map dating to 1623 shows a rectangular structure, probably of similar size to Burgh Castle, with projecting bastions at each corner suggesting a late-3rd century foundation. Walton le Dale (SD551282) The site is located on the major north/south road from Carlisle to Wroxeter, 3 km south of the road from Kirkham to Ribchester, and eventually York. An extensive area of industrial settlement is located near the junction of the former course of the River Darwen with the River Ribble (Belisama fluvis). At this point the river is tidal and sea-going ships would have been able to dry out on the tidal mud flats. A late 1st century phase of development comprised a north/south road of ground gravel, associated with a number of rectilinear enclosures, together with hearth/furnace bases indicating ironsmelting activity. These structures were destroyed by fire c.120 and immediately replaced by a series of rectangular buildings, at right angles to the road, again with clear evidence for extensive iron smelting and other industrial processes. In the late 2nd/early 3rd century, these buildings were succeeded by new structures on the same general plan, again with extensive evidence for industrial processes, including clay-lined pits suggesting evidence for tanning. A pottery kiln is dated to the later 3rd century and there is limited pottery evidence for 4th century occupation. The high number of silver coins, samian pottery and amphorae suggest, particularly in the 2nd century, a military-run industrial centre.

Watercrook (NY515907) - Aluana The late-Flavian (c.90) turf and timber fort was located, on the southern outskirts of Kendal, on a meander bend of the River Kent, to the north of Morecambe Bay and presumably had a role in checking seaborne attacks from that area. There are waterfalls on the river to the south, but high tides reach to Levens Hall, so supply by sea is a possibility. The 1.6 ha. fort lay 9 km to the west of the main Roman road to the north, but there was also a minor road leading north from Lancaster, which passed a small military site in Hincaster and from there to Watercrook. After the initial foundation (c.90-100), there was a possible break in occupation c.120 followed by the construction of a stone wall to improve the rampart c.135-45. There was a further break in occupation early in the 3rd century, with the fort remaining unmanned until the late 260s or early 270s. A fresh coin of 320, found in the top of a ditch fill, suggests demilitarization at about that time. The vicus was located on both sides of the road that headed south-east out of the fort. Narrow strip houses about 7m in width were originally built of timber and were later replaced in stone.

Ward Law (NY024669) Situated on the east bank of the River Nith, near its outlet into Luce Bay, the 0.35 ha. camp has views over the Solway Firth. A known Roman road, leading in the direction of Stranraer, on Loch Ryan (Ptolemy’s Rerigonius Sinus), is to the east.

158

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  It has been suggested that the site served for the transfer of products from water to land transport. However, if the River Douglas followed a similar course to the present day, it pursued a winding route for some 30km to the Ribble estuary, with little evidence for Roman activity in that area.

Weston under Penyard (ST634235) - Ariconium Pre-Flavian pottery and a number of Claudian coins suggest the existence of a fort and, although the archaeological evidence is slight, a large military presence, perhaps a vexillation base, similar to that at Clyro is possible. The site lies at a point where the road by which supplies and men would have entered the area from across the Severn meets a major river valley. With a strong military presence, Ariconium would be strategically well-placed, effectively blocking Silurian raids into the Forest of Dean and across the lower reaches of the Severn.

Wilderspool (SJ618867) In the area now known as Wilderspool, on the route of the road from Northwich to Preston and at the lowest crossing point on the River Mersey, an area extending for some 300 m along the banks of the River Mersey became a centre for the production of pottery, mortaria, glass, leather and metal working. There were a number of large timber-framed buildings, each within its own courtyard, at least one with painted wall plaster and a hypocaust. There may have been a Roman fort located to the north of the river controlling the river crossing at Latchford, this remains to be proven, but a tile-stamp indicates a legionary presence.

The small town of Ariconium became a major centre for iron smelting and forging, nearby slag tips covered over 80 ha. (this may be compared with the largest example from the Weald, which covered 2.8 ha.). The site is situated 4 km from the River Wye; a road leads to Mitcheldean and then to Gloucester and the River Severn, so supply of military materials such as weapons, and nails for fort construction, to a wide area was possible. It has been estimated that around 80 furnaces could have been in operation at Ariconium and other nearby smelting sites, and that a labour force of 400 would have been required to operate them, in addition to the smiths operating the forges. A number of settlements, e.g. Ruardean, Drybrook, Huntley and Symonds Yat, have produced evidence of smithing, but not of forging, and it is probable that the blooms were supplied to Ariconium, for manufacture into iron products.

The settlement was occupied from the late 1st to the 4th century. Coin loss was greatest during the reign of Trajan, which has 60 examples, followed by 33 Domitianic issues and 25 Hadrianic, indicating this as the peak of industrial activity. The finds include bronze cavalry harnesses and an armour clip, and buildings prefabricated in the barrack-block style, suggest that Wilderspool was a manufacturing centre under military control. There was large scale production of pottery, and a high proportion of Wilderspool mortaria has been discovered at numerous forts along Hadrian’s Wall, with also a wide distribution across England, north Wales and southern Scotland during the 2nd century. Large numbers of lead pieces have been found indicating that lead-working was carried out, and glass-making has been indicated by the discovery of a crucible containing black glass waste and slag. There is also evidence of copper alloy working, in the form of unfinished copper objects, copper ingots, crucibles, casting moulds and slag. It is likely that the bronze processed at Wilderspool used copper ore mined on Alderley Edge, which lies c 35 km to the south–east.

An excavation 1 km north of the iron working site revealed two small fortlets, each of about 0.2 ha., one overlying the other and therefore not contemporary. Each fortlet could have accommodated a century of soldiers and the excavators concluded that they should be viewed as modest-sized administrative centres, staffed with a minimum of manpower and supervising the production, supply and distribution of artefacts and forgeable iron, to service military forts west of the Wye. Wigan (SD583056) - Coccium Wigan is now generally accepted as the site of the settlement of Coccium, which occupied land adjacent to the River Douglas. Excavations in the centre of Wigan revealed significant evidence of Roman occupation dating to the 1st and 2nd centuries. Amongst the finds were scorched clay, iron slag, charcoal and large quantities of coal, indicating extensive metalworking activity and this is underpinned by the fact that high-quality coal seams outcrop throughout the Wigan area. A large bathhouse dating to the mid-2nd century, with two hypocaust rooms has been located, together with a probable plunge pool. Close by was a workshop with a series of hearths where lead was processed. The stone walling was, in some cases, over one metre thick suggesting the building complex was at least two storeys high. It is now believed that the bathhouse complex was connected to a much larger building which had a colonnade running around it and that the complex represents a large mansio.

Winchester (SU482293) - Venta Belgarum Situated on the River Itchen 21 km north of Southampton Water, Winchester was the civitas capital of the Belgae and, by the middle of the 2nd century, had become the fifth largest city in Roman Britain. The centre of an important communication network, roads lead towards Silchester, Chichester, Old Sarum, Lake Farm and Mildenhall, covering virtually all points of the compass except to the east. Unusually, there is, as yet, no clear evidence for a military presence; unplanned civilian settlement of the valley bottom commenced c. 55. The city, based on a regular street grid pattern, was founded towards the end of the 1st century and all the usual trappings of a civitas capital, forum, basilica, etc. following in quick succession. Earthwork defences were constructed, in the late Neronian or early Flavian period but their extent has not yet been determined. Towards the end of the 2nd century, the defences were reconstructed in

159

JAMES ELLIS JONES

  Probably in the 3rd century, an iron foundry with at least six smelting hearths was established, and the large quantity of slag used for road surfacing indicates that the processing of ore, brought up river from the Forest of Dean, was a major activity in the town. It is possible that lead from mines near Felindre, over 20 miles above Leintwardine, and iron ore from mines above Ludlow was carried down to forges in the Worcester area in barges. The absence of late period coins indicates probable abandonment in the late -3rd/early -4th century.

earth and timber, enclosing an area of some 57 ha. and, in the early-3rd century, the rampart was fronted by a stone wall with external bastions added in the 4th century. During the 4th century, there is evidence for both urban and suburban development and, whilst part of the forum appears to have been abandoned, town houses continued to be remodelled and were clearly occupied into the 5th century, suggesting a still functioning Romano-British community. However, grave goods and burial rites suggestive of Germanic origin indicate the presence of foederati from the middle of the 4th century and, by the middle of the 5th century, these "barbarians" appear to have become the dominant element of the population.

Wroxeter (SJ564086) – Viroconium Wroxeter lies on the road along which the army moved its troops and supplies for the final battle against Caractacus (wherever that might be precisely located), and would have been a convenient location for a base from which to conduct operations. The earliest military presence was probably at Eaton Constantine, 6 km to the southeast, where there are marching camps and a large 4 ha. fort. A normal sized auxiliary fort is situated 1 km to the south on the bank of the Severn. A 20 ha. fortress constructed about 58 by legio XIV Hispana; pottery associated with the military levels indicates continued occupation until about 90. After the final withdrawal of the legio XIV Hispana from Britain in AD 69, the new garrison became legio XX Valeria Victrix, previously based at the Kingsholm site at Gloucester. The latest military buildings are different in character from the normal fortress layout, and it is likely that the function had changed to become a depot for administration, storage and training, while the legion was actively engaged in the northern campaigns under Agricola.

Recent excavations of wooden structures in the intertidal zone at St Denys, near to the point where the Itchen joins Southampton Water, indicate riverfront development, including shoreline revetment and a probable jetty, suggesting a trans-shipment facility for the importation of goods intended for Winchester. It is probable that Caen stone used in the construction of Winchester Abbey arrived by water and Chesil Street takes its name from the site of the gravel (chesil as in Chesil Beach) bank on which boats were grounded for loading or unloading during the medieval period. Winteringham (SE944212) - Ad Abum? Situated on the southern bank of the Humber Estuary, Winteringham was the southern terminal of the ferrycrossing to Brough. Ermine Street follows a resolutely straight line southwards for some 55 km until it reaches Lincoln and a further road leads south-easterly towards Horncastle.

With the removal of the military presence the site was handed over to the civil authorities and, unlike other redundant fortresses such as Colchester, Lincoln and Gloucester that became coloniae, was made into an administrative capital (civitas) and given a certain amount of local autonomy. The city of Viroconium Cornoviorum grew to be the 4th largest city in Britain in the 2nd century, with a defended area of 90 ha. Following the visit of Hadrian to Britain in 122, the city was much increased in size, and provided with one of the largest and most sophisticated civic centres in the country, with a large forum-basilica and extensive public baths. A wellexecuted inscription of 130 commemorates the erection of the forum. Following successive periods of growth and decline in the 2nd, 3rd and 4th centuries, a significant resurgence of activity took place after 400, at a time when the Roman administration began to withdraw from Britain, and 5th century continuity, of a kind, is now established.

Pre-Flavian ware indicates a Claudian fort, probably established during the governorship of Ostorius Scapula. A civilian settlement covered an extensive area and, in 1696, De la Pryme wrote that "The old Roman way has come straight from Lincoln thither. There are great foundations dug and ploughed up hard by this way near the Humber, which I take to have been some old beach made by the Romans to bring and secure their shipps in, because that it encompasses a great piece of land and is warp up. I also saw an old coin or two of the Roman Empire that had been found there." Worcester (SO385255) Lying on the River Severn, the town was engaged in the metal industry on a substantial scale, making use of processed ore from the Forest of Dean. The town was situated on a major road and a navigable river with easy access to the Western Seaways, and was therefore able to take advantage of an integrated transport system. The river was probably fordable; the existence of a Roman bridge over the river has yet to be proved, although it seems a reasonable proposition. The tactical importance of the river crossing suggests an early fort; coins of Augustus, Tiberius and Claudius have been found, together with two 1st-century ditches that may have been of military origin, but no certain structural evidence has been located.

The function as a civitas capital was probably in decline by the mid-4th century, as shown by the abandonment in disrepair of the baths and basilica complex in c.350. Commercial activity, however, continued long enough for wheeled ruts to have been worn into collapsed columns.. Some postholes at Wroxeter were found to contain sherds of Mediterranean amphorae of 5th century date, but the overall chronology remains uncertain. Remnant magnetic dating of the last firing of a bread oven gave a date of 160

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  between 490 and 550. Post-Roman Wroxeter has more of the appearance of a power base for an individual rather than a town. The tombstone of an Irish chieftain is known, perhaps belonging to the early years of the 6th century. It has been suggested that the control of the area was subject to religious, not secular, control and may have been the seat of an early bishopric, as a result of either an Irish importation or local opportunism.

York is located at the confluence of the River Ouse with the smaller River Foss, and lies not far above the conjectured tidal limit in Roman times, some 24 km from the Humber Estuary. The course of the rivers in Roman times is not altogether certain, the Ouse probably has not changed a great deal, but the Foss was substantially altered by the creation of the King’s Fish Pool in the 11th century, and by canalisation in more recent times. There is considerable evidence of overseas trade at York during the late-2nd and early-3rd centuries, but despite considerable efforts to establish the location of the quays, their whereabouts are not yet known.

As aerial photography had failed to indicate any sign of buildings in the eastern segment within the defensive walls, Wroxeter had variously been characterised as either a failed town or a garden city. However, recent extensive geophysical survey has shown a density of wooden buildings, extending right up to the walls. It is clear that the overall density of occupation means that the accepted picture of Viroconium as an under-populated city is incorrect. York (SE603521) - Eboracum The 20 ha. legionary fortress at York was built in 71, to house legio IX Hispana who, in about 120, were probably transferred to the Netherlands and replaced by legio VI Victrix remained in garrison until the early-5th century. Pottery attributable to the Claudian-Neronian period suggests the existence of an earlier fortlet, perhaps lying beneath the later fortress. Roads lead north-west to Catterick and the north, east to the Yorkshire coast, southeast to the Humber, southwards to Lincoln and trans-Pennine roads lead to Ribchester and Manchester. The Vale of York provides an entirely lowland route between the southern and northern areas of Britain, and the site is therefore of considerable military importance, on occasion serving as an Imperial headquarters. Septimius Severus set up the Imperial Court in York whilst campaigning in Scotland and died in the city in 211 and, in 306 Constantius I died in York and his son, Constantine (The Great) was acclaimed as Emperor by the army. York became capital of Britannia Inferior (Lower Britain) during the early- 3rd century and, in the 4th century, the city became the military base of the Dux Britanniarum. Rebuilding of the early timber-and-turf fortress commenced in the Trajanic-Hadrianic period with the construction of stone towers and gates-houses. In the late2nd or early-3rd century, work began on a stone-built wall with a new ditch and rampart but there appears to have been the hiatus in construction, lasting for perhaps 50 to 100 years, with the defensive circuit not being completely rebuilt in stone until the mid-3rd/late-4th century. The range of buildings within the fortress, praetorium, principia, barracks, bath-house, etc. appear to have been completed before the end of the 2nd second century. The methods of provision for the massive quantities of stone needed for the construction of the walls and buildings are dealt with in some detail in Chapter 10.

161

JAMES ELLIS JONES

  Table by location

From river

Via

Origin

1st Mil

Abertanant Alcester Aldborough Ambleside Annanfoot Axminster Badbury Barochan Hill Barry Bath Bawtry Baylham House Beckfoot Bertha Berwick-on-Tweed? Binchester Bishopston Bitterne Blackwardine Blennerhasset Bowness-on-Solway Bradwell on Sea Brancaster Brecon Gaer Brithdir Brough Brough-on-Humber Bryn-y-Gefeiliau Burgh by Sands Burgh Castle Burrow Walls Caer Gybi Caerhun Caerleon Caernarfon Caerphilly Caersws Caerwent Caister on Sea Caistor St Edmund Calstock Cambridge Camelon Canterbury Cardean Cardiff Cargill Carlisle Carmarthen Carnforth Carpow Carriden Carzield Castle Hill Castleford Catterick Charterhouse Chelmsford Chepstow Chester Chester-le-Street Chichester Clifford Clyro Colchester Corbridge Cowbridge Cramond Dalswinton

Severn Alne Ure Rothay Annan Axe Stour Coastal Coastal Avon Idle Gipping Coastal Almond Coastal Wear Coastal Itchen Lugg Ellen Coastal Coastal Coastal Usk Wnion Trent Estuary Llugwy Coastal Yare Coastal Coastal Conwy Usk Coastal Rhymey Severn Nedern Coastal Tas Tamar Cam Carron Stour Dean Coastal Isla Eden Tywi Coastal Tay Forth Nith Trent Aire Swale Avon/Axe? Chelmer Wye Dee Wear Coastal Wye Wye Colne Tyne Thaw Forth Nith

Severn Avon Humber Windermere Annan Seaton Solent Coast Coast Avon Idle Orwell Coast Tay Coast Wear Coast Solent Lugg Coast Coast Coast Coast Usk Coast Trent Humber Conwy Coast Yare Coast Coast Conwy Usk Coast Taff ? Severn Coast Coast Yare Tamar Cam Forth Wantsum Tay Coast Tay Eden Tywi Coast Tay Forth Nith Trent Aire Humber Avon Blackwater Severn Dee Wear Coast Severn Severn Colne Tyne Thaw Forth Nith

Ft Ft Ft Ft Camp Ft Ft Ft

X X X X X X X

2nd Mil

X X

3rd Mil

X X

4th Mil

X X

Civil

Def S Civ Set

Isurium Brigantum Glannoventa

Set Set

Moridunum Vindocladia

X X

X

X

? Def S Set Set Set

X X

X

X

Set

Vinovia

Def S Set

Clausentum

Def S

Maia Othana Branodunum Cicucium

Ft Ftlt Ft Ft Ft Ft Ft Ft Ft

X X X X X ? X

Ft Ft Ft Ft Ft Ftlt

X X

X

X X

X X

X X

X

X X X X

Ft Ft Ft Ft Ft Ft Ft Ftrs Ft Ft Ft

X X X

X X X

X

X

X X X X X X X

X X X X

Def S

X

Set Set

X

X

X

X

Set Civ Def S Civ Set Def S

X

X X X X X X

X X X X X

Ft Ft Ft ? Ft Ft Ft Ft Ft Ft Ft

X X X X X X X X X

Ft Ft Ft Ft Ft Ft Ft Ft Ft Ftrs Ft Ft Ft Ft Ft Ft

X

X X X X X

Ft Ft

X

X

Set Set Def S Civ

Aquae Sulis Combretovium Bibra

Crococalana Petuaria Aballava Garianum Axelodunum Canovium Isca Segontium

Venta Silurum Venta Icenorum Duroliponte

X Civ X ? X X

X ? X

X X X X

X

X

X

X X

X X

X

162

Lugovalium/Carvetiorum Moridunum

Velunia

X X

Duro.Cantiacorum Tameia

Set Civ Civ Set

X X X X X

Roman Name

X

X X

Def S Set Def S Set Civ

Margidunum Lagentium Cateractonium Vebaracum ? Caesaromagus

Set Set Def S

Deva Concangis Noviomagus Regnorum

Col Def S Set

Camulodunum Coria Bovium

Cst

Grid Ref

W W E W W S S W W W E E W E E E W S W W W E E W W E E W W E W W W W W W W W E E S E E E E W E W W W E E W E E E W E W W E S W W E E W E W

SJ2420 SP0857 SE4066 NY3703 NY1765 SY2997 ST9603 NS4169 ST0966 ST7565 SK6693 TM1152 NY0948 NO0926 NT9956 NZ2131 NS4172 SU4313 SO5356 NY1941 NY2262 TM0308 TF7844 SO0029 SH7718 SK8358 SE9326 SH7457 NY3259 TG4704 NY0030 SH2482 SH7770 ST3390 SH4862 SO1688 SO0492 ST4790 TG5212 TG2303 SX4369 TL4459 NS8681 TR3141 NO2846 SO1876 NO1637 NY3956 SN4120 SD4971 NO2017 NT0280 NX9681 SK7041 SE4225 SE2299 ST5055 TL7006 ST5394 SJ4066 NZ2751 SU8604 SO2446 SO2243 TL9925 NY9864 SS9974 NT1977 NX9384

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Doncaster Dorchester Dorchester on Ths Dover Droitwich Drumburgh Ebchester Elginhaugh Erglodd Exeter Filey Fingrinhoe Wick Fleetwood Forden Gaer Gatehouse of Fleet Girvan Mains Glenlochar Glenluce Gloucester Godmanchester Goldsborough Hardknott Healam Bridge Heronbridge Holt Horncastle Huntcliff Ilchester Inchtuthil Inveresk Irvine Bay ? Ixworth Kelvedon Kenchester Kentchurch Kingsholm Kirkbride Kirkham Lake Farm Lancaster Lantonside Learchild Lease Rigg Leicester Leintwardine Lincoln Littleborough Llandeilo Llandovery Llansantfraid-y-M..n Loch Ryan London Long Sandall Longthorpe Lostwithiel Loughor Lurg Moor Lydney Lympne Manchester Martinhoe Maryport Meols Middlewich Monmouth Moresby Nantstallon Neath Netherby Newcastle Newton Kyme Newton on Trent Northwich

Don Frome Thames Coastal Salwarpe Coastal Derwent Esk Scot Dyfi Exe Coastal Roman Coastal Severn Fleet Coastal Dee Scot Luce Severn Westwater Coastal Esk Swale Dee Dee Bain Coastal Yeo Tay Forth Coastal Lark Blackwater Wye Monow Severn Wampool Ribble Stour Lune Nith Aln Esk Soar Teme Witham Trent Tywi Tywi Vymwy Coastal Thames Don Nene Fowey Llwchwr Coastal Severn Coastal Irwell Coastal Coastal Coastal Dane Wye Coastal Camel Neath Esk Tyne Wharfe Trent Weaver

Don Frome Thames Coast Salwarpe Coast Tyne Esk Dyfi Exe Coast Colne Coast Severn Fleet Coast Dee Luce Severn Westwater Coast Esk Ouse Dee Dee Bain Coast Yeo Tay Forth Coast Great Ouse Blackwater Wye Wye Severn Wampool Ribble Stour Lune Nith Aln Esk Soar Teme Witham Trent Twyi Twyi Severn Coast Thames Humber Nene Fowey Llwchwr Coast Severn Coast Irwell Coast Coast Coast Dane Wye Coast Camel Neath Esk Tyne Wharfe Trent Weaver

Ft Ft Ft Ft Ft Ft Ft Ft Ftlt Ftrs Ftlt Ft ? Ft Ftlt Camp Ft Camp Ftrs Ft Ftlt Ft Ft

X X X

X Civ Def S

X X X X X X X

X

X Def S

X X

X X

X

X

X

X

X

Danum Durnovaria Portus Dubris Salinae Congavata Vindomora

X Civ

Isca Dumnoniorum

Set

Lavobrinta Lucopibia ?

Col Def S

Glevum Durovigutum

X X X X X X X X

X

X

X X

X Mediobogdum

X

Ftlt Ftlt Ft Ftrs Ft ? Ft Ft

X

X X

Set Set Set Def S

Bannovalum

Def S

Lindinis

X

X X X X X X

Ft Ftrs Ft Ft Ftrs Ft Ftlt Ft Ft

X X X X X X

Vindogara Sinus Set Def S Def S

X X

Set

X

X

X X

X X X X

Ft Ftrs

X

X

X

X

Ft Ft Ft ? Ft Ft Ftrs Ft Ft Ftlt

X X X X

X X

X X X X

Canonium Magnis

Set

Civ Set Civ Set

Ratae Branogenium Lindum Segelocum

Set

X

X

X

X

X X

X X

X

X X

X X

X

X

X

Mun?

Regionius Sinus Londinium

Leucarum

X Set

Ft Ftlt Ft Ft Ft Ft Ft Ft Ft Ft Ft Ftrs Ft

X X X X X X X X X X X X

X X X

X X

X

163

X

X

X X

X

Set

Portus Lemanis Mamcumium Alauna

Set Set Set Set

Salinae Blestium Gabrosentum

Set Set Set Set

Nidum Castra Exploratum Portus Aelius Praesidium ?

Set

Condate

E S E E W W E E W S E E W W W W W W W E E W E W W E E W E E W E E W W W W W S W W E E E W E E W W W W E E E S W W W S W W W W W W W W W W E E E W

SE5703 SY6990 SU5794 TR3141 SO9063 NY2659 NZ1055 NT3267 SN6590 SX9192 TA1281 TM0419 SD4833 SO2098 NX5957 NX1999 NX7364 NX1956 ST8318 TL2470 NZ8315 NY2101 SE3384 SJ4163 SJ4054 TF2569 NZ8622 ST5222 NO1239 NT3472 NS3820 TL9369 TL8618 SO4442 SO4223 ST8319 NY2357 SD4332 SY9899 SD4761 NY0166 NU1011 NZ8104 SK5804 SO4074 SK9771 SK8283 SN6122 SN7635 SJ2220 ??????? TQ3281 SE6067 TL1597 SX1061 SS5697 NS2973 ST6102 TR1134 SJ8397 SS6649 NY0337 ST6102 SJ7065 ST5012 NX9821 SX0367 SS7497 NY3971 NZ2563 SE4545 SK8237 SJ6573

JAMES ELLIS JONES

  Old Barrow Old Kilpatrick Osmanthorpe Outerwards Papcastle Pen Lwyn Pennal Pentrehyling Pen-y-Gaer Pevensey Piercebridge Plymouth Portchester Prestatyn Ravenglass Ravenscar Reculver Ribchester Richborough Roall Rochester Roecliffe Rossington Ruthwell Scarborough Sea Mills Shadwell Shapwick South Shields Springhead Sracathro St Albans St Madoes Staines Stanwix Stracathro Stranraer ? Stretton Grandison Sudbrook Tarbock Thorpe Tomen-y-Mur Topsham Trawscoed Usk Wall Town Wallsend Walton Castle Walton le Dale Ward Law Water Newton Watercrook Weston-u-Penyard Wigan Wilderspool Winchester Winteringham Worcester Wroxeter York

Coastal Clyde Greet Coastal Derwent Rhiedol Dyfi Severn Rhiangoll Coastal Tees Coastal Coastal Coastal Coastal Coastal Coastal Ribble Coastal Aire Medway Ure Torne Coastal Coastal Avon Thames Stour Tyne Ebbsfleet Esk Scot Ver Tay Thames Eden N. Esk Coastal Frome Coastal Ditton Trent Dwrd Exe Ystwth Usk Severn Tyne Coastal Ribble Nith Nene Kent Wye Douglas Mersey Itchen Humber Severn Severn Ouse

Coast Clyde Trent Coast Derwent Rhiedol Dyfi Severn Rhiangoll Coast Tees Coast Coast Coast Coast Coast Coast Ribble Coast Aire Medway Ure Don Coast Coast Avon Thames Stour Tyne Thames Esk Thames Tay Thames Eden Esk Coast ? Wye Coast Mersey Trent Dwrd Exe Ystwth Usk Severn Tyne Coast Ribble Nith Nene Kent Wye Douglas ? Mersey Solent Humber Severn Severn Ouse

Ftlt Ft Ftrs Ftlt Ft Ft Ft Ft Ft Ft Ft ? Ft

X X X X X X X X

X X X X X X

X

X

X

X

Set

Derventio

Set Set

X X X

X X

X

X

Set Set

Anderitum Morbium Portus Ardaoni

Set Ft Ftlt Ft Ft Ft Ft Ft Ft Ft Camp Ftlt Ftlt

X

X X X X X

X X X X X X X

X

X

X

Ft Ft

X X

X

X

Ft Ft Ftlt Ft? Ft Ft ? Ft Ftlt

X X X X

X X X

X

X X X X X

X X X X X

Set

Regulbium B.Veteranorum Rutupae

Def S

Durobrivae

Set Set

Portus Abonae

Set Set

Arbeia Vagniacis

Mun

Verulamium

Set Def S

Pontibus Uxelodunum

X

X

X X

X X X X

Itunocelum

Set

Ft Ft Ft Ft Ftrs Ft Ft Ft

X X X X X X

X

X

Camp Ft Ft Ftlt

X X X X

X X

X

X

Ft

X

X

X

X

Ftrs Ftrs

X X

X X

X X

X X

Set Def S

Ad Pontem

Set X X X

X X

X X

Set Set Set

Burrium Segedunum

Set

164

Def S Set Set Set Set Civ Set Set Civ Col

Dubrivae Avalana Ariconium Coccium Venta Belgarum

Vir. Cornovium Eburacum

W W E W W W W W W S E S S W W E E W E E E E E W E W E S E E E E E E W E W W W W E W S W W W E E W W E W W W W S E W W E

SS7849 NS4673 SK6756 NS2366 NY1031 SN6580 SH7000 SO2593 SO1621 TQ6404 NZ2115 SX4853 SU6204 SJ0681 SD0895 NZ9801 TR2269 SD6434 TR3260 SE5625 TQ7468 SE3866 SK6399 NY1067 TA0589 ST5575 TQ3581 ST9402 NZ3667 TQ6172 N06165 TL1307 NO2019 TQ0371 NY4057 NO6165 ????? SO6343 ST5087 SJ4689 SK7550 SH7038 SX9688 SN6772 SO3700 SO6978 NZ3066 TM3235 SD5528 NY0266 TL1297 NY5190 ST6423 SD5805 SJ6186 SU4829 SE9421 SO8555 SJ56408 SE6052

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Table by coast

From river

Via

Aldborough Bawtry Baylham House Bertha Berwick-on-Tweed? Binchester Bradwell on Sea Brancaster Brough Brough-on-Humber Burgh Castle Caister on Sea Caistor St Edmund Cambridge Camelon Canterbury Cardean Cargill Carpow Carriden Castle Hill Castleford Catterick Chelmsford Chester-le-Street Colchester Corbridge Cramond Doncaster Dorchester on Ths Dover Ebchester Elginhaugh Filey Fingrinhoe Wick Godmanchester Goldsborough Healam Bridge Horncastle Huntcliff Inchtuthil Inveresk Ixworth Kelvedon Learchild Lease Rigg Leicester Lincoln Littleborough London Long Sandall Longthorpe Newcastle Newton Kyme Newton on Trent Osmanthorpe Piercebridge Ravenscar Reculver Richborough Roall Rochester Roecliffe Rossington Scarborough Shadwell South Shields Springhead Sracathro

Ure Idle Gipping Almond Coastal Wear Coastal Coastal Trent Estuary Yare Coastal Tas Cam Carron Stour Dean Isla Tay Forth Trent Aire Swale Chelmer Wear Colne Tyne Forth Don Thames Coastal Derwent Esk Scot Coastal Roman Westwater Coastal Swale Bain Coastal Tay Forth Lark Blackwater Aln Esk Soar Witham Trent Thames Don Nene Tyne Wharfe Trent Greet Tees Coastal Coastal Coastal Aire Medway Ure Torne Coastal Thames Tyne Ebbsfleet Esk Scot

Humber Idle Orwell Tay Coast Wear Coast Coast Trent Humber Yare Coast Yare Cam Forth Wantsum Tay Tay Tay Forth Trent Aire Humber Blackwater Wear Colne Tyne Forth Don Thames Coast Tyne Esk Coast Colne Westwater Coast Ouse Bain Coast Tay Forth Great Ouse Blackwater Aln Esk Soar Witham Trent Thames Humber Nene Tyne Wharfe Trent Trent Tees Coast Coast Coast Aire Medway Ure Don Coast Thames Tyne Thames Esk

Origin

1st Mil

2nd Mil

3rd Mil

4th Mil

Civil

Roman Name

Cst

Grid Ref

Ft Ftlt Ft Ft Ft Ft Ft Ft

X X X X ? X

X

X

X

Civ Set Set

Isurium Brigantum

X

X X

X X X

Set

Ft Ft Ft

X

X

X X X

X X X

Vinovia Othana Branodunum Crococalana Petuaria Garianum

Ft ? Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ftlt Ft Ft Ftlt Ft

X X X X X X

E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E E

SE4066 SK6693 TM1152 NO0926 NT9956 NZ2131 TM0308 TF7844 SK8358 SE9326 TG4704 TG5212 TG2303 TL4459 NS8681 TR3141 NO2846 NO1637 NO2017 NT0280 SK7041 SE4225 SE2299 TL7006 NZ2751 TL9925 NY9864 NT1977 SE5703 SU5794 TR3141 NZ1055 NT3267 TA1281 TM0419 TL2470 NZ8315 SE3384 TF2569 NZ8622 NO1239 NT3472 TL9369 TL8618 NU1011 NZ8104 SK5804 SK9771 SK8283 TQ3281 SE6067 TL1597 NZ2563 SE4545 SK8237 SK6756 NZ2115 NZ9801 TR2269 TR3260 SE5625 TQ7468 SE3866 SK6399 TA0589 TQ3581 NZ3667 TQ6172 N06165

Ftlt Ftrs Ft Ft Ft Ft Ft

X X

? X X X

? X

X

X

X

X

X X X

X X

X

Venta Icenorum Duroliponte

Def S Set Def S Civ Set Col Def S

Duro.Cantiacorum Tameia

Velunia Margidunum Lagentium Cateractonium Caesaromagus Concangis Camulodunum Coria Danum

Def S X X

X X

X X

X

X

X

X

X X X X

Portus Dubris Vindomora

Def S

Durovigutum

Set Def S

Bannovalum

Set Def S

Canonium

Civ Civ Set Mun?

Ratae Lindum Segelocum Londinium

Set Set

Portus Aelius Praesidium ?

