Palaeohistoria 25 (1983): Institute of Archaeology, Groningen, the Netherlands [First edition] 9781000108330, 1000108333, 9781000131062, 1000131068, 9781000151480, 1000151484, 9781003079248, 1003079245

Table of contents :
Cover......Page 1
Half Title......Page 2
Title Page......Page 4
Copyright Page......Page 5
Table of Contents......Page 6
Introduction......Page 7
A.E.van Giffen as archaeozoologist......Page 8
Some notes on recent and pre- and protohistoric fishing gear from Northwestern Europe......Page 14
The molluscs of the dwelling mound Gomolava, Yugoslavia. An environmental investigation on and near Gomolava......Page 62
Spoolde. Worked and unworked antlers and bone tools from Spoolde, De Gaste, the IJsselmeerpolders and adjacent areas......Page 84
The animal remains of Tell Sweyhat, Syria......Page 138
Farming on the banks of the river Aa. The faunal remains and bone objects of Paddepoel 200 B.C.-250 A.D.......Page 152
A meal for the dead. Animal bone finds in Roman graves......Page 190
Medieval horses from Utrecht (Jan Meijenstraat)......Page 202
The zoological reference collection of the Biologisch-Archaeologisch Instituut in Groningen. With appendix by A.T.Clason, R.J.Kosters & T.P.Jacobs: The catalogue of the reference collection of the Biologisch-Archaeologisch Instituut......Page 218

Citation preview

PALAEOHISTORIA

PALAEOHISTORIA

ACTA ET COMMUNICATIONES

INSTITUTI BIO-ARCHAEOLOGICI

UNIVERSITATIS GRONINGANAE

25

A.A.BALKEMA / ROl'fERDAM / BOSTON / 1983

Editorial staff' Mette Bierma, J.W.Boersma, H.R.Roelink & J.D.van der Waals Address: Biologisch-Archaeologisch Instituut, Poststraat 6, 9712 ER Groningen, Netherlands

ISSN 0552-9344

ISBN 90 61916062

© 1986 A.A.Balkema, P.O.Box 1675, 3000 BR Rotterdam, Netherlands

Distributed in USA & Canada by: A.A.Balkema Publishers, P.O.Box 230, Accord, MA 02018

Printed in the Netherlands

CONTENTS

Introduction

VI

A.T.CLASON

1

A.E.van Giffen as archaeozoologist DICK C.BRINKHUIZEN

Some notes on recent and pre- and protohistoric fishing gear from Northwestern Europe

7

JAN WILLEM EGGINK

The molluscs of the dwelling mound Gomolava, Yugoslavia. An environmental investigation on and near Gomolava

55

A.T.CLASON

Spoolde. Worked and unworked antlers and bone tools from Spoolde, De Gaste, the IJ sselmeerpolders and adjacent areas

77

H.BUITENHUIS

The animal remains of Tell Sweyhat, Syria

131

EGGEKNOL

Farming on the banks of the river Aa. The faunal remains and bone objects of Paddepoel 200 B.C.-250 A.D.

145

R.C.G.M.LAUWERIER

A meal for the dead. Animal bone finds in Roman graves

183

W.PRUMMEL

Medieval horses from Utrecht (Jan Meijenstraat)

195

A.T.CLASON

The zoological reference collection of the Biologisch-Archaeologisch Instituut in Groningen With appendix by A. T. Clason, R.J.Kosters & T.P.Jacobs: The catalogue of the reference collection of the Biologisch-Archaeologisch Instituut

V

211

INTRODUCTION

Volume 25 of Palaeohistoria celebrates the 25th anniversary ofthe Archaeozoology Department in the Biologisch-Archaeologisch Instituut of the Groningen State University. A. E. van Giffen, the Institute's founder, began with ar­ chaeozoology 70 years ago; but it was in 1959 that the B.AJ. appointed its first full-time research worker in this field. Since then the Department has developed so as to include

several research workers, a number of technical assistants, and research students of prehistory and/or archaeozoology. This volume, to which former students as well as present researchers have contributed, gives an impression ofthe Department's current research activity. To it has been added, on micro-fiches, a catalogue of its recent reference collection.

VI

A.E.VAN GIFFEN AS ARCHAEOZOOLOGIST

A.T.Clason CONTENTS 1. INTRODUCTION 2. VAN GIFFEN: 1908-1913 3. 1913: DIE FAUNA DER WURTEN 4. VAN GIFFEN AFTER 1913 5. VAN GIFFEN IN THE FIFTIES AND LATER 6. ACKNOWLEDGEMENTS 7. REFERENCES 8. KEYWORDS

1

2

A.T.CLASON

1. INTRODUCTION The founder of the Biologisch-Archaeologisch Instituut (B.A.I.) of the State University of Groningen, the late Prof. Dr. A.E.van Giffen, started his scientific career at the beginning of the 20th century as a student of biology at the University of Groningen in the Neth­ erlands. In 1913 he concluded his doctoral research with a thesis concerning the subfossil faunal remains collected from pre- and pro­ tohistoric settlements in the clay region of the Northern Netherlands. When I was reading Van Giffen's thesis anew I was struck by the fact that he approached archaeozoological problems in a way that is, at least partly, still relevant today. From 1920 to 1954 Prof. van Giffen was director of the B.A.I. On the recommendation of his successor, Prof. Dr. H.T.Waterbolk, systematic archaeozoological research was star­ ted in Groningen in 1959. This was 25 years ago. So far, little attention has been paid to the archaeozoological work of Van Giffen and to his interest in zoology in general (Clason, 1973). The object of this paper is to call attention to this part of Van Giffen's scientific career. 2. VAN GIFFEN: 1908-1913 In 1908, Van Giffen, when still a student, was appointed by the Centraal Bureau ter versprei­ ding van kennis van Groningen en omgelegen streken (Central Bureau for the knowledge of the province of Groningen and surrounding areas) as an assistant to supervise the com­ mercial excavation of the terp Dorkwerd north of Groningen (Van Giffen, 1913a; 1922). He started to collect regularly animal remains and other finds from Dorkwerd and other terpen in the provinces of Groningen and Friesland in the north of the Netherlands. The terpen were at that time excavated for commercial purposes. Through the ages they had been heightened by layers of cow dung and clay sods, which proved to be excellent manure for the poor sandy soils of the province of Dren­ the, south of Groningen. The earliest terpen started their growth in the 6th century B.C. The terpen only lost their function of protec­

ting man and beast against the sea after the building of sea-dikes started in the 11th cen­ tury A.D. (Waterbolk & Boersma, 1976). The results of a study of the construction of the terp of Dorkwerd and of archaeological finds from that terp other than bones were published by Van Giffen (1913a) as part of his doctoral thesis Die Fauna der Wurten. He included also a description ofthe faunal remains from other terpen. In 1912 Van Giffen was under pressure to finish his research and to obtain his doctoral degree so that he could be appointed as keeper of the Rijksmuseum van Oudheden in Leiden. As a result, the book was shortened and the part about the domestic animals was not included. It has never been published as part II, as was originally the intention. 3.1913: DIE FAUNA DER WURTEN Van Giffen chose this title for his thesis very deliberately. In the 19th century Riitimeyer (1862) had published the results of his study of the animal remains collected from the Swiss lake dwellings in a book entitled Die Fauna der Pfahlbauten der Schweiz. By adopting the title Die Fauna der Wurten Van Giffen wanted to stress the fact that in his opinion it was as important to study and publish the faunal remains from the terpen area as those from the Swiss lake villages. It was also his intention to compare both faunas. Up to 1913, C.Broekema (1910) and L.Broe­ kema (1908; 1909a; 1909b; 1910; 1912) had published short articles on the remains of the aurochs, cattle, horse, sheep and pig from the terpen area in the agricultural journal Cultura. Bakker (1909) used the cattle remains from the terpen for his hypothesis of the descent of Dutch cattle. Unlike these authors, Van Giffen intended to give a survey of all the domestic species ofthe terpen area and to place them in a European geographical and histo­ rical context. Van Giffen expected the study of subfossil animal bones from the terpen to contribute to a better understanding of the life and work of our predecessors, but also to shed light on the descent and inter-relatedness of our do­ mestic animals in the past. This could give indications on the relationship of recent do­ mestic animals with animals of the past, both

A.E. van Giffen as archaeozoologist

wild and domesticated, and also on the dis­ tribution of domestic animals and the cultures they belonged to. Wenn wir uns jetzt noch die Frage vorlegen, was von einer Bearbeitung der erwahnten Fauna zu erwarten ist, so ist es klar, dasz wir dadurch ein besseres Bild gewinnen werden von dem Tun und Treiben unserer Vorfahren. Auch kann sie dazu beitragen, ein wenig mehr Licht zu werfen auf und in das dunkle Ab­ stammungsproblem unserer Haustiere und auf ihre iilteren verwandtschaftlichen Beziehungen. Diese Kenntnis kann wiederum fUr die jetzige Viehzucht von Nutzen sein. Umgekehrt konnen wir aber aus der Verwandtschaft mit rezenten Formen einerseits, mit alteren zahmen und wilden Tieren derselben oder anderer Gegenden andererseits, ein Bild, wenigstens einige Anhaltspunkte, gewinnen von den Wander­ strassen dieser Tiere, wie auch von der Herkunft der Kultur, zwischen deren Resten sie aufgefunden wor­ den sind. (p.49)

Van Giffen saw in 1913 the discipline of archaeozoology, or Haustierkunde (the study of domestic animals) as it was called then, as an independent discipline alongside archae­ ology and philology. In dieser Weise kann dann auch die Haustierkunde als selbstandige Wissenschaft neben Archaeologie und Philologie betrieben werden und mit dazu dienen, das schwierige Problem der prae- und protohistorischen Kulturstromungen, Handelsbeziehungen und V61ker­ wanderungen zu losen. In jener Weise wird man sich eine neue selbstandige Hilfswissenschaft fUr die Pal­ ethnographie denken konnen. Vorausgesetzt wird natiirlich dabei, dasz diese Wissenschaft, soviel wie moglich, unabhangig arbeitet von ihren eben genann­ ten Schwestern. (Fur die Kulturpflanzen gilt natiirlich dasselbe.) (p.49)

A new aspect in the work of Van Giffen is that in his opinion it is necessary to add to the description of the animal remains the measurements that could be taken of the bones. These measurements should be clearly defined and always be the same. This necessity still exists in 1983. Van Giffen used the mea­ surements published by Hue (1907) for this purpose. Die Masze habe ich als tatsachliches Material voran­ gestellt, wei! dieselben dann auch von anderen ver­ wertet und ausgearbeitet werden konnen, und in dieser Weise eine moglichst genaue Kontrolle stattfinden kann. In dieser Beziehung muss ich es sehr bedauern, dasz die Arbeit von Prof. Windle and Herrn Hum-