Set

Morbium

X

X X X X X X

X X

Ftrs Ft Ft Ftrs Ft Ft Ftrs Ftrs Ft Ftlt Ft Ft Ft Ft Ft Ft Ftlt

Def S Civ Def S Civ

X

X X

Def S Civ

X

X

X X

Combretovium

X

X X X X X X X X

X

Ft

X

X

X

X

X X X X X

Ft

X

X

165

X X X X

X X X X

X

X

X

X

Regulbium Rutupae Def S

Durobrivae

Set Set Set

Arbeia Vagniacis

JAMES ELLIS JONES

  St Albans St Madoes Staines Stracathro Thorpe Wallsend Walton Castle Water Newton Winteringham York

Ver Tay Thames N. Esk Trent Tyne Coastal Nene Humber Ouse

Thames Tay Thames Esk Trent Tyne Coast Nene Humber Ouse

Ft Ftlt Ft? Ft Ft Ft Ft Ft Ft Ftrs

X X X

Axminster Badbury Bitterne Calstock Chichester Dorchester Exeter Lake Farm Lostwithiel Lympne Pevensey Plymouth Portchester Shapwick Topsham Winchester

Axe Stour Itchen Tamar Coastal Frome Exe Stour Fowey Coastal Coastal Coastal Coastal Stour Exe Itchen

Seaton Solent Solent Tamar Coast Frome Exe Stour Fowey Coast Coast Coast Coast Stour Exe Solent

Ft Ft Ft Ft Ft Ft Ftrs Ftrs Ft

X X X X X X X X X

Abertanant Alcester Ambleside Annanfoot Barochan Hill Barry Bath Beckfoot Bishopston Blackwardine Blennerhasset Bowness-on-Sol Brecon Gaer Brithdir Bryn-y-Gefeiliau Burgh by Sands Burrow Walls Caer Gybi Caerhun Caerleon Caernarfon Caerphilly Caersws Caerwent Cardiff Carlisle Carmarthen Carnforth Carzield Charterhouse Chepstow Chester Clifford Clyro Cowbridge Dalswinton Droitwich Drumburgh Erglodd Fleetwood Forden Gaer Gatehouse of Fleet Girvan Mains

Severn Alne Rothay Annan Coastal Coastal Avon Coastal Coastal Lugg Ellen Coastal Usk Wnion Llugwy Coastal Coastal Coastal Conwy Usk Coastal Rhymey Severn Nedern Coastal Eden Tywi Coastal Nith Avon/Axe? Wye Dee Wye Wye Thaw Nith Salwarpe Coastal Dyfi Coastal Severn Fleet Coastal

Severn Avon Windermere Annan Coast Coast Avon Coast Coast Lugg Coast Coast Usk Coast Conwy Coast Coast Coast Conwy Usk Coast Taff ? Severn Coast Coast Eden Tywi Coast Nith Avon Severn Dee Severn Severn Thaw Nith Salwarpe Coast Dyfi Coast Severn Fleet Coast

Ft ? Ft Ft Ft

Ft Ft Ft Camp Ft

X X X X

Mun

Verulamium

X

Set

Pontibus

X X X

X X X

Def S Set

Ad Pontem Segedunum

X X

Def S Set Col

Dubrivae

X X

Set Set Def S Set Def S Civ Civ

Moridunum Vindocladia Clausentum

X X

X X

X X

X

X

X

X X X

X X X

X

X

Eburacum

Noviomagus Reg. Durnovaria Isca Dumnoniorum

Portus Lemanis Anderitum Set

X X

Portus Ardaoni Set Civ

Venta Belgarum

X X X X

X

Ft Ft Ft

X X

Ft Ft Ft Ftlt Ft Ft Ft Ft Ft Ftrs Ft Ft Ft

X X X X X X

X X X X X

X X X X X

Ft Ft Ft

X X X

X X X

Ft Ft Ft Ftrs Ft Ft

X X X X X

X X

X

X

X

Def S Set

Glannoventa

X

? Def S Set

Aquae Sulis Bibra

Set

Ft Ft Ft Ftlt ? Ft Ftlt Camp

X X X X X X

X X

X X

Def S Set Set

Maia Cicucium

X X X X X X

X X X

Def S

Aballava Axelodunum

X

Set Set

X

X

X X

X

Set Civ Set Civ Civ Set

Canovium Isca Segontium

Venta Silurum Lugo./Carvetiorum Moridunum

X

X X X X X X X

X

X

X X

X

X

X

166

X

X

Set

Vebaracum ?

Set

Deva

Set

Bovium

Def S

Salinae Congavata

Set

Lavobrinta Lucopibia ?

E E E E E E E E E E

TL1307 NO2019 TQ0371 NO6165 SK7550 NZ3066 TM3235 TL1297 SE9421 SE6052

S S S S S S S S S S S S S S S S

SY2997 ST9603 SU4313 SX4369 SU8604 SY6990 SX9192 SY9899 SX1061 TR1134 TQ6404 SX4853 SU6204 ST9402 SX9688 SU4829

W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W

SJ2420 SP0857 NY3703 NY1765 NS4169 ST0966 ST7565 NY0948 NS4172 SO5356 NY1941 NY2262 SO0029 SH7718 SH7457 NY3259 NY0030 SH2482 SH7770 ST3390 SH4862 SO1688 SO0492 ST4790 SO1876 NY3956 SN4120 SD4971 NX9681 ST5055 ST5394 SJ4066 SO2446 SO2243 SS9974 NX9384 SO9063 NY2659 SN6590 SD4833 SO2098 NX5957 NX1999

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Glenlochar Glenluce Gloucester Hardknott Heronbridge Holt Ilchester Irvine Bay ? Kenchester Kentchurch Kingsholm Kirkbride Kirkham Lancaster Lantonside Leintwardine Llandeilo Llandovery Llansantfraid Loch Ryan Loughor Lurg Moor Lydney Manchester Martinhoe Maryport Meols Middlewich Monmouth Moresby Nantstallon Neath Netherby Northwich Old Barrow Old Kilpatrick Outerwards Papcastle Pen Lwyn Pennal Pentrehyling Pen-y-Gaer Prestatyn Ravenglass Ribchester Ruthwell Sea Mills Stanwix Stranraer ? Stretton Grandison Sudbrook Tarbock Tomen-y-Mur Trawscoed Usk Wall Town Walton le Dale Ward Law Watercrook Weston-u-Penyard Wigan Wilderspool Worcester Wroxeter

Dee Scot Luce Severn Esk Dee Dee Yeo Coastal Wye Monow Severn Wampool Ribble Lune Nith Teme Tywi Tywi Vymwy Coastal Llwchwr Coastal Severn Irwell Coastal Coastal Coastal Dane Wye Coastal Camel Neath Esk Weaver Coastal Clyde Coastal Derwent Rhiedol Dyfi Severn Rhiangoll Coastal Coastal Ribble Coastal Avon Eden Coastal Frome Coastal Ditton Dwrd Ystwth Usk Severn Ribble Nith Kent Wye Douglas Mersey Severn Severn

Dee Luce Severn Esk Dee Dee Yeo Coast Wye Wye Severn Wampool Ribble Lune Nith Teme Twyi Twyi Severn Coast Llwchwr Coast Severn Irwell Coast Coast Coast Dane Wye Coast Camel Neath Esk Weaver Coast Clyde Coast Derwent Rhiedol Dyfi Severn Rhiangoll Coast Coast Ribble Coast Avon Eden Coast ? Wye Coast Mersey Dwrd Ystwth Usk Severn Ribble Nith Kent Wye Douglas ? Mersey Severn Severn

Ft Camp Ftrs Ft Ftlt Ft ?

X X X

X Col X Set Set Def S

X X X

Def S Ft Ftrs Ft Ft Ft Ftlt Ft Ft Ft Ft ? Ft Ftlt

X X X X X

Ft Ftlt Ft

X X X

Ft Ft Ft Ft Ft Ft Ft Ftlt Ft Ftlt Ft Ft Ft Ft Ft

X X X X X X

X X X X X X X

Glevum Mediobogdum

Lindinis Vindogara Sinus Magnis

Set X

X

Set

X

X

Set

Branogenium

Set

X

X

X

X

X

X

X

X

Regionius Sinus Leucarum

X

X X X X X X X

X X X X X

X X X

X X X X X X X X X

Set Set

Mamcumium Alauna

Set Set Set Set

Salinae Blestium Gabrosentum

X

X

X X

X

Set Set Set

Nidum Castra Exploratum Condate

X

X

Set

Derventio

Set Set

X

Set Ft Ft Camp Ftlt Ft ? Ft Ftlt

X X X X X X

X X

X X

X X

Set

Itunocelum B.Veteranorum

X X

X

X

Set Def S

Portus Abonae Uxelodunum

X X

Set Set

Ft Ft Ftrs Ft

X X X X

X X

Camp Ft Ftlt

X X X

X X

X

X

Ftrs

X

X

X

X

X

167

Set Set Set

Burrium

Set Set Set Set Set Civ

Avalana Ariconium Coccium

Vir. Cornovium

W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W W

NX7364 NX1956 ST8318 NY2101 SJ4163 SJ4054 ST5222 NS3820 SO4442 SO4223 ST8319 NY2357 SD4332 SD4761 NY0166 SO4074 SN6122 SN7635 SJ2220 ??????? SS5697 NS2973 ST6102 SJ8397 SS6649 NY0337 ST6102 SJ7065 ST5012 NX9821 SX0367 SS7497 NY3971 SJ6573 SS7849 NS4673 NS2366 NY1031 SN6580 SH7000 SO2593 SO1621 SJ0681 SD0895 SD6434 NY1067 ST5575 NY4057 ????? SO6343 ST5087 SJ4689 SH7038 SN6772 SO3700 SO6978 SD5528 NY0266 NY5190 ST6423 SD5805 SJ6186 SO8555 SJ56408

JAMES ELLIS JONES

  Table by river

From river

Via

Origin

1st Mil

Castleford Roall Bertha Learchild Alcester Annanfoot Bath Sea Mills Charterhouse Axminster Horncastle Kelvedon Cambridge Nantstallon Camelon Chelmsford Old Kilpatrick Barochan Hill Barry Beckfoot Berwick-on-Tweed? Bishopston Bowness-on-Solway Bradwell on Sea Brancaster Burgh by Sands Burrow Walls Caer Gybi Caernarfon Caister on Sea Cardiff Carnforth Chichester Dover Drumburgh Filey Fleetwood Girvan Mains Goldsborough Huntcliff Irvine Bay ? Loch Ryan Lurg Moor Lympne Martinhoe Maryport Meols Moresby Old Barrow Outerwards Pevensey Plymouth Portchester Prestatyn Ravenglass Ravenscar Reculver Richborough Ruthwell Scarborough Stranraer ? Sudbrook Walton Castle Colchester Caerhun Middlewich Cardean Chester Heronbridge

Aire Aire Almond Aln Alne Annan Avon Avon Avon/Axe? Axe Bain Blackwater Cam Camel Carron Chelmer Clyde Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Coastal Colne Conwy Dane Dean Dee Dee

Aire Aire Tay Aln Avon Annan Avon Avon Avon Seaton Bain Blackwater Cam Camel Forth Blackwater Clyde Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast Coast ? Coast Coast Colne Conwy Dane Tay Dee Dee

Ft Ft Ft Ft Ft Camp Ft Ftlt Ft Ft

X X X X

X X X X X

Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft

X ?

Ftlt Ft Ft Ftlt Ftlt Ft ? Ft

3rd Mil

4th Mil

Civil

Roman Name

Cst

Grid Ref

Set

Lagentium

E E E E W W W W W S E E E W E E W W W W E W W E E W W W W E W W S E W E W W E E W W W S W W W W W W S S S W W E E E W E W W E E W W E W W

SE4225 SE5625 NO0926 NU1011 SP0857 NY1765 ST7565 ST5575 ST5055 SY2997 TF2569 TL8618 TL4459 SX0367 NS8681 TL7006 NS4673 NS4169 ST0966 NY0948 NT9956 NS4172 NY2262 TM0308 TF7844 NY3259 NY0030 SH2482 SH4862 TG5212 SO1876 SD4971 SU8604 TR3141 NY2659 TA1281 SD4833 NX1999 NZ8315 NZ8622 NS3820 ??????? NS2973 TR1134 SS6649 NY0337 ST6102 NX9821 SS7849 NS2366 TQ6404 SX4853 SU6204 SJ0681 SD0895 NZ9801 TR2269 TR3260 NY1067 TA0589 ????? ST5087 TM3235 TL9925 SH7770 SJ7065 NO2846 SJ4066 SJ4163

X X Def S

X X X X X

Ft Ft ? Ft Ft Ft Ft Ft

Ft Ft Ft Ftlt ? Camp Ftlt Ftlt ? ? Ftlt

2nd Mil

X

Def S Set Set Set Def S Def S Def S

Aquae Sulis Portus Abonae Vebaracum ? Moridunum Bannovalum Canonium Duroliponte

Civ

Caesaromagus

? Set

Bibra

X X

X

X

X X X

X

X X

X

X

X

X

X

X

X X

X X X X X X

X X X X X X

X X

X X

X X

X X X

X

X X

X X

Def S

Def S

Set Def S Set Set Def S

Maia Othana Branodunum Aballava Axelodunum Segontium

Noviomagus Regnorum Portus Dubris Congavata

X

X

X X

Vindogara Sinus Regionius Sinus X

X X

X X

X

X

X

X

X

X

X

X

Portus Lemanis Alauna Set Set

Gabrosentum

X X Anderitum Set X

X

Portus Ardaoni Set

Ft Ftlt Ft Ft Camp Ftlt ? Ftlt Ft Ft Ft Ft Ft Ftrs

X

X X X X X X X X X

X X X X

X X X X

X

X

X

X

Itunocelum Regulbium Rutupae

X Col X X

X

X

X

168

Set X

Set Set

Camulodunum Canovium Salinae Tameia Deva

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Holt Glenlochar Ebchester Papcastle Tarbock Doncaster Long Sandall Wigan Tomen-y-Mur Erglodd Pennal Springhead Carlisle Stanwix Blennerhasset Hardknott Lease Rigg Netherby Elginhaugh Sracathro Brough-on-Humber Exeter Topsham Gatehouse of Fleet Carriden Cramond Inveresk Lostwithiel Dorchester Stretton Grandison Baylham House Osmanthorpe Winteringham Bawtry Manchester Cargill Bitterne Winchester Watercrook Ixworth Bryn-y-Gefeiliau Loughor Glenluce Blackwardine Lancaster Rochester Wilderspool Kentchurch Stracathro Neath Caerwent Longthorpe Water Newton Carzield Dalswinton Lantonside Ward Law York Pen-y-Gaer Pen Lwyn Caerphilly Kirkham Ribchester Walton le Dale Fingrinhoe Wick Ambleside Droitwich Abertanant Caersws Forden Gaer Gloucester Kingsholm Lydney

Dee Dee Scot Derwent Derwent Ditton Don Don Douglas Dwrd Dyfi Dyfi Ebbsfleet Eden Eden Ellen Esk Esk Esk Esk Scot Esk Scot Estuary Exe Exe Fleet Forth Forth Forth Fowey Frome Frome Gipping Greet Humber Idle Irwell Isla Itchen Itchen Kent Lark Llugwy Llwchwr Luce Lugg Lune Medway Mersey Monow N. Esk Neath Nedern Nene Nene Nith Nith Nith Nith Ouse Rhiangoll Rhiedol Rhymey Ribble Ribble Ribble Roman Rothay Salwarpe Severn Severn Severn Severn Severn Severn

Dee Dee Tyne Derwent Mersey Don Humber Douglas ? Dwrd Dyfi Dyfi Thames Eden Eden Coast Esk Esk Esk Esk Esk Humber Exe Exe Fleet Forth Forth Forth Fowey Frome Wye Orwell Trent Humber Idle Irwell Tay Solent Solent Kent Great Ouse Conwy Llwchwr Luce Lugg Lune Medway Mersey Wye Esk Neath Coast Nene Nene Nith Nith Nith Nith Ouse Rhiangoll Rhiedol Taff ? Ribble Ribble Ribble Colne Windermere Salwarpe Severn Severn Severn Severn Severn Severn

Ftlt Ft Ft Ft

X X X

X X X X

Ft Ft

Set X X

X X

X

X

Ft Ftlt Ft

X X X

X X X

Ft Ft Ft Ft Ft Ft Ft Ft Ft Ftrs Ft Ftlt Ft Ft Ft Ft Ft Ft Ft Ftrs Ft Ftlt Ft Ft Ft

X

X X

X X X X X X X X X X

Ft Ft Ft Ft Camp

X X X X X

X

Ft Ft

X X

X

Ft Ft Ft

X X X

Ftrs Ft Ft Ft Ftlt Camp Ftrs Ft Ft Ft Ft Ft

X X

X X X X X X X

Ft Ft Ft Ft Ft Ft Ftrs Ftrs

X X X X X X X X

X X

X X

Set Set

Vindomora Derventio Danum

X

Set

Coccium

Set Set Civ Def S

Vagniacis Lugovalium/Carvetiorum Uxelodunum

X X X X X X X X X X

X X

X

X

X

X X X X

Mediobogdum

X

Set

Castra Exploratum

Civ Civ Set

Petuaria Isca Dumnoniorum Lucopibia ? Velunia

X X

X Civ Set Set

X

X

X

X

X ?

X ?

X

X

X

X

X

X

X

Set Set Set Def S Civ Set Set

Durnovaria Combretovium

Mamcumium Clausentum Venta Belgarum Avalana

X

X

X

Leucarum Set Set Def S Set

Durobrivae

Set Civ

Nidum Venta Silurum

Def S

Dubrivae

X X X X X X X X X

X

X

Col

Eburacum

X

X

Set Set

B.Veteranorum

X

X

X

Set Def S

Glannoventa Salinae

X X

X X

X X

Set Set Col

Lavobrinta Glevum

Set

169

W W E W W E E W W W W E W W W W E W E E E S S W E E E S S W E E E E W E S S W E W W W W W E W W E W W E E W W W W E W W W W W W E W W W W W W W W

SJ4054 NX7364 NZ1055 NY1031 SJ4689 SE5703 SE6067 SD5805 SH7038 SN6590 SH7000 TQ6172 NY3956 NY4057 NY1941 NY2101 NZ8104 NY3971 NT3267 N06165 SE9326 SX9192 SX9688 NX5957 NT0280 NT1977 NT3472 SX1061 SY6990 SO6343 TM1152 SK6756 SE9421 SK6693 SJ8397 NO1637 SU4313 SU4829 NY5190 TL9369 SH7457 SS5697 NX1956 SO5356 SD4761 TQ7468 SJ6186 SO4223 NO6165 SS7497 ST4790 TL1597 TL1297 NX9681 NX9384 NY0166 NY0266 SE6052 SO1621 SN6580 SO1688 SD4332 SD6434 SD5528 TM0419 NY3703 SO9063 SJ2420 SO0492 SO2098 ST8318 ST8319 ST6102

JAMES ELLIS JONES

  Pentrehyling Wall Town Worcester Wroxeter Leicester Canterbury Badbury Lake Farm Shapwick Catterick Healam Bridge Calstock Caistor St Edmund Carpow Inchtuthil St Madoes Piercebridge Leintwardine Dorchester on Ths London Shadwell Staines Cowbridge Rossington Brough Castle Hill Littleborough Newton on Trent Thorpe Corbridge Newcastle South Shields Wallsend Carmarthen Llandeilo Llandovery Aldborough Roecliffe Brecon Gaer Caerleon Usk St Albans Llansantfraid Kirkbride Binchester Chester-le-Street Northwich Godmanchester Newton Kyme Lincoln Brithdir Chepstow Clifford Clyro Kenchester Monmouth Weston-u-Penyard Burgh Castle Ilchester Trawscoed

Severn Severn Severn Severn Soar Stour Stour Stour Stour Swale Swale Tamar Tas Tay Tay Tay Tees Teme Thames Thames Thames Thames Thaw Torne Trent Trent Trent Trent Trent Tyne Tyne Tyne Tyne Tywi Tywi Tywi Ure Ure Usk Usk Usk Ver Vymwy Wampool Wear Wear Weaver Westwater Wharfe Witham Wnion Wye Wye Wye Wye Wye Wye Yare Yeo Ystwth

Severn Severn Severn Severn Soar Wantsum h Solent Stour Stour Humber Ouse Tamar Yare Tay Tay Tay Tees Teme Thames Thames Thames Thames Thaw Don Trent Trent Trent Trent Trent Tyne Tyne Tyne Tyne Tywi Twyi Twyi Humber Ure Usk Usk Usk Thames Severn Wampool Wear Wear Weaver Westwater Wharfe Witham Coast Severn Severn Severn Wye Wye Wye Yare Yeo Ystwth

Ft Ft

X X

X X

X

Ftrs

X

X

X

Ft Ft Ftrs Ft Ft Ft Ft

X X X X X

X

Ft Ftrs Ftlt Ft Ft Ft Ft

X X

X

X

X X X

X X

X X X

X X

X

X

X

X

X

X

X X

Ft?

X

Ft

X

Ft

X

X

X

Set Set Set Civ Civ Civ Set

Def S Set Set Civ

Cateractonium

Set Set Def S Mun? Set Set Set

Morbium Branogenium

Def S Def S Set

Crococalana Margidunum Segelocum

Def S Def S Set Set Set Civ

Ad Pontem Coria Portus Aelius Arbeia Segedunum Moridunum

Set Civ

Isurium Brigantum

Set Set Set Mun

Cicucium Isca Burrium Verulamium

Vinovia Concangis Condate Durovigutum Praesidium ? Lindum

Ftrs Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ft Ftrs Ftrs Ft Ft Ft Ft Ft Ft Ft Ft Ftrs Ftlt Ft Ft Ft

X X X X X

X X X X

X

Set Set Set Set Def S Set Civ Set

Ft Ftlt Ft Ft Ft

X X

X X

Def S Set Set

X X X X X X X X X X X X X X X

X X X X X X X X X

X

X

X X

X X

X

X

X X

X X

X X

X X X X

X X

X X

X X X

X Def S

X

170

Vir. Cornovium Ratae Duro.Cantiacorum Vindocladia

Venta Icenorum

Londinium Pontibus Bovium

Magnis Blestium Ariconium Garianum Lindinis

W W W W E E S S S E E S E E E E E W E E E E W E E E E E E E E E E W W W E E W W W E W W E E W E E E W W W W W W W E W W

SO2593 SO6978 SO8555 SJ56408 SK5804 TR3141 ST9603 SY9899 ST9402 SE2299 SE3384 SX4369 TG2303 NO2017 NO1239 NO2019 NZ2115 SO4074 SU5794 TQ3281 TQ3581 TQ0371 SS9974 SK6399 SK8358 SK7041 SK8283 SK8237 SK7550 NY9864 NZ2563 NZ3667 NZ3066 SN4120 SN6122 SN7635 SE4066 SE3866 SO0029 ST3390 SO3700 TL1307 SJ2220 NY2357 NZ2131 NZ2751 SJ6573 TL2470 SE4545 SK9771 SH7718 ST5394 SO2446 SO2243 SO4442 ST5012 ST6423 TG4704 ST5222 SN6772

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Appendix 2 Admiralty Tidal Stream Atlas – Irish Sea and Bristol Channel  

          

 

 

171

JAMES ELLIS JONES

   

           

 

172

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Appendix 3 Stowage factors

Stowage Factors (cubic meters per tonne)

 

Tin ingots Lead pigs Lead ingots Iron pigs Iron ore Lead ore Marble blocks Granite slabs Marble slabs Sand Ivory Iron scrap Water Oak Wheat Wheat in bags Barley Hides, baled and dry Elm Barley in bags Loose stone Pine Ivory in cases Wine in casks Wine in cases Olive oil in casks Earthenware in crates Tiles Peat or turf, dry Tiles in crates Hides, dry Hay in bales Porcelain in crates/ cases/baskets/tubs Wool in bales

0.22 0.22 0.28 0.28 0.33 .36 0.42 0.45 0.50 0.53 0.84 0.98 1.00 1.00 1.18 1.34 1.36 1.39 1.39 1.45 1.50 1.52 1.53 1.62 1.67 1.67 1.70 2.13 2.30 2.41 2.79 3.34 3.34 5.55

to

0.28 0.31 0.33 0.33 0.42 0.39 0.47 0.50 0.56 0.56 0.98 0.98 1.00 1.31 1.34 1.50 1.50 1.67 1.78 1.67 1.78 1.84 1.67 1.78 1.95 1.73 2.13 2.27 2.30 2.41 4.18 4.46 5.57 6.00

Figures are for bulk stowage, except where noted. . These stowage factors apply to the relatively large and obstruction-free holds of modern vessels: factors for vessels in antiquity would probably have been at the higher end of the given ranges. (S. McGrail - Sources: Lewis 1962; Thomas 1983)

173

JAMES ELLIS JONES

  Appendix 4 Relative usage of Antonine Itinerary Roads  

 

      

174

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Bibliography Abdy, R 2002. A survey of coin finds from the Antonine Wall. Britannia 33:, 189–217 Aberg, A. & Lewis, C. (eds.) 2000. The Rising Tide. Archaeology and Coastal Landscapes. Oxbow Books. Oxford. Adams, C. & Laurence R. (eds.). 2001. Travel and geography in the Roman Empire. Routledge. London & New York. Adams, J. 2001. Ships and boats as archaeological source material. World Archaeology 32.3: 292-310. Adams, J.N. 1977. Roman military pay from Caesar to Diocletian. Journal of Roman Studies, 84: London. 113–23. Adamson, H. 1774. The Muses Threnodie, Cant, J. (ed.). Perth. Adamson, H.C. 1979. The Roman fort at Bertha, in Breeze, D.J. (ed.). 33–5. Adamson, |H.C. & Gallagher, D.B. 1986. The Roman fort at Bertha, the 1973 excavation. Proceedings of the Society of Antiquaries of Scotland, 116: 195–204. Alarcao, J de. 1988. Roman Portugal. Aris & Phillips. Warminster. Alcock, J.P. 2001. Food in Roman Britain. Tempus. Brimscombe Port. Aldcroft, D.H. & Freeman, M.J. (eds.) 1983. Transport in the Industrial Revolution. Manchester University Press. Aldhouse-Green, M. & Webster, P. (eds.). 2002. Artefacts and archaeology : aspects of the Celtic and Roman world. University of Wales. Cardiff. Alföldy, G., Dobson B. & Eck. W. 2000. Kaiser, Heer und Gesellshaft in der Römischen Kaiserzeit: Gedenschrift für Eric Birley. Franz Steiner. Zurich. Allan, T.M. 1994. The Roman Route Across the Northern Lake District: Brougham to Moresby. Centre for Regional Studies, University of Lancaster. Allason-Jones, L. 2001. Material culture and identity, in James, S. & Millett, M. (eds.). 19–25. Allen, J.H. 1991. Excavations at Brompton Fort. Britannia, 22: 246–50. Allen, J.R.L. 1991. Saltmarsh accretion and sea level movement. Journal of the Geological Society, London, 148: 485–94. Allen, J.R.L. 1996b. The sea bank on the Wentlooge Level, Gwent: date of setback from documentary and pottery evidence. Archaeology in the Severn Estuary, 7: 67–84. Allen, J.R.L. 1997b. Romano-British and early Medieval pottery scatters on the alluvium at Hill and Oldbury, Severn Estuary Levels. Archaeology in the Severn Estuary, 8: 67–81. Allen, J.R.L. 1999. Magor Pill Multi-period Site: The Romano-British Pottery, and Status as a Port. Archaeology in the Severn Estuary, 9: 75–97. Allen, J.R.L. 2000. Two probable wooden jetties at Cone Pill, Lydney Level, Gloucestershire in Rippon (ed.). 132–4. Allen, J.R.L. 2005. Roman and mediaeval-early modern building stones in South East Wales: the Sudbrook sandstone and the limited conglomerate (Triassic). The Monmouthshire Antiquity, 21: 21-44. Allen, J.R.L. 2002. Intertidal archaeology at Old Passage, Aust, Gloucestershire: a sketch of the surviving evidence. Archaeology in the Severn Estuary 13: 53–64 Allen, J.R.L. & Fulford, M.G. 1990b. Romano-British wetland reclamation at Longney, Gloucestershire, and evidence for the early settlement of the inner Severn Estuary. Antiquaries Journal, 70/2: 288–326. Allen, J.R.L. & Fulford, M.G. 1996. The distribution of South-East Dorset Black Burnished Category 1 pottery in Britain. Britannia, 27: 223–81. Allen, J.R.L. & Fulford, M.G. 1999. Fort building and military supply along Britain’s Eastern Channel and North Sea Coasts: the later second and third centuries. Britannia, 30: 163–84. Allen, J.R.L. 1990. The Severn Estuary in South West Britain: its retreat under marine transgression and fine sediment regime. Sedimentary Geology, 66: 13–28. Allen, J.R.L., Bradley, R.J. et al., 1997. The archaeological resource: chronological overview, in Fulford, M. et al. (eds.). 103–53. Allen, M.J. & Gardiner, J. 2000. Our Changing Coast: A Survey of the Intertidal Archaeology of Langstone Harbour, Hampshire, CBA Research Report 124. York. Allison, P., Fairbairn, A.S., Ellis, S.J.R. & Blackall. C.W 2005. 'Extracting the social relevance of artefact distribution in Roman military forts. Internet Archaeology 17. Anderson A.C. & Anderson, A.S. 1981. Roman Pottery Research in Britain and Europe, British Archaeological Reports International Series. 123. Anderson, J.D. 1992, Roman Military Supply in North-east England. An alternative to the Piercebridge Formula. British Archaeology Report 224. Oxford. Anderson, J.K. 1970. Military theory and practice in the age of Xenophon. University of California Press. Berkeley. Applebaum, S. 1967. Roman Frontier Studies. Tel Aviv University. Tel Aviv. Applebaum, S. 1976. Observations on the economy of the villa at Bignor. Britannia, 6: 118–132.

175

JAMES ELLIS JONES

  ApSimon, A.M. 1965. The Roman Temple on Brean Down, Somerset. Proceedings of the University of Bristol Speleological Society, 10: 195–258. Arenson, S. 1990. The encircled sea. Constable. London. Arnold, C.J & Davies, J.L. 2000. Roman and Early Medieval Wales. Sutton . Stroud. Ashby, T. et al. 1910–13. Excavations at Caerwent, Monmouthshire, on the site of the Romano-British City of Venta Silurum. Archaeologia,Vols. 57–64. London. Aston, M. & Burrows, I. (eds.) 1982. The Archaeology of Somerset. Somerset County Council. Taunton. Aston, M. & Iles, R. (eds.) 1987. The Archaeology of Avon. Avon County Council. Bristol. Atkinson, M. & Preston, S.J. 1998. The late Iron Age and Roman settlement at Elms Farm, Heybridge, Essex, excavations 1993-5: an interim report. Britannia 29: 85–110. Ayers, B. 2003. Norwich, A Fine City. Tempus. Stroud. Austen, P.S. 1991. Bewcastle and Old Penrith: A Roman Outpost Fort and a Frontier Vicus. Cumberland and Westmorland Antiquarian and Archaeological Society Research Series, No 6. Kendal. Bagshaw, R.W. 1979. Roman Roads. Shire. Aylesbury. Bagwell, P.S. 1974. The transport revolution from 1770. Batsford. London. Bailey, G.B. 1994. The provision of fort-annexes on the Antonine Wall. Proceedings of the Society of Antiquaries of Scotland, 124: 299-314 Bailey, G.B. 1996. Stream crossings on the Antonine Wall, Proc. Soc. Antiq. Scot. 126: 347–69. Bailey, G.B., Dickson, C., Miller, J. & Webster, P.V. 2000. Excavations on the Roman temporary camps at the Three Bridges, Camelon, Falkirk. Proceedings of the Society of Antiquaries of Scotland, 130: 299–314 Barker, P. 1983. The Montgomery Area. Archaeological Journal, 138: 7–10. Barrowman, R. C. 2007. Excavations at Tintagel Castle, Cornwall, 1990-1999. Society of Antiquaries, 74: London. Bass, G.F. & van Doornink, F.H. 1982. Yassi Ada. College Station. Texas University. Bass, G.F. (ed.) 1972. A History of Seafaring based on Underwater Archaeology. Thames & Hudson. London. Bateman, N. & Milne, G. 1983. A Roman harbour in London: excavations and observations near Pudding Lane, City of London 1979-82. Britannia, 14: 207–26. Baum, J. & Robinson, D.J. 2002. Deva Victrix Restored: The Application of Computer 3-D Modelling Techniques in the Reconstruction of Roman Chester, in Carrington (ed.). 113–7. Beagrie, N. 1989. The Romano-British pewter industry. Britannia 20: 169–191. Becker, T. 2006. Women in Roman forts - lack of knowledge or a social claim? Archaeological Dialogues 13: 36–38. Bédoyère, G. de la. 1991. The Buildings of Roman Britain. Batsford. London. Bédoyère, G. de la. 1999. Companion to Roman Britain. Tempus. Brimscombe Port. Bédoyère, G. de la. 2003. Eagles over Britain: The Roman Army in Britain. Tempus. Stroud. Bédoyère, G. de la. 2006. Roman Britain: A New History. Thames & Hudson. London. Bédoyère, G. de la. 1989. The Finds of Roman Britain. Batsford. London. Belfiglio, V.J. 2001. A Study of Ancient Roman Amphibious and Offensive Sea-Ground Task Force Operations. Edwin Mellen. Lewiston. Bell, M. 1990. Brean Down. Excavations 1983–1987. English Heritage. London. Bell, M. 1994. Field Survey and excavation at Goldcliff 1994. Archaeology in the Severn Estuary, 1994: 115–44. Bellhouse, R. & Richardson, G.G.S. 1975. A Roman site at Kirkbride. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 75: 58–90. Bellhouse, R.L. 1955. The Roman fort at Burrow Walls near Workington. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 55: 30–45. Bellhouse, R.L. 1988. Roman Sites on the Cumberland Coast: Hadrian’s Wall. The fort at Bowness-on-Solway, a reappraisal. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 88: 33–54. Bellhouse, R.L. and Richardson G.G.S. 1975. The Roman Site at Kirkbride, Cumberland. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 75: 58–90. Bennett, J. 1982. Legionary Speed of March before the Battle with Boudicca. Britannia, 17: 359–62. Bennett, J. 1985. Sea Mills: The Roman Town of Abonae. City of Bristol Museum Monograph No. 3: Bristol. Bewley, R. & Macleod, D. The discovery of a Roman fort at Roall Manor Farm, North Yorkshire. Britannia, 24: 243–47 Bick, D. & Boon, G. 1993. Dolaucothi again. Britannia, 24: 247–9. Bidwell, P.T. 1979. The Legionary Bath-House and Basilica and Forum. Exeter Archaeological Reports, Vol. 1. Exeter.