3

preys: 'On some Cranial and Dental Characters of the Domestic Dog' nur die relativen Masze statt der absoluten bringt, und ich dieselbe deswegen nicht fUr meine Zwecke brauchen konnte, wiewohlsie sehr viele Masze von rezenten Rassen enthaIt, und sich sonst zur Vergleichung mit den heutigen Hunden sehr geeignet hatte. Da nun in der Literatur die osteometrischen An­ gaben sehr verschieden dargestellt worden sind, und die Masze nur selten nach einem bestimmten Plan gegeben werden (vergleiche aber Woldrich (104)), und demnach Vergleichung sehr zeitraubend ist, halte ich es fUr zweckmaszig, mich bei meiner Arbeit nach dem, diese Lucke ausfiillenden, Buche von Edmond Hue iiber 'Osteometrie des Mammiferes' (13) zu richten und genau anzugeben, welche die genommenen Masze sind. Zu diesem Zwecke habe ich Zeichnungen, ent­ sprechend den von ihm gegebenen Figuren, anfertigen lassen, und darin iiberal! die namlichen Buchstaben eingetragen, es sei denn, dasz ich Masze gebraucht habe, welche er nicht angibt. (p.63)

The drawings of bones with indications of the measurement points were not published in Die Fauna der Wurten, but were intended to be included in part II. These drawings have nQt been found again. Another new idea was that, though at the time it was generally known which domestic animals were kept by the inhabitants of the terpen region, less was known, or nothing at all, about the animals themselves. Van Giffen calls this the problem of variability (p.54). To get a better understanding of this va­ riability Van Giffen used statistics for the evaluation of metric data. He used the statistics worked out by the geneticist Dr. T.Tammes (1907) for her research on the flax stem (Linum sativum). Anders wird aber die Aufgabe, wenn man sich nicht fragt, welche Haustiere in der Wurtenzeit vorkommen, sondern wie dieselben beschaffen waren. Diese Frage der Losung naher zu bringen, stellt an den Unter­ sucher schon bedeutend hohere Anspruche. (p.55)

and 1m Foigenden werden wir mit einer statistischen Bearbeitung der Haustierreste anfangen und sehen, ob diese Methode auch etwas ans licht zu bringen vermag. Dieselbe wurde meines Erachtens bis jetzt zu wenig verwertet, auch in solchen Fallen wo die Fiille des Materiales ihre Anwendung gestattet haben wiirde; ja bei den praehistorischen Knochen fand sie bis jetzt wohl gar keine Verwendung. (p.57)

and

4

A.T.CLASON Wir wollen versuchen, aus der statistischen Zusam­ menstellung verschiedener Merkmale die Variabilitat kennen zu lernen und sehen ob diese auf bestimmte Mittelwerte hindeutet oder nicht, und ob bestimmte Groszen verschiedener Merkmale zusammenhangen, korrelativ verwandt sind oder nicht. Daraus wollen wir dann wieder versuchen, ein Urteil zu gewinnen tiber das Vorhanden sein oder fehlen konstanter Haustiertypen, denen in der Literatur (ofters nur nach quantitativen Maszunterschieden), bestimmte Namen zugelegt sind und der Wert von Rassen oder Spezies zuerkannt worden ist. (p.57) Die statistische Zusammenstellung, das Berechnen der Frequenzkurve, Mediane, des Quartils u.s.w., das Darstellen von Korrelationstafeln machte ich in der Weise, wie es Fraulein Dr. T. Tammes in ihrem Buche tiber den Fl~chsstengel auseinandersetzt (106, S.32, 33,39 u.f.). Uberdies habe ich auch noch graphische Darstellungen verwertet; namentlich habe ich, nach­ dem ich den Maximum- und Minimumvarianten fUr ein bestimmtes Merkmal herausgesucht und die Dif­ ferenz in eine bestimmte Anzahl Intervalle verteilt hatte, tiber jedem Intervalle in Reihenfolge die Rang­ nummer der dazugehorigen Knochen oder Schadel gestellt. Wir konnen dann aus diesen Darstellungen verschiedener Merkmale der nahmlichen Ueberreste ersehen, ob die Knochen (respektive Schadel) immer an derselben Stelle in der graphischen Darstellung stehen oder nicht. Wir k6nnen also wahrnehmen, ob der Knochen oder das Schadelmerkmal einer Aus­ nahme oder einem Mittelwerte entspricht,' ob die Maximum- und Minimumvarianten eines bestimmten Merkmafes auch als solche fUr ein anderes Merkmal desselben Knochens oder Schadels gelten oder nicht u.s.w. Die Variabilitat korrespondierender Merkmale fUr gleichartige Knochen festzustellen, ist also die erste Aufgabe, danach kommt es aber darauf an, auszu­ machen, wie diese Variabilitatsgrenzen sich denjeni­ gen der in der Literatur beschriebenen Formen und der wilden Verwandten gegentiber verhalten, oder umgekehrt. (pp.63-64)

As already mentioned above, Van Giffen did not carry out this intention in Die Fauna der Wurten. He only discussed some of the results of his research. In 1913 Van Giffen mentions that, according to him, the cattle from the terpen were pro­ bably of monophyletic descent. Neither in the form of the horn cores nor in the length of the metapodia was Van Giffen able to find the four different forms: Primigenius, Brachy­ ceros, Aceratos and Brachycephalus, that were at that time recognized as belonging to pre­ history. He ends his discussion as follows: Liegt nun der Gedanke nicht nahe, der Brachyceros sei eine Ktimmerform des Primigenius, und der Ace­ ratos (hornless, A.T.C.) eine des brachyceren Typus, und die Abstammung sei monophyletisch? (p.62)