176

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Bidwell, P.T. 1993. The Roman Fort of Arbeia at South Shields. Tyne and Wear Museums. Newcastle. Bidwell, P.T. 1997. Roman Forts in Britain. English Heritage. London. Bidwell, P.T. (ed.). 1999. Hadrian’s Wall 1989–1999. Newcastle-upon-Tyne Bidwell, P.T. 2001. A Probable Roman Shipwreck on the Herd Sand at South Shields, The Arbeia Journal 6-7: 1-21 Bidwell, P.T. (ed.) 2008a. Understanding Hadrian’s Wall. South Shields. Bidwell, P.T. 2009. The earliest occurrences of baths at auxiliary forts, in Hanson, W. 55–62. Bidwell, P.T. and Hodgson, N. 2009 The Roman army in Northern England. Newcastle upon Tyne. Bidwell, P.T. & Holbrook, N.M. 1989. Hadrian’s Wall Bridges. English Heritage. London. Bidwell, P.T., Snape, M. & Croom, A. 1999. Hardknott Roman Fort, Cumbria. . Kendal. Bidwell, P. T., and Speak, S. C. 1994. Excavations at South Shields Roman Fort, Vol. I. Society of Antiquaries of Newcastle upon Tyne. Newcastle upon Tyne. Biggins, J.A. & Taylor, D.J.A. 2004. The Roman fort and vicus at Maryport: geophysical survey, 2000–2004, in Wilson, R.J.A. and Caruana, I. (eds.) 2004. 102–34 Bill, J. & Clausen, B. 1999. Maritime Topography and the Medieval Town. National Museum of Denmark. Birley, A. 1997. Hadrian: the restless Emperor. Routledge. London. Birley, A. 2002. Garrison life at Vindolanda: a band of brothers. Tempus. Brimscombe Port. Birley, A.R. 1981b. The Fasti of Roman Britain. Oxford. Birley, A.R. 1997. Supplying the Batavians at Vindolanda, in Waateringe et al. (eds.). 273–80. Birley, A.R. 2000. Senators as Generals, in Alföldy, G. Dobson, B. & Eck. W. (eds.). 97–120. Birley, E. 1961. Research on Hadrian’s Wall. Kendal Press. Kendal. Birley, E. 1958. The Roman fort at Ravenglass. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 58: 14–30. Birley, E. 1958. The Roman fort at Brough-under-Stainmore. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 58: 31–56. Birley, E. 1982. Veterans of the Roman Army in Britain and elsewhere. Ancient Society 13: London. 265–76. Birley, E., Dobson, B. & Jarrett, M. 1974. Roman Frontier Studies 1969. Proceedings of the VIIIth International Conference. Cardiff. Birley, E.B. 1948. The Roman fort at Moresby. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 48: 42–72. Birley, E.B. 1953. Roman Britain and the Roman Army. Titus Watson. Kendal. Birley, E.B. 1954. The Roman fort at Netherby. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 54: 6–39. Birley, E.B. 1957. The Roman fort at Watercrook. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 57: 13–17. Birley, E.B. 1959. The Roman fort at Ravenglass. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 59: 31–56. Birley, E.B. 1963. Roman Papcastle. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 63: 96–125. Birley, R. 1963. The Roman Legionary Fortress at Carpow, Perthshire. The Scottish Historical Review, 42: 126–134. Birley, R. 1973. Civilians on the Roman Frontier. Graham. Newcastle-upon-Tyne Birley, R. 1977. Vindolanda: a Roman frontier post on Hadrian’s Wall. Thames & Hudson. London. Bishop, M.C. & Coulston, J.C.N. 1993. Roman military equipment from the Punic Wars to the fall of Rome. Batsford. London. Bishop, M.C. (ed.) 1985. The Production and Distribution of Military Equipment. British Archaeological Report 275. Oxford. Bishop, M.C. 1985. The military fabrica and the production of arms in the early Principate, in Bishop, M.C. (ed.), 1–42. Bishop, M.C. 1999. Praesidium: social, military, and logistical aspects of the Roman army’s provincial distribution during the early Principate, in Goldsworthy & Haynes (eds.). 111–18. Bishop, M. C. (ed.) 2002. Roman Inveresk: Past, Present and Future. Armatura. Duns. Bishop, M.C., Allen, D. et al. 2005. A New Flavian Military Site at Roecliffe, North Yorkshire. Britannia, 36: 135–223 Black, E.W. 1984. The Antonine Itinerary: aspects of government in Roman Britain. Oxford Journal of Archaeology, 3.3: 109–21. Black, E.W. 1995. Cursus Publicus. The infrastructure of government in Roman Britain. British Archaeology Report International Series 241. Oxford. Black, E.W. 1998. How many rivers to cross? Britannia, 31: 306–7.

177

JAMES ELLIS JONES Blackman, D.J. (ed.). 1973a. Maritime Archaeology. Butterworth. London. Blagg, T. & Millett, M. (eds.) 1990. The early Roman Empire in the West. Oxbow. Oxford. Blagg, T.F.C. 1976. Tools and Techniques of the Roman Storeman. Britannia, 7: 152. Blagg, T.F.C. & King, A.C. (eds.). 1984. Military and Civilian in Roman Britain: Cultural relationships in a frontier province. British Archaeology Report 136. Oxford. Blagg, T.F.C. 1990. Building Stone in Roman Britain, in Parsons, D. (ed.). 33–50. Blair, J. (ed.) 2007. Waterways and Canal-Building in Mediaeval England. Oxford University Press. Oxford. Blockley K. 1988. Forden Gaer. Britannia, 19: 417. Blockley, K. 1989a. Excavations on the Romano-British settlement at Ffrith, Clywd, 1967–9. Flintshire Historical Society Journal, 32: 135–66. Blockley, K. 1989b. Prestatyn, 1984–5. An Iron Age Farmstead and Romano-British Industrial settlement in North Wales. Oxford. Blood, K. & Pearson, T. 2004. The Roman fort at Ravenglass: a survey by RCHME. in Wilson, R.J.A. & Caruana, I. (eds.), 95-101. Blue, L., Hocker, F. & Englert, A. (eds.). 2006. Connected by the Sea. Proceedings of the Tenth International Symposium on Boat and Ship Archaeology. Roskilde 2003 Oxbow. Oxford. Blue, L.K. 2006a. Sewn boat timbers from the Medieval Islamic port of Quseir al-Qadim on the Red Sea coast of Egypt, in Blue, L.K., Hocker, F.& Englert, A. (eds.). Bond, J. 2007. Canal Construction in the Early Middle Ages: An introductory Review, in Blair, J. 153–204 Boe G. de 1978. Roman boats from a small river at Pommeroeul, Belgium, in Taylor, J. du P & Cleere, H. (eds.). 22-30. Bonnal, General H. 1907. Sadowa. Translated from the French by C.F. Atkinson. Hugh Rees. London. Bond, C.J. 2000 The reconstruction of the Medieval landscape: the estates of Abingdon Abbey. Landscape History, 1: 59–75. Boon G.C. 1945 The Roman Site at Sea Mills 1945–46. Transactions of the Bristol and Gloucester Archaeological Society, 66: 258–295. Boon, G.C. 1949. A Claudian Origin for Sea Mills. Transactions of the Bristol and Gloucester Archaeological Society, 68: 184–188. Boon, G.C. 1978. Excavations on the site of a Roman Quay at Caerleon, and its significance, in Boon, G.C. (ed.). 24–36. Boon, G.C. 1980. Caerleon and the Gwent Levels in early historic times, in Thomson, F.H. (ed.). 24–36. Boon, G.C. 1987b. A decorated Arretine sherd from Steepholm, Bristol Channel. Antiquaries Journal, 67: 375–6. Boon, G.C. & Brewer, R.J. 1982. Two Central Gaulish Bottles from Pennal, Mer. and Early Roman Movements in Cardigan Bay. Bulletin Board of Celtic Studies, 29: 363–7. Boon, G.C. (ed.) 1978. Monographs & Collections: Volume 1, Roman Sites. Cambrian Archaeological Association. Cardiff. Boon, G.C. 1969. Caernarfon, in Jarrett, M.G. (ed.). 59–64. Boon, G.C. 1977a. A Greco-Roman anchor stock from North Wales. Antiquaries Journal, 57: 10–30. Boon, G.C. 1977b. The Porth Felen anchor-stock. International Journal of Nautical Archaeology 6.3: 239–42. Boon, G.C. 1987. The Legionary Fortress of Caerleon – Isca. National Museum of Wales. Cardiff. Booth, P., Dodd, A. et al. 2007. The archaeology of the gravel terraces of the upper and middle Thames. Oxford. Bourne, William, d. 1583. A regiment for the sea and other writings on navigation, edited by E.G.R. Taylor. Hakluyt Society. Cambridge. Bowles, C.R. 2007. Rebuilding the Britons: The postcolonial archaeology of culture and identity in the late antique Bristol Channel region. BAR British Series 452. BAR Publishing. Oxford Bowen, E.G. 1970. Britain and the Western Seaways, in Moore, D. (ed.). 13–28 Bowman, A.K. 1994. (2003 rev.) Life and letters on the Roman frontier: Vindolanda and its people. British Museum Press. London. Bowman, A.K. & Thomas, J.D. 1984. Vindolanda: the Latin writing tablets. Sutton. Gloucester. Bowman, A.K. & Thomas, J.D. 1996. New writing tablets from Vindolanda. Britannia, 27: 299–328. Boyes, G. 2006. Local Government and Inland Navigation. The Railway and Canal Historical Society. 15–20 Bradley, R. 1997. Rock Art and the Prehistory of Atlantic Europe: Signing the Land. Routledge. London. Branigan, K. & Fowler, P.J. 1976. The Roman West Country. David & Charles. Exeter. Branigan, K. 1977. Gatcombe: the excavation and study of a Romano-British villa estate, 1967–1976. British Archaeological Report 44. BAR Publishing. Oxford. Braudel, F. 1949. The Mediterranean and the Mediterranean world in the age of Phillip II. University of California Press. California.

178

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Breeze, D.J. 1974 The Roman Fortlet at Bamburgh Mill, Dumfrieshire. Britannia, 5: 130–162. Breeze, D.J. (ed.). 1979. Roman Scotland: Some Recent Excavations. Edinburgh. Breeze, D.J. 1982a. The Northern Frontiers of Roman Britain. Batsford. London. Breeze, D.J. 1984. Demand and Supply on the Northern Frontier, in Miket, R. & Burgess, C. (eds.) 264–86 Breeze, D.J. 1987. The logistics of Agricola’s final campaign. Talanta, 18/19: 7–22. Breeze, D.J. 1988. Why did the Romans fail to conquer Scotland? Proceedings of the Society of Antiquaries of Scotland 118: 3–22. Breeze, D.J. 1988. The Roman Army in Cumbria. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 88: 9–22. Breeze, D.J. 1997. Regiments and frontiers: patterns of distribution on rivers and artificial frontiers, in Waateringe et al. 61–6. Breeze, D.J. 2000. Supplying the army, in Alfoldy, G., Dobson, B., & Eck, W. (eds.). 59–64. Breeze, D.J. 2002. Overview, in Bishop, M.C. (ed.) Breeze, D.J. 2004b. Roman military sites in the Cumbrian coast, in Wilson, R.J.A. & Caruana, I. (eds.), 66–94. Breeze, D.J. 2006. Roman Scotland: Frontier Country. Batsford. London. Breeze, D.J. 2006b. J. Collingwood Bruce’s Handbook to the Roman Wall. Society of Antiquaries of Newcastle upon Tyne. Newcastle. Breeze, D.J. 2008. Edge of Empire. Rome's Scottish Frontier. The Antonine Wall. Birlinn. Edinburgh. Breeze, D.J. 2009. Introduction/Conclusions, in Breeze, D.J. et al. 1–8; 59–62. Breeze, D.J. & Dobson, B. 1969. The development of the Mural Frontier in Britain from Hadrian to Caracalla. Proceedings of the Society of Antiquaries of Scotland. Edinburgh. 109–121. Breeze, D.J. & Dobson, B. 1985. Roman military deployment in North England. Britannia, 16: 1ff. Breeze, D.J. & Dobson, B. 2000. Hadrian’s Wall (fourth revised edn.). Penguin. Harmondsworth. Breeze, D. J. & Maxfield, V.A. (eds.). 1989. Service in the Roman Army. Edinburgh University Press. Breeze, D.J., Thoms, L.M. and Hall, D.W. (eds.) 2009. First Contact: Rome and Northern Britain. TAFAC Monograph Series No 7, Perth Brewer, R.J. 1991. Caerwent – Venta Silurum: A Civitas Capital, in Burnham, B.C. & Davies, J.L. 75–85. Brewer, R.J. 1993. Caerwent Roman Town. CADW. Cardiff. Brewer, R.J. (ed.) 2000. Roman Fortresses and their Legions. Society of Antiquaries. National Museums and Galleries of Wales. Cardiff. Briden, C. 1997. York as a Tidal Port. Yorkshire Archaeological Journal, 69: 165–71. Briggs, S.C. 1988. The location and recognition of metal ore in pre-Roman and Roman Britain and their contemporary exploitation, in Ellis Jones, J. (ed.). 106–114. Brigham, T. 1990a. The later Roman waterfront in London, Britannia, 21: 140–1. Brigham, T. & Watson, B. 1998. Regis House: The Romans erect their port. Current Archaeology, 158: 44–7. Brigham, T. with Woodger, A. 2001. Roman and Medieval townhouses on the London waterfront: excavations at Governor’s house, City of London. MuLAS Monograph 9. London. Brill, R. & Wampler, J. 1967. The Isotope Studies of Ancient Lead. American Journal of Archaeology, 71: 63–77. Brindley, A.L. & Lanting, J.N. 1991. A boat of the Mediterranean tradition in Ireland: preliminary note. International Journal of Nautical Archaeology, 20.1: 69–70. Britnell, J. 1989. Caersws Vicus, Powys. Excavations at the Old Primary School, 1985–6. British Archaeological Report 205. Oxford. Britnell, J. 1991. Settlement and industry in north-east Wales, in Burnham, B.C. & Davies J.L. (eds.) 130–7. Brodersen, K. 2001. The presentation of geographical knowledge for travel and transport in the Roman world; itineraria non tantum adnota sed etiam picta, in Adams, C. & Laurence, R. (eds.). 7–21. Brown, A.G. 1997. Alluvial Geoarchaeology; Floodplain archaeology and environmental change. Cambridge University Press. Cambridge. Brunning, R. & Farr-Cox, F. 2005. The River Siger rediscovered: LIDAR survey and relic landscape on the Somerset Claylands. Archaeology in the Severn Estuary 16: 7–15 Bruce, J. Collingwood. 1867. The Roman Wall: A description of the mural barrier of Northern England. Longman. London. Buckland, P.C. & Sadler, J. 1990. Ballast and Building Stones: A Discussion, in Parsons, D. (ed.). 114–125. Bulleid, A. & Gray, H. St. G. 1911. The Glastonbury Lake Village, Antiquarian Society, 11–17: Glastonbury. Burford, A. 1960. Heavy Transport in Classical Antiquity. Economic History Review, 31/1: 1–18. Burnham, B. & Burnham, H. 1991. Pumpsaint, Dyfed; the excavations of 1989, in Burnham, B.C. & Davies J.L. 49–56. Burnham, B. & Burnham, H. 2004. Dolaucothi-Pumpsaint: Survey and Excavations at a Roman Gold-Mining Complex 1987–1999. Oxbow. Oxford.

179

JAMES ELLIS JONES

  Burnham, B.C. 1979. Pre-Roman and Romano-British Urbanism? Problems and Possibilities, in Burnham, B.C. & Johnson, H.B. (eds.). 255–272. Burnham, B.C. & Davies, J.L. 1990. Conquest, Coexistence and Change. Trivium 25: Lampeter. Burnham, B.C. & Davies, J.L. (eds.). 2010. Roman Frontiers in Wales and The Marches. Comisiwn Brenhinol Henebion Cymru. Aberystwyth. Burnham, B.C. & Johnson, H.B. (eds.) 1979. Invasion and Response; the case of Roman Britain. British Archaeological Report 73. Oxford. Burnham, B.C. & Wacher, J. 1990. The Small Towns of Roman Britain. Batsford. London. Burnham, C.P. 1989. The coast of south-east England in Roman times, in Maxfield, V. (ed.). 12–17. Burrow, I. 1981. Hillfort and hill-top settlement in Somerset in the First to Eighth Centuries A.D. British Archaeological Report 91. Oxford. Burton, A. 1982. The Changing River. Gollanz. London. Burwash, D. 1947. English Merchant Shipping. 1460-1540. University of Toronto Press. Bushe-Fox, J.P. 1949. Fourth Report on the Excavations of the Roman Fort at Richborough, Kent. London University Press. London. Callender, J.G. 1919. Notes on the Roman Remains at Grassy Walls and Bertha. Proceedings of the Society of Antiquaries of Scotland, 53: 137–52. Callender, M. 1965. Roman Amphorae. London University Press. London. Camden, W. 1637. Britannia. (Translated by P. Holland). London University Press. London. Campbell, D.B. 2003. The Roman Siege of Burnswark. Britannia 34: 19–33. Campbell, E. 2007. Continental and Mediterranean imports to Atlantic Britain and Ireland, AD 400-800. CBA Research Report 157. Council for British Archaeology. York. Campbell, E. 1996. The Archaeological Evidence for External Contacts: Imports, Trade and Economy in Celtic Britain, AD 400–800, in Dark, K. (ed.), 83–96. Carney, T.F. 1975. The Shape of the Past: Models and Antiquity. Lawrence. Kansas. Carr, G. et al. (eds.). 2003. Theoretical Roman Archaeology Conference (12th Canterbury) Oxbow. Oxford. Carreras Monfort, C. 1994. A macro-economical and spatial analysis of long-distance exchange: the amphora evidence from Roman Britain. Unpublished Ph.D. thesis. Carreras Monfort, C. 1998. Britannia and the imports of Baetican and Lusitanian amphorae. Journal of Iberian Archaeology. Carreté, J-M., Keay, S. & Millett, M. 1995. A Roman Provincial Capital and its Hinterland: the survey of the territory of Tarragona, Spain, 1985–1990: Journal of Roman Archaeology Supplementary series 15. Ann Arbor, Michigan. Carrington, P. 1985. The earliest evidence for lead mining in Flintshire. Cheshire Archaeological Bulletin, 10: 102–5. Carrington, P. 1994. Chester. English Heritage. London. Carrington, P. 1996 (ed.). Where Deva spreads her wizard stream. Chester Archaeology. Chester. Carrington, P. 2002. The Exploration of Roman Chester 1962–1999: Retrospect and Prospect in Carrington, P. (ed.). 7–24. Carrington, P. 2002. Deva Victrix: Roman Chester Re-assessed. Chester Archaeological Society. Chester. Carrington, P. 2008. Fortress, canabae and hinterland. Unpublished working paper , accessed 21-11-08. Carrington, P. 2008. Feeding the wolf in Cheshire: models and (a few) facts, in Stallibrass S. & Wilson, R. (eds.) Caruana, I. D. 1992. Carlisle: Excavation of a Section of the Annexe Ditch of the First Flavian Fort, 1990. Britannia 23: 45–109. Caruana, I. D. 1997. Maryport and the Flavian conquest of North Britain, in Wilson, R. J. A. (ed.) 40–52 Caruana, I. 2004. Timber towers on the Solway frontier? in Wilson, R.J.A. & Caruana, I. (eds.). 184–5. Caruana, I. 2004b. The cemetery at Beckfoot Roman for, in Wilson, R.J.A. & Caruana, I. (eds.). 124–73. Carver, M. 1990. Pre-Viking traffic in the North Sea, in McGrail, S. (ed.) 117–25 Carver, M. 2005. Sutton Hoo. A Seventh-century Princely Burial Ground and its Context. British Museum Press. London. Casey, P.J. 1969. Caerhun, in Jarrett M.G. (ed.). 56–59. Casey, P.J. 1969. Forden Gaer, in Jarrett M.G. (ed.). 85–88. Casey, P.J. & Davies, J.L. with Evans, J. 1993. Excavations at Segontium Roman Fort, 1975–79. Council for British Archaeology Research Report 90. London. Casey, P.J. & Reece, R. (eds.) 1974. Coins and the Archaeologist. British Archaeological Report 4. Oxford. Casey, P.J. (ed.). 1979. The End of Roman Britain. British Archaeological Report 71. Oxford. Casey, P J. 1984. Roman coinage of the fourth century in Scotland, in Miket, R and Burgess, C (eds.). 295–304. Casey, J. 2011. Who built Carpow? A Review of Events in Britain in the Reigns of Commodus and Septimius Severus. Britannia, 41: 225–235. Casson, L. 1965. Harbour and River Boats of Ancient Rome. Journal of Roman Studies, 55: 31–39.

180

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Casson, L. 1974. Travel in the Ancient World. Allen & Unwin. London. Casson, L. 1995. Ships and seamanship in the Ancient World. John Hopkins University Press. Baltimore. Casson, L. 1989. The Periplus Maris Erythraei. Princeton. Castledine, A. 1991. Environmental archaeology and the Roman period in Wales: recent work, in Burnham, B.C. & Davies J.L. (eds.), 172–184. Chalmers, G. Caledonia Vol I: 18. Charlesworth, D. 1965. Excavations at Papcastle 1961–2. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 65: 102–14. Charlesworth, M.P. 1924. Trade-routes and Commerce of the Roman Empire. Cambridge University Press. Cambridge. Chevallier, R. 1976. Roman Roads. Batsford. London. Christensen, A. E. (ed.) 1996. The Earliest Ships: The Evolution of Boats into Ships. Conway. London. Christie, P.M.L. 1993. Chysauster and Carn Euny. English Heritage. London. Chitwood, P. 1991. Lincoln’s ancient docklands: the search continues, in Good, F.L. et al. (eds.). Clare, T. 2004. Coastal change and the western end of Hadrian’s Wall, in Wilson, R.J.A. & Caruana, I. (eds.) 39–51. Clark, R. 1991. Hallgate, Wigan: Excavation Report. Greater Manchester Archaeological Unit, unpublished report. Clarke S. & Wise, R. 1999. Evidence for extramural settlement north of the Roman fort at Newstead (Trimontium), Roxburghshire. Proceedings of the Society of Antiquaries of Scotland 129: 373-391. Cleere, H. 1977. The Classis Britannica, in Maxfield, V.A. (ed.). 18–22. Cleere, H. 1978. Roman harbours in Britain south of Hadrian’s Wall, in Taylor J. Du P & Cleere, H. (eds.), 36–40. Cleere, H.F. 1975. The Roman iron industry in the Weald and its connections with the Classis Britannica. Archaeological Journal, 131: 171–99. Cleere, H.F. & Crossley, D. 1985. The Iron Industry of the Weald. Leicester University Press. Leicester. Codrington, T. 1918. Roman Roads in Britain. Sheldon Press. London. Cole, A. 2007. The Place-Name Evidence for Water Transport in Early Medieval England, in Balair, J. 55–84 Coles, B. and Coles, J. 1986. Sweet Track to Glastonbury; the Somerset Levels in prehistory. Thames and Hudson. London. Coles, J. 1984. The archaeology of wetlands. Edinburgh University Press. Edinburgh. Coles, J., Fenwick, V. and Hutchinson, G. (eds.) 1993. A Spirit of Enquiry: Essays for Ted Wright. National Maritime Museum. London. Collingwood Bruce, J. 1863. The Wallet-book of the Roman Wall: A Guide to Pilgrims journeying along the Barrier of the Lower Isthmus. Longman. Newcastle-upon- Tyne. Collingwood, R.G. 1924. The Fosse. Journal of Roman Studies, 14: 252–6. Collingwood, R.G. 1929. Town and Country in Roman Britain. Antiquity, 3: 261–76. Collingwood, R.G. & Myres, J.N.L. 1937. Roman Britain and the English Settlements. Oxford University Press Oxford. Collingwood, R.G. & Richmond, I.A. 1969. The Archaeology of Roman Britain. Methuen. London. Collingwood, R.G. & Wright, R.P. 1995. The Roman Inscriptions of Britain Volume 1 Inscriptions in stone. Edited by Frere, S.S. & Tomlin, R.S.O. Sutton. Stroud. Collingwood, R.G. 1928. Roman Ravenglass. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 28: 353–66. Collingwood, R.G. & Taylor, M.V. 1934. Roman Britain in 1933. Journal of Roman Studies, 24: 220–1. Collins, R. 2006. Late Roman frontier communities in northern Britain : a theoretical context for the end of Hadrian's Wall, in Croxford, B. et al.. Collis, J.R. 1979. City and State in pre-Roman Britain, in Burnham, B.C. & Johnson, H.B. (eds.). 223–40. Colls, D. et al. 1977. L’épave Port-Vendres II et le commerce de la Bétique à l’époque de Claude. Archaeonautica, 1: 1–145. Colvin, H.M. (ed.) 1963. The History of the King’s Works. HMSO. London. Connell, J. 1975. Forden Gaer, Vicus. Archaeology in Wales, 15: 51. Cook, J. & Rowley, T. (eds.) 1985. Dorchester through the ages. Oxford University Department of External Studies. Oxford Cook, N.C. 1971. An iron anchor, probably Roman, from Blackfriars. The Antiquaries Journal 51: 318. Cool, H.E.M & Baxter, M.J. 2002. Exploring Romano–British finds assemblages, Oxford Journal of Archaeology 21/4: 365–80 Cool, H. 2006. Eating and Drinking in Roman Britain. Cambridge University Press. Cambridge. Cooley, R. 1993. The Unknown Fleet: The Army’s Civilian Seamen in Peace and War. Alan Sutton. London. Coppack, G. 1990. Abbeys and Priories. Batsford. London. Corney, M. 1997. The origins and development of the Small Town of Cunetio, Mildenhall, Wiltshire. Britannia, 28: 337–349. Corney, M. 2002. The Roman Villa at Bradford on Avon. The Investigations of 2002. Ex Libris Press.

181

JAMES ELLIS JONES

  Cottam, S. et al. (eds.). 1994. Theoretical Roman Archaeology Conference (4th Durham). Oxbow. Oxford. Cracknell, B. E. 2005. Outrageous Waves: Global Warming and Coastal Change in Britain through Two Thousand Years. Chichester. Craddock, P.T. & Hook, D.R. 1987. Ingots from the sea: The British Museum collection of ingots. International Journal of Nautical Archaeology, 16.3: 201-6. Crawford, O.G.S. 1949. The Topography of Roman Scotland. Cambridge. Crew, P. 1991. Late Iron-age and Roman iron production in Wales, in Burnham, B.C. & Davies J.L. 150–160. Croft, R. & Aston. M. 1993. Somerset from the air. Somerset County Council. Taunton. Crosby, A.G. 1999. The Chester Archaeological Society: the first one hundred and fifty years 1849–1999. Chester Archaeological Society. Chester Crossley, D.W. 1969. Pen-y-Gaer, in Jarrett (ed.). 108–110. Crowther, J. 1991. Palaeosols and Sediments, in Burnham, B.C. & Davies J.L. (eds.). 185–193. Croxford, B., Goodchild, H., Lucas, J. & Ray, N. (eds.) 2006. TRAC 2005 : Proceedings of the fifteenth annual Theoretical Roman Archaeology Conference, Birmingham 2005. Oxbow. Oxford Crummy, P. 1997. City of Victory. Colchester Archaeological Trust. Oxford. Crummy, P. 2003. Colchester’s town wall, in Wilson, P (ed.) 44–52 Cunliffe, B.W. 1966. The Somerset Levels in the Roman Period, in Thomas, C. (ed.) Council for British Archaeology Research Report 7. 68–73. Cunliffe, B.W. (ed.) 1968. Fifth Report on the Excavations of the Roman Fort at Richborough, Kent. Society of Antiquaries. Cunliffe, B.W. 1968. The British Fleet, in Cunliffe, B.W. (ed.). 255–272. Cunliffe, B.W. 1975. Excavations at Portchester Castle, Vol. 1. Society of Antiquaries of London 32. Cunliffe, B.W. 1978. Hengistbury Head. Elek. London. Cunliffe, B.W. 1980. The Evolution of Romney Marsh: a Preliminary Statement, in Thompson, F.H. (ed.) Cunliffe, B.W. 1984. Relations between Britain and Gaul in the First Century B.C. and Early First Century A.D, in Macready, S. & Thompson, F.H. (eds.). 3–23. Cunliffe, B.W. 1988. Mount Batten: a Prehistoric and Roman Port: Oxford University Committee for Archaeology. Oxford. Cunliffe, B.W. 1988. Greeks, Romans and barbarians: spheres of interaction. Batsford. London. Cunliffe, B.W. 1991. Iron Age Communities in Britain (3rd edn.). Routledge & Kegan Paul. London. Cunliffe, B.W. 1997. Amorica and Britain: cross-Channel relationships in the late First Millennium BC. Oxford University Monograph 45. Oxford. Cunliffe, B.W. 1998. Fishbourne Roman Palace. Tempus. Stroud. Cunliffe, B.W. 2001. Facing the Ocean: The Atlantic and its peoples, 8000 BC to AD1500. Oxford University Press. Oxford. Cunliffe, B.W. 2002. The Extraordinary Voyage of Pytheas the Greek. Penguin. London. Cunliffe, Sir B. 2008. Europe Between the Oceans: Themes and Variations: 9000 BC–AD 1000, New Haven and London: Yale University Press. Cunliffe, B.W. and de Jersey, P. 1997. Armorica and Britain: Cross-Channel Relationships in the late first millennium BC. Oxford University Committee for Archaeology. Oxford. Curchin, L.A. 1991. Roman Spain: Conquest and Assimilation. Routledge. London. Curle, J. 1911. A Roman Frontier Post and its People: the Fort of Newstead in the Parish of Melrose. Glasgow University Press. Dacre, J.A. 1985. An excavation on the Roman fort at Stanwix, Carlisle. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 85: 53–69. Dalwood, H. & Edwards, R. 2004. Excavations at Deansway, Worcester, 1988–89: Romano-British Small Town to Late Medieval City. Council for British Archaeology Research Report 139. York. Dannell, G.B. 2006. Samian cups and their uses, in Wilson, R.J.A. Darby, M.C. 1983. The Changing Fenland. Cambridge University Press. Cambridge Dark, K. & P. 1997. The Landscape of Roman Britain. Sutton. Stroud. Dark, K.R. (ed.) 1996. External contacts and the economy of late Roman and post-Roman Britain. Boydell. Woodbridge. Dark, K.R. 1994. Civitas to Kingdom. Leicester University Press. Leicester. Dark, P. 2000. The environment of Britain in the first millennium AD. Duckworth. London. Darvill, T. & Gerrard, C. 1994. Cirencester: Town and Landscape. Cotswold. Cirencester. Darvill, T. & Holbrook, N. 1994. The Cirencester area in the prehistoric and early Roman periods, in Darvill, T. & Gerrard, C. (eds.). 47–56. Darwin, A. 1976. Canals and Rivers of Britain. Dent. London. Davidson, A. (ed.) 2002. The coastal archaeology of Wales. Council for British Archaeology Research Report 131. York.

182

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Davies, C., Jones, G.B.D & Putnam, W.G. 1968/71. Excavations at Caersws. Montgomeryshire Collections. Davies, E. H. 2002. Roads in Roman Britain. Tempus. Brimscombe Port. Davies, J.L. 1969. Abergavenny, in Jarrett, M. (ed.). 45–46. Davies, J.L. 1983. Coinage and settlement in Roman Wales and the Marches: some observations. Archaeologia Cambrensis, 132: 78–94. Davies, J.L. 1984a. Excavations at Trawscoed Roman Fort, Dyfed. Bulletin of the Board of Celtic Studies, 31: 259–92. Davies, J.L. 1984b. Soldiers, Peasants and Markets in Wales and the Marches, in Blagg, T.F.C. & King, A.C. (eds.). 93–127. Davies, J.L. 1990. Segontium Roman Fort. CADW. Cardiff. Davies, J.L. 1991. Military Vici: Recent research and its significance, in Burnham, B.C. & Davies J.L. (eds.). 65–74. Davies, J.L. 1991. Roman Military Deployment in Wales and the Marches from Pious to Theodosius I, in Maxfield, V. and Dobbs, B. (eds.). 52–57. Davies, J.L. 1997. Native Producers and Roman Consumers: The mechanisms of military supply in Wales from Claudius to Theodosius, in Waateringe et al. (eds.). 267–72. Davies, R. 1987. Conquest, coexistence and change in Wales 1063–1415. University of Wales Press. Cardiff. Davies, R.W. 1989. Service in the Roman Army. Edinburgh University Press. Edinburgh. Davies, R.W. 1971. The Roman military diet. Britannia, 2: 122–42. Davies, W. 1979. Roman settlements and post-Roman estates in south-east Wales, in Casey, P.J. (ed.). 153–73. Davis, A.S. 1984. The Roman Road from Richborough to London (An estimation of the construction times for a Tactical and Strategic Road). Royal School of Military Engineering. Chatham. Dean, M. 1984. Evidence for possible pre-historic and Roman wrecks in British waters. International Journal of Nautical Archaeology, 13.1: 78–9. Deurne, M.J. & Branigan, K. 1995. The use of coal in Roman Britain. Antiquaries Journal, 75: 71–105. Defoe, D. 1748. A Tour Through the whole Island of Great Britain. abridged and edited by Furbank, P.N. and Owens, W.R. 1991. Yale University Press. New Haven. Denison, S. (ed.). 1999. Roman harbour found at Wroxeter. British Archaeology. Issue 50. 1. Desicatas. A. (ed.). 1973. Current Research in Romano-British Pottery. C.B.A Research Report 10. London. Devoy, J.J.N. 1990. Controls on coastal and sea-level changes and the application of archaeological-historical records to understanding recent patterns of sea-level movement, in McGrail, S. (ed.). 17–26. Dobat, E. 2009. The Gask ‘system’ in Perthshire: the first artificial frontier in line in the Roman Empire? in Breeze, D.J. et al. 39–48. Dobson, B. & Mann, J.C. 1973. The Roman army in Britain and Britons in the Roman army. Britannia, 4: 191–205. Dobson, B. 1980. Agricola’s life and career. Scottish Archaeological Forum, 12: 1–13. Dobson, B. 1986. The Function of Hadrian’s Wall. Archaeologia Aeliana, 14: 1–30. Dobson, B. 2000. The primipilares in the Army and Society, in Alföldy, G. Dobson, B. & Eck. W. 139–152. Dobson, B. 2008. Moving the goal posts, in Bidwell, P.T. 5–9 Dobson, D.P. 1937. Excavations, Sea Mills, near Bristol. Transactions of the Bristol and Gloucester Archaeological Society, 59: 330–32. Dobson, D.P. & Walker, F. 1939. Excavations at Sea Mills, near Bristol, 1938. Transactions of the Bristol and Gloucester Archaeological Society, 61: 202–23. Donaldson, G.H. 1988. Signalling Communications and the Roman Imperial Army. Britannia 19: 349–56. D’Olier, B. 1972. Subsidence and sea-level rise in the Thames Estuary. Philological Transactions of the Royal Society London , 272: 121-30. Dore, J.N. & Gillam, J.P. 1979. The Roman Fort at South Shields. Society of Antiquaries of Newcastle upon Tyne. Dore, J. & Greene, K. (eds.). 1977. Roman Pottery Studies in Britain and Beyond: Papers presented to John Gillam, July 1977. British Archaeological Report Supplementary Series 30. Oxford. Dore, J. & Wilkes, J. 1999. Excavations directed by J. D. Leech and J. J. Wilkes on the site of a Roman fortress at Carpow, Perthshire, 1967-79. Proceedings of the Society of Antiquaries of Scotland, 129: 481–575. Dornier, A. 1967. Was there a Coastal Limes in Western Britain in the Fourth Century?, in Applebaum, S. 14–20. Došanić, S. 2000. Army and Mining in Moesia superior, in Alföldy, G. Dobson, B. & Eck. W. (eds.). 343–64. Dove, C.E. 1970. The first British navy. Antiquity, 45: 15–20. Driver, T. 2002. Roman road west of Carmarthen: the road to Menapia, after all? Archaeology in Wales 42: 60–1. Drummond, S. K. & Lynn, H. N. 1994. The Western frontiers of Imperial Rome. M. E. Sharpe. Drury, P.J. (ed.) 1982. Structural reconstruction: Approaches to the interpretation of the excavated remains of building. British Archaeological Report 110. Oxford. Duckham, B.F. 1972. The Inland Waterways of East Yorkshire. East Yorkshire Historical Society. York. Dudley D.R. 1974. Tacitus and the Roman Conquest of Wales, in Birley, E., Dobson, D. and Jarrett, M. (eds.). 27–33. Dugdale, W. 1662. History of Imbankling and Drayning. Dumayne, L. 1994. The Effect of the Roman Occupation on the Environment of Hadrian's Wall: A Pollen Diagram from Fozy Moss, Northumbria. Britannia, 25: 217–224.