At present the aurochs is considered as the progenitor of domesticated cattle (Herre & Rohrs, 1973). The skull of a wild pig - Sus sero/a, from the terp Cornjum northwest of Leeuwarden in the province of Friesland was compared with two skulls of the Indian wild pig (Sus vittatus from Sumatra) because it was consi­ dered by a few as the wild parent species for Sus palustris domesticus. Van Giffen came to the conclusion that the skull of Sus vittatus differed from those of the wild pigs described by Riitimeyer, and that the skull from Corn­ jum was comparable to the Swiss lake pig skulls. He intended to return to this compa­ rison when he discussed the remains of do­ mestic pigs. Although Van Giffen had only one skull and an upper jaw available for comparison he ended his description by de­ claring that den riesigen Dimensionen der Wurtenwildschweines auf dessen zahlreiches Vorkommen in der damaligen Zeit mit einiger Sicherheit schliesen zu dtirfen. (p.123)

4. VAN GIFFEN AFTER 1913 In Van Giffen's later archaeozoological work, only data of dog skulls were statistically evalu­ ated. In a paper from 1915-16 he found that the dogs from the terpen area (c. 600 B.c. ­ 1000 A.D.) formed a homogeneous group, comparable to a small group of dogs with the same variation in measurements from Bronze Age lake settlements in Switzerland. The ma­ jority of the dog skulls from the Bronze Age and Neolithic Swiss lake settlements were small­ er, however. These smaller dogs had, according to Van Giffen, characteristic features of the jackal - Canis aureus, and must have developed from this species. The larger Bronze Age dogs in Switzerland and those of the terpen must, however, have had characteristics of the wolf - Canis lupus, and must have had this species as progenitor. Van Giffen thus chose for a polyphyletic descent of the dog. In 1983 the descent is considered to be monophylitic, with the wolf - Canis lupus, as wild progenitor (Herre & Rohrs, 1973; Clutton-Brock, 1981). Although more archaeozoological publica­ tions followed later (Van Giffen, 1913b; 1915­

A.E. van Giffen as archaeozoologist

16; 1917; 1927; 1928), Van Giffen soon became a full-time archaeologist. In 1921, on the oc­ casion of the official opening of the B.A.I., Van Giffen (1922) mentions three tasks for the new Institute: first, the study of organic remains both zoological and botanical, second, the study of prehistory, third, the study of geology. In the lecture with which Van Giffen, in 1930, took up his post as a lecturer at the Groningen State University, zoology is only mentioned in an indirect way when he addresses his former director J.P. van Bemmelen and promised to redeem his debt of honour (probably meaning the publication of part II of Die Fauna der Wurten). 5. VAN GIFFEN IN THE FIFTIES AND LATER Waterbolk (1973; 1975-1976) mentions the fact that Van Giffen always had an open mind concerning new developments that could be useful for archaeological research. The use­ fulness of pollen analysis for dating purposes and the reconstruction of past vegetations and climates was fully recognized by Van Giffen. In 1950, he stimulated the physicist De Vries to start Cl4-dating in Groningen and he engaged Van Zeist in 1953 for pollenanalytical research. A zoologist was not appointed by Van Giffen although the work of Kuhn (1935; 1937; 1946), Hescheler (1920), Hescheler & Rueger (1940; 1942) in Switzerland, and Nobis (1955) in Germany had laid the basis for the research of faunal complexes from single sett­ lements. In the fifties, it was in the first place the German veterinarian Boessneck (1956; 1958) who saw the importance of this kind of archaeozoological research. Together with his collaborators and students he undertook research on a large scale on subfossil faunal material from numerous settlements, stimu­ lating others to do the same. In 1952 Van Giffen described the animal remains collected already in 1927 (Van Giffen, 1950) in the Bolleveen near Zeijen in the province of Dren­ the, as a single complex. In 1953 he again published' some data concerning prehistoric dogs. Also, after his retirement in 1954, he showed an interest in archaeozoological re­ search. Could it be that Van Giffen still con­ sidered himself active in the field of zoological

5

research until the last years of his directorship? When in 1952 the Vereniging voor Zoog­ dierkunde en Zoogdierbescherming (V.Z.Z.) (Society for the Study and Protection of Mam­ mals) was founded, Van Giffen became a member of this new society and attended the meetings regUlarly. The V.Z.Z. appointed Van Giffen as a Honorary member during the winter meeting of 27th January 1968, at Gouda, for his contributions to the study of zoology. In October, 1970, the members of the V.Z.Z. met in Groningen and visited the terpen region afterwards. On this occasion, in Ezinge Van Giffen, at the age of 86, gave a lecture on his excavations of this terp in the early 1930's, standing before the steep profile left after the excavations.

6. ACKNOWLEDGEMENTS The manuscript was corrected by Mrs. B.M. van der Meulen-Melrose. The text was typed by Ms. H. Klaassens.

7. REFERENCES Bakker, L., 1909. Sudien uber die Geschichte, den heutigen Stand und die Zukunft des Rindes und seiner Zucht in den Niederlanden mit besonderer kritischer Beruck­ sichtigung der Arbeitsweise des Niederliindischen Rind­ viehstammbuches. Diss. Bern. Boessneck, J., 1956. Tierknochen aus spiitneolithischen Siedlungen Bayerns (= Studien an vor- und fruhge­ schichtlichen Tierresten Bayerns 1). Munchen. Boessneck, J., 1958. Zur Entwicklung vor- und fruhge­ schichtlicher Haus- und Wildtiere Bayerns im Rahmen der gleichzeitigen Tierwelt Mitteleuropas (= Studien an vor- und fruhgeschichtlichen Tierresten Bayerns 2). Munchen. Broekema, C., 1910. Overblijfselen van Bos taurus pri­ migenius (Rutimeyer) in de terpen. Cultura 22, pp. 475-477. Broekema, L., 1908. Eene kleine bijdrage tot de kennis van de fauna onzer terpen. Cultura 20, pp. 722-724. Broekema, L., 1909a. De overblijfselen van paarden in onze terpen. Cultura 21, pp. 688-695. Broekema, L., 1909b. Verdere waarnemingen over de fauna onzer terpen in Friesland en Groningen. Cultura 21, pp. 57-59. Broekema, L., 1910. De schapen der vroegere bewoners onzer terpen. Cultura 22, pp. 136-150. Broekema, L. 1912. Overblijfselen van varkens uit onze Friessche terpen. Cultura 24, pp. 71-77. Clason, A.T., 1973. In memoriam Prof. Dr. A.E. van Giffen. Lutra 15, pp. 13-15.

6

A.T.CLASON

C1utton-Brock, J., 1981. Domesticated animalsfrom early times. London. Giffen, A.E. van, 1913a. Die Fauna der Wurten. 1. Teil. Diss. Groningen. Giffen, A.E. van, 1913b. lets over terpen en den terp­ hondo Handelingen van het 14de Nederlandsch Na­ tuur- en Geneeskundig Congres, pp. 468-481. Giffen, A.E. van, 1915-1916. De statistische methode in het huisdiervraagstuk. Tijdschrift der Nederland­ sche Dierkundige Vereeniging (2e serie) 14, pp. 48­ 60. Giffen, A.E. van, 1917. Statistisch-bio1ogische gegevens voor de pa1ethnografie. Handelingen van het 16de Nederlandsche Natuur- en Geneeskundig Congres, pp. 256-258. Giffen, A.E. van, 1922. Het Biologisch-Archaeologisch Instituut en zijn taak. Groningen. Giffen, A.E. van, 1927. Het oudste huisdier en de pa1ethno1ogie. Verslag van de gewone vergadering der Afd. Natuurkunde van de K.A. W. 36, pp. 1287-1297. Giffen, A.E. van, 1928. De Woerd, genaamd de "Ker­ kenhof' te Ressen, gem. Ressen-Bemme1 (with M.A. Eve1ein). J. v.T. 11-12 (1926-1928), pp. 16-29. Giffen, A.E. van, 1930. Archaeologische verkenmerken. Openbare voordracht. Groningen. Giffen, A.E. van, 1950. De nederzettingsoverblijfse1en in het Bolleveen en de versterking, de zgn. "le­ gerp1aats" aan het Witteveen op het Noordse Veld, beide bij Zeijen, gem. Vries. Nieuwe Drentse Volks­ almanak 68, pp. 89-123. Giffen, A.E. van, 1952. Het Bolleveen bij Zeijen, gem. Vries (naschrift). Nieuwe Drentse Volksalmanak 70, pp.89-108. Giffen, A.E. van, 1953. De v66rhistorische vuursteen­ exp10itatie bij Rijckholt in Nederlands Limburg. La brochure publiee par la Societe Royale Beige d'An­ thropologie et de Prehistoire 1, pp. 1-6. Herre, W. & M. Rohrs, 1973. Haustiere - zoologisch gesehen. Stuttgart. Hesche1er, K., 1920. Beitrage zur Kenntnis der Pfah1­ bautenfauna des Neo1ithikums (Die Fauna der Pfah1­ bauten in Wauwy1ersee). Vierteljahresschrift der Na­ turforschenden Gesellschaft in Zurich 65, pp. 248­ 332. Hesche1er, K. & J. Rueger, 1940. Die Wirbe1tierreste aus den Pfah1bauten des Ba1deggersees nach den Grabungen von 1938 und 1939. Vierteljahresschrift der Naturforschenden Gesellschaft in Zurich 87, pp. 59-70. Hesche1er, K. & J. Rueger, 1942. Die Reste der Haustiere

aus den neolithischen Pfah1baudorfern Ego1zwi1 2 (Wauwi1ersee, Kt. Luzern) und Seematte-Ge1fingen (Ba1deggersee, Kt. Luzern). Vierteljahresschrift der Naturforschenden Gesellschajt in Zurich 87, pp. 383­ 486. Hue, E., 1907. Musee Osteologique. Etude de la fauna quarternaire, osteometrique des mammiferes. Paris. Kuhn, E., 1935. Die Fauna des Pfah1baues Obermei1en am Zurichsee. Vierteljahresschrift der Naturforschen­ den Gesellschaft in Zurich 80, pp. 65-154. Kuhn, E., 1937. Die Fauna der Wallsied1ung im Borscht (Endneo1ithikum - 1a Tene). Aus dem zo%gischen Museum der Universitiit Zurich, pp. 4-42. Kuhn, E., 1946. Die Tierreste des bronzezeitlichen Pfah1­ baues in der B1eiche von Arbon (Kt. Thurgau), Bericht der schweizerischen Pa1aonto1ogischen Ge­ sellschaft. Eclogae geologicae Helvetiae 39, pp. 364­ 366. Nobis, G. 1955. Die Entwick1ung der Haustierwelt Nordwest- und Mitte1deutsch1ands in ihrer Bezie­ hung zu 1andschaftlichen Gegebenheiten. Peter­ manns Geographische Mitteilungen 1, pp. 2-7. Rutimeyer, L., 1862. Die Fauna der Pfahlbauten der Schweiz. Basel. Tammes, T., 1907. Der Flachsstengel. Eine Stat. Anat. Monographie (= Natuurkundige Verhandelingen van de Hollandse Maatschappij der Wetenschappen VI, 4e st). Waterbo1k, H.T., 1973. A.E. van Giffen, Noordhorn 14 March 1884 - Zwolle 31 May 1973. Palaeohistoria 15, pp. 7-12. Waterbo1k, H.T., 1975-1976. Albert Egges van Giffen. Noordhorn 14 maart 1884 - Zwolle 31 mei 1973. Jaarboek van de Maatschappij der Nederlandse Let­ terkunde te Leiden pp. 122-153. Waterbo1k, H.T. & J.W. Boersma, 1976. Bewoning in v66r- en vroeghistorische tijd. In: W.J. Formsma et al. (eds.), Historie van Groningen, stad en land. Groningen, pp. 13-74.