183

JAMES ELLIS JONES Duncan-Jones, R. !982. The Economy of the Roman Empire: quantitative studies. Cambridge University Press. Cambridge. Dymond, D.P. 1961. Roman Bridges and Dere Street, County Durham, with a general appendix on the evidence for bridges in Roman Britain. Archaeological Journal, 118: 136–64. Dyson, A.G. 1981. The terms ‘quay’ and ‘wharf’ and the early Medieval waterfront, in Milne, G. & Hobley, B.(eds.). 37–8. Eardley-Wilmot, E. 1990. New light on old travel routes; Combwich Causeway and the Herapath. Proceedings of the Somerset Archaeological and Natural History Society, 134: 187–91. Eckoldt, M. 1984. Navigation on small rivers in Central Europe in Roman and Medieval times. International Journal of Nautical Archaeology, 13.1: 3–9. Eddison, J. 2000. Romney Marsh: survival on a frontier. Tempus, Stroud. Eddison, J., Gardiner, M. & Long, A. 1998. Romney Marsh: environmental change and human occupation in a coastal lowland. Oxford Edwards, B.J.N. 1981. Ribchester. Ribchester Museum Trust. Ribchester. Edwards, J. & Hindle, B. 1991. The transportation system of Medieval England and Wales. Journal of Historical Geography, 19: 123–34. Edwards, N. & Lane, A.(eds.). 1988. Early Medieval Settlements in Wales, AD 400–1100. Bangor and Cardiff. Eglinton, E. 1982. The Last of the Sailing Coasters. Her Majesties Stationery Office. London. Elkington, H.D.H. 1976. The Mendip Lead Industry, in Branigan, K. & Fowler, P.W. (eds.). 183–99, 230–34. Ellis, P. 1987. Sea Mills, Bristol: the 1965–1968 excavations at the Roman Town of Abonae. Transactions of the Bristol and Gloucestershire Archaeological Society, (1987): 15–108. Ellis, P. & White, R.H. (ed.) 2006. Wroxeter Archaeology. Excavation and Research on the Defences and in the Town, 1968–1992. Shropshire Archaeology and History, vol. 78. Ellmers, D. 1978. Shipping on the Rhine in the Roman period: the pictorial evidence, in Taylor, J. & Cleere, H. (eds.). 1–14. Engles, D. 1978. Alexander the Great and the Logistics of the Macedonian Army. University of California Press. Berkeley. Erdcamp, P. 1998. Anger and the Sword: warfare and food supply in Roman Republican Wars. (264–30 B.C.). Geiben. Amsterdam. Erdkamp, P. (ed.) 2002. The Roman Army and the Economy. Blackwell. London. Erdkamp, P. 2007. A Companion to the Roman Army. Blackwell. London. Erdkamp, P. 2005. The Grain Market in the Roman Empire. Cambridge University Press. Cambridge Erim, K.T. & Reynolds, J. 1973. The Aphrodosias Copy of Diocletian’s Edict on Maximum Prices. Journal of Roman Studies, 63: 99–100. Esmonde Cleary, A.S. 1989. The Ending of Roman Britain. Batsford. London. Evans, E. 2000. The Caerleon canabae: excavations in the civil settlement, 1984–90. Society for the Promotion of Roman Studies. Britannia Monograph Series 16: London. Evans, E., Dowdell, G. & Thomas, H.J. 1985. A third-century maritime establishment at Cold Knap, Barry, South Glamorgan. Britannia, 16: 57–125. Evans, J. & Skull, C. 1990. Fieldwork on the Roman Fort Site at Blennerhassett. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 90: 127–38. Evans, J.G. et al (eds.). 1975. The effect of man on the landscape: the Highland Zone. Council for British Archaeology Research Report 11. London. Evans, J. 2006. The Pottery, in Masser, P. Everard, C.E. 1980. On sea-level changes, in Thomson, F.H. (ed.). 1–23. Farr, G. 1970. Ships and Harbours of Exmoor. Exmoor Press. Dulverton. Farr, G. 1954. Somerset Harbours, including the Port of Bristol. Christopher Johnson. London. Fasham, P.J. & Hawkes, J.W. 1985. Reading Abbey, a Study of the Horizontal and Vertical Movement of an Inland Waterway, in Herteig, A.E. (ed.). 133–7. Faulkner, N. 2000. The Decline and Fall of Roman Britain. Tempus. Stroud. Fenwick V.H. & Gale, A. 1988. Historic Shipwrecks: Discovered, Protected, Investigated. Tempus. Stroud. Fenwick, V. 1978. The Graveney Boat. British Archaeological Report 73. Oxford. Ferris, I. 1995. Shoppers’ paradise: consumers in Roman Britain, in Rush, P. (ed.). 132–40. Fincham, G. 2002. Landscapes of imperialism: Roman and native interaction in the East Anglian Fenland. British Archaeological Report 338. BAR Publishing. Oxford. Finley, M. 1973. The ancient economy. Chatto & Windus. London. Fisher, H.E.S.(ed.) 1971. Ports and Shipping in the North . University of Exeter. Exeter. Fitzpatrick, A.P. 1989.The submission of the Orkney Islands to Claudius: new evidence? Scottish Archaeological Review, 6: 24–33 Fitzpatrick, A.P. 1999. The Roman bridge at Piercebridge. Britannia, 30: 111–32. Fletcher, H.L.V. 1969. The Coasts of Wales. Robert Hale. London.

184

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Fleming, G. 1902. Mules in the Army. Military uses of Mules. Encyclopaedia Britannica. 9th Edn. Flohn, H. 1967. Klimaschwankungen in historischer Zeit Braunschweig (in Rudloff, H. von, 1967). Flower, C.T. 1915. Public Works in Medieval Law. Selden Society. London. Forcey, C. et al. (eds.). 1998. Theoretical Roman Archaeology Conference (7th Nottingham). Oxbow. Oxford. Ford, B. 2003. Coarse pottery, in Holmes, N. 58–86. Ford, W. 1839. A Description of Scenery in the Lake District. Thurnham. London. Forster, R. H. 1907. The bridge, in Woolley, C. L. (ed.) 177–80. Fortescue, J.W. 1930. Royal Army Service Corps: Early History of Supply and Transport in the British Army. The University Press. Cambridge. Fox, A., 1955. Some evidence for a Dark Age Trading site at Bantham, near Thurlestone, South Devon. Antiquaries Journal, 35: 55-67. Fox, A. 1974. New Light on the Military Occupation of South West England, in Birley, E. et al. (eds.). 84. Fox, A. & Ravenhill, W. 1965. Old Burrow and Martinhoe. Antiquity, 39: 253–8. Fox, A. & Ravenhill, W. 1972. The Roman Fort at Nantstallon, Cornwall. Britannia, 3: 56–111. Fradley, M. 2009. The Field Archaeology of the Romano-British Settlement at Charterhouse-on-Mendip. Britannia,40: 99–122 Frank, J. 1997. The Humber Keels. Humber. Hull. Frere, S.S. 1978. Britannia - A History of Roman Britain (3rd Edn.). Routledge. London. Frere, S.S. 1984a. Excavations at Dorchester on Thames, 1963. Archaeological Journal, 141: 91-174 Frere, S.S. & St Joseph, J.K. 1983. Roman Britain from the Air. Cambridge University Press. Cambridge. Frere, S. & Fulford, M. 2001. The Roman Invasion of AD 43. Britannia, 32: 45–55. Friel, I. 1985. The good ship: ships, shipbuilding and technology in England, 1200–1520. British Museum Press. London. Friel, I. 2003. Maritime History of Britain and Ireland c. 400 – 2001. British Museum Press. London. Fryer, J. 1973. The harbour installations of Roman Britain, in Blackman, D.J. (ed.). 261–73. Fuentes, N. 1991. The mule of a soldier. Journal of Roman Equipment Military Studies, 2: 65–99. Fulford, M.G. & Hodder, I. 1974. A regression analysis of some later Romano-British pottery: a case study. Oxoniensia, 39: 26–33. Fulford. M.G. 1978. The interpretation of Britain’s late Roman trade; the scope of Medieval historical and archaeology analogy, in Taylor, J. & Cleere, H. (eds.). 59–69. Fulford, M.G. 1989. A Roman Shipwreck off Normour, Isle of Scilly. Britannia, 20: 245–249. Fulford, M.G. 1989. The economy of Roman Britain, in Todd, M. (ed.). 175–202. Fulford, M.G. 1991. Britain and the Roman Empire: the Evidence for Regional and Long Distance Trade, in Jones, R.F.J. (ed.). 35– 47. Fulford, M.G. 1996. The second Augustan legion in the west of Britain. National Museum of Wales. Cardiff. Fulford, M.G. 2000. The organisation of legionary supply: the Claudian invasion of Britain, in Brewer, R.J. (ed.). 41–50. Fulford, M.G. 2006. Corvées and civitates, in Wilson, R.J.A. Fulford, M.G. 2007. Coasting Britannia: Roman Trade and Traffic Around the Shores of Britain, in Gosden, C. et al. (eds.). 54–74. Fulford, M.G. & Allen, J.R.L. 1992. Iron-making at the Chesters Roman Villa, Woolaston. Britannia, 23: 159–215. Fulford, M.G. et al. 1992. The Medieval quay at Woolaston Grange, Gloucestershire. Transactions of the Bristol and Gloucestershire Archaeological Society, 110: 101–127. Fulford, M. & Huddleston, K. 1990. The current state of Romano-British pottery studies. English Heritage. London. Fulford, M.G., Champion, T. & Long, A. (eds.) 1997. England’s Coastal Heritage: a Survey for English Heritage and the RCHME. English Heritage Archaeological Report 15. London. Fuller, Maj. Gen. J.F.C. 1965. Julius Caesar, Man, Soldier and Tyrant. De Capo Press. London. 315–6 Funari, P.P.A. 1996. Dressel 20 Inscriptions from Britain and the Consumption of Spanish Olive Oil. British Archaeological Report 250. Oxford. Gabriel, R.A. & Metz, K.S. 1996. A Short History of War. The evolution of Warfare and Weapons. U.S. Army War College. Pennsylvania. Gaffney, V.L., White, R.H. & Goodchild, H. 2007 Wroxeter, the Cornovii, and the urban process: Final Report on the Wroxeter Hinterland Project 1994–1997. Journal of Roman Archaeology, Supp Ser. 68. Galliou, P. 1982. Les amphores tardo-républicaines découvertes dans l’ouest de la France et les importations de vin italien à la fin de l’Age de Fer. Fasc. I du Corpus des amphores découvertes dans l’ouest de la France. Brest. Gardiner, M. 2007. Hythes, Small Ports and Other Landing Places in Later Medieval England, in Blair, J. 85–109 Gardiner, R. (ed.) 1995. The Age of the Galley. Mediterranean Oared Vessels since Pre-classical Times. Conway. London.

185

JAMES ELLIS JONES

  Garnsey, P. 1988. Famine and food supply in the Graeco-Roman world: responses to risk and crisis. Cambridge University Press. Cambridge. Garnsey, P. & Saller, R. 1987. The Roman Empire: economy, society and culture. Duckworth. London. Garnsey, P., Hopkins, K. & Whittaker, C.R. (eds.). 1983. Trade in the Ancient Economy. Chatto & Windus. London. Gauld, W.W. 1990. Vegetius on Roman scout-boats. Antiquity, 64: 402–6. Gaunt, G.D. & Buckland, P.C 2003. Sources of building materials in Roman York, in Wilson & Price (eds.) 135–143, Gentry, A.P. 1976. Roman military stone-built granaries in Britain. British Archaeological Report 32. Oxford. Gerrard, J. 2008. Feeding the Roman army from Dorset: pottery, salt and the Roman state, in Stallibrass S. & Wilson, R. (eds.) Gibb, A. 1903. New Measurement of the Wall of Antoninius Pius. The Scottish Antiquary Vol 17, 68; 188–97 Gibbins, D. 1996. Roman shipping and trade: a Mediterranean perspective, in Carrington, P. (ed.). 24–39. Gifford, E. & J., 1995. The sailing characteristics of Saxon ships. International Journal of Nautical Archaeology, 24: 121–131. Gildas (translated by Williams, H. 1899). De Excidio et Conquestu Britanniae. Cywrodorion Record Series. Cardiff. Gillam, J. & Greene, K. 1981. Roman Pottery and the Economy in Anderson A.C. & Anderson, A.S. (eds.). note 34, 14, 17. Gillam, J. 1973. Sources of Pottery found on Northern Military Sites, in Desicatas, A. (ed.). Gillam, J.P. 1974. The Coarse Pottery, in Breeze, D.J. 159 Gillmer, T. 1979. The capability of the single square rig sail, in McGrail, S. (ed.). 167–83. Glover, M. 1977. Wellington’s Army in the Peninsula 1808–1814. David & Charles. Newton Abbot. Going, C.J. 1992. Economic ‘Long Waves’ in the Roman Period: a reconnaissance of the Romano-British Ceramic Evidence. Oxford Journal of Archaeology, 11 (1): 93–117 Goldsworthy, A. 1996. The Roman Army at War, 100 BC – AD 200. Clarendon Press. Oxford. Goldsworthy, A. 2000. Roman Warfare. Cassell. London. Goldsworthy, A. 2008. The Late Roman Army: asking the awkward questions. Hadrianic Society Bulletin. New Series, Vol. 3: 25–32. Goldsworthy, A. 2009. The Fall of the West: the Slow Death of the Roman Superpower . Weidenfield & Nicolson. London Goldsworthy, A. & Haynes, I. 1999. The Roman Army as a Community. Journal of Roman Archaeology, Supplementary Series 34. Portsmouth, R.I. Good, G.L., Jones, R.H. & Ponsford, M.W. 1991. Waterfront Archaeology. Council for British Archaeology Research Report 74. London. Goodburn, R. 1979. Chedworth: the Roman villa. National Trust. London Gosden, C. et al. (eds.) 2007. Communities and Connections: Essays In Honour Of Barry Cunliffe. Oxford University Press. Oxford. Gould, R.A. 2000. Archaeology and the social history of ships. Cambridge University Press. Cambridge. Graham, F. 1989. Dictionary of Roman Military Terms. Butler & Butler. Rothbury. Grahame, M. 1998. Redefining Romanisation: material culture and the question of social continuity, in Roman Britain, in Forcey, C. et al. (eds.) Grainge, G. 2002. The Roman Channel Crossing of AD 4: The constraints on Claudius’s naval strategy. British Archaeological Report 332. Oxford. Grainge, G. 2005. The Roman Invasions of Britain. Tempus. Stroud. Grainge, G. 2006. Double tides in the Wantsum: fact or fiction? Archaeologia Cantiana, 126: 381-91. Green, C. 1996. Trows and the Severn coastal trade. Archaeology in the Severn Estuary 6: 97–113. Green, C. 1997. The Forest ports of the Severn Estuary. Archaeology in the Severn Estuary 7: 107–113. Green, C. 1999. Severn Traders. Black Dwarf. Lydney, Green, C. 1943. Glevum and the Second Legion, Evidence of the Pottery, Metal Objects etc. Journal f Roman Studies, 33: 15–29. Green, M.J. 1995. The Celtic World. Routledge. London. Greene, K. 1979, invasion and Response: Pottery and the Roman Army, in Burnham, B.C. & Johnson, H.B. (eds.). 99–108. Greene, K. 1986. The Archaeology of the Roman Economy. Batsford. London. Greene. K. 1978. Roman trade between Britain and the Rhine provinces: the evidence of pottery to AD 250, in Taylor, J. & Cleere, H. (eds.). 52–58. Greenhill, B. 1988. The Evolution of the Wooden Ship. Batsford. London. Greenhill, B. 1995. The Archaeology of Boats and Ships. Conway. London. Greep, S.J.(ed.) 1993. Roman Towns: The Wheeler Inheritance. Council for British Archaeology Research Report 93. York.

186

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Gregory, K.J. 1977. River Channel Changes. John Wiley. Chichester. Gregory, K.J. & Walling, D.E. 1973. Drainage Basin Form and Process; a geomorphological approach. Arnold. London. Gregory, K.J. (ed.) 1997. Fluvial Geomorphology of Great Britain. Chapman & Hall. London. Gregory, K.J., Lewin, J. & Thornes, J.B. (eds.). 1987. Palaeohydrology in Practice: A River Basin Analysis. Wiley. Chichester. Gregory, R. 2001b. Living on the Frontier: Iron Age–Roman Transitions in South-West Scotland. in Higham, N. 35–49. Gresham, C.A. 1969. Pennal, in Jarrett M.G. (ed.). 104–106. Griffith, F.M. 1985. A nemeton in Devon? Antiquity, 59: 121–4. Griffith, F.M. 1997. Developments in the study of Roman military sites in South West England, in Waateringe et al. (eds.). 361–8. Griffith, F.M. 1984a. Roman Military Sites in Devon: Some Recent Discoveries. Proceeding of the Devon Archaeological Society, 42: 11–32. Griffiths, D, Philpott, R A & Egan G. 2007. Meols: the archaeology of the north Wirral coast. Discoveries and observations in the 19th and 20th centuries, with a catalogue of collections. Oxford Institute of Archaeology. Oxford University School of Archaeology Monograph. 68. Griffiths, D. 1996. The maritime economy of the Chester region in the Anglo-Saxon period, in Carrington, P. (ed.). 49–61. Grimes, W. 1968. The excavation of Roman and Medieval London. London. Grimes, W.F. 1930. Holt, Denbighshire: the works-depot of the Twentieth Legion at Castle Lyons. Y Cymmrodor 41. Groenman–van Waateringe, W. 1997. Classical authors and the diet of Roman soldiers: true or false?, in Groenman– van Waateringe, W. et al. 261–5. Groenman–van Waateringe, W. van Beek, B.L., Willems W.J.H. and Wynia, S.L. (eds.) 1997. Roman frontier studies 1995 Oxbow. Oxford. Grove, J. 2002. Reclamation and Utilisation of the Upper Axe Valley during the Roman Period. Archaeology in the Severn Estuary,13. 65–87. Guest, P. 2010. Coins from the Roman frontier in Wales, in Burnham, B.C. & Davies, J.L. (eds.). 24–32 Habesch, D. 2001. The Army's Navy - British Military Vessels and their History since Henry VIII. Chatham Publishing. London. Hall, D. 1992. The Fenlands Project no. 6: The Western Cambridgeshire Fenlands. East Anglian Archaeology, 56: 42. Hall, M. 2009. Preface, in Breeze, D.J. et al. vii Hall, R.A. 1991. The waterfronts of York, in Good, F.L. et al. (eds.). Hall, R. 1996. Ports on the east and south coasts in the Anglo-Saxon era: an overview, in Carrington, P. (ed.). 40–8. Hanson, W. S. 1980a. The first of Roman occupation of Scotland, in Hanson, W.S. & Keppie, L.J.F. 15–44 Hanson, W.S. 1978a. The Organisation of Roman Military Timber-supply. Britannia, 9: 293–305. Hanson, W.S. 1996. Forest clearance and the Roman Army. Britannia, 27: 354–60. Hanson, W.S. & Keppie, L.J.F. (eds.) 1980. Roman Frontier Studies 1979: papers presented to the Twelfth International Congress of Roman Frontier Studies. Oxford. Hanson, W.S. & MacInnes, L. 1991. Soldiers and Settlement in Wales and Scotland, in Jones, R.F.J. (ed.). 85–92. Hanson, W.S. & Maxwell, G.S. 1983. Rome’s North West Frontier: The Antonine Wall. Edinburgh University Press. Edinburgh. Hanson, W.S. 1991. Agricola and the conquest of the North. Batsford. London. Hanson, W.S. 2009a The fort at Elginhaugh and its implications for Agricola's role in the conquest of Scotland, in Breeze et al. 49–58 Harbison, P. 1988. Pre-Christian Ireland: From the first settlers to the early Celts. Guild Publishing. London. Harding, A. F. 1982a. Climatic change in later prehistory. Edinburgh University Press. Edinburgh. Harris, A. 2003. Byzantium, Britain and the West: the archaeology of cultural identity, AD 400–650. Tempus Publishing, Stroud. Harris, B. (ed.). 1987. A history of the County of Chester 1. (Victoria History of the Counties of England). Oxford University Press. London. Harris, Rev. H. 1763. Observations on the Julia Strata and on the Roman stations, forts and camps, in the counties of Monmouth, Brecknock, Caermarthen and Glamorgan. Archaeologia, Volume II. 2 Hart, C. 2004. The coast from South Shields to Whitburn. The Arbeia Society. Newcastle. Hart, C.E. 1967. Archaeology in Dean. Bellows. Gloucester. Hartgroves, S. & Smith, J. 2008. A Second Roman Fort is Confirmed in Cornwall. Britannia, 39: 237–9. Hartley, B.R. 1972. The Roman occupation of Scotland; the evidence of Samian Ware. Britannia, 3: 1–55 Hartley, B. & Wacher, J. (eds.). 1983. Rome and her Northern Provinces. Allan Sutton. Gloucester. Hartley, K. 1973. The Marketing and Distribution of mortaria, in Desicatus, A. Hartley, K. 1975. Mortarium Stamps, in Robertson, Scott, & Keppie (eds.). 142–47 Hartley, K. 1977. Two major industries producing mortaria in the first century AD, in Dore, K. & Greene, K. 5–17

187

JAMES ELLIS JONES Hassall, M. 1978. Britain and the Rhine provinces: epigraphic evidence for Roman trade, in Taylor, J. du P. & Cleere, H. (eds.). 41–48. Haselgrove, C.C. 1984b. Romanisation before the Conquest: Gaulish Precedents and the British Consequences, in Blagg, T.F.C. & King, A.C. (eds.). 5–63. Hassall, M. 1980. Altars, curses and other epigraphic evidence, in Rodwell (ed.) 79–89 Hassall, M. 2000. Pre-Hadrianic legionary dispositions in Britain, in Brewer, R.J. (ed.). 51–68. Hassan, F.A. 1981. Demographic Archaeology. Academic Press. London. Hatcher, J. 1993. The history of the British coal industry: Volume 1: before 1700. Oxford University Press. Oxford. Haverfield, F. 1912. The Romanisation of Britain (2nd Edn.) Oxford University Press. Oxford. Hawkins, A.B. 1973. Sea level changes around South West England, in Blackman, D.J. (ed.). 67–87. Haywood, J. 1991. Dark Age naval power: a re-assessment of Frankish and Anglo-Saxon seafaring activity. Routledge. London. Haselgrove, C.C. 1989. The later Iron Age in southern Britain and beyond, in Todd, M. (ed.) Haselgrove, C.C.& Moore, T. (eds.) 2007, The Later Iron Age in Britain and Beyond. Oxbow. Oxford. Heighway, C.M. & Parker, A.J. 1982. The Roman Tilery at St Oswald’s Priory, Gloucester. Britannia, 13: 25–78. Helbaek, H. 1964. The Isca Grain, a Roman plant introduction in Britain. New Phytologist, 63. Henderson, C. G. 1991. The development of Exeter Quay, in Good, F.L. et al. (eds.). Henderson, J.C. 2007. The Atlantic Iron Age: Settlement and Identity in the First Millennium BC. Routledge. Henig, M. 2002. Tales from the Tomb: Sculpture in Roman Chester, in Carrington, P. (ed.). 75–8. Henig, M. & Ross, A. 1998. A Roman Intaglio Depicting a Warship from the Foreshore at King’s Reach, Winchester Wharf, Southwark. Britannia, 29: 325-327. Henig, M. & Booth, P. 2000. Roman Oxfordshire. Sutton Publishing. Stroud. Herring, P. 2000. St Michael's Mount Archaeological Works, 1995-8. Cornwall Archaeological Unit. Cornwall County Council. Herring, P. 1994. Stratton Hundred, Rapid Identification Survey. Cornish Archaeology, 33: 235–7. Herson, J. 1996. Canals, railways and the demise of the port of Chester, in Carrington, P. (ed.). 75–89. Herteig, A.E.(ed.). 1985. Conference on Waterfront Archaeology in North European Towns No2. Historisk Museum. Bergen. Heslop, D. 1987. The excavation of an Iron Age Settlement at Thorpe Thewles, Cleveland, 1980–82. London. Heywood, B. 1969. Neath, in Jarrett, M.G. (ed.). 98–101. Heywood, B. 1969. Pennydarren, in Jarrett, M.G. (ed.). 106–8. Heyworth, A & Kidson, C. 1982. Sea level changes in South West England and Wales. Proceedings of the Geologists Association, 93/I: 91–111. Hibbert, F.A. 1980. Possible Evidence for Sea-level Change in the Somerset Levels, in Thompson, F.H. (ed.). 103–5. Higham, N.J. & Jones, G.B.D. 1985. The Carvetii. Stroud. Higham, N.J. & Jones, G.D.B. 1975. Frontier, forts and farmers. Archaeological Journal, 132: 16–45 Higham, N.J. 1993. The Origins of Cheshire. Manchester University Press. Manchester. Higham, N. (ed.). 2001. Archaeology of the Roman Empire: a tribute to the life and works of Professor Barri Jones. BAR Publishing. Higham, R. (ed.). 1987. Security and Defence in South West England before 1800. Exeter Studies in History 19. Exeter. Hill, J.D. 2001. Romanisation, gender and class: recent approaches to identity in Britain and their possible consequences, in James & Millett (eds.). 12–18. Hill, P.R. (ed.). 2002. Polybius to Vegetius : essays on the Roman army and Hadrian's Wall presented to Brian Dobson to mark his 70th birthday : Hadrianic Society. Durham. Hill, P. 2006. The Construction of Hadrian’s Wall. Tempus. Stroud. Hinchliffe J. and Williams J.H. 1992. Roman Warrington. Manchester. Hind, J.G.F. 1974. Agricola’s fleet and the Portus Trucculensis. Britannia, 5: 285–8. Hind, J.G.F. 1989. The invasion of Britain in AD43 – an alternative strategy for Aulus Plautius. Britannia, 23: 1–21. London. Hingley, R. 1989. Rural Settlement in Roman Britain. Seaby. London. Hoare, P. et al. 2002. Reused Bedrock ballast in King's Lynn Town Walls. Medieval Archaeology, 46: 91-105. Hobbs, J.S. 1859. The British Channel Pilot from the Downs to Bristol, etc. Charles Wilson. London. Hobley, A.S. 1989. The Numismatic Evidence for the Post-Agricolan Abandonment of the Roman Frontier in Northern Scotland. Britannia, 20: 69–74. Hobley, B. & Milne. G. 1981. Waterfront Archaeology in Britain and Northern Europe. Council for British Archaeology Research Report 41. Oxford. Hobley, B. 1982. Roman Military structures at ‘The Lunt’ Roman Fort; Experimental Simulations 1966–77, in Drury, P.J. (ed.). 227–74. Hockmann, O. 1988. Late Roman river craft from Mainz, Germany. Local Boats: 23–4. Hodder, I. 1974. Some marketing models for Romano-British pottery. Britannia, 5: 340–59.

188

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Hodder, I.R. 1979. Pre-Roman and Romano-British Tribal Economies, in Burnham, B.C. & Johnson, H.B. (eds.). 189–196. Hodder, I.R. & Millett, M. 1980. Romano-British Villas and Towns: a systematic analysis. World Archaeology, 12(1): 69–76. Hodgson, G. B. 1903. The Borough of South Shields from the Earliest Period to the Close of the Nineteenth Century. Newcastle upon Tyne. Hodgson, N. 1995. Were there two Antonine occupations of Scotland? Britannia 26: 29–49 Hodgson, N. 1997. Relationships between Roman river frontiers and artificial frontiers, in Waateringe et al (eds.). 61–6. Hodgson, N. 2000. The Stanegate: A Frontier Rehabilitated. Britannia, 31: 11–22. Hodgson, N. 2008. After the Wall-Periods: what is our historical framework for Hadrian's Wall in the 21st century? in Bidwell, P. (ed.). 11–24. Hodgson, N. 2009. Hadrian’s Wall 1999–2009: A summary of Excavation and Research. Cumberland and Westmorland Antiquarian and Archaeological Society and Society of Antiquaries of Newcastle upon Tyne. Kendal. Hoffmann, B. 2002. Where have all the Soldiers gone? Some thoughts on the Presence and Absence of soldiers in Fourth-Century Chester, in Carrington, P. (ed.). 79–88. Hoffmann, B. 2008. A new look at legions and their hinterland. Hadrianic Society Bulletin. New Series Vol. 3: 42–45. Hoffmann, B. 2009. Cardean: the changing face of the Flavian fort in Scotland, in Breeze, D.J. et al. 29–32. Hogg, A.H.A. 1979. Invasion and Response: The Problems in Wales, in Burnham, B.C. & Johnson, H.B. (eds.). 285–98. Hogg, A.H.A. 1985. Native Settlement in Wales, in Thomas, C. (ed.). 28–38. Hogg, A.H.A. 1969. Pen Llystyn, in Jarrett, M.G. (ed.). 101–4. Holbrook, N. & Bidwell, P.T. 1991. Roman Finds from Exeter. Exeter Archaeological Report 4. Exeter. Holbrook, N. 2001. Coastal Trade around the South-East Peninsula of Britain in the later Roman period: A summary of the evidence. Proc. Devon Archaeological Soc, 59: 149–58. Holder, P. 2004. Roman place-names on the Cumbrian coast, in Wilson & Caruana, I. (eds.). 52–65. Holder, P.A. 1982. The Roman Army in Britain. Batsford. London. Hohlfeldrer, R.L. 2008. The Maritime World of Ancient Rome. Memoirs of the American Academy in Rome. Supplementary Volume VI. Ann Arbor. Hollinrake, C. & Hollinrake, N. 1991. A late Saxon monastic enclosure ditch and canal, Glastonbury, Somerset. Antiquity, 65: 117–8. Hollinrake, C. & Hollinrake, N. 1992. The abbey enclosure ditch and a late Saxon canal: rescue excavations at Glastonbury, 1984–1988. Proceedings of the Somerset Archaeological and Natural History Society, 136: 7. 3–94. Holmes, N. M. McQ. 1999. Coins, in Dore, J. & Wilkes, J. 528–35. Holmes, N. 2003. Excavation of Roman sites at Cramond, Edinburgh. Society of Antiquaries of Scotland Monograph 23. Edinburgh Hooke, D. 2007. Uses of Waterways in Anglo-Saxon England in Blair, J. 37–54 Hope, V.M. 1997. Words and Pictures: the interpretation of Romano-British tombstones. Britannia, 27: 245–58. London Hopkins, K. 1980. Taxes and trade in the Roman Empire (200BC–AD400). Journal of Roman Studies, 71: 1. Hopkins, K. 1983. Introduction in Garnsey et al. (eds.). ix–xxv. Horden, P. & Purcell, N. 2000. The Corrupting Sea: A Study of Mediterranean History. Blackwell. Oxford. Hornell, J. 1946. Water Transport. Cambridge University Press. Cambridge. Horton, M.L. 1997. Britain and its International Position, in Keen, L. 12–13. Hoskins, W.G. 1955. An Elizabethan provincial town in Plumb, J.H. (ed.) Hoskins, W.G. 1955. The making of the English landscape. Hodder & Stoughton. London. Houston, G.W. 1988. Ports in Perspective: Some Comparative Materials on Roman Merchant Ships and Ports. American Journal of Archaeology, 92: 553–64. Howard-Davis, C. and Leah, M. 1999. Excavations at St Nicholas Yard, Carlisle, 1996–7. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 99: 89–115. Howard-Davis, C.L.E. & Buxton, K. 2000. Roman Forts in the Fylde: Excavations at Dowbridge, Kirkham 1994. University of Lancaster. Lancaster. Hudson, J.D. & Sutherland, D.S. 1990. The Geological Description and Identification of Building Stones: Examples from Northamptonshire, in Parsons, D. (ed.). 16–32. Hughes, G. 2003. A Roman fort at Dinefwr Park, Llandeilo: An interim statement on a geophysical survey by Stratascan. Carmarthenshire Antiquary. 34: 144–7. Hulme, M & Barrow, E. 1997. Climates of the British Isles: present past and future. Routledge. London. Hunter, F. & Collard, M. 1997. The Cramond Lioness. Current Archaeology 155: 404–7.