8. KEYWORDS The Netherlands, Groningen, archaeozoology, terpen area, Wurten, Bos prim igen ius, B. Brachy­

ceros, B. aceratos, B. brachycephalus, Sus scrota,

S. palustris, S. Vittatus, Canis aureus, C. lupus,

H austierkunde.

SOME NOTES ON RECENT AND PRE- AND PROTOHISTORIC FISHING GEAR FROM

NORTHWESTERN EUROPE Dick C.Brinkhuizen CONTENTS 1. INTRODUCTION

2. RECENT FISHING GEAR AND FISHING TECHNIQUES 2.1. A few recent fishing techniques not requiring the use of any fishing gear 2.2. Fishing methods using actively operated fishing gear 2.3. Passively-operating fishing gear 2.4. Auxiliary fishing equipment 3. ARCHAEOLOGICAL FINDS OF FISHING GEAR 3.1. The method of approach 3.2. The fish-spear 3.3. The gorge 3.4. The fish-hook 3.5. The weir 3.6. The wickerwork fish-trap 3.7. The fishing net 4. SUMMARY AND CONCLUSIONS 5. ACKNOWLEDGEMENTS 6. NOTES 7. REFERENCES 8. KEYWORDS

7

8

D.C.BRINKHUIZEN

1. INTRODUCTION Objects are sometimes found that are interpre­ ted by the archaeologist as fishing gear or parts there of. For various reasons finds of items of fishing gear are very rare in direct relation to the remains of their victims. In recent years large quantities of fish remains have been retrieved with more precise excavation me­ thods, but the fishing gear with which these fish were caught are often absent or are un­ recognizable as such. Sometimes the investiga­ tion of fish remains provides data from which it can be deduced what kind of fishing gear may have been used. Ethnographical sources can also be of use in providing relevant data. Such sources can only be made use of, however, on a certain condition. The natural environment described in the ethnographical source concer­ ned must be more or less equivalent to the . natural environment ascertained for the pre­ historic settlement. From a combination of archaeological, pedological, botanical, zoologi­ cal and ethnographical data it is then permis­ sible to attempt reconstructions of prehistoric fishing gear. The first part of this article deals with a number of fishing methods and types of fishing gear known from ethnographical sources. Much of this information has been obtained from Sirelius: Ober die Sperrjischerei bei den Finnisch­ Ugrischen Valkern (1906) and Die Volkskultur Finnlands. Jagd und Fischerei (1934). In these works the fishing tackle and fishing techniques used by the Lapps, Woguls, Ostjaks and Syrjans are described in detail. The last three of these peoples live immediately east of the Ural Moun­ tains along the rivers that flow into the Ob basin. The aquatic environment in this region is freshwater and contains a standing popula­ tion of i.a. pike, perch, burbot and members of the carp family (Cyprinidae) and a migrant popUlation mainly consisting of whitefishes (Coregonidae). This ichthyofauna is compara­ ble to that of the freshwater environment of Northwestern Europe. The second part of this article deals with a variety of pre- and protohistoric fishing gear from Northwestern Europe.

2. RECENT FISHING GEAR AND FISH­ ING TECHNIQUES A characteristic feature of primitive cultures is that social organization, food acquisition, medicine and religion have not become iso­ lated elements of cultural identity. No diffe­ rentiation of these elements has taken place. Even in present-day developed societies it is still possible to trace remnants of cultural traditions that remind us of more primitive societies. According to Van Doorn (1971), in Western Europe this is still valid to a limited extent for inland and sea fishery, notably with regard to the following points. a. The trade and its associated jargon are handed down from father to son; there is no official occupational training. b. The fisherman is dependent on the na­ tural breeding patterns of the fish . c. The equipment handed down and the working ~ethods remain the same, with a few exceptions. d. There is little contact between the fish­ erman and other trades. The fisherman ma­ kes his own equipment. He carves his knitting­ needles out of wood, makes his own nets and weaves his own wickerwork fish-traps. e. Forms of taboo are very abundant among fisher-folk and sailors. f. Because automatization and mechaniza­ tion play hardly any role in freshwater fishery, the terminology of the freshwater fisherman includes many words that have undergone little or no differentiation. Thus there exists in this terminology e.g. a close relation be­ tween bird and fish nomenclature. In non-agrarian societies of the past and present, gathering, hunting and fishing are essentially the only means of existence. There is no great distinction between these three. It could be said that hunting and fishing are methods of gathering that require the use of tools. An evolution in fishery technique is evident, from fishing for individual fish to catching fish en masse. Thus we see the development of fishing gear from those forms in which one fish can be caught at a time, to those forms in which many fish can be caught simulta­ neously. As for fishing methods, a correspon­

Recent and pre- and protohistoric fishing gear

ding distinction can be made between active and passive fishing. What I mean by active fishing is that the fisherman searches for the fish or lures it towards himself and then tries to catch it. With passive fishing the fisherman only has to collect the fish that have become caught in his traps or nets. Those types of fishing gear with which only one fish at a time can be caught generally involve active fishing, while those types of gear designed to catch many fish simultaneously involve passive fish­ ing, although it is not always easy to make a sharp distinction between these two types of gear.

2.1. A few recent fishing techniques not re­ quiring the use of any fishing gear These techniques are very simple, and often give good results. The use of such techniques by prehistoric fishermen will not have left any traces for the archaeologist to discover. Never­ theless I shall briefly describe some of these fishing techniques.

2.1.1. Catching fish with one's bare hands Many fish, even the swift and elusive trout, can be caught with one's bare hands. With this method one must be able to see or feel the fish.

2.1.2. Stunning fish with a stick or cudgel A fish that can be seen in shallow water can be struck dead with a stick or cudgel. The Syrjans had a certain method of fishing that was practised in the autumn. When ice forms over a surface of water, the oxygen distribution in the water changes and the fish come up to just under the ice. They are visible through the thin ice if there is no covering layer of snow. A blow is struck on the ice immediately above the fish by means of a stone, attached to a wooden haft. The pressure wave generated in this way stuns the fish. Since thin ice cannot take the weight of an adult man, boys practised

9

this method of fishing (Sirelius, 1934). 2.1.3. Catching fish by making the water turbid In small pools with a muddy bottom the fisherman stirs up the silty substratum with a stick, so that the water becomes turbid. The fish thus suffer a shortage of oxygen, float up to the surface and can be lifted out of the water. 2.1.4. Poisoning Orall known fishing methods fishing with the aid of poison is the most effective. After being poisoned almost all fish come to the surface, more or less in a state of torpor, and from there they can be lifted out of the water. In Asia and South America many fish poisons are known to fishermen. Rotenone, which is prepared from the roots of plants of the genus Derris, is used at the present time in biological research concerned with fishery. This poison is effective at very low concentrations (0.5 I per 1,000,000 I water) and is harmless to people. Formerly pOisons were also used in the Netherlands. Such poisons were freely available and among Dutch fishermen were known as koggelbonen (Van Doorn, 1971). Also in ancient times in Europe various poi­ sons derived from plants were used in catching fish. An extract of the flowers of the great mullein was used by the Greeks. The Romans used the juice of cyclamens (Muus & Dahlstmm, 1968). 2.2. Fishing methods using actively operated fishing gear With these methods specially made equipment is used by the fisherman who is an active participant in the process of catching fish. The types of gear include the stick fitted with a hook, the rod with a bob, the stick with a noose, the fish-spear (fish harpoon, leister), the gorge, and the fish-hook. A number of these are in evidence in various archaeological contexts.

10

D.C.BRINKHUIZEN

2.2.1. The stick with hook or gaff A bent nail is fixed to a stick. A piece of bait is put on the nail. The baited stick is held in the water until the fisherman feels a fish biting. Then he suddenly flings up the stick with the fish on it out of the water (Barthel, 1977). The fisherman can also use a non-baited stick. When he sees his prey, he carefully shoves the stick with hook under the fish and drags the fish out on to dry land. This is the way in which the Kwakiutl, who lived on the coast of British Columbia, caught salmon when these fish swam up the rivers en masse (Forde, 1934). This method was also used in winter by the Woguls to catch burbot at night with the aid of artificial light (Sirelius, 1934). 2.2.2. The rod with a bob The fisherman threads together a number of worms from end to end on a string and makes a clew out of them. This clew of worms, or bob as it is called, is hung on a string, that is fixed to the end of a rod measuring about 1 m in length. With the aid of a piece of string a stone is attached to the lower surface of the bob (fig. 1). The weighted bob is let down into the water as far as the bottom. It is then repeatedly raised a few decimetres and allowed to sink again. The worms and the movement of the bait serve to attract eels. These bite into the bob. The fisherman feels this and lifts the line with the eel attached carefully out of the water. This method of fishing is carried out at night. 2.2.3. The stick with a noose With this method a noose of copper wire or other material, attached to the end of a stick, was carefully put into the water around the body of a fish spotted by the fisherman. Then the noose was suddenly pulled tight an the fish was flung on to dry land. This method was used in the Netherlands to catch pike. A few remains of pieces of wood from a peat­ bog near Oberdorla (East Germany), dating from Roman times, have been interpreted by Barthel (1977) as including a stick with a hook and a stick used for fishing with a noose.

Together with the latter the remains of a noose of horse-hair were found. 2.2.4. The fish-spear The basic form of this kind of fishing gear is a long straight wooden stick with a forked extremity. In the recent past the Lapps used this kind of fishing gear to catch arctic charr (Sirelius, 1934). Elsewhere we find fish-spears with separate points (prongs). These prongs are often ser­ rated or barbed on one or both edges. They are made out of bone or metal and are fixed to the wooden shaft by means of cord. The most highly developed metal fish-spears have a socket into which the wooden shaft is fitted. Figure 2 shows a number of examples of recent fish-spears. There are also fish-spears with detachable prongs. When a fish is hit the prong comes loose from the shaft, but remains connected to it, at the end of a long line. We then call the fish-spear a fish-harpoon and the prong the harpoon-point. Fish are also caught by means of bows and arrows. The fishing arrow that was used by the Ostjaks and Woguls consisted of a long shaft without any feathers and with a heavy double-pointed arrow-head (Sirelius, 1934). Fishing with a fish-spear (fish-harpoon, fish­ ing arrow) is a well-known practice almost all over the world. The fish-spear is used in places where the water is clear. When the fisherman sees his prey, he tries to spike it with his fish-spear or to shoot it using his bow and arrow. These techniques require a certain skill, as the fish is situated in a different spot in the water to where it appears to be, on account of diffraction. A correction therefore has to be made by the fisherman. If the fisherman with his fish-spear wishes to exploit a local population of non-migratory fish, he will have to search for his prey. A freshwater, non-migratory fish that is eminent­ ly suitable for catching with the aid of a fish­ spear is the pike. As these large predators frequent shallow water close to the bank or shore, either when spawning or when lying in wait, motionless, for their prey, they easily fall victim to the fisherman armed with his spear. As this species is attracted to light, it