189

JAMES ELLIS JONES

  Hunter, F. 2002. Finds from Inveresk – glimpses of Roman frontier life, in Bishop, M.C. (ed.). 70–80. Hurst, H.R. 1988. Gloucester (Glevum), in Webster, G. (ed.). 48–73. Hurst, H.R. 1999. Civic space at Glevum, in Hurst, H.R. (ed.). 113–25. Hurst, H.R. (ed.) 1999. The coloniae of Roman Britain: new studies and a review. Journal of Roman Archaeology. Portsmouth. Hurst, H.R. 1985. Kingsholm. Gloucestershire Archaeological Publications. Gloucester. Hurst, H.R. 1986. Gloucester: the Roman and later defences. Gloucester Archaeological Publications. Gloucester. Hussey, D. 2000. Coastal and River trade in pre-Industrial England: Bristol and its region. 1680–1730. Exeter University Press. Exeter. Hutchinson, G. 1994. Medieval Ships and Shipping. Leicester University Press. Leicester. Hutchinson, R.W. 1957. Fore and Aft Rigging in the Roman Empire. Antiquity, xxxi, 121: 39. Hydrographer of the Navy, 1974. Tidal Stream Atlas: Irish Sea and Bristol Channel. Admiralty. Taunton. Hydrographer of the Navy, 1974. Channel Pilot. Admiralty. Taunton. Hydrographer of the Navy, 1927. West Coasts of England and Wales Pilot. Admiralty. Taunton. Hydrographer of the Navy, 1974. West Coasts of England and Wales Pilot. Admiralty. Taunton. Hyland, A. 1990. Equus. The Horse in the Roman World. Batsford. London. Hyland, A. 1993. Training the Roman Cavalry: From Arrian’s Ars Tactica. Grange. London. Ireland, S. 1996. Roman Britain: a Sourcebook. Routledge. London. Jackson, G. 1983. The History and Archaeology of Ports. Tadworth. James, H. 1978. Excavations at Church Street, Carmarthen, in Boon, G.C. (ed.). 65-103. James, H. 1992. Excavations in Roman Carmarthen 1978–1990. Dyfed Archaeological Trust. Carmarthen. James, H. 2003. Roman Carmarthen: Excavations, 1978–1993. Britannia Monograph Series 20. London James, H.J. 1993. Roman Carmarthen, in Greep, S.J. (ed.). 93–8. James, S. 1984. Britain and the late Roman Army, in Blagg, T.F.C. & King, A.C. (eds.). 161–86. James, S. 2001. 'Soldiers and civilians: identity and interaction in Roman Britain, in James, S. & Millett, M. (eds.) James, S. 1999. The Atlantic Celts. Ancient People or Modern Invention. British Museum Press. London. James, S. 2002. Writing the legions; the past, present and future of Roman military studies in Britain. Archaeological Journal 159: 1–58. James, S. & Millett, M. (eds.). 2001. Britons and Romans: advancing the archaeological agenda. British Archaeology Research Report 125. York. Jarrett, M.G. 1958. The pre-Hadrianic occupation of Roman Maryport. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society 58: 63–67. Jarrett, M.G. 1976. Maryport, Cumbria: A Roman fort and its garrison. Cumberland and Westmorland Antiquarian and Archaeological Society extra series 22. Kendal. Jarrett, M.G. & Wrathmell, S. 1981. Whitton: An Iron Age and Roman farmstead in South Glamorgan. University of Wales. Cardiff. Jarrett, M.G. (ed.) 1969. The Roman Frontier in Wales. University of Wales Press. Cardiff. Jarrett, M.G. 1969. Brecon Gaer, in Jarrett M.G. (ed.). 48–51. Jarrett, M.G. 1969. Cardiff, in Jarrett M.G. (ed.). 70–3. Jarrett, M.G. 1969. Clyro, in Jarrett M.G. (ed.). 77–80. Jarrett, M.G. 1969. Llandovery, in Jarrett M.G. (ed.). 95–6. Jarrett, M.G. 1969. Llanio, in Jarrett M.G. (ed.). 97–8. Jarrett, M.G. 1969. Pumpsaint, in Jarrett M.G. (ed.). 110–1. Jarrett, M.G. 1969. Tomen-y-Mur, in Jarrett M.G. (ed.). 111–3. Jarrett, M.G. 1994a. Early Roman Campaigns in Wales. National Museum of Wales. Cardiff. Johnson, D.E. (ed.). 1977. The Saxon Shore. Council for British Archaeology Report 18. London. Johnson, J.S. 1980. Later Roman Britain. Grafton. London. Johnson, P. 2003. Late Roman economic systems: their implication in the interpretation of social organisation, in Carr, G. et al. (eds.). 101–12. Johnson, R. W. 1895. The making of the Tyne: a record of fifty years progress. Scott. London. Johnstone, C. 2008. Commodities or logistics? The role of equids in Roman Supply networks, in Stallibrass S. & Wilson, R. (eds.) Jones, A.H.M. 1964. Later Roman Empire. Blackwell. Oxford. Jones, A.H.M. 1974. The Roman Economy: Studies in Ancient Economic and Administrative History. Blackwell. Oxford. Jones, E.T. 2000. River Navigation in Medieval England. Journal of Historical Geography, 26: 60–75. Jones, G,D.B. & Shotter, D.C.A. 1988. Roman Lancaster. Manchester, Jones, G.D.B. 1968. The Romans in the North. Northern History 3. 1–26. Jones, G.C. & Price, J. 1985. Excavations at The Wiend, Wigan 1982-4. Greater Manchester Archaeological Journal, 1: 25–33. Jones, G.D.B. 1969. Caersws, in Jarrett M.G. (ed.). 66–70.

190

THE MARITIME LANDSCAPE OF ROMAN BRITAIN Jones, G.D.B. 1969. Carmarthen, in Jarrett M.G. (ed.). 73–4. Jones, G.D.B. 1980. Archaeology and Coastal Change in the North, in Thompson, F.H. (ed.). 87–102. Jones, G.D.B. 1980. Archaeology and Coastal Change in the North-West, in Thompson, F.H. (ed.) Jones, G.D.B. 1990. Llantsantfraid-y-Mechain. Britannia, 19: 417–9 Jones, G.D.B. 1984. ‘Becoming different without knowing it’, in Blagg T.F.C. & King, A.C. (eds.). 75–92. Jones, G.D.B. 1990. The Emergence of the Tyne-Solway Frontier, in Maxfield, M.A. & Dobson, M.J. (eds.). Roman Frontier Studies, 1989, 98–107. Exeter. Jones, G.D.B. 1991. North-east Wales: the empty quarter, in Burnham, B.C. & Davies J.L.(eds.). 123–9. Jones, G.D.B. 2004. Excavations on a coastal tower: Campfield Tower 2b, Bowness-on-Solway in Wilson, R.J.A. & Caruana, I. (eds.). 174–83. Jones, G.D.B. & Mattingly, D. 1990. An Atlas of Roman Britain. Blackwell. Oxford. Jones, G.D.B. & Maude, K. 1987. The Solway frontier 1984: Roman Frontier Studies. Manchester Archaeological Bulletin, 1: 3–9. Jones, G.D.B. & Reynolds, P. 1978. Roman Manchester: the Deansgate Excavations 1978. University of Manchester. Jones, J. Ellis 2009.The Maritime and Riverine Landscape of the West of Roman Britain: Water Transport on the Atlantic Coasts and Rivers of Britannia. British Archaeological Report 493. BAR Publishing. Oxford. Jones, M. 1984. Society and Settlement in Wales and the Marches 500 BC to AD 1100. British Archaeological Report 121. Oxford. Jones, M. 1989. Agriculture in Roman Britain: The dynamics of change, in Todd, M. (ed.). 127–34. Jones, M.E. 1979. Climate, nutrition and disease: an hypothesis of Romano-British population, in Casey, P.J. (ed.). 231–51. Jones, M.E. 1996. The End of Roman Britain. Cornell University Press. New York. Jones, M.J. 1999. Roman Lincoln: Changing Perspectives in Hurst, H.R. (ed.). 101–112. Jones, M.J. 2002. Roman Lincoln: Conquest, Colony and Capital. Tempus. Stroud. Jones, M.K. 1982. Crop Production in Roman Britain, in Miles, D. (ed.). 97–108. Jones, R.F.J. 1991. The urbanisation of Roman Britain, in Jones, R.F.J. (ed.). 53–66. Jones, R.F.J. (ed.) 1991. Britain in the Roman Period: Recent Trends. Collis. Sheffield. Jones, R. H. 2009a. Chasing the Army: the problems of dating camps, in Breeze, D J, et al. , 20–7 Jones, T. 1984. A Steady Trade: a boyhood at sea. Granada. London. Jones, W.H. & Jackson, J.E. 1907. Bradford-upon-Avon. Library Press. Bradford-on-Avon. Jongman, W. 1991. The economy and society of Pompeii. Geiben. Amsterdam. Junkelman, M. 1986. Die Legionen des Augustus: Der romische Soldat im archaologische Experiment. Von Zabern. Mainz. Junkelmann, M. 1997. Panis Militaris. Panis militaris. Die Ernährung des römischen Soldaten oder der Grundstoff der Macht. Philipp von Zabern. Mainz. Kamm, A. 2004. The Last Frontier: The Roman Invasions of Scotland. Tempus. Stroud. Keay, S. & Williams, D. 2005. A Digital Resource. http://ads.ahds.ac.uk/catalogue/archive/amphora_ahrb_2005/index.cfm Keay, S.J. 1998. The archaeology of early Roman Baetica. Journal of Roman Archaeology. Supplementary Series 29. Keegan, J. 1994. A history of Warfare. Pimlico. London. Keen, J. 1979. The Little Russel Wreck. Unpublished information sheet. Guernsey. Keen, L. 1997. Almost the Richest City: Bristol in the Middle Ages. British Archaeological Association. London. Kendal, R. 1996. Transport Logistics associated with the building of Hadrian’s Wall. Britannia 27: 129–152. London. Kent, J.P.C. 1979. The end of Roman Britain: the literary and numismatic evidence reviewed, in Casey, P.J. (ed.). 15–27. Kepler, J. 1979. The maximum duration of trading voyages from various parts of Europe to London c. 1577 Mariner’s Mirror, 65: 265–8 Keppie, L. 1982. The Antonine Wall 1960-1980. Britannia 13: 91–111 Keppie, L. 1986. Scotland’s Roman Remains. James Thin. Edinburgh. Keppie, L.J.F. 1989. Beyond the Northern Frontier: Romans and Natives in Scotland, in Todd, M. (ed.). 61–74. King, A. & Henig, M. 1981. The Roman West in the Third Century. British Archaeological International Series 109. Oxford. King, A.C. 1984. Animal Bones and the Dietary Identity of Military and Civilian Groups in Roman Britain, Germany and Gaul, in Blagg, T.F.C. & King, A.C. (eds.). 187–217. King, A.C. 1999. Animals and the Roman Army: the evidence of animal bones, in Goldsworthy, A. & Haynes, I. (eds.) 139–149 King, J.W. 1897. The Pilot’s Handbook for the English Channel. Admiralty. London. King, P. 2006. The River Teme and Other Midland Navigations. The Railway and Canal Historical Society. 348–355

191

JAMES ELLIS JONES

  Kissack, K. 1978. The River Wye. Dalton. Knight, J.K. 1988. Caerleon Roman Fortress. CADW. Cardiff. Knighton, D. 1998. Fluvial Forms and Processes; A New Perspective. Arnold. London. Kolb, A. 2001. Transport and travel on the Column of Trajan, in Adams, C. & Laurence, R. (eds.) . 95–105. Lamb, H.H. 1995. Climate, History and the Modern World (2nd edn.) Routledge. London. Lane Fox, A. 1870. On the Threatened Destruction of the British Earthworks near Dorchester. Journal of the Ethnological Society of London. Vol. 2. Langdon, J. 1986. Horses, oxen and technological innovation: the use of draught animals in English farming from 1066 to 1500. Cambridge University Press. Cambridge. Langdon, J. 1993. Inland water transport in Medieval England. Journal of Historical Geography, 19: 1–11. Langdon, J. 2000. Inland water transport in Medieval England – the view from the mills: a response to Jones. Journal of Historical Geography, 26: 75–82. Langdon, J. 2007. The Efficiency of Inland Water Transport in Medieval England, in Blair, J. 110–132 Langdon, M. & Fowler, P.J. 1971. Excavations at Crandon Bridge, Puriton. 1971. Proceedings of the Somerset Archaeological & Naturalist Society, 115: 53–4. Langouet, L. 1984. Alet and cross-Channel trade, in Macready, S & Thomson, F.H. (eds.). 66–77. Laughton, J. 1996. The port of Chester in the later Middle Ages, in Carrington, P. (ed.). 66–71. Laurence, R. 2001. The creation of geography: an interpretation of Roman Britain, in Adams, C. & Laurence, R. (eds.). 67–94. Laycock, S. 2008. Britannia - the failed state: Tribal conflict and the end of Roman Britain. Tempus. Stroud. Le Roux, P. 1994. 'La questione municipale nel I secolo d.C. l’esempio spagnolo’,Epigrafia e territorio 3. Leach, P. (ed.) 1994. Ilchester: Archaeology, Excavations and Fieldwork to 1984. Sheffield Excavation Report. Sheffield. Leech, R.H. 1977a. Romano-British rural settlement in South Somerset and North Dorset. Unpublished Ph.D. thesis (Bristol). Lehmann, 1978. The flat-bottomed Roman boat from Druten, Netherlands, International Journal of Nautical Archaeology, 7.4: 259–67. Leland, J. 1774. Itineries (edited, re-arranged and rendered into contemporary English by John Chandler 1998). Alan Sutton. Stroud. Leslie, A. 2002. Excavations on the “amphitheatre” and other areas east of Inveresk fort, in Bishop, M.C. (ed.) 41–52 Lethbridge, T.C. 1952. Boats and Boatmen. Thames & Hudson. London. Lewin, J. 1987. Historical river channel changes, in Gregory, K.J., Lewin, J. & Thornes, J.B. (eds.). 161–70. Lewin, J. 1992. Alluvium sedimentation style and archaeological sites: the Lower Vrnwy Valley, Wales, in Needham, S. & Macklin, M.G. (eds.). 103–110. Lewis, J.M. 1969. Caerphilly, in Jarrett (ed.). 64–5. Lewis, M.J.T., Slatcher, W.N. & Jarvius, P.N. 1969. Flashlocks on English waterways: a survey. Industrial Archaeology, 6: iii: 209–53. Lewis, S. (ed.) 1848. A Topographical History of England. London. Livens, R.G. 1972. Litus Hibernicum, in Pippidi, D.M. 333–40 Livingstone, H. 1995. In the footsteps of Caesar; Walking Roman Roads in Britain. Dial House. Littlehampton. Liszka, T.R. & Walker, L.E.M. (eds.) 2001. The North Sea World in the Middle Ages, Dublin.. Lloyd Jones, M. 1984. Society and settlement in Wales and the Marches 500 BC–AD 1100. British Archaeological Report 121. Oxford. Lyne, M. 1999. Roman ships’ fittings from Richborough. Journal of Roman Military Equipment Studies 7: 147-9. Lyne, M.A.B. & Jefferies, R.S. 1979. The Alice Holt/Farnham Roman pottery industry. Council for British Archaeology Research Report 30: London Lysons, S. 1807. Excavations at Caerhun. Archaeologia, 16: 127–34 Lysons, S. 1817. Coins from Kingsholm. Reliquaiae Britannico Romanae ii. Macdonald, G. & Curle, A.O. 1929. The Roman fort at Mumrills, near Falkirk. Proceedings of the Society of Antiquaries of Scotland 63: 396–575. Macdonald, G. 1932. Notes on the Roman forts at Old Kilpatrick and Croy Hill', Proceedings of the Society of Antiquaries of Scotland, 66: 243-96

Macdonald, G. 1934. The Roman Wall in Scotland. Oxford. Macfarlane, W. 1723. Geographical Collections relating to Scotland, I. Scottish Historical Society. Edinburgh Mackenzie, E. 1827. The River Tyne, Historical Account of Newcastle-upon-Tyne: including the Borough of Gateshead. 736–44 Mackinder, Sir H.J. 1902. Britain and the British Seas. Appleton and Co. New York. Macready, S. & Thompson, F.H. 1984. Cross-Channel Trade between Gaul and Britain in the Pre-Roman Iron Age. Society of Antiquaries, Occasional Paper IV. London. Maitland, W. 1757. History and Antiquities of Scotland. London. Manley, J. 2002. AD 43 The Roman Invasion of Britain: a reassessment. Tempus. Brimscombe Port. 192

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Mann, J.C. 1974a. The frontiers of the Principate, in Temporini, H. (ed.). 508–33. Mann, J.C. 1985. Two ‘Topoi’ in the Agricola. Britannia, 16: 21–4. Mann, J.C. 1989. Birdoswald to Ravenglass. Britannia, 20. 75–80. Mann, J.C. 1988. The history of the Antonine Wall – a reappraisal. Proceedings of the Society of Antiquaries of Scotland 118: 131–7 Mann, J.C. 1996. Britannia and the Roman Empire. Aldershot. Mann, J.C. & Jarrett, M.G. 1967. The Division of Britain. Journal of Roman Studies 57: 61–64 Manning, W.H. 1969. Usk, in Jarrett M.G. (ed.). 116–8. Manning, W. H. 1976. The Conquest of the West Country, in Branigan, K. & Fowler, P. (eds.). 15–41. Manning, W.H. 1975. Economic influences on land use in the military areas of the Highland Zone during the Roman Period, in Evans, J.G. et al. (eds.). Manning, W.H. 1979. The Native and Roman contribution to the development of Metal Industries in Britain, in Burnham, B.C. & Johnson, H.B. (eds.). 111–122. Manning, W.H. 1981. Report on the Excavations at Usk 1965–76: The Fortress Excavations 1968–71. Board of Celtic Studies of the University of Wales. Cardiff. Manning, W.H. 1993. Report on the Excavations at Usk 1965–76: The Roman Pottery. Board of Celtic Studies of the University of Wales. Cardiff. Manning, W.H. 2000. The fortresses of legio XX, in Brewer, R.J. (ed.). 69–82. Manning, W.H. 2001. Roman Wales. University of Wales Press. Cardiff. Manning, W.H. 2002. Roman Fortress Studies: Present Questions and Future Trends, in Carrington, P. (ed.). 1–6. Manning, W.H. 2003. The Defences of Roman Caerwent, in Wilson, P. (ed.). 69–82. Manning, W.H. & Scott, I.R. 1979. Roman Timber Military Gateways. Britannia, 19: 19–62. Manning, W.L. 1997. Ptolemy and the pre-Flavian military sites of Britain, in Waateringe et al. (eds.). 33–40. Marcus, G.J. 1953. Factors in early Celtic navigation. Etudes Celtiques, 6: 312–27. Marcus, G.J. 1980. The Conquest of the North Atlantic. Woodbridge. London. Margary, I.D. 1967. Roman Roads in Britain (rev. edn.). J. Baker. London. Markey, M. 1997. An inscribed stone anchor from Dorset. International Journal of Nautical Archaeology, 26.2: 127–32. Markey, M., Wilkes, E. & Darvill, T. 2002. Poole Harbour: an Iron Age port, Current Archaeology, 181: 7–11. Marriner, N. & Morhange, C. 2006. Geoarchaeological evidence for dredging in Tyre’s ancient harbour, Levant. Quarternary Research 65: 164–171 Marsden, P. 1980. Roman London. Thames & Hudson. London. Marsden, P. 1990. A re-assessment of Blackfriars Ship 1, in McGrail. S. (ed.) 66–74 Marsden, P. 1977. Celtic Ships of Europe, in McGrail, S. (ed.) Marsden, P. 1994. Ships of the Port of London; first to eleventh centuries AD. English Heritage. London. Marsden, P. 1997. Ships and Shipwrecks. Batsford. London. Marsden, P. (ed.) 1987. The Roman Forum site in London. Batsford. London Markey, M. 1991. Two stone anchors from Dorset. International Journal of Nautical Archaeology, 20.1: 47-51. Martin, A.T. 1886. On the Roman Road between Bath and Caerwent. Proceedings of the Clifton Antiquarian Club, ii: 63–4. Martin, A.T. & Tratman, E.K. 1923. Excavations at Sea Mills, Bristol. Transactions of the Bristol and Gloucester Archaeological Society, 23: 193–201. Martin, C. 1992. Water Transport and the Roman Occupations of North Britain, in Smout, T.C. (ed.). 1–34. Martin, C.J.M. 2000. A Maritime Landscape in East Fife, in Aberg, A. & Lewis, C. (eds.). 39–50 Martin, C.J.M. 2001. Seafaring and Trade in East Fife, in Liszka, T.R. & Walker, L.E.M. (eds.). 164–174 Martin-Kilcher, S 1987. Die romischen Amphoren aus Augst und Kaiseraugst; Ein Beitrag zur romischen Handels - und Kulturgesh. Forschungen. Augst. Martins, C. 2003. Becoming consumers: looking beyond wealth as an explanation for villa variability, in Carr, G. et al. (eds.) 84–100. Marvell, A. 1991. Recent work on the Neronian fortress at Usk, 1986–8, in Burnham, B.C. & Davies J.L. (eds.) 19–26. Marvell, A.G. 1999. Excavations at Usk 1986–1988. Britannia, 27: 51–110. Marvell, A,G. & Heywood, B. 1992. Excavations at Neath. Bulletin of the Board of Celtic Studies, 39: 171-298. Marvell, A.G. & Owen-John, H.S. 1997. Leucarum: Excavations at the Roman auxiliary fort at Loughor, West Glamorgan 1982–84 and 1987–88. Britannia Monograph Series 12. CADW. Cardiff. Mason, D.J.P. 1987. Chester: the canabae legionis. Britannia, 18: 143–168 Mason, D.J.P. 1988b. Prata legionis in Britain. Britannia, 19: 163–89. Mason, D.J.P. 2002. The Town and Port of Roman Chester, in Carrington, P. (ed.). 53–74. Mason, D.J.P. 2002. The Construction and Operation of a Legionary Fortress: Logistical and Engineering Aspects, in Carrington, P. (ed.). 89–112. Mason, D.J.P. 2002. The Foundation of the Legionary Fortress: The Flavians and Imperial Symbolism, in Carrington, P. (ed.). 33–52. 193

JAMES ELLIS JONES

  Mason, D.J.P. 2002. The Heronbridge Research Project: First Interim Report. Chester. Mason, D. J. P. 2002. Roman Chester : city of the eagles. Tempus. Brimscombe Port. Mason, D.J.P. 2003. Roman Britain and the Roman Navy. Tempus. Stroud. Masser, P. & Evans, J. 2005. Excavations within the vicus settlement at Burgh by Sands, 2002. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 5: 31–63. Masser, P. 2006. Cramond Roman Fort: evidence from excavations at Cramond Kirk Hall, 1998 and 2001. Scottish Archaeological Internet Report 20. Edinburgh. Matthews, J. 2009. A Roman road from Loxton towards Banwell through the Lox-Yeo valley. CBA South West. Matthews K.J. et al. 1996. Excavations at Chester, the evolution of the heart of the city: investigations at 3–15 Eastgate Street 1990/1. Chester City Council. Chester. Matthews, K. 1996. Iron Age sea-borne trade in Liverpool Bay, in Carrington, P. (ed.). 12–23. Mattingly, D. J. 2006. An Imperial possession: Britain in the Roman Empire, 54 BC–AD 409. Penguin. London. Mattingly, H. & Sydenham, E.A. et al (eds.) 1923. The Roman Imperial Coinage. Spink. London. Maxfield, V.A. 1979. Camelon 'South Camp': excavations 1975-79, in Breeze, D.J. (ed.). 28-32. Maxfield, V.A. (ed.). 1989. The Saxon Shore: A Handbook. Exeter Studies in History 25. Exeter. Maxfield, V.A. 1986. Roman forts and their garrisons. Britannia, 17: 59–72. Maxfield, V.A. 1987. The Army and the Land in the Roman South West, in Higham (ed.). 1–26. Maxfield, V.A. 1989. Conquest and Aftermath, in Todd, M. (ed.). 19–29. Maxfield, V.A. 1995. Soldier and Civilian; Life beyond the Ramparts. Roman Legionary Museum. Caerleon. Maxfield, V. & Dobson, G. (eds.). 1991. Roman Frontier Studies 1989: 15th International Congress of Roman Frontier Studies. University of Exeter. Exeter. Maxwell, G.S. 1989. The Romans in Scotland. Mercat Press. Edinburgh. Maxwell, G.S. 1998. A gathering of eagles: scenes from Roman Scotland. Edinburgh Maxwell, G.S. 1990. A Battle Lost: Romans and Caledonians at Mons Graupius, Edinburgh. May, T. 1987. Economic and Social History of Britain, 1760–1970. Longman. London. McCarthy, M.R. 2002. Roman Carlisle and the Lands of the Solway. Tempus. Stroud. McCarthy, M. 2003. Luguvalium (Carlisle); a civitas capital on the northern frontier, in P. Wilson (ed.). The Archaeology of Roman Towns. 145–55. Oxbow Books . Oxford. McCarthy, M. 2005. Social Dynamics on the Northern Frontier of Roman Britain. Oxford Journal of Archaeology, 24: 47–71 McCord, N. & Tait, J. 1978.'Excavations in the northern Annexe of the Roman Fort at Camelon, near Falkirk, 1961-3. Proceedings of the Society of Antiquaries of Scotland, 109: (1977-8), 151-65. McDonald, G. 1932. Notes on the Roman Forts at Old Kirkpatrick and Croy, and on a Relief of Jupiter Dolichenus. Proceedings of the Society of Antiquaries of Scotland, Lxvi: 220 ff. McGrail, S. 1978. Logboats of England and Wales. National Maritime Museum. Greenwich. McGrail, S. 1979. Archaeology of Medieval Ships and Harbours in Northern Europe. British Archaeological Report International Series 66. Oxford. McGrail, S. 1981. The Ship: Rafts, Boats and Ships from Prehistoric Times to the Medieval Era. Her Majesties Stationery Office. London. McGrail, S. 1981. The Brigg ‘raft’ and her prehistoric environment. British Archaeological Report 89. Oxford. McGrail, S. 1983. Cross-Channel seamanship in the late first millennium BC. Oxford Journal of Archaeology, 2: 299–337. McGrail, S. (ed.). 1984. Aspects of Maritime Archaeology and Ethnography. National Maritime Museum. Greenwich. McGrail, S. 1985. Early Landing Places, in Herteig, A.E. (ed.). 12–18. McGrail, S. 1989. The shipment of traded goods and of ballast in antiquity. Oxford Journal of Archaeology, 8.3: 353–8. McGrail, S. (ed.) 1990. Maritime Celts, Frisians and Saxons. Council for British Archaeology Research Report 71. Oxford. McGrail, S. 1992. Replicas, reconstructions and floating hypotheses, International Journal of Nautical Archaeology, 21: 353–5. McGrail, S. 1993. Medieval boat and ship timbers from Dublin. Royal Irish Academy. Dublin. McGrail, S. 1995. Celtic Seafaring and Transport, in Green, M.J. (ed.). 254–84. McGrail, S. 1996. Ancient Boats in Europe: the archaeology of water transport to AD 1500. Longman. London. McGrail, S. & Roberts, O. 1999. A Romano-British boat from the shores of the Severn Estuary. The Mariners Mirror, Vol. 85, 2: 133–146. McGrail, S. 2001. Boats of the World. Oxford University Press. Oxford. McGrail, S. 2005. Foreword, in Grainge 2005. 7 McKee, E. 1996. Working Boats of Britain: their shape and purpose. Conway. London. McNeil, R. & Roberts, A.F. 1987. A Plank Tank from Nantwich. Britannia, 18: 287–95.

194

THE MARITIME LANDSCAPE OF ROMAN BRITAIN McWhirr, A. 1981. Roman Gloucestershire. Alan Sutton. Stroud. McWhirr, A. 1986. Houses in Roman Cirencester. Cirencester Excavation Committee. Cirencester. Meadows, K. et al. (eds.). 1997. Theoretical Roman Archaeology Conference (6th Sheffield). Oxbow. Oxford. Meijer, F. & van Nifj, O. 1992. Trade, transport and society in the ancient world: a sourcebook. Routledge. London. Middleton, P.S. 1979. Army Supply in Roman Gaul: a hypothesis for Roman Britain, in Burnham, B.C. & Johnson, H.B. (eds.). 81–99. Middleton, P. 1983. The Roman Army and Long Distance Trade, in Garnsey, P. & Whittaker, C.R. (eds.). 75-83. Miket, R. & Burgess, C. (eds.) 1984. Between and Beyond the Walls: Essays on the Prehistory and History of North Britain in Honour of George Jobey. Edinburgh Miles, D. (ed.). 1982. The Romano-British Countryside: Studies in rural settlement and economy. British Archaeological Report 103. Oxford. Miles, D. 1984. Archaeology at Barton Court Farm, Abingdon, Oxon. Council for British Archaeology Research Report 50. Oxford. Miller, L. 1986. The Roman quay at St. Magnus House, London :excavations at New Fresh Wharf, Lower Thames Street, London, 1974–78. London & Middlesex Archaeological Society. London. Miller, S.N. 1928. The Roman Fort at Old Kilpatrick. Society of Antiquaries of Scotland. Edinburgh. Millet, M. 1979. The dating of Farnham pottery. Britannia, 10: 121–38. Millett, M.J. 1984. Forts and the origins of towns: cause or effect? in Blagg & King (eds.) 65–75. Millett, M.J. 1990. The Romanisation of Britain. Cambridge University Press. Cambridge. Millett, M.J. 1995. Roman Britain. Batsford. London. Millett, M. & McGrail, S. 1987. Archaeology of the Hasholme logboat Archaeological Journal, 144: 69–155. Millett, M. & Wilmott, T. 2003. Rethinking Richborough, in Wilson, P. (ed.) 184–194. Milne, G. 1985. Roman Waterfront Archaeology in London, in Herteig, A.E. (ed.). 31–45. Milne, G. 1985. The port of Roman London. Batsford. London. Milne, G. 1990. Maritime traffic between the Rhine and Roman Britain: A preliminary note, in McGrail, S. (ed.). 82–4. Milne, G. 1993. The Rise and fall of Roman London, in Greep, S.J. (ed.). 11–15. Milne, G. 1995. Roman London. English Heritage. London. Milne, G. 1995. Foreshore Archaeology. Royal Commission for Historic Monuments England. London. Milne, G. 1996. Blackfriars ship 1: Romano-Celtic, Gallo-Roman or Classis Britannica. International Journal of Nautical Archaeology, 25: 234–6. Milne, G. with Bates, M. & Webber, M.D. 1997. Problems, Potential and Partial Solutions: An Archaeological Study of the Tidal Thames, England. World Archaeology. 29: 130–146 Milne, G. 2000. A Roman Provincial Fleet: The Classis Britannica reconsidered, in Oliver, G. et al. (eds.). 127–31. Milne, G. & Hobley, B. 1981. Waterfront Archaeology in Britain and Northern Europe. Council for British Archaeology Report 41. Mitchell, S. 1976. Requisitioned Transport in the Roman Empire. Journal of Roman Studies, 66: 106–131. Moir, D.M. 1860. The Roman Antiquities of Inveresk. Edinburgh. Mommsen, T. et al. 1985. The Digest of Justinian. University of Pennsylvania Press. Philadelphia. Monaghan, J. 1991. Pottery from marine sites around Guernsey. Journal of Roman Pottery Studies, iii: 63–69. Moore, D. (ed.). 1970. The Irish Sea Province in Archaeology and History. Cambrian Archaeological Association. Cardiff. Morris, C.D., Batey, C.E., Brady, K., Harry, R, Johnson, P.G & Thomas, A.C. 1999. Recent Work at Tintagel. Medieval Archaeology, 43: 206–15. Morris, E.L. 1994, Production and Distribution of Pottery and Salt in Iron Age Britain: a Review. Proceedings of the Prehistoric Society, 60: 371–393 Morris, E.L. 2007. Making magic: later prehistoric and early Roman salt production in the Lincolnshire fenland, in Haselgrove, C. & Moore, T. (eds.) 430–43. Morris, F.M. 2010. North Sea and Channel Connectivity during the Late Iron Age and Roman Period (175/150 BC – AD 409). BAR International Series 2157. BAR Publishing. Oxford. Morris, J. 1982. Londinium: London in the Roman Empire. Weidenfeld & Nicolson. London. Morris, P. 1979. Agricultural Buildings in Roman Britain. British Archaeological Report 70. Oxford. Morrison, J. 1980. The Ship: Long and Round. Her Majesties Stationery Office. London. Morrison, J. (ed.). 1995. The Age of the Galley; Mediterranean Oared Vessels since Pre-Classical Times. Conway. London. Muckelroy, K. 1978. Maritime Archaeology. Cambridge University Press. Cambridge. Muckelroy, K. (ed.). 1980. Archaeology Underwater. McGraw-Hill. New York & London. Murphy, J.P. 1977. Ora Maritima or Description of the Seacoast. Chicago. Murphy, K. 2002. The archaeological resource: chronological overview to 1500 AD, in Davidson, A. (ed.) 5–64.