Recent and pre- and protohistoric fishing gear

can also be speared at night using artificial light (Sirelius, 1934). Pike can be caught all the year round. For the fisherman with his spear they are most vulnerable, however, from the beginning of their breeding season (Feb­ ruary) until late in the autumn. When the water temperature drops they search for deeper water. When the water is frozen over, they can be speared through blow-holes in the ice. Also other species of freshwater, non-mi­ gratory fish (e.g. bream) can be speared when they are spawning in shallow water close to the bank or shore. If the fisherman with his fish-spear is ex­ ploiting a migratory population, there is no need for him to search for his prey. He simply waits at a suitable spot until the fish swim past. Fish species that fall into this category include salmon and sturgeon. The large salmon arrive from the sea at the mouths of rivers and swim upstream together in large numbers in order to spawn. Then the water of clear rivers looks red on account of the great number of salmon in their spawning colour. The clearly visible, large fish thus form an easy prey for the fisherman with his spear. Also sturgeon can be caught using fish­ spears. In the Fraser, a river in British Co­ lumbia, fishermen used to harpoon sturgeon from their canoes (Lord, quoted by Clark, 1948). Inhabitants of the basins of southward flo­ wing rivers in the Urals also used to harpoon sturgeon. This took place in winter, when the rivers were frozen over. A large number of fishermen together smashed holes in the ice, and tried to harpoon the sturgeon that had been startled by the noise. When the fishermen had fished part of the river in this way, they went further downstream and repeated their performance (Mohr, 1952). Also in the Netherlands the fish-spear is known to have been used. Its use is a thing of the past, however, as it is now forbidden by law. It was used formerly for catching flounders and eels. The flounder occurs abundantly in the Wad­ den Sea. When this fish is startled its escape path in the water is marked by clouds of disturbed bottom-sediment.. From these clouds of sediment the fisherman is able to

11

locate his prey, and spike it with his fish-spear. The fisherman could also simply stand on top of the fish, so as to hold it down, and remove it from under his foot. This method is called bottrappen in Dutch (literally: flounder-tread­ ing). The fish-spear that was used in the Nether­ lands for catching eels is known as an aalschaar or elger in Dutch, equivalent to the English term eel-spear (fig. 3). The fisherman stood on the bank or in a small boat and thrust the eel-spear haphazardly into the mud. On lifting the spear out of the water he could see whether any eels had been caught between the prongs. 2.2.5. The gorge

The gorge is a short stick-like object, cylin­ drical in the middle and pointed at both ends. It may be made of wood, bone or iron. Gorges come in varying sizes. They may be only a few centimetres long, or as much as 10-15 centimetres. In the middle of the gorge (at its centre of gravity) a line is attached. To prevent the line from slipping off, a groove, or a few shallow grooves or a notch may be cut in the surface of the gorge (fig. 4). The line is thrown into the water with the gorge hidden inside some kind of bait. If the fisherman feels a fish biting, he gives the fish time to swallow the bait completely. Then the fisherman pulls hard on the line so as to make it taut. In this way the gorge is pulled into a position per­ pendicular to the line and thus becomes stuck in the throat of the fish. The line with gorge can also be used for passive fishing. The fisherman fastens the loose end of the line on to something and goes away. When a fish swallows the baited gorge and swims away, the line is pulled taut and the gorge similarly becomes stuck in the throat of the fish. The size of the gorge, the strength of the line and the kind of bait used are determining factors for the species of fish that may be caught. Partly because of its behaviour and the shape of its snout, the pike is a very suitable species of fish for catching with the aid of a gorge. The gorge is then baited with a small

D .C.BRINKHUIZEN

12

lig.1

lig. 2

fig.3

lig.5 lig.4

lig.6

lig . 7

Recent and pre- and protohistoric fishing gear

fish. The fish used as bait is partly slit open along the ventral surface, and the fisherman inserts the gorge into this aperture. Gorges are in fact used for a variety of purposes. Until recently gorges were used to catch wildfowl on Lake Constance (Clark, 1948). Also Sirelius (1934) mentions the use of gorges in catching waterfowl. At the same time he states that elsewhere in the world gorges are used to catch crocodiles. 2.2.6. The fish-hook

If we talk about a fish-hook, then this piece of equipment must consist of three distinguish­ able elements. These are: the hook-shaft, the hook-bend and the hook-point. The hook-shaft is the more or less straight part of the hook to which a line is attached. At the end of the hook-shaft there is usually some kind of modification for the attachment of the line. This may be in the form of a thickening of the shaft, a hole, one or more transverse notches in the shaft or one or more grooves cut into the shaft and running round it. The hook-shaft passes into the hook-bend, that may be regularly curved or angular. The hook-bend passes into the hook-point. This is more or less parallel to the hook-shaft. The hook-point is shorter than the hook-shaft and has a pointed extremity, that can penetrate the body of the fish. This extremity may be fitted with a barb to prevent the fish escaping. For all types of fish-hook it is important that they have a good penetrating capacity, i.e. when in use the hook-point must point in the right direction for it to be able to penetrate the fish easily. At the same time the hook must be of the optimal shape, so as to minimize any chance of the fish escaping. This brings us to the concept of 'angle of grip' (Lekholm, 1951). If we let a line with a fish-

Fig. 1. The rod with a bob.

Fig. 2. Examples of fish-spears (after Sirelius, 1934): a,

b. Greenland; c. Russia; d. Finland.

Fig. 3. The eel-spear (after De Groot & Schaap, 1973).

Fig. 4. The gorge (schematic).

Fig. 5. The fish-hook and its mode