195

JAMES ELLIS JONES Murphy, P. 2009. The English Coast: a History and a Prospect. Continuum. London. Muskett, G. & Georgiadis, M. (eds.). 2002. The Seas in Antiquity. University of Liverpool. Liverpool. Musson, C. 1972. Two Winters at the Breidden. Current Archaeology, 10: 263–7. Musson, C. 1994. Wales from the Air. Royal Commission for Archaeological & Historic Monuments - Wales. Cardiff. Musson, C.R. 1981. Prehistoric and Romano-British settlement in Powys and Shropshire. Archaeological Journal, 138: 5–7. Mytum, H.C. 1982. Rural Settlement of the Roman Period in North and East Wales, in Miles, D. (ed.). 313–35. Naish, J. 1985. Seamarks: their history and development. Stanford Maritime. London. Nash, S. 1972/3. A deep-water inlet at Highbridge: a precis of a paper. Proceedings of the Somerset Archaeological and Natural History Society, 117: 97–101. Nash-Williams, V.E. 1954. The Roman Frontier in Wales. University of Wales Press. Cardiff. Nash-Williams, V.E. 1939. An Early Iron Age coastal camp at Sudbrook, near the Severn Tunnel, Monmouthshire. Archaeologia Cambrensis, 94: 42–79. NAVIS I & NAVIS II database. http://www2.rgzm.de/navis/home/frames.htm Nayling, N. (ed.) 1998. The Magor Pill Medieval wreck. Council for British Archaeology Research Report 115. York. Nayling, N. Maynard, D.R. & McGrail, S. 1994. Barland’s Farm, Magor, Gwent: a Romano-Celtic boat. Nayling, N. & McGrail, S. 2004. Barland’s Farm Romano-Celtic Boat. Council for British Archaeology Research Report 138. York. Needham, S. & Macklin, M.G. (eds.). 1992. Alluvial Archaeology in Britain. Oxbow Monograph 27. Oxford. Neu, S. 1985. The Rhine Frontage of Roman Cologne, in Herteig, A.E. (ed.). 151–6. Nightingale, K.R. 1954. Trial excavations at Sea Mills, 1954. Transactions of the Bristol and Gloucester Archaeological Society, 73: 70–72. North, F.J. 1964. The Evolution of the Bristol Channel. National Museum of Wales. Cardiff. O’Connor, C. 1993. Roman Bridges. Cambridge University Press. Cambridge. O’Dwyer, S. 1937. The Roman Roads of Wales. Montgomery Publishing. Newtown. O’hEailidhe, P. 1992. The Monk’s Boat – a Roman period relic from Lough Lene, Co. Westmeath, Eire. International Journal of Nautical Archaeology, 21.2: 185–90. Ogilvie, R.M. & Richmond, I.A. 1967. Cornelii Taciti de Vita Agricolae. Oxford University Press. Oxford. Oliver, G. et al. (eds.). 2000. The Sea in Antiquity. British Archaeological Report, International Series 899. Oxford. Oppenheim, M.M. 1896. A history of the administration of the Royal Navy and of merchant shipping in relation to the Navy. Vol. 1, From MDIX to MDCLX, with an introduction treating of the preceding period. John Lane. London. Oppenheim, M.M. 1926. The maritime history of Devon. Exeter University Press. Exeter. Ordnance Survey 1956. Map of Roman Britain (3rd Edn.). Chessington. Ordnance Survey 1978. Map of Roman Britain (4th Edn.). Southampton. Ordnance Survey 2001. Map of Roman Britain (5th Edn.). Southampton. Osborn, G. 2006. Hadrian’s Wall and its People. Bristol Phoenix Press. Bristol. Ottaway, P. 1993. Roman York. Batsford/English Heritage. London. Owen-John, H. 1991. Recent excavations at the Roman auxiliary forts at Loughor and Neath, West Glamorgan, in Burnham, B.C. & Davies J.L. (eds.). 40–8. Pahl, R.E. 1975, Whose city? And further essays on urban society. Harmondsworth. London. Parfitt, K. 1993. The Dover Boat. Current Archaeology, 133: 4–8. Palmer, C. 2009. Reflections on the Balance of Traditional Sailing Vessels. International Journal of Nautical Archaeology, 38.1: 90–6. Parker, A.J. 1973. Evidence provided by underwater archaeology for Roman trade in the Western Mediterranean, in Blackman, D. (ed.). 361–80. Parker, A.J. 1980. Mediterranean wreck sites and Classical seafaring, in Muckelroy, K. (ed.). 32–60. Parker, A. J. and Price, J., 1981. Spanish exports of the Claudian period: the significance of the Port-Vendres II wreck reconsidered. International Journal of Nautical Archaeology, 10: 221–8. Parker, A.J. 1984. Shipwrecks and Ancient Trade in the Mediterranean. Archaeological Review from Cambridge, 3: 99–113. Parker, A.J. 1990. Classical Antiquity: the maritime dimension. Antiquity, 64: 335–46. Parker, A.J. 1992a. Cargoes, containers and stowage: the ancient Mediterranean, international Journal of Nautical Archaeology, 21: 89–100. Parker, A.J. 1992b. Ancient Shipwrecks of the Mediterranean and Roman Provinces. British Archaeological Report International Series 580. BAR Publishing, Oxford. Parker, A.J. 1995. Maritime cultures and wreck assemblages in the Graeco-Roman world. International Journal of Nautical Archaeology, 24.2: 87–95. Parker, A.J. 1996. Sea transport and trade in the ancient Mediterranean, in Rice, E.E. (ed.). 97–109. Parker, A. J. 2001. Maritime Landscapes. Landscapes, 1: 22–41. Parker, A.J. 2008. Artefact Distributions and Wreck Locations: the archaeology of Roman commerce, in Hohlfelder (ed.). 177–196. 196

THE MARITIME LANDSCAPE OF ROMAN BRITAIN Parkhouse, J. 1981b. Cowbridge. Current Archaeology, 81: 308–9. London Parkhouse, J. 1982a. Excavations in Cowbridge. Annual Report, Glamorgan-Gwent Archaeological Trust 1981–2: 71–2. Cardiff. Parkhouse, J. 1991. Cowbridge: Village or ‘Small Town’?, in Burnham, B.C. & Davies J.L. (eds.). 94–102. Parsons, D. 1990. Stone: Quarrying and Building in England AD 43~1525. Phillimore. Chichester. Parsons, D. 1990. The Stone Industry in Roman, Anglo-Saxon and Medieval England, in Parsons, D. (ed.). 1–15. Partridge, C.R. 1975. Braughing, in Rodwell, W. & Rowley, T. Peacock, D.P.S. 1978. The Rhine and the problems of Gaulish wine in Roman Britain, in Taylor, J du P & Cleere, H. (eds.). 49–51. Peacock, D.P.S. 1982. Pottery in the Roman world: an ethnoarchaeological approach. Longman. London. Peacock, D.P.S. & Williams, D.F. 1986. Amphorae and the Roman Economy. Longman. London. Peacock, D. & Blue, L. (eds.). 2006a. Myos Hormos—Quseir al-Qadim: Roman and Islamic Ports on the Red Sea. Volume 1:Survey and Excavations 1999–200. Oxford. Pearson, A.F. 2002. The Roman Shore Forts: Coastal Defences of Southern Britain. Tempus. Stroud. Pearson, A.F. 2003. The Construction of the Saxon Shore Forts. British Archaeological Report, British Series 349. BAR Publishing. Oxford. Pearson, A.F. 2006. Piracy in Late Roman Britain: a perspective from the Viking Age. Britannia , 37: 337–53. Pearson, A.F. 2006. The Work of Giants. Tempus. Stroud. Pearson, A.F. 2005. Barbarian Piracy and the Saxon Shore: a reappraisal. Oxford Journal of Archaeology, 24 (1): 73–88. Peddie, J. 1987, Invasion: The Roman Conquest of Britain. Guild Publishing. London. Peddie, J. 2004. The Roman War Machine. Sutton Publishing. Stroud. Pelham, H.F. 1896. Arrian as Legate of Cappodocia. English Historical Review, Vol. 11, 44: 625–640. Peña J. T. 2007. Roman pottery in the archaeological record. Cambridge University Press. Cambridge. Penhallurick, R.D. 1986. Tin in Antiquity: Its Mining and Trade throughout the Ancient World with Particular Reference to Cornwall. Institute of Metals. London Perkins, D. 2006. Prehistoric Maritime Traffic in the Dover Straits and Wantsum. Archaeologia Cantiana. Petts, G.E. 1980. Identification and Interpretation of River Channel Changes. Loughborough Press. Loughborough. Phang, S.E. 2008. Roman Military Service: Ideologies of Discipline in the Late Republic and Early Principate. Cambridge University Press. Cambridge. Phillips, C. 2006. Portages in Early Medieval Scotland. The Great Glen roué and the Forth-Clyde Isthmus in Westerdahl, C. 191–8 Phillips, C.W. (ed.) 1970. The Fenland in Roman Times. Royal Geographical Society, Ser. 5. Philp, B.J. 1981. The Excavation of the Roman Forts of the Classis Britannica at Dover 1970–77. Kent Monograph Series No 3. Dover. Philp, B. 1996. The Roman Fort at Reculver. Kent Archaeological Rescue Unit. Philpott, R. (ed.) 2004. The Romano-British Period. North-west Region Archaeological Research Framework, Romano-British Resource Assessment. Liverpool. Philpott, R.A. 1991a. Merseyside in the Roman Period. Journal of the Merseyside Archaeological Society, 7: 61–74. Pike, R.J. & Langdon, M. 1981. A Romano-British site at Combwich Passage. Report of the Bridgwater and District Archaeological Society, 1: 4–13. Pippidi, D.M. 1972. D’Études sur les Frontières Romaines Editura Academiei Bucureşti România. Bucharest. Pirazzoli, P.A. 1991. World Atlas of Holocene Sea-Level Changes. Elsevier. Amsterdam. Pitts, L.F. & St. Joseph, J.K. 1985. Inchtuthil; the Roman legionary fortress excavations, 1952–65. Britannia Monograph Series 6. London. Place, G. 1996. Parkgate and Ireland, in Carrington, P. (ed.). 72–4. Platt, C. 1976. The English Medieval Town. Book Club Associates. London. Plumb, J.H. 1955. Studies in Social History: a tribute to G.M. Trevelyan. Longmans, Green. London. Porteous, J. D. 1977 . Canal ports : the urban achievement of the canal age . Academic Press. London. Potter, C.W. & Johns, C. 1992. Roman Britain. British Museum Press. London. Potter, T.W. 1975. Excavations at Bowness-on-Solway, 1973. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 73: 29–57. Potter, T.W. 1977a. Excavations on the Roman Fort at Watercrook, 1975. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 77: 49–52. Potter, T. W. 1979. Romans in England: Excavations at the Roman forts of Ravenglass, Watercrook and Bowness on Solway. Cumberland and Westmorland Antiquarian and Archaeological Society, Research Series I. Kendal.

197

JAMES ELLIS JONES Poulter, J. 2009. Surveying Roman Military Landscapes across Northern Britain. The Planning of Roman Dere Street, Hadrian’s Wall and the Vallum, and the Antonine Wall in Scotland. British Archaeological Report 492. BAR Publishing. Oxford. Pounds, N.J.G. 1990. Buildings, Building Stones and Building Accounts in England, in Parsons, D. (ed.). 228–37. Pownall, T. 1778. Memoire on the Roman Earthenware fished up within the Mouth of the River Thames. Archaeologia, 5: 282-90. Pratt, E.A. 1912. A History of Inland Transport and Communication in England. Kegan Paul, Trench, Trubner & Co. London. Price, J. & Wilson, P.R. (eds.) 1988. Recent research in Roman Yorkshire. Studies in honour of Mary Kitson Clark British Archaeological Report 193. BAR Publishing. Oxford. Priestley, J. 1831. Historical Account of the Navigable Rivers, Canals and Railways of Great Britain. Longman, Rees, Orme, Brown and Green. London. Proudfoot, E.V.W. 1980a. Camelon Native Site. Proceedings of the Society of Antiquaries of Scotland 109: 112– 28. Pryme, A. de la. 1700. A Letter from the Reverend Mr Abraham de la Pryme, to the Very Reverend Dr G. D. of Y. and F. R. S. concerning Some Roman Antiquities in Lincolnshire'. Philosophical Transactions, 22:. 561–567 Pucci, G. 1983. Pottery and trade in the Roman period, in Garnsey, P. et al. 105–117. Putnam, W.G. 1969. Caer Gybi, in Jarrett (ed.). 135–7. Quinnell, H. 1986. Cornwall during the Iron Age and Roman Period. Cornish Archaeology, 25: 111–34. Quinnell, H. 2004. Trethurgy: Excavations at Trethurgy Round, St Austell: Community and Status in Roman and post-Roman Cornwall. Cornwall County Council. Truro. Rae, A. & Rae, V. 1974. The Roman fort at the Cramond, Edinburgh: excavations 1954-1966. Britannia 5: 163–224 Rahtz, P., Woodward, A. et al. 1992. Cadbury Congresbury: a Late/post Roman Hilltop. British Archaeological Report 223. Rahtz, P. 1986. Post Roman Avon, in Aston, M. & Iles, R. (eds.). 73–83. Rahtz, P. 2000. Cannington cemetery : excavations 1962-3 of prehistoric, Roman, post-Roman, and later features at Cannington Park Quarry, near Bridgwater, Somerset. Society for the Promotion of Roman Studies. London. Rea, J.T. 1902. How to estimate, being the analysis of builders prices. London. Reddaway, T.F. & Ruddock, A.A. 1969. The accounts of John Balsall, purser of the Trinity of Bristol, 1480–1. Camden Miscellany, 23: 1–27. London. Reddé. M. 2000. Les Marins, in Alföldy, G. Dobson, B. & Eck. W. 179–92. Reece, R. 1979. Roman Monetary Impact, in Burnham, B.C. & Johnson, H.B. (eds.). 211–20. Reece, R. 1980. Town and country: the end of Roman Britain. World Archaeology 12.1: 77–92. Reece, R. 1987. Coinage in Roman Britain. Seaby. London. Reece, R. 1988. My Roman Britain. Cotswold. Cirencester. Reece, R. 1991. Roman Coins from 140 Sites in Britain. Cotswold. Cirencester. Reece, R. 1994. 353, 367 or 357? Splitting the difference or taking a new approach? Britannia, 25: 236–8. Reece, R. 1995. Site finds in Roman Britain. Britannia, 26: 179–206. Reece, R. 1997. The Future of Roman Military Archaeology. Tenth Annual Caerleon Lecture. National Museums and Galleries of Wales. Cardiff. Reece, R. 1999. Colonia in context: Glevum and the civitas Dobunorum in Hurst, H.R. (ed.). 73–87. Reece, R. 2002. The Coinage of Roman Britain. Tempus. Brimscombe Port. Reynolds, P.J. 1979. Iron Age Farm. The Butser Experiment. London. Reynolds, P.J. 1992. Crop yields of the Pre-historic Cereal types Emmer and Spelt: the Worst Option. Butser Ancient Farm Occasional Papers 2: 17–28. Reynolds, P.K.B. 1930. Excavations on the site of the Roman fort at Caerhun: fourth interim report. Archaeologia Cambrensis, 85: 74–102. Rhodes, E. 2007. Identifying Human Modification of Water Channels, in Blair, J. 133–152 Rice, E.E. (ed.) 1996. The Sea and History. Sutton. Stroud. Richardson, A. 2004. Granaries and Garrisons in Roman Forts. Oxford Journal of Archaeology 23: 429–442, Richardson. C. 1990. The Fell Pony. J.A. Allen. London. Richmond, I.A. 1922. Ptolemaic Scotland. Proceedings of the Society of Antiquaries of Scotland 56: 288–301. Richmond, I.A. 1935. Trajan’s army on Trajan’s Column. Papers of the British School at Rome 13: 1–40. Richmond, I.A. 1943. Recent discoveries in Roman Britain from the air and in the field. Journal of Roman Studies, 33: 53. Richmond, I.A. 1945. The Sarmatae, Bremetennacum Veteranorum and the Regio Bremetennacensis. Journal of Roman Studies, 35: 14–29. Richmond, I.A. 1949. The Roman road from Ambleside to Ravenglass. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 49: 15–31. Richmond, I.A. 1963. Roman Britain (2nd Edn.). Pelican. Harmondsworth. 198

THE MARITIME LANDSCAPE OF ROMAN BRITAIN Richmond, I.A. 1966. Industry in Roman Britain, in Wacher, J. (ed.). 76–86. Richmond, I.A. 1980. A Roman fort at Inveresk, Midlothian, (ed. Hanson, W.S.). Proceedings of the Society of Antiquaries of Scotland 110: 286–304. Richmond I.A. & Crawford O.G.S. 1949. The British section of the Ravenna Cosmography. Archaeology, 93: 1–50. Richmond, I.A. (ed.). 1958. Roman and Native in North Britain. Nelson. Edinburgh. Rickman, G. 1971. Roman Granaries and Store Buildings. Cambridge University Press. Cambridge. Rickman, G. 1980. The corn supply of Ancient Rome. Oxford University Press. Oxford. Riley, D.N. 1974. The end of Ermine Street at the south shore of the Humber. Britannia, 5: 375–7. Riley, H. & Wilson-North, R. 2001. The field archaeology of Exmoor. English Heritage. Swindon. Riley, W. & Gomme, L. 1912. Ship of the Roman period discovered on the site of the new County Hall. London. Rippon, S. 1995c. Roman settlement and salt production on the Somerset coast: the work of Samuel Nash – a Somerset archaeologist and historian 1913–1985. Somerset Archaeological and Natural History Society, 139: 99–117. Rippon, S. 1996a. The Gwent Levels: the exploitation of a wetland landscape. Council for British Archaeology Research Report 105. Oxford. Rippon, S. 1997a. The Severn Estuary: landscape evolution and wetland reclamation. Leicester University Press. Leicester. Rippon, S. 2000. The Transformation of Coastal Wetlands. Oxford University Press. Oxford. Rippon, S. 2008. Coastal Trade in Roman Britain: the Investigation of Crandon Bridge, Somerset, a Romano-British Trans-shipment Port beside the Severn Estuary. Britannia, 39: 85–144. Rippon, S.J. (ed.). 1990–2002. Archaeology in the Severn Estuary, Volumes 1–10. Severn Estuary Levels Research Committee. Exeter. Rivet, A.L.F. 1964. Town and Country in Roman Britain, (2nd edn.) Hutchinson. London. Rivet, A.L.F. & Smith, C. 1979. The Place-names of Roman Britain. Batsford. London. Robertson, A.S. 1970. Roman Finds from Non-Roman Sites in Scotland. Britannia, 1: 198–226. Robertson, A.S. 1970. Roman Coins found in Scotland, 1961–70. Proceedings of the Society of Antiquaries of Scotland : 113–168 Robertson, A.S. 1975. Coarse Ware, in Robertson, Scott, & Keppie (eds.). 138–167. Robertson, A.S. 2000. An inventory of Romano-British Coin Hoards. London. Robertson, A.S. & Keppie, L. 2001. The Antonine Wall: A handbook to the surviving remains. Glasgow Archaeological Society. Glasgow. Robertson, A.S., Scott, M.E. & Keppie, L.J.F. 1975. Bar Hill: a Roman Fort and its Finds. British Archaeological Report 16. Oxford. Robinson, A.H.W. 1962. Marine Cartography in Britain. A History of the Sea Chart to 1855. Leicester University Press. Leicester. Robinson, D.M. 1980. Cowbridge: The archaeology and topography of a small market town in the Vale of Glamorgan. Gwent-Glamorgan Archaeological Trust. Swansea. Rodriguez, J.R. 2002. Baetica and Germania. Notes on the concept of Provincial Independence in the Roman Empire, in Erdkamp, P. (ed.) J.C. Geiben. Amsterdam. Rodwell, W. & Rowley, T. 1975. The Small Towns of Roman Britain. British Archaeological Report 15. Oxford. Rodwell, W.J. (ed.), 1980. Temples, churches and religion in Roman Britain. Batsford. London. Rogers, I. 1996. The Conquest of Brigantia and the development of the Roman Road System in the North-West. Britannia, 27: 365–8. . Rogers, I. 1998: ‘Walton le Dale’, Britannia, 29, 388–90 Rolt, L.T.C. 1969. Navigable Waterways. Longman. London. Roth, J.P. 1998. The logistics of the Roman army at war: (264 BC – AD 235). Brill. Leiden, Boston, Koln. Rowley, T. 1985. Roman Dorchester, in Cook, J. & Rowley, T. (eds.) Roxan, M.M. 1997. Settlement of veterans of the auxillia - a preliminary study, in Waateringe et al. (eds.). 483–92. Roxan, M.M. 2000. Veteran Settlement of the Auxillia in Germany, in Alföldy, G. Dobson, B. & Eck. W. 307–26. Roy, W. 1793. The Military Antiquities of the Romans in Britain. London. Rudloff, H. von, 1967. Die Shwankungen und Pendelungen des Klimas in Europa seit dem Beggin del regelmaessigen Instrumenten-Beobachtunge Braunschweig Rule, M. & Monaghan, J. 1993. A Gallo-Roman Trading Vessel from Guernsey: The Excavation and Recovery of a Third Century Shipwreck. Guernsey Museum Monograph No 5. Guernsey. Rush, P. 1995a. Economy and space in Roman Britain, in Rush, P. (ed.). 141–7. Rush, P. (ed.). 1995. Theoretical Roman Archaeology Conference (2nd Bradford). Avebury. Aldershot. Ran, M. (ed.) 1991. The Illustrated Archaeology of Ireland. Town House and Country House. Dublin. Rye, H. 1900. Rievaulx Abbey, its canals and building stones. Archaeological Journal, 57: 69–88. Saddington, D.B. 1990. The Origin and Nature of the German and British Fleets. Britannia, 21: 223–232.

199

JAMES ELLIS JONES

  Sauer, E.W. 2005a. Inscriptions from Alcester: Vespasian’s base of the Second Augustan Legion (?). Britannia, 36: 101–33 Salway, B. 2001. Travel, itineria and tabellaria, in Adams, C. & Laurence, R. (eds.). 22–66. Salway, P. 1965. The frontier people of Roman Britain. Cambridge University Press. Cambridge. Salway, P. 1980. The Lincolnshire Car Dyke. Navigation or drainage? Britannia, 11: 337–8. Salway, P. 1991. Roman Britain. Oxford University Press. Oxford. Salway, P. 1993. The Oxford Illustrated History of Roman Britain. Oxford University Press. Oxford. Salzman, L. F. 1967. Building in England down to 1540. Oxford. Sauer, E. 2002. The Roman Invasion of Britain (AD 42) in Imperial Perspective: A response to Frere and Fulford. Oxford Journal of Archaeology 24:4. 333–363. Sawyer, P.H. & Thacker, A.T. 1987. Domesday Survey, in Harris (ed.). 293–391. Sawyer, P.H. (ed.). 1999. The Oxford Illustrated History of the Vikings. Oxford University Press. Oxford. Sedgley, J.P. 1975. The Roman Milestones of Britain. British Archaeological Report 18. Oxford. Sealey, P.R. & P.A. Tyers. 1989. Olives from Roman Spain: A Unique Amphora Find in British Waters, Antiquaries Journal 69.1: 53-72.. Sealey, P.R. 1995. New light on the salt industry and Red Hills of prehistoric and Roman Essex. Essex Archaeology and History 26: 65–81. Sealey, P.R. 2009. New Light on the Wine Trade with Julio-Claudian Britain. Britannia, 40: 1–40 Selkirk, R. 1983. The Piercebridge Formula: A Dramatic New View of Roman History. Patrick Stevens. Cambridge. Selkirk, R. 1995. On the Trail of the Legions. Anglia. Norwich. Selkirk, R. 2001. Chester-le-Street and its place in history. Casdec. Birtley. Sellers, M. 1917. The York Mercers and Merchant Adventurers. Surtees Society. York. Severin, T. 1983. The Brendon Voyage. Arena. London. Sheppard, J. 1958. The Draining of the Hull Valley. East Yorkshire Local History Society. Hull. Sherratt, A. 1996. Why Wessex? The Avon Route and River Transport in Later British Prehistory. Oxford Journal of Archaeology, 15(1): 211–34. Shirley, E.A.M. 1996. The Building of the Legionary Fortress at Inchtuthil. Britannia, 27: 111–8. Shirley, E.A.M. 2001. Building a Roman legionary fortress. Tempus. Brimscombe Port. Shotter, D.C.A. 1973. Numeri Barcariorum: A note on RIB 601. Britannia, 4: 206–209. Shotter, D.C.A. 1975. Coin evidence and the northern frontier in the second century AD. Proceedings of the Society of Antiquaries of Scotland 110: 81–91. Shotter, D.C.A. 1978. A Hoard of Antoniniani from Hackensall Hall Farm. Lancashire Archaeological Journal, Vol.1: 47–52. Shotter, D.C.A. 1993. Coin-loss and the Roman Occupation of England, British Numismatic Journal, 63: 1–19. Shotter, D.C.A. 1993. Roman Coins from England: First Supplement. Lancaster. Shotter, D.C.A. 2000a. The Roman Conquest of the North. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 100: 33–53. Shotter, D.C.A. 2000b. Roman Coins from England: 2nd Supplement. Lancaster. Shotter, D.C.A. 2001. Petillius Cerealis in northern Britain. Northern History 36.2: 189–98. Shotter, D.C.A. 2002. Chester: Early Roman Occupation, in Carrington, P. (ed.). 25–32. Shotter, D.C.A. 2004. The Cumberland coast and the evidence of Roman coin loss, in Wilson, R.J.A. & Caruana, I. (eds.). 195–204. Shotter, D.C.A. 2004. Romans and Britons in North West England (4th Edn.). Centre for Regional Studies. Lancaster University. Shotter, D.C.A. 2007. How the North West was won. Hadrianic Society Bulletin. New Series Vol. 2: 5–12. Shotter, D.C.A. 2008. From conquest to frontier in the North West, in Bidwell, P. (ed.). 105–112. Shotter, D.C.A. 2009. When did the Romans invaded Scotland?, in Breeze, D.J. et al. 15–20. Shrubsole, G.W. 1887. The Traffic between Deva and the Coast of North Wales in Roman Times. Chester Archaeological and Historical Society Journal, N.S. 1: 71–113. Sibbald, R. 1707. Historical Inquiries. London. Sim, D. 1992. The manufacture of disposable weapons for the Roman army. Journal of Roman Military Equipment Studies, 3: 105–19. Simmons, B.B. 1979. The Lincolnshire Car Dyke. Navigation or drainage? Britannia, 10: 183–96. Simmons, B.B. 1980. The Iron Age and Roman coasts around the Wash. in Thompson, F.H. (ed.) Simpson, F. G. & I. A. Richmond. 1952. The Roman fort at Drumburgh.. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society 52: 9–14. Singleton, F.J. 1980. Kirhkam, A short history. Lytham St Annes. Sitwell, N.H.H. 1981. Roman Roads of Europe. Casswell. London. Sitwell, N.H.H. 1984. The world the Romans knew. Hamish Hamilton. London. Slade, W.J. 1959. Out of Appledore. Percival Marshall. London. Smith, A.H.V. 1996. Provenance of coals from Roman sites in U.K. counties bordering the River Severn and its estuary and including Wiltshire. Journal of Archaeological Science, 23: 373–389. 200

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Smith, A.H.V. 1997. Provenance of coal from Roman sites in England and Wales. Britannia 28: 297–324. Smith, D.E., Cullingford, R.A. & Firth, C.R. 2000. Patterns of isostatic land uplift during the Holocene: evidence from mainland Scotland. The Holocene 10,4: 489–501. Smith, D.I. & Stopp, P. 1978. The River Basin. Cambridge University Press. Cambridge. Smith, J.T. 1997. Roman villas; a study in social structure. Routledge. London. Smith, N.A.F. 1977. Roman Canals. Transactions of the Newcomen Society, 49: 75–86. Smith, R.A. 1909. The wreck on Pudding Pan Rock, Herne Bay, Kent. Proceedings of the Society of Antiquaries 21: 395–414 Smith, S. 1945. A Roman dupondius from the base of Aust Cliff. Transactions of the Bristol and Gloucester Archaeological Society, 66. Smith, W. 1588. The Particular Description of England. London. Smout, T.C. (ed.). 1992. Scotland and the Sea. John Donald Publishers. Edinburgh. Snape, M.E. 2003. A horizontal-wheeled Watermill of the Anglo Saxon period at Corbridge, Northumberland, and is River Environment. Archaeologia Aeliana 32: 37–72. Solley, W.T.J. 1966. Roman Remains from the Severn Bridge Approach at Aust. Transactions of the Bristol and Gloucester Archaeological Society, 85: 36–44. Sommer, C.S. 1984. The military Vici in Roman Britain: aspects of their origins, their location and layout, administration, function and end. British Archaeological Report 129. Oxford. Sommer, C.S. 2006. Military vici in Roman Britain revisited, in Wilson, R.J.A. (ed.). 95–146. Southern, P. 1989. The Numeri of the Roman Imperial Army. Britannia, 20: 81–141. Southern, P. 1990. Signals versus Illumination on Roman Frontiers. Britannia, 21: 233–242. Southern, P. 2007. The Roman Army. A social and institutional history. Oxford University Press. Oxford. Spain, R.J. 1985. Romano-British watermills. Archaeologia Cantiana, 100: 101–28. Spaul, J. 2002. Classes Imperii Roman: an epigraphic examination of the men of the Imperial Roman Navy. Nectoreca Press. Andover. Speidel, P. 2000. Who fought at the front?, in Alföldy, G. Dobson, B. & Eck. W. 473–82. Spiedel, M.P. 1987. The Chattan War, the Brigantian Revolt and the Loss of the Antonine Wall. Britannia, 18: 233–38. Spurrell, F.C.J. 1885. Early Sites and Embankments on the Margins of the Thames Estuary. Archaeological Journal, 42: 269-302. St. Joseph, J.K.S. 1969. Discoed, in Jarrett M.G. (ed.). 84–5. St. Joseph, J.K.S. 1969. Trawscoed, in Jarrett M.G. (ed.). 84–5. St. Joseph, J.K.S. 1969. Coed-y-Caerau, in Jarrett M.G. (ed.). 81. St. Joseph, J.K.S. 1969. Brithdir, in Jarrett M.G. (ed.). 130–2. St. Joseph, J. K. 1978. A Roman Camp near Girvan, Ayrshire, Britannia, 9: 397–401. Stallibrass, S. & Thomas, R. (eds.) 2008. Feeding the Roman Army: The Archaeology of Production and Supply in NW Europe. Oxbow. Oxford. Stammers, M. 1996. What's in a picture? Pictorial evidence for ships and harbours, in Carrington, P. (ed.). 90–4. Stanford, S.C. 1969. Leintwardine, in Jarrett M.G. (ed.). 91–5. Stanford, S.C. 1974. Native and Roman in the central Welsh Borderland, in Birley, E., Dobson, B. and Jarrett, M. (eds.). 44–61. Stanford, S.C. 1980. The Archaeology of the Welch Marches. Collins. London. Starr, C. 1989. The Influence of Sea Power on Ancient History. Oxford University Press. Oxford. Starr, G.C. 1960. The Roman Imperial Navy. Ithaca. New York. Steane, J. 1996. Oxfordshire. Pimlico County History Guide. Statistical Account. 1795. The Statistical Account of Scotland, Perthshire. Steer, K.A. 1949. The Roman Fort at Whitemoss, Renfrewshire. Proceedings of the Society of Antiquaries Scotland, lxxiii: 40–42. Steffy, J.R. 1982. In retrospect and an analysis of the hull, in Bass, G.F. & Doorninck, F.H. 82–6. Stephens G.R. 1987. A Severan Vexillation at Ribchester. Britannia, 18: 239–42. Stevens, C. 1950. The Roman town at Ilchester. Archaeological Journal, 107: 94–5. Stevenson, W.I. 1973. Some aspects of the geography of the Clyde tobacco trade in the 18th century. Scottish Geographical Journal, 89; 1. 19-35. Straker, V. 1987. Carbonised cereal grain from first-century London: a study for importation and crop processing, in Marsden, P. (ed.). 151–5. Strang, A 1997. Explaining Ptolemy's Roman Britain. Britannia 28: 1–30. Strickland, T.J. 1995. The Romans at Wilderspool: the Story of the First Industrial Development on the Mersey. Warrington. Strickland, T.J. 1999. What kind of community existed at Chester during the hiatus of the 2nd c.? in Goldsworthy & Haynes (eds.). 105–110.

201

JAMES ELLIS JONES

  Strickland, T. 2001. Roman Middlewich: A Story of Roman and Briton in Mid Cheshire. Roman Middlewich Program. Middlewich. Stuart, R. 1853. Caledonia Romana. Edinburgh & London. Summers, D. 1973. The Great Ouse. History of a Navigation. David & Charles. Newton Abbot Swain, H. 1986. A note on the Romano-British and other finds from Carr House Sands, Seaton Carew, Cleveland. Northern Archaeology 7.2: 31-34. Swann, V.G. 1978. Pottery in Roman Britain. Shire Archaeology. Avebury. Swan, V.G. 1997. Vexillations and the garrisons of Britannia in the second and early third centuries: a ceramic view-point, in Waateringe et al. (eds.). 289–94. Swan, V.G. 1988. Comments on Inveresk Ware, in Thomas, G.D. 167–71. Swan, V. G. 2008. Builders, suppliers and supplies in the Tyne–Solway region and beyond, in Bidwell, P. (ed.). 49–82. Swann, V.G. & Philpott, R.A. 2000. Legio XX VV and Tile Production at Tarbock, Merseyside. Britannia, 31: 57–8. Sylvester, D. 1969. The rural landscape of the Welsh Borderland. Macmillan. London. Symonds, M. 2007. Built for purpose: early fortlet use in Britain and Germany. Hadrianic Society Bulletin, New Series Vol.2: 39–47. Symonds, M. 2007 The design and purpose of Roman fortlets in the north-western frontier provinces of the Empire. Unpublished D.Phil. thesis. University of Oxford. Tatton-Brown, T. 1974. Excavations at the Custom House site, City of London, 1973. Transactions of the London & Middlesex Archaeological Society, 25: 117–219. Tatton-Brown, T. 1980. Camelon, Arthur’s O’on and the main supply base for the Antonine Wall. Britannia, 11; 340–3 Taylor, C. 1979. Roads and Tracks of Britain. Dent. London. Taylor, J. 2001. The Isle of Portland: an Iron Age port-of-trade. Oxford Journal of Archaeology 20: 187–205. Taylor, J. 2007. An Atlas of Roman Rural Settlement in England. Council for British Archaeology Research Report 151. York. Taylor, J. du P & Cleere, H. (eds.) 1978. Roman shipping and trade: Britain and the Rhine provinces. Council for British Archaeology Research Report 24. Oxford. Taylor, M.V. (ed.). 1912. Extracts from the MS Liber Luciani de laude cestrae. Record Society Lancashire and Cheshire, 64. Tayor, E.G.R.A (ed.) 1963. A regiment for the sea: and other writings on navigation. Hakluyt Society. Cambridge. Tchernia, A. 1983. Italian wine in Gaul at the end of the Republic, in Garnsey, P. et al. 87–104. Tchernia, A. 1986. Le Vin de L’Italie Romaine: Essai D’Histoire Economique D’Apres les Amphores. Ecole Francais de Rome. Rome. Temin, P. 2001. A market economy in the early Roman Empire. Journal of Roman studies, 91: 169–81 Temporini, H. (ed.) 1974. Aufstieg und Niederganger Römischen Welt II. Berlin. Thew, N. 1994. Geology and Geoarchaeology in the Yeo Valley at Ilchester, in Leach, P. (ed.). 38–43. Thomas, C. (ed.). 1966. Rural Settlement in Roman Britain. Council for British Archaeology Research Report 7. London. Thomas, C. 1972a. The Irish Settlements in Post Roman Western Britain. Journal of the Royal Institution of Cornwall, 6: 260–5 Thomas, C. 1981. A Provisional List of Imported Pottery in Post-Roman Western Britain and Ireland. Institute of Cornish Studies Special Report 7. Camborne. Thomas, C. 1985. Exploration of a drowned landscape: Archaeology and History of the Isles of Scilly. Batsford. London. Thomas, C. 1988a. The context of Tintagel: a new model for the diffusion of post-Roman Mediterranean imports. Cornish Archaeology, 27: 7-26. Thomas, C. 1988b. Tintagel Castle. Antiquity, 62: 421-34. Thomas, C. 1990. Gallici nautae de Galliarum provinciis: a sixth/seventh century trade with Gaul reconsidered. Medieval Archaeology, 34: 1–26. Thomas, C. 1993. English Heritage book of Tintagel : Arthur and archaeology. Batsford. London. Thomas, G.D. 1988. Excavations at the Roman civil settlement at Inveresk, 1976–77. Proceedings of the Society of Antiquaries of Scotland 118: 139–176 Thomas, R. & Stallibrass, S. 2008. For starters: producing and supplying food to the army in the Roman north-west provinces, in Stallibrass S. & Wilson, R. (eds.). Thompson, F.H. 1965. Roman Chester. Cheshire Community Council. Chester. Thompson, F.H. 1980. Archaeology and Coastal Change. Society of Antiquaries. Throckmorton, P. (ed.). 1996. The Sea Remembers: Shipwrecks and Archaeology. Chancellor. London. Thrupp, S.L. 1948. The Merchant Class of Medieval London. Ann Arbor. Michigan. Tilley, A. 1994. Sailing to windward in the ancient Mediterranean, international Journal of Nautical Archaeology, 23: 209–313. Timby, J. 1990. Severn Valley Ware: A Re-assessment. Britannia, 21: 243–252. 202

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Timby, J. 1999. Pottery supply to Gloucester colonia in Hurst, H.R. (ed.). 40–1 Timescape Archaeological Surveys 2000. Archaeological Survey at Maryport Roman Fort and Vicus. Geophysical Survey Report no. 23-03-00. Tipping, R. Tisdall, E. 2005. The landscape context of the Antonine Wall: a review of the literature. Proceedings of the Society of Antiquaries of Scotland 135: 443–469. Todd, M. 1976. The vici of Western England, in Branigan, K. & Fowler, P.J. (eds.). 99–119. Todd, M. 1981. Roman Britain: 55bc–400ad, (2nd Ed.) Fontana. London. Todd, M. 1984. Hembury (Devon): Roman troops in a Hillfort. Antiquity, 58: 171–4. Todd, M. 1985. The Roman Fort at Bury Barton, Devonshire. Britannia, 16: 49–55. Todd, M. 1987. The South West to AD 1000. Longman. London. Todd, M. 1993. An Arretine Vessel from Hembury. Devon Archaeological Society Proceedings, 51: 17. Todd, M. (ed.). 1989. Research on Roman Britain: 1960– 89. Britannia Monograph 11. Toft, L.A. 1992. Roman Quays and Tide Levels. Britannia, 23: 249–54. Tomlin, R.S.O. 1998. Roman Manuscripts from Carlisle: the Ink-written Tablets. Britannia, 29: 31–84. Tomlin, R.S.O. 2000. The legions in the Late Empire, in Brewer, R.J. (ed.). 159–182. Toms, G. 1969. Excavations at Llanymenech. Shropshire News Letter, 36: 20–1. Tooley, M.J. 1980. Theories of Coastal Change in North-west England, in Thompson, F.H. (ed.). 74–86. Tooley, M.J. 1990. Sea-level and coastline change during the last 5000 years, in McGrail, S. (ed.). 1–16. Tratman, E.K. 1963. Some ideas on Roman Roads in Bristol and North Somerset. Proceedings of the University of Bristol Spelaeological Society, 9, 3: 159–76. Turnbull, P. 1996. The supposed Roman harbour at Maryport. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 96: 233–5. Twigg, G.D. 2005. Salt making sites in Cheshire. Cheshire Local History Society. Tyers, P. A. 1996. Roman Pottery in Britain. Batsford. London. Tylecote, R.F. 1976. A History of Metallurgy, institute of Materials. London. Unger, R.W. (ed.). 1994. Cogs, Caravels and Galleons: The Sailing Ship 1000–1650. Conway. London. Unger, R.W. 1980. The Ship in the Medieval Economy 600–1600. Croom Helm. London Van Creveld, M. 1989. Technology and War: from 2000 BC to the Present. Free Press. New York. Van der Noort, R. 2004. The Humber Wetlands. The Archaeology of a Dynamic Landscape. Windgather Press. Macclesfield. Viatores, T. 1964. Roman Roads in the South-east Midlands. Gollanz. London. Villar, R. 1984. Merchant Ships at War: The Falklands Experience. Naval Institute Press. Waateringe, W. Groenman-van et al. (eds.) 1997. Roman Frontier Studies 1995: Proceedings of the XVIth International Congress of Roman Frontier Studies. Oxbow Monograph 91. Oxford. Waateringe, W. Groenman-van. 1997. Classical authors and the diet of Roman soldiers: true or false? in Waateringe et al. (eds.). 261–6. Wacher, J. (ed.). 1966. The Civitas Capitals of Roman Britain. Leicester University Press. Leicester. Wacher, J. 1978. Roman Britain. Dent. London. Wacher, J. 1979. The Coming of Rome. Routledge, Kegan, Paul. London. Wacher, J. (ed.) 1987. The Roman World. Routledge & Keegan Paul. London. Wacher, J. 1989. Cities from the Second to Fourth Centuries, in Todd, M. (ed.). 91–114. Wacher, J. 1995. The Towns of Roman Britain (2nd Edn.). Batsford. London. Wacher, J. & McWhirr, A.D. 1982. Early Roman Occupation at Cirencester. Cirencester Excavation Committee. Cirencester. Waddell, J. 1992. The Irish Sea in Prehistory. Journal of Irish Archaeology, 6: 29–40 Waddelove, A.C. & E. 1990. Archaeology and research into sea level during the Roman era. Britannia, 21: 253–266. Wadmore, B. 1920. The Earthworks of Bedfordshire. Bedford Standard Newspaper. Bedford. Wainwright, G. J. 1971. The excavation of a fortified settlement at Walsesland Rath, Pembrokeshire. Britannia, 2: 48–108. Walls, E. 1927. The Bristol Avon. Arrowsmith. Bristol. Walters, B. 1992. Ancient Dean and the Wye valley. Thornhill Press. Cheltenham. Walsh, M. 2002. Pudding Pan: A case study on the enhancement of the Romano-British maritime record, in Muskett, G. & Georgiadis, M. (eds.) 76-87. Walthew, C.V. 1975. The Town House and Villa House in Roman Britain. Britannia, 6: 189–205 Walton, K. 1974. A geographer’s view of the sea. Scottish Geographical Magazine, 90: 4–13 Ward, M. 1998. Some finds from the Roman works-depot at Holt. Studia Celtica, 32: 43–84. Ward, S. 1996. The course of the River Dee at Chester, in Carrington, P. (ed.). 4–11. Ward, S.W. 1988. Excavations at Chester, 12 Watergate Street 1985: Roman headquarters building to Medieval Row. Chester City Council. Chester. Ward, S.W. et al. 1994. Excavations at Chester, Saxon occupation within the Roman fortress: sites excavated 1978–1981. Chester City Council. Chester. Waters, B. 1947. Severn Tide. J.M. Dent. London. 203

JAMES ELLIS JONES Watson, G.R. 1969. The Roman Soldier. Cornell University Press. Ithaca. Watson, M. & Musson, C. 1993. Shropshire from the Air. Shropshire Books. Shrewsbury. Watts, D.J. 1995. A Lead Tank Fragment from Brough, Notts. (Roman 'Crococalana'), Britannia, 26: 318–322. Weatherhill, C. 1985. The Ships of the Veneti; a fresh look at the Iron Age tin ships. Cornish Archaeology, 24: 163–169. Weatherill, J. 1954. Rievaulx Abbey: the stone used in its building, with notes on the means of transport, and a new study of the diversion of the River Rye in the twelfth century. Yorkshire Archaeological Journal, 38: 333–54. Webster, G. 1953. The Lead Mining Industry in North Wales in Roman Times. Journal of the Flintshire Historical Society, 13: 5–33. Webster, G. 1955. A note on the use of coal in Roman Britain. Antiquaries Journal, 35: 199–217. Webster, G. 1970. The Military Situations in Britain between AD 43 and 71. Britannia, 1: 179–97. Webster, G. 1975. The Cornovii. Duckworth. London. Webster, G. 1981. Rome against Caractacus. Batsford. London. Webster, G. 1983. The Welsh Marches and the Early Roman Campaigns. Archaeological Journal, 138: 1–5. Webster, G. 1985. The Roman Imperial Army. Black. London. Webster, G. (ed.). 1988. Fortress into City: the Consolidation of Roman Britain. Batsford. London. Webster, P.V. 1976. Severn Valley Ware: A Preliminary Study. Transactions of the Bristol and Gloucestershire Archaeological Society. 18–46. Webster, P.V. 1991. Pottery and trade in Roman Wales, in Burnham, B.C. & Davies J.L. (eds.). 138–149. Webster, P.V. 1991. The first Roman fort at Cardiff, in Burnham, B.C. & Davies J.L. (eds.). 35–9. Webster, P.V. 1991. Pottery supply to the Roman . Journal of Roman Pottery Studies, 4: 11–18. Webster, P.V. 2010. Pottery, chronology and the forts of Roman Wales, in Burnham, B.C. & Davies, J.L. 33–36. Welfare, H. & Swan, V. 1998. Roman Camps in England - The Field Archaeology. Her Majesties Stationery Office. London. Weller, J. et al. 2004. Roman Traction Systems. http://www.humanist.de/rome/rts/index.html Wells, C.M. 1997. The daughters of the Regiment: sisters and wives in the Roman army, in Waateringe et al. (eds.). 571–4. Weerd, M.D. 1978. Ships of the Roman period at Zwammerdam/Nigrum Pullum, Germania Inferior, in Taylor, J. du P. & Cleere, H. (eds.). 15-21. Westerdahl, C. 1992. The Maritime Cultural Landscape. International Journal of Nautical Archaeology, 21,1: 5–14. Westerdahl, C. 1994. Maritime Cultures and Ship Types: Brief Comments on the Significance of Maritime Archaeology, international Journal of Nautical Archaeology 23, 4: 265–270. Westerdahl, C. (ed.). 2006. The Significance of Portages. British Archaeological Report, International Series 1499. BAR Publishing. Oxford. Wheeler, R.E.M. 1925. Prehistoric and Roman Wales. Clarendon Press. Oxford. Wheeler, R.E.M. 1929. The Roman Lighthouses at Dover. Archaeological Journal, 186: 29–47. Wheeler, R.E.M. and T.V. 1932. Report on the Excavation of Prehistoric, Roman and post-Roman Site in Lydney Park, Gloucestershire. Society of Antiquarians Report 9. Whimster, R. 1989. The Emerging Past. Royal Commission for Historic Monuments England. London. Whitcomb, D. S. & Johnson, J. H. 1979. Quseir al-Qadim 1978. Preliminary Report. American Research Centre in Egypt, Cairo. Whitcomb, D. S. & Johnson, J. H., 1982a. Quseir al-Qadim 1980. Preliminary Report. American Research Centre in Egypt, Vol. 7. Malibu. Whitcomb, D. S. & Johnson, J. H., 1982b, Season of Excavation at Quseir al-Qadim. American Research Centre in Egypt, Newsletter 120. 24–30. White, D.A. 1960. Litus Saxonicum: the British Saxon Shore in scholarship and history. Winconsin. White, K.D. 1984. Greek and Roman technology. Cornell University Press. Ithaca, N.Y. White, R.B. 1978. Excavations at Brithdir, near Dolgellau, in 1974, in Boon, G.C. (ed.). 35–62. White, R.H. 1999. A Roman Harbour at Wroxeter. British Archaeology, 50: 5. White, R.H. 2007. Britannia prima: The Romans in the West of Britain. Tempus. Stroud. White, R.H. & Barker, P. 1998. Wroxeter; Life and Death of a Roman City. Tempus. Stroud. White, R. & Dalwood, H. 2005. Archaeological assessment of Wroxeter. English Heritage. White, R.H., & Gaffney, V.L. 2003. Resolving the paradox: the work of the Wroxeter Hinterland Project, in P. Wilson, R.J.A. (ed.). Whittaker, C.R. 1983. Late Roman trade and traders, in Garnsey, P. et al. 163–180 Whittaker, C.R. 1990. The Consumer City Revisited: the vicus and the city. Journal of Roman Archaeology, 3 110–18.

204

THE MARITIME LANDSCAPE OF ROMAN BRITAIN Whittick, G.C. 1982. The earliest Roman lead mining on Mendip and in North Wales: a reappraisal. Britannia, 13: 113–23. Whitwell, B. 1988. Late Roman settlement on the Humber and Anglian beginnings, in Price, J. & Wilson, P.R. (eds.) Wilkes, J.J. 2000. Army and Society in Roman Dalmatia, in Alföldy, G. Dobson, B. & Eck. W. 327–42 Willan, T.S. 1964. River Navigation in England 1600 – 1750. Cass. London. Willan, T.S. 1967. The English Coasting Trade. Manchester University Press. Manchester. Willan, T.S. 1976. The inland trade: Studies in English internal trade in the sixteenth and seventeenth centuries. Manchester University Press. Manchester. Williams, D. & Carreras, C. 1995. North African Amphorae in Roman Britain: A Reappraisal. Britannia, 26: 231–252. Williams, D.F. & Peacock, D.P.S. 1983. The importation of olive-oil into Iron Age and Roman Britain. Madrid. Williams, G.R. 1968. Chester and the Northern Marches. Longman. London.. Williams, G. et al., 1988. Recent Archaeological Work on Merlin’s Hill, Abergwili. Carmarthenshire Antiquaries, 24: 5–13. Williams, G. 1991. Recent work on rural settlement in Wales, in Burnham, B.C. & Davies J.L. (eds.). 112–122. Williams, M. 1970. The Draining of the Somerset Levels. Cambridge University. Cambridge. Willis, S.H. 2007. Sea, coast, estuary, land and culture in Britain during the Iron Age, in Haselgrove, C.C. & Moore, T. (eds.) 105–29. Wilmott, T. 2001. Birdoswald Roman Fort: 1800 years on Hadrian’s Wall. Tempus. Stroud. Wilmot, T. 2008. The Vallum: how and why: a review of the evidence, in Bidwell, P. (ed.). 119–128. Wilson, A. 2003. Roman and Native in Dumfriesshire. Transactions of the Dumfries and Galloway Natural History and Archaeology Society, 77: 103 –60 Wilson, D. 1863. The Prehistoric Annals of Scotland. London and Cambridge. Wilson, P.R. 1991. Aspects of the Yorkshire Signal Stations, in Maxfield, V.A. & Dobson, B. Wilson, P. (ed.). 2003. The Archaeology of Roman Towns: Essays in honour of John S. Wacher. Oxbow. Oxford. Wilson, P. 2003. The Roman towns of Yorkshire: 30 years on, in Wilson, P. (ed.) 258–69 Wilson, P. 2009. Holding the line? The Humber frontier and the expansion into Yorkshire reconsidered, in Breeze, D.J. et al. 9–14 Wilson, P. & Price J. (eds.) 2003. Aspects of Industry in Roman Yorkshire and the North. Oxbow. Oxford. Wilson, R.J.A. (ed.) 1997. Roman Maryport and its Setting. Essays in memory of Michael J. Jarrett. Kendal. Wilson, R.J.A. 2002. Roman Remains in Britain. Constable. London. Wilson, R.J.A. 2004, introduction: the Roman frontier on the Solway, in Wilson, R.J.A. & Caruana, I. (eds.). 19 –38 Wilson, R.J.A. 2006. Urban defences and civic status in early Roman Britain, in Wilson, R.J.A. (ed.) 1–48. Wilson, R.J.A. (ed.). 2006. Romanitas: Essays on Roman Archaeology in Honour of Shepherd Frere on the Occasion of his Ninetieth birthday. Oxbow Books. Oxford. Wilson, R.J.A. and Caruana, I. (eds.) 2004. Romans on the Solway: Essays in Honour of Richard Bellhouse. Trustees of the Senhouse Roman Museum, Maryport Wintjes, J. 2007. The Classis Britannica - aspects of the history of Roman naval units in North Western Europe. Hadrianic Society Bulletin, New Series Vol. 2: 5–12. Witcher, R. 1997. Roman roads that reshaped the land. British Archaeology, 27: 7. Wolfson, S. 2008. Tacitus, Thule and Caledonia: The achievements of Agricola’s navy in their true perspective. British Archaeological Report 459. BAR Publishing. Oxford. Wooding, J.M. 1996a. Cargoes in trade along the Western Seaboard., in Dark, K.R. (ed.). 67–82. Wooding, J.M. 1996b. Communication and Commerce along the Western Sea-lanes: AD 400–800. British Archaeological Report International Series 654. Oxford. Woodiwiss, S. 1992 (ed.). Iron Age and Roman Salt Production in the Medieval Town of Droitwich. Council for British Archaeology Research Report 81. London. Woodward, D. 1996. The port of Chester in context 1500 to 1800, in Carrington (ed.). 61–5. Woolley, C.L. 1907. Corospitum: provisional report on the excavations in 1906. Archaeol Aeliana, 3: 3. Wooliscroft, D.J. 1988. The Outpost System of Hadrian’s Wall. Transactions of the Cumberland Westmorland Antiquarian and Archaeological Society, 88: 23–28. Wooliscroft, D.J. 1994. Reading the Tea Leaves, signalling as a means of prophesy on Roman Frontiers, in Cottam, S. et al. (eds.). 124–35. Wooliscroft, D.J. 1994. Signalling and Design of the Cumberland Coast. Transactions of the Cumberland and Westmorland Antiquarian and Archaeological Society, 94: 55–64. Wooliscroft, D.J. 1996. Signalling and Design of the Antonine Wall. Britannia, 27: 153–78. Wooliscroft, D.J. 2001. Roman Military Signalling. Tempus. Stroud. Woolliscroft, D.J. 2002a. The Roman frontier on the Gask Ridge, Perth and Kinross. An interim report on the Roman Gask Project 1995–2000. British Archaeological Reports 335. BAR Publishing. Oxford. Wooliscroft, D.J. & Jones, G.D.B. 2004. Excavations on the Cumberland coast at Silloth, and that Fingland Rigg, 1994, in Wilson, R.J.A. & Caruana, I.D. (eds.). 186–194.

205

JAMES ELLIS JONES

  Wooliscroft, D.J. & Hoffmann, B. 2006. Rome’s First Frontier, The Flavian Occupation of Northern Scotland. Tempus. Stroud. Wooliscroft, D. 2008. Signalling on Roman frontiers, in Bidwell, P. (ed.). 91–98. Wooliscroft, D. 2009. 79 AD and all that: when did the Romans reach Perthshire? in Breeze, D.J. et al. 33–8. Worsam, B.C. & Tatton-Brown, T.W.T. 1990. The Stone of the Reculver Columns and the Reculver Cross, in Parsons, D. (ed.). 51–69. Wright, E. 1990. The Ferriby Boats. Sea-craft of the Bronze Age. Routledge. London. Wright, R.P. 1964. An imperial inscription from the Roman fortress at Carpow, Perthshire. Proceedings of the Society of Antiquaries of Scotland 97: 202–5. Wright, R.P. 1974. Carpow and Caracalla. Britannia 5: 289–292. Yeo, C.A. 1946. Land and Sea Transportation in Imperial Italy. Transactions of the American Philological Society, 77: 221–44. Zienkiewicz, J.D. 1993. Excavations at the Scamnum Tribunorum at Caerleon: the Legionary Museum site 1983–5. Britannia, 24: 27–140. Zienkiewicz, J.D. 1986. The legionary fortress baths at Caerleon. National Museum of Wales. Cardiff

                                                           206

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Index of places Aardensburg, 75 Abergavenny, 122, 124, 133, 143, 151, 157, 183 Abertanant, 27, 85, 98, 124 Aberystwyth, 71, 150, 156 Abingdon, 91 Actium, 4, 61, 62, 63 Adriatic, 61, 86 Adur, River, 82 Aegean, 23 Afon Ceiriog, River, 10 Afon Twyi, 88 Afon Wnion, 27, 127 Africa, 47, 61, 111 Aire, River, 82, 93, 103 Alcester, 100, 138 Aldborough, 124, 141, 149, 153, 162, 165, 170 Alderley Edge, 98, 159 Alet, 20, 21, 22, 192 Alexandria, 30 Alexandrina, 61 Algeria, 39 Allendale, 57 Allington, 6, 38, 101 Alps, 89 Als, 76 Alston Moor, 98 Altenburg, 61 Ambleside, 28, 55, 124, 141, 198 Amesbury, 81 Amisia, River, 86 Amorica, 19, 109, 182 Anchormen, River, 33 Anglesey, 21, 53, 54, 70, 71, 73, 77, 83, 98, 118, 129, 130 Annan, 57 Annan Hill, 124 Annan, River, 6, 27, 124, 125, 127 Annanfoot, 124 Antonine Wall, 54, 56, 57, 58, 104, 106, 107, 126, 134, 135, 150, 152, 187, 199, 201 Appledore, 26 Ardoch, 58 Arles, 21 Astigiti, 22 Atlantic, 1, 2, 4, 19, 20, 21, 22, 23, 24, 25, 26, 32, 33, 39, 40, 45, 109, 110, 111, 116, 129, 180, 182, 193 Aude, River, 20, 22, 109 Aust, 84 Aust Cliff, 201 Austria, 113 Avignon, 92 Avon, 26 Avon, River, 6, 20, 27, 81, 83, 101, 134, 154, 176, 181, 191, 198, 200, 203 Axe, River, 45, 91 Axminster, 125, 162, 166, 168 Aylmer Hall, 90 Ayrshire, 57 Badgworth, 13, 100

Baetica, 22, 25, 32, 110, 111, 112, 191, 199 Bagendon, 47 Baghdad, 67 Bainbridge, 55, 56 Balbuildy, 47 Balkan, 50 Balmuildy, 58, 106 Bandon, 26 Banwell, 100 Bar Hill,, 106, 110 Barland's Farm, 3, 27, 33, 34, 35, 88, 107, 196 Barnstaple, 53 Barochan Hill, 125 Barry, iv, 78, 103, 125, 162, 166, 168, 184, 186 Barton Court, 91 Barygaza, 23 Bath, 30, 83, 101, 102, 122, 125, 154, 162, 166, 168, 176, 193 Bawtry, 125, 162, 165, 169 Bay of Biscay, 21, 24, 25 Bay of Fundy, 45 Baylham House, 125, 162, 165, 169 Beachley, 84 Beachy Head, 27 Beaumaris, 73, 83 Beaumaris Castle, 83 Beckfoot, 56, 85, 126, 129, 180 Bedriacum, 62, 114 Belerion, 98 Belgium, 113 Benwell, 56, 113 Berkeley, 94, 118, 175, 184 Berrow, 91 Bertha, 5, 58, 85, 126, 132, 143, 162, 165, 168, 175, 180 Berwick upon Tweed, 2 Berwick-on-Tweed, 126, 162, 165, 168 Beverley, 93 Bewcastle, 56 Bewdley, 28, 94 Bideford, 26 Binchester, 126, 151, 162, 165, 170 Birdlip Hill, 17 Birdoswald, 56, 193 Bishops Castle, 97 Bishopton, 126 Bithynia, 87 Bitterlees West, 111 Bitterne, 126, 162, 166, 169 Black Pastor quarries, 103 Black Sea, 73 Blackdown Hills, 99 Blackfriars, iv, 3, 6, 33, 35, 36, 37, 38, 39, 40, 42, 43, 44, 65, 76, 82, 101, 104, 105, 107, 195 Blackpool, 12, 144 Blackwardine, 127 Blackwater, 100, 134, 143, 162, 163, 165, 168 Blaen-cwm-bach, 71 Blakeney, 24 Bleadney, 91 Blennerhasset, 55, 127 Bodiam, 113 Bodiham, 11

207

Bodysgallen, 101, 129 Bordeaux, 20, 24, 25, 81, 83, 94, 109, 112 Boroughbridge, 93 Boston Spa, 102 Bothwellhaugh, 5, 85 Boulogne, 4, 20, 61, 62, 64, 75, 114 Bourne, 90 Bourne-Morton, 90 Bowes, 55, 56 Bowness, 55, 56, 85, 127, 176, 191, 197 Bowness-on-Solway, 127, 176, 191, 197 Bradwell, 59, 102, 127, 162, 165, 168 Brampton Bryan, 71 Brampton Old Church,, 55 Brancaster, 59, 60 Braughing, 11, 69, 107 Brean Down, 78 Brecon, 51 Brecon Gaer, 54, 124, 127, 129, 130, 136, 149, 190 Brede, 113 Bredon Hills, 99 Brent, 91 Bridgnorth, 94 Bridgwater, 45, 88, 142, 197 Bridlington, 5, 84 Brigg, iii, 33, 34, 194 Bristol, i, ii, 1, 3, 7, 24, 29, 30, 31, 35, 45, 94, 98, 103, 123, 129, 130, 131, 132, 134, 142, 146, 147, 148, 157, 176, 178, 183, 184, 185, 190, 191, 192, 193, 196, 198, 201, 203, 204 Bristol Channel, 5, 21, 25, 26, 28, 33, 35, 45, 66, 69, 70, 77, 78, 84, 88, 95, 100, 103, 107, 110, 154 Britanny, 148 Brithdir, 10, 27, 71, 84, 104, 127, 151, 156, 201, 204 Brittany, 19, 20, 21, 24, 39, 41, 62, 76, 111, 113, 114 Brockworth, 118 Broighter, 41 Brompton, 97, 98, 151, 175 Brough on Fosse, 128 Brough on Humber, 11 Brougham, 28, 55 Brough-on-Humber, 128, 162, 165, 169 Brough-on-Noe, 51, 55 Brough-under-Stainmore, 55 Brownrigg, 111 Broxtowe, 54, 55 Brue, River, 91, 100 Brunton, 103 Bryn-y-Gefeiliau, 122, 128, 129, 146, 156 Burcombe, 99 Burcot, 85, 93 Burges, 109 Burgh by Sands, 127 Burgh Castle, 59, 102, 128, 131, 158, 162, 165, 170 Burgh-by-Sands, 55, 85 Burnham-on-Sea, 13 Burrow Walls, 85, 128, 150, 176

JAMES ELLIS JONES

  Burwell Old Lode, 90 Bury Barton, 53, 203 Buxton, 55 Bykers Dyke, 91 Byzantium, 74 Cadbury Congresbury, 115, 198 Cadiz, 110 Caer Gybi, 54, 60, 77, 84, 129, 198 Caerhun, 84, 101, 104, 128, 129, 135, 156, 180, 192, 198 Caerleon, iii, 7, 10, 12, 13, 15, 17, 22, 32, 43, 50, 51, 54, 57, 71, 78, 79, 84, 95, 100, 101, 103, 104, 108, 111, 117, 122, 124, 127, 129, 130, 131, 135, 148, 149, 154, 157, 178, 184, 192, 194, 198, 206 Caermote, 55 Caernarfon, 43, 51, 52, 54, 67, 73, 84, 97, 101, 104, 110, 117, 122, 130, 135, 156, 178 Caernarfonshire, 94 Caerphilly, 130, 192 Caersws, 27, 54, 85, 88, 106, 107, 130, 140, 146, 151, 179, 183, 190 Caerwent, 6, 12, 17, 54, 82, 101, 102, 103, 111, 119, 129, 130, 131, 133, 154, 156, 176, 179, 193 Caister-on-Sea, 102, 128, 131 Caistor, 59, 60 Caistor St Edmund, 131, 136, 143, 162, 165, 170 Caldicot, 12, 101, 129 Caledonia, 54, 63, 114 Calstock, 53 Cam, River, 11, 29, 85, 90, 93 Cambridge, iv, 5, 11, 29, 81, 85, 90, 93, 125, 131, 141, 146, 162, 165, 168, 178, 179, 181, 182, 184, 185, 186, 189, 192, 195, 196, 197, 199, 200, 201, 202, 205 Cambridgeshire, 90, 187 Camel, River, 53, 149 Camelon, 5, 58, 85, 131, 134, 156, 162, 165, 168, 176, 194, 198, 202 Cannington, 115 Canterbury, 39, 42, 89, 108, 132, 147, 152, 153, 154, 162, 165, 170, 180 Cape Wrath, 64 Capel Bangor, 150 Cappodocia, 73 Car Dyke, 81, 89, 90 Carcassone, 109 Cardean, 5, 85, 132, 162, 165, 168, 189 Cardiff, 52, 54, 59, 60, 66, 67, 70, 72, 77, 78, 79, 84, 104, 130, 132, 145, 147, 149, 154, 175, 177, 178, 179, 183, 184, 185, 186, 190, 192, 193, 195, 196, 197, 198, 204 Cardiff Roads, 26 Cardigan, 24, 178 Cargill, 5, 85, 132, 162, 165, 169 Carhayes, 98 Carkin Moor, 55 Carlisle, 5, 17, 28, 54, 55, 56, 59, 85, 103, 104, 105, 106, 110, 111, 124, 127, 132, 133, 144, 149, 150, 152, 155, 158, 182, 189, 194, 203

Carmarthen, 17, 24, 51, 54, 71, 84, 88, 99, 104, 127, 129, 132, 133, 146, 147, 149, 183, 190, 191 Carn Euny, 118, 181 Carnforth, 133 Carngoon Bank, 100 Carnsore Point, 21 Carpow, 51, 58, 59, 64, 114, 126, 133, 154, 162, 165, 170, 177, 180, 183, 206 Carr Rocks, 115 Carrawburgh, 56 Carriden, 5, 58, 85, 133, 162, 165, 169 Carron, River, 85 Carthage, 4, 20, 61, 116 Carvoran, 55, 56 Carzield, 134 Castle Collen, 54 Castle Douglas, 5, 85 Castle Hill, 134, 162, 165, 170 Castlecary, 58 Castleford, 56, 108, 134, 149, 153, 162, 165, 168 Castlehill Boothby, 55 Castleshaw, 55 Castlesteads, 56 Catterick, 29, 56, 134, 141, 151, 153, 161, 162, 165, 170 Caucasus, 62, 114 Channel Islands, 3, 21, 27, 39 Chapel Hill, 9 Charterhouse, 97, 98, 134 Cheddar, 6, 111, 119, 134 Chedworth, 103, 186 Chelmer, River, 69 Chelmsford, 69, 108, 134, 136, 143, 162, 165, 168 Chepstow, 24, 79, 135 Cherbourg, 39 Cheshire, iv, 52, 70, 88, 98, 100, 101, 104, 122, 129, 130, 135, 148, 150, 180, 188, 202, 203 Cheshire Plain, 52, 88 Chester, iv, 4, 8, 13, 14, 21, 22, 24, 30, 31, 32, 51, 52, 54, 56, 57, 66, 70, 71, 79, 81, 88, 94, 95, 98, 99, 101, 103, 104, 105, 111, 113, 114, 117, 122, 129, 130, 135, 142, 146, 148, 150, 156, 176, 180, 182, 187, 188, 189, 192, 193, 194, 200, 201, 202, 203, 204, 205 Chesterholm, 118 Chester-le-Street, 135, 149, 162, 165, 170 Chesters, 18, 56, 97, 99, 103, 185 Chew Valley, 103 Chichester, 25, 65, 69, 70, 100, 103, 126, 135, 159, 162, 166, 168, 182, 187, 197 Chigwell, 69 Christchurch, 21 Cirencester, 6, 17, 47, 101, 103, 106, 108, 118, 141, 143, 144, 182, 198, 203 Clayhanger, 53 Clifford, 127, 129, 136, 143 Clwyd, River, iv, 91, 97, 129 Clyde, 54, 57, 58, 59, 85, 114

208

Clyde, River, 4, 5, 6, 27, 64, 69, 94, 125, 126, 135, 146, 150, 199 Clydesdale, 57 Clyro, 54, 122, 127, 136, 143, 159, 190 Coalpit Heath, 103 Coed Cefn Pwll-Du, 97 Colchester, 39, 69, 100, 101, 103, 106, 107, 110, 136, 141, 160 Cold Knap, 78 Collingham, 102, 103 Colne, 100, 136, 139, 155, 162, 163, 165, 168, 169 Cologne, 61, 75 Colwyn Castle, 120 Combe Martin, 99 Cone Pill, iv, 84, 175 Coniston, 98 Conway, 71, 73, 181, 185, 186, 194, 195, 203 Conwy, River, 6, 27, 115, 128, 129 Corbridge, 19, 51, 55, 56, 83, 107, 114, 136, 139, 145, 153, 162, 165, 170, 201 Corduba, 22 Cornwall, 8, 12, 20, 21, 23, 24, 40, 53, 72, 94, 98, 99, 100, 110 Cos, 108 Côte d’Or, 112 Cotentin, 20 Cotswolds, 52, 101, 131, 143 County Hall, 33, 35, 36, 199 County Kilkenny, 76 Courtmacsherry, 26 Cowbridge, 137, 162, 166, 170, 197, 199 Cowley, 118 Cramond, 5, 58, 59, 64, 85, 106, 108, 114, 137, 162, 165, 169, 189, 194, 198 Crandon Bridge, 108, 192, 199 Cree, River, 57 Crick, 82 Cudmore Farm, 53 Cullompton, 53 Cumberland, 28, 51 Cumbria, 98, 103 Cyzicus, 74 Dalswinton, 134, 137 Dane, River, 100, 148, 150 Danube, 50, 59, 62, 74, 114 Danube, River, iii, iv, 4, 63, 64, 67, 68, 110, 135, 147 Dee, River, 5, 8, 10, 13, 14, 21, 27, 57, 70, 81, 85, 94, 95, 97, 101, 105 Dee. River, iv, 6, 130, 135, 140, 142, 148, 156, 203 Deeping Gate, 89 Deganwy, 115 Denbighshire, 53 Denmark, 30, 76, 89 Denver, 90 Derby, 28, 29 Derbyshire, 53 Dere Street, 17, 18, 55, 56, 106 Derwent, River, 28, 29 Derwydd-bach, 71 Devon, 8, 14, 53, 72, 85, 94, 99 Ditton Brook, 105, 156

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Dneiper, 28 Dogger Land, 8 Dolaucothi, 98, 176, 179 Don, River, 28, 90, 91 Doncaster, 134, 137, 146, 153, 163, 165, 169 Dorchester, 28, 78, 108, 125, 137, 138, 142, 163, 165, 166, 169, 170, 181, 185, 192, 199 Dordogne, 109 Dorset, 8, 20, 28, 105, 112, 175, 186, 192, 193 Dour, River, 80, 138 Dover, 4, 15, 21, 43, 45, 59, 69, 74, 75, 80, 84, 113, 114, 138, 196, 197, 204 Dover Strait, 45 Downham Market, 90 Downton, 81 Draethen, 97 Droitwich, 27, 30, 85, 99, 100, 110, 138, 205 Drumanagh, 57 Drumburgh, 85, 138, 163, 166, 168, 200 Dublin, 40, 57, 94, 194, 199 Dumbuck, 94, 150 Dumfries, 55, 85 Dungeness, 11 Dunkirk, 75 Dunwich, 14 Durham, i, 97, 126, 182, 184, 188 Dvina, River, 28 Dyfi, River, 6, 10, 27, 151 Dylife, 98 Earith, 11 East Anglia, 9 East Brent, 11 East Grimstead, 81 East Hattersley, 82 Ebbw, River, 97 Ebchester, 55, 56, 138, 163, 165, 169 Eden, River, 5, 6, 27, 85, 106, 132, 138 Edinburgh, 30, 179, 181, 183, 187, 191, 194, 195, 199, 201 Egypt, 23, 30, 47, 86, 116 Elginhaugh, 122, 139, 143, 163, 165, 169, 187 Ellen, River, 6, 27, 127, 148 Elsack, 55 Ely, 11 Emms, River, 65, 72 English Channel, 65 English Channel, 3, 4, 5, 7, 8, 11, 22, 24, 27, 39, 62, 67, 68, 69, 74, 75, 76, 109, 114, 115 Ermine Street, 16, 17, 31 Esk, River, 85 Essex, 81, 100 Esuris, 51 Exe Estuary, 53 Exe, River, 25, 53, 94 Exeter, 94, 112, 125, 138, 139, 140, 156, 163, 166, 169, 176, 178, 184, 188, 189, 190, 191, 194, 196, 199 Exmoor, 99, 184, 199 Faeroes, 23 Falklands, 67

Falls of Campsie Linn, 143 Falmouth, 72 Farnham, 102 Fen Causeway, 90 Fengate, 89 Fenland, 9, 89, 90 Fens, 90 Ferriby, 34, 206 Ffridd, 97 Fife, 100, 115 Fife Ness, 42 Filey, 60, 139, 163, 165, 168 Fingringhoe, 139 Firth of Forth, 5, 107, 114, 131, 137, 143 Fishbourne, 53, 112 Fiskerton, 34 Flanders, 43 Fleetwood, 5, 55, 85, 139, 140 Flint, 97 Flintshire, 98 Folkestone, 11 Forden Gaer, 27, 54, 85, 88, 104, 130, 140, 178, 180, 181 Forest of Dean, 72, 98, 99, 103, 105, 107, 113 Forest of Dean,, 157, 160 Fort Augustus, 28 Forth, 42, 54, 57, 58, 59, 85, 106, 114, 115 Foss Dyke, 27, 29, 93 Foss, River, 83, 90, 93 Fosse Dyke, 81, 85 Fowey, River, 53 Freestone Hill, 76 Frocester, 118 Frome, River, 98, 156 Fylde, 12 Gades. See Cadiz Galloway, 5, 55, 57, 69, 85, 156, 205 Garonne, River, 5, 20, 22, 109, 115 Gascony, 73 Gatcombe, 83, 94, 103, 178 Gatehouse of Fleet, 57, 140, 156 Gateshead, 55, 140, 192 Gaul, 1, 2, 3, 7, 47, 50, 62, 65, 69, 74, 75, 83, 86, 89, 97, 98, 103, 109, 110, 111, 112, 114, 115 German Sea, 69 Germany, 1, 62, 74, 75, 112, 114 Gibraltar, 22, 23 Gier, River, 109 Gironde, River, 23, 109 Girvan, 57, 140, 201 Girvan Mains, 140 Gittisham, 53 Glasgow, 94, 150, 182, 199 Glasson Dock, 95 Glastonbury, 91, 97, 179, 181, 189 Glenlochar, 5, 57, 85, 140 Glenluce, 57, 140 Glouceater, 6 Gloucester, iv, 10, 17, 27, 28, 29, 79, 80, 81, 83, 85, 86, 91, 93, 94, 100, 106, 111, 118, 119, 130, 135, 138, 140, 141, 143, 144, 148, 154, 156, 159, 160, 178, 183, 187, 188, 190, 193, 196, 201, 203 Gloucestershire, 53

209

Godmanchester, 90, 131, 141, 145, 153, 163, 165, 170 Goldborough, 60 Goldcliff, 12, 101, 176 Goldsborough, 141, 163, 165, 168 Goodwin Sands, 9, 25 Goole, 27 Gore Sands, 45 Gower, 76, 100 Great Dunmow, 69 Great Glen, 28 Great Orme, 98, 115 Great Ouse, River, 5, 11, 29 Great Witcombe, 103 Great Yarmouth, 31 Greatchesters, 56 Greece, 61, 112 Greenland, 23 Greenock, 94 Greenwich, 27 Greta Bridge, 108 Gripps, River, 100 Guadalquivir, 111 Guadalquivir, River, 110, 111, 112 Guernsey, iv, 3, 35, 38, 39, 115, 195, 199 Gulf of Moriban, 41 Gwent, 34, 101, 175, 178, 196, 197, 199 Hadrian's Wall, 4, 5, 6, 10, 17, 18, 22, 24, 32, 54, 55, 56, 58, 78, 80, 84, 88, 102, 103, 106, 107, 113, 117, 127, 128, 133, 135, 138, 144, 148, 159, 176, 177, 179, 181, 183, 191, 205 Halifax, 31 Haltonchesters, 56 Haltwhistle Burn, 55 Hampshire, 102 Hamworthy, 53 Hardknott, 28, 104, 124, 141, 152, 177 Harlow, 107 Hartshill, 89, 106 Hasholme, 3, 33, 34, 195 Hastings, 11, 60 Hatfield Chase, 91 Hatherthwaite, 28 Hayton, 55 Healam Bridge, 141, 163, 165, 170 Hembury, 53, 203 Hengistbury Head, 20, 21, 182 Herd Hill, 111 Herd Sand, 42, 115 Herefordshire, 53 Herne Bay, 115 Hernor, 23 Heronbridge, 81, 142, 194 Hertfordshire, 11, 93 Hexham, 148 Hibernia, 40 High Rochester, 56 High Street, 28, 124 Highbridge, 13, 91, 100, 196 Highland Zones, 4, 73 Hispalis, 22 Hjortspring, 76 Hod Hill, 51 Holderness, 14

JAMES ELLIS JONES

  Holkham Bay, 2 Holland, 4, 64 Holt, 51, 66, 105, 130, 142, 156, 163, 167, 168, 187, 192, 203 Holyhead, iii, 60, 66, 67, 77, 129 Horncastle, 142, 160, 163, 165, 168 Horsey, 2 Housesteads, 56 Hull, 31, 33, 93, 185, 200 Humber, 161, 185 Humber, River, 14, 16, 27, 31, 33, 34, 90, 91, 100, 106, 112 Humberside, 106 Huntcliff, 60, 138, 142, 163, 165, 168 Huntspill, River, 13, 96, 100 Huntstanton, 2, 5, 84 Hurlestone Point, 72 Hvarf, 23 Hyssi Portus, 74 Hythe Inlet, 11 Iberia, 6, 24, 25, 109, 110, 111, 119 Iceland, 23, 64 Ictis, 98 Ide, 53 Idle, River, 91, 93 Ierne, 23 Ilchester, 28, 31, 78, 81, 88, 125, 138, 139, 142, 163, 167, 170, 192, 201, 202 Ilkley, 47, 55, 56, 147 Inchtuthil, 5, 47, 57, 85, 98, 105, 107, 126, 132, 142, 143, 163, 165, 170, 197, 200 India, 63, 100 Ingolmells, 100 Inveresk, 5, 58, 85, 139, 143, 163, 165, 169, 177, 190, 192, 195, 199, 202 Ionian Sea, 64 Ireland, iv, 20, 23, 33, 39, 40, 41, 42, 51, 57, 69, 76, 77, 115, 129, 131, 179, 180, 185, 187, 190, 197, 199 Irish Sea, ii, 2, 5, 7, 20, 21, 22, 25, 26, 28, 32, 56, 66, 69, 77, 88, 110, 115, 116, 123, 129, 144, 190, 195 Irthing, River, 10 Irvine, 69, 143 Irving, River, 56 Isla, River, 5, 85 Isle of Bute, 69 Isle of Grain, 89 Isle of Thanet, 9, 21, 89 Isle of Wight, 39, 40, 42, 53, 115 Isleham Sluice, 90 Italy, 61, 62, 86, 108, 109, 112, 114 Ixworth, 143, 163, 165, 169 Kelvedon, 136, 143, 163, 165, 168 Kenchester, 101, 122, 124, 127, 136, 143, 144, 156 Kendal, 18, 30, 31 Kennet, River, 81, 83 Kent, 17, 89, 100, 106 Keswick, 98 Keynsham, 83 Khartoum, 67 Killegrew Round, 98 Killerton, 53 Kimmeridge, 20 King Street, 55

King’s Fish Pool, 83 King’s Lynn, 9, 85, 93 King's Lynn, 5 Kingsclere, 102 Kingscote, 118 Kingsholm, 85, 122, 129, 141, 144, 160, 190, 192 Kingsroad, 26 Kingston-Upon-Hull, 93 Kinsale, 26 Kirby Thore, 55 Kirkbride, 55, 56, 85, 126, 128, 144, 176 Kirkcudbright, 5, 85 Kirkham, 12, 55, 85, 139, 140, 144, 158, 189 Knaresborough, 103 Konnigratz, 48 La Coruña, 21, 24 Lake District, 55 Lake Farm, 144, 155, 159, 163, 166, 170 Lake House Farm, 11 Lake Moaris, 87 Lamnaios, 23 Lancashire, 5, 85, 103, 104 Lancaster, 35, 55, 56, 57, 60, 66, 77, 85, 94, 95, 103, 105, 140, 145, 158, 175, 189, 190, 200 Land’s End, 1, 21, 24, 25, 72 Lantonside, 145 Lapford, 53 Lark, River, 90, 93 lbion, 23 Le Yaudet, 22 Lea, River, 29, 69, 82, 93 Learchild, 126, 145, 163, 165, 168 Lease Rigg, 5, 84, 145, 163, 165, 169 Leeds, 103 Léguer, River, 22 Leicester, 31, 90, 134, 145, 146, 163, 165, 170, 181, 182, 190, 199, 203 Leintwardine, 54, 143, 160, 201 Leith, 30 Lérida, 40 Lerwick, 64 Lesmahago, 148 Leven, River, 28 Lewis, 69 Ley Pill, 99 Lezoux, 115 Lillebone, 113 Lilstock, 14, 15, 94 Lincoln, 27, 29, 34, 54, 81, 83, 85, 89, 90, 93, 128, 137, 142, 145, 146, 149, 153, 156, 160, 161, 163, 165, 170, 181, 191 Lincolnshire, 89, 90, 100 Linley, 97, 151 Lisbon, 25 Little Russel, 42, 115 Littleborough, 81, 146, 163, 165, 170 Littlechester, 29 Littlecote, 81 Littleport, 11 Liverpool, 24, 105, 156, 194, 197 Lizard Point, 21 Llandeilo, 54, 146, 147, 163, 167, 170, 189

210

Llandovery, 54, 88, 127, 133, 146, 190 Llandudno, 101, 129 Llanidloes, 88 Llanio, 133, 146, 151, 157, 190 Llantsantfraid, 98 Llantwit Major, 103 Llanymenech, 87, 203 Llanymynech, 98, 124 Lleyn, 20, 71, 76 Llugwy, River, 6, 27, 128 Loch Ness, 28 Loch Oinch, 28 Loch Ryan, 69, 140, 156, 158 Loire, River, 20, 22, 42, 63, 86, 109, 115 Londinium, 12, 25, 36, 37 London, iii, iv, 2, 3, 7, 9, 12, 15, 17, 21, 24, 25, 28, 30, 31, 33, 35, 36, 37, 38, 42, 65, 67, 68, 76, 82, 85, 92, 93, 99, 101, 106, 110, 111, 112, 113, 117, 124, 125, 132, 134, 135, 136, 138, 141, 146, 153, 154, 155, 163, 165, 170, 175, 176, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 193, 194, 195, 196, 197, 198, 199, 200, 201, 202, 203, 204, 205, 206 Long Sandall, 55, 146, 153, 163, 165, 169 Longships, 24 Longthorpe, 90, 108, 146, 158, 163, 165, 169 Lostwithiel, 10, 53, 146, 163, 166, 169 Loughor, 8, 70, 77, 78, 84, 101, 104, 122, 125, 132, 133, 146, 149, 163, 167, 169, 193, 196 Low Mire, 111 Lower Machen, 97 Lugg, River, 5, 127 Lune, River, 6, 27, 57, 95, 145 Lurg Moor, 146 Lusitania, 51 Lydia, 30 Lydney, 10, 69, 70, 99, 147, 163, 167, 169, 175, 186, 204 Lympne, 11, 59, 132, 147, 163, 166, 168 Lympsham, 91 Lynmouth, 26 Lynn, 29 Lyon, 7, 20, 62, 92, 112, 114 Magor Pill, 27, 78, 175, 196 Maidstone, 37, 38 Mainz, 68, 188, 191 Majabigwaduce, 73 Malta, 24 Malvern, 99 Mancetter, 89, 106 Manchester, 28, 55, 103, 107, 124, 144, 147, 150, 152, 181, 188, 190, 191, 205 Manningford Bruce, 81 Mardyke, 69 Marseilles, 2, 23, 109 Marshfield, 103 Martin Mere, 12

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Martinhoe, 8, 77, 85, 144, 148, 185 Maryport, 56, 66, 85, 104, 105, 113, 126, 148, 150, 177, 190, 203, 205 Massachusetts, 73 Mawbray, 111 Mawddach, River, 10, 127 Meare Pool, 91 Mediterranean, 1, 2, 6, 19, 20, 22, 23, 24, 25, 31, 33, 36, 38, 39, 42, 43, 44, 45, 62, 63, 65, 76, 78, 87, 94, 98, 108, 109, 110, 111, 112, 115, 116, 119, 161, 179, 180, 185, 186, 189, 195, 196, 202 Medway, 6 Medway, River, 29, 37, 38, 100, 101, 102 Melandra Castle, 55 Mendips, 30, 53, 97, 98, 99, 113 Meols, 14, 21, 116, 148, 163, 167, 168, 187 Mersey, River, 6, 27, 42, 98, 105, 115, 147, 156, 159, 201 Metchley, 100, 138 Meuse, River, 86 Mevagissey Bay, 53 Mictis, 98 Middlewich, 54, 100, 148, 202 Mile End, 90 Milford Haven, 54, 71 Military Way, 17, 55, 56, 58 Mill Stream, 12 Millengen, 27 Misenum, 4, 61, 67 Moesica, 59, 61, 74, 114 Monks’ Lode, 90 Monmouth, 79, 135, 148 Monmouthshire, 99 Mons Graupius, 57, 64, 87 Monte Testaccio, 61, 110 Montgomery, 97, 176, 196 Moresby, 56, 64, 85, 129, 149, 175, 177 Morton Fen, 90 Morwenstow, 149 Mosel, River, 22 Mount Batten, 21, 53, 82, 98, 182 Mounts Bay, 72 Mull, 69 Mull of Kintyre, 69 Mumrills, 51, 106 Muncaster, 105 Murray Firth, 57 Mylae, 61 Myos Hormos, 78 Nantes, 109 Nantstallon, 53, 98, 108, 110, 149, 185 Nantwich, 194 Neath, 54, 70, 71, 77, 78, 84, 104, 127, 132, 147, 149, 188, 196 Nedern Brook, 102 Nene, River, 85, 89, 90, 93, 100, 101, 106, 118 Nether Denton, 55, 56 Netheravon, 81 Netherby, 5, 56, 85, 149, 177 Netherlands, 47, 119, 192 Netown-on-Trent, 55 New Cut, 94, 95

New Guy's House, 36, 65 New Hythe, 14 Newbrough, 55 Newcastle, 25, 30, 106, 177 Newcastle upon Tyne, 149, 177, 183, 189 Newfoundland, 45 Newlyn, 12, 72 Newnham, 94 Newport, iv, 13, 26, 95 Newstead, 19, 51, 55, 58, 110, 182 Newton Kyme, 149, 163, 165, 170 Newton on Trent, 149, 163, 165, 170 Newton Tracey, 53 Nicomedia, 87 Nidd, River, 103 Nijmegen, 75 Nile, River, 67, 86, 87 Nith, River, 6, 27, 57, 85, 134, 137, 145, 158 Nithdale, 5, 85 Norfolk, 90 Nornour, 42, 115 North Kyme, 89 North Sea, 4, 8, 20, 22, 25, 26, 27, 39, 64, 65, 69, 74, 75, 89, 101, 175 North Tawton, 53 Northumberland, 97, 148, 201 Northwich, 54, 100, 147, 148, 150, 159 Norway, 23, 25 Norwich, 31 Nottingham, 28 Nova Libica, 61 Nydam, 76 Nyland, 91 Oberstim, 68 Okehampton, 53 Old Burrow, 77, 148, 185 Old Carlisle, 55 Old Church, 55, 56 Old Church Bampton, 55 Old Head of Kinsale, 26 Old Kilpatrick, 5, 58, 85, 94, 125, 126, 150 Old Passage, 84 Old Penrith, 55, 56 Old West River, 90 Old Winteringham., 31 Omdurman, 67 Ongar, 69 Orford, 14 Orkney, 64, 69, 73 Osmanthorpe, 54, 55, 150, 164, 165, 169 Ostia, 17, 40, 61, 62, 83, 140 Ouse, River, 31, 83, 85, 90, 91, 93, 103, 161 Outerwards, 150 Outwell, 29 Ouze Deep, 115 Oxford, iv, 27, 91, 94, 106, 118, 175, 177, 178, 179, 180, 181, 182, 183, 184, 185, 186, 187, 188, 189, 190, 191, 192, 194, 195, 196, 199, 200, 201, 202, 203, 204, 205 Padstow, 24 Pannonia, 62, 114 Pannonica, 59, 61, 67, 114

211

Papcastle, 85, 105, 127, 148, 150, 177, 181 Parkgate, 94, 95 Parma, 86 Parrett, River, 6, 27, 78, 88, 94, 142 Parys Mountain, 98 Pearl Harbour, 73 Peckforton Hills, 98 Pembrokeshire, 5, 54, 85 Pen Llwyn, 10, 84, 150 Penarth, 101 Pengwryd, 71 Pennal, 10, 51, 71, 84, 104, 130, 150, 151, 156, 157, 178, 187 Pennines, 55, 135 Penobscot, 73 Penrhos, 53 Pentland Firth, 64 Pentre, 97 Pentrehyling, 27, 85, 151 Pen-y-Cochbren, 104 Pen-y-Crocbren, 98 Pen-y-Gaer, 124, 151, 182 Penzance, 72 Perpignan, 25 Perry Barr, 106 Perth, 143 Peterborough, 11, 85, 89, 90, 93 Pevensey, 30, 59, 74, 102, 151, 164, 166, 168 Philippi, 1 Piercebridge, 18, 151, 152, 164, 165, 170, 175, 184, 200 Pilrow Cut, iv, 91 Placentia, 86 Ploumanach, 113 Plumpton Rocks, 103 Plymlimon, 98 Plymouth, 21, 53, 82, 98, 99 Po, River, 86 Pompeii, 108 Pontica, 61, 74 Pool Quay, 5, 88, 91 Poole, 20, 88, 100, 106, 142 Porcupine Bank, 20, 42 Porlock, 72 Porlock Weir, 78 Port Einon, 100 Port Vendres, 112 Portage Bay, 28 Portage County, 28 Portage la Prairie, 28 Portage-du-Fort, 28 Portchester, 59, 74, 127, 132, 152, 164, 166, 168, 182 Porth Felen,, 20 Portishead, 26 Portskewett, 82 Portsmouth, 100 Portugal, 41 Portus, 40 Portus Setantiorum, 140 Port-Vendres, 25, 181, 196 Prestatyn, 97, 104, 129, 152, 164, 167, 168, 178 Pudding Lane, 82 Pudding Pan, 9, 25, 42, 115 Pumpsaint, 146, 179 Purbeck, 100, 101, 103

JAMES ELLIS JONES

  Puriton, 96, 192 Quentovic, 89 Quernmore, 105 Radstock Basin, 103 Rainsbury, 53 Ramsbury, 81 Rance, 22 Ravenglass, 28, 51, 56, 66, 85, 124, 141, 152, 177, 178, 181, 193, 197, 198 Ravenna, 4, 61, 67, 86, 127, 130, 199 Ravenscar, 60, 152, 164, 165, 168 Raw Dykes, 90 Reach Lode, 90 Reading, 27, 102, 184, 205 Reculver, iii, 8, 9, 59, 60, 66, 89, 102, 131, 132, 147, 152, 164, 165, 168, 197, 206 Red Hills, 100 Red Sea, 78, 86 Remagen, 27 Rheidol, River, 10 Rhine, River, 3, 4, 5, 7, 20, 21, 22, 26, 30, 36, 55, 59, 61, 63, 65, 66, 67, 68, 72, 74, 75, 86, 88, 110, 111, 113, 115, 129, 184, 186, 188, 195, 196, 197, 202 Rhineland, 50, 109 Rhodes, 108, 109 Rhone, River, 5, 22, 26, 86, 109, 110 Rhône, River, 92, 109, 111 Rhos-on-Sea, 110 Rhuddlan, iv, 91 Rhymey, River, 97 Rhyn Park, 10 Ribble, River, 6, 12, 27, 57, 107, 140, 144, 152, 158, 159 Ribchester, 54, 55, 56, 114, 139, 144, 147, 152, 153, 158, 184, 201 Ribe, 89 Richborough, iii, 9, 17, 21, 51, 59, 60, 64, 65, 67, 69, 70, 89, 102, 112, 113, 132, 138, 153, 164, 165, 168, 180, 182, 183, 192, 195 Richmond, 28 Risca, 97 Rise How, 111 Risingham, 56 Roall, 55, 153, 164, 165, 168, 176 Rochester, 132, 147, 153, 155, 164, 165, 169 Roding, River, 69 Roecliffe, 55, 124, 153, 164, 165, 170, 177 Rome, 58, 61, 62, 74, 83, 86, 87, 108, 110, 113, 114 Romney Marsh, 11, 100 Roodee, 79 Rooksbridge, 11 Rossington, 55, 125, 153, 164, 165, 170 Rothay, River, 28 Rouen, 83 Rudchester, 56 Ruhr, River, 5 Rumney Great Wharf, 99 Runcorn, 101, 129 Runnymede, 29 Ruthwell, 153

Ryde, 42, 115 Rye, 26, 27 Saint George’s Channel, 26 Salwarpe, River, 100, 138 Santander, 25 Sarmatia, 62, 114 Saxon Shore, 2, 4, 5, 54, 56, 59, 60, 66, 78, 94, 101, 102 Scalesceugh, 105 Scandinavia, 76 Scarborough, iii, 8, 15, 60, 154, 164, 165, 168 Scheldt, River, 27 Scilly, 12, 21, 24, 42, 115 Sea Mills, 83, 98, 101, 134, 144, 154, 176, 178, 184, 193 Seaton, 14 Seine, River, 20, 21, 22 Seiont, River, 6, 27, 130 Severn Bridge, 84 Severn Estuary, 12, 26, 78, 82, 84 Severn, River, iv, 5, 6, 10, 12, 13, 25, 27, 28, 29, 31, 33, 34, 35, 70, 77, 79, 80, 84, 85, 86, 87, 88, 89, 91, 93, 94, 99, 100, 102, 104, 106, 107, 124, 127, 130, 131, 134, 135, 138, 140, 141, 144, 154, 156, 159, 160, 175, 176, 186, 194, 196, 199, 201, 203, 204 Shadwell, 154, 164, 165, 170 Shapwick, 155, 164, 166, 170 Shepton Mallet, 106 Shetland, 23, 64, 69 Shetland Islands, 64 Shrewsbury, 31, 94, 204 Shropshire, 5, 53, 78, 97, 103 Sicily, 61 Siger, River, 11 Silchester, 31, 101, 102, 103, 110, 125, 135, 138, 155, 159 Sindercome, 99 Skegness, 2, 5, 84, 100 Skinners Quay, 95 Skye, 69 Slack, 55 Soar, River, 90 Solent, 60, 75, 100 Solway, 55, 56, 57, 64, 66, 69, 85, 104, 111 Solway Estuary, 127 Solway Firth, 55, 57, 64, 66, 69, 125, 126, 132, 134, 148, 153, 158 Solway Frontier, 132, 191 Somerset, 26, 53, 78, 88, 90, 91, 94, 96, 97, 100, 103, 108, 115 Somerset Levels, 11, 13, 88, 91, 100, 108, 181, 182, 188, 205 South Cadbury, 115 South Shields, 5, 7, 12, 22, 47, 55, 56, 59, 66, 78, 104, 108, 111, 112, 114, 122, 132, 135, 148, 153, 154, 157, 164, 165, 170, 177, 183, 187, 189 Southampton, 31, 196 Southerndown, 101 Southport, 12 Southwold, 2 Spain, 47, 52, 61, 62, 63, 97, 111, 112, 114, 115

212

Sprat Beach, 78 Springhead, 155, 164, 165, 169 Spurn, 14 Spurn Head, 27 St Albans, 153, 154, 164, 166, 170 St Austell Bay, 53 St David’s, 5, 21, 54, 71, 85 St George’s Channel, 23 St Ives, 5, 24, 72, 85, 93 St Madoes, 133, 154, 164, 166, 170 St Michael’s Mount, 24, 98 St Monans, 100 St Neots, 11 St Peter Port, 3, 33, 35, 36, 37, 38, 39, 40, 42, 43, 44, 95, 105, 112 St. Peter Port, 115 Staines, 155, 164, 166, 170 Stainmore Pass, 55 Stanegate, 17, 55, 56, 128, 138, 144, 189 Stanwix, 133, 155, 182 Star, 103 Steep Holm, 21, 72, 110 Stirling, 5, 85, 156 Stogursey, 14 Stoke Hill, 53 Stour, River, 9, 20, 89 Stracathro, 156, 164, 166, 169 Strageath, 58 Straits of Dover, 27, 75 Straits of Gibraltar, 2, 23, 25, 110, 111, 115 Stranraer, 140, 156, 158 Strasbourg, 3 Stretham, 82 Stretton Grandison, 156 Sudan, 67 Sudbrook, 82, 84, 101, 102, 131, 156, 196 Suessa, 19 Suffolk, 14 Sunderland, 60, 94, 95 Surrey, 102 Sutton Pool, 82 Swale, River, 28, 29, 103 Swansea, 71 Sweden, 76 Syria, 62, 114 Syriaca, 61 Tadcaster, 102 Tamar, River, 82 Tarbock, 105, 156, 202 Tattenhall, 101, 129 Tavy, River, 82 Taw, River, 53, 99 Tay, River, 5, 27, 42, 58, 64, 85, 114, 115, 143 Teme, River, 5 Tewkesbury, 94 Thames Estuary, 25, 27, 42 Thames, River, 2, 3, 5, 10, 13, 17, 20, 25, 26, 27, 28, 29, 35, 36, 37, 42, 62, 69, 82, 85, 87, 89, 91, 93, 94, 100, 102, 106, 113, 115, 176, 178, 181, 192, 195 The Gut, 12 The Marches, 1, 53, 101, 104, 180 The Rumps, 21 The Wash, 27, 60

THE MARITIME LANDSCAPE OF ROMAN BRITAIN

  Thevesdale, 102 Thorpe, 149, 150, 156, 164, 166, 170, 188 Throp, 55 Thule, 64 Tiber, River, 61, 88 Tigris, River, 66, 67 Tintagel, iv, 116, 195 Tiverton, 53 Tomen-y-Mur, 71, 84, 104, 128, 129, 130, 151, 156, 190 Tone, River, 94 Topsham, 25, 53, 139, 156, 164, 166, 169 Torksey, 85, 90, 93 Torridge, River, 53 Toulouse, 109 Towcester, 108 Trawscoed, 10, 71, 84, 104, 150, 151, 156, 183, 201 Trebarveth, 100 Trebia, 86 Trefeglwys, 98 Trent, River, 10, 54, 81, 85, 90, 91, 93, 146 Troutbeck, 55 Trucculum, 64 Tunisia, 6, 111, 119 Turkey, 44, 67, 116 Tweed, River, 55 Twitchen, 99 Tyne, River, 5, 12, 26, 27, 30, 42, 55, 56, 62, 66, 78, 97, 106, 114, 115, 177, 179, 191 Tynemouth, 66 Tywardreath Bay, 53 Uley, 118 Upwell, 11 Ure, River, 93, 103 Urr, River, 6, 27, 57 Ushant, 20, 21, 24, 73 Usk, 6, 10, 13, 27, 34, 54, 79, 88, 95, 101, 103, 122, 124, 127, 129, 148, 151, 157, 159, 193 Usk, River, 26 Utrecht, 75 Vakkenburg, 75 Valéry-sur-Somme, 113 Venafrum, 19 Verulamium, 103 Vesuvius, 51, 113, 134 Vienne, 109

Vindolanda, 6, 18, 19, 24, 31, 47, 52, 55, 56, 62, 99, 104, 110, 112, 114, 118, 177, 178 Vindonissa, 51 Volga, River, 28 Vyrnwy, River, 10 Wall Town, 85, 157, 158, 164, 167, 170 Wallsend, 55, 56, 78, 154, 157, 158, 164, 166, 170 Waltham Abbey, 81 Walton Castle, 2, 59, 147, 158, 164, 166, 168 Walton le Dale, 57, 107, 140, 144, 158 Walton-le-Dale, 55, 199 Wantsum Channel, 9, 89 Ward Law, 57, 145, 158 Ware, River, 69, 93, 106, 112, 118, 199, 203, 204 Warrington, 99 Wash, 2, 9, 11, 81, 91 Washing Well, 55, 158 Washing Well, Whickham, 55 Washingborough, 89 Washington, 28 Watchet, 103 Water Newton, 81, 141, 158 Watercrook, 11, 55, 56, 158, 177, 197 Waterfoot, 57 Watling Street, 107 Weald, 11, 99, 159, 181 Weaver, River, 100, 150 Well Creek, 29 Well Stream, Channel, 11 Welland, River, 81, 89, 100, 101, 146 Welshpool, 10, 88, 91 Wensley, 55 Went, River, 91 Wessex, 20 West Dean, 81 West Indies, 52 West Stockwith, 91 West Water, River, 11 Western Approaches, 22, 32 Westminster, 17, 35 Weston under Penyard, 148, 159 Weston-super-Mare, 26 Weston-under-Penyard, 98 Wexford, 21 Wharfe, River, 93, 102 Whitby, 5, 60, 84

 

213

Whitchurch, 54, 81, 142, 148 Whitley Castle, 97 Whitstable, 27, 42, 115 Whittington, 53 Wicken Lode, 90 Wigan, 55, 103, 147, 150, 159, 181, 190 Wilderspool, 55, 106, 159, 201 Willowford, 10, 18 Winchester, 126, 135, 159, 160, 164, 166, 169, 188 Windermere, Lake, 28 Winteringham, 31, 160, 164, 166, 169 Wirksworth, 28 Wirral, 21, 116, 187 Wisbech, 11, 85, 93 Wisbech Ouse, River, 11 Witham, 34 Witham, River, 81, 89, 90 Wiveliscombe, 53 Woolaston, 84, 99, 185 Worcester, 27, 29, 30, 85, 94, 98, 99, 107, 156, 160, 182 Worcestershire, 53, 73, 103 Wreay, 55 Wrexham, 97 Wroxeter, 5, 6, 27, 31, 54, 57, 78, 85, 100, 101, 105, 106, 107, 115, 135, 138, 140, 143, 150, 158, 160, 161, 184, 185, 204, 205 Wrynose Pass, 28, 124, 141 Wye, River, 6, 27, 33, 77, 79, 88, 101, 135, 136, 143, 144, 148, 156, 159, 192, 203 Yantlet, 89 Yare, River, 31 Yarmouth, 42, 115 Yassı Ada, 43, 44 Yeo, River, 26, 28, 30, 81, 88, 134, 142, 202, 206 York, 4, 11, 47, 54, 55, 59, 60, 62, 66, 83, 85, 91, 93, 102, 103, 108, 109, 114, 117, 147, 152, 158, 161, 175, 180, 182, 186, 190, 191, 195, 196, 200, 201, 202, 203 Yorkshire, 5, 14, 47, 82, 84, 85, 90, 93, 139, 141, 142, 152, 154, 161, 176, 177, 179, 183, 198, 200, 204, 205 Yssel, River, 86 Ystwyth, River, 10, 98, 156 Zwammerdam, 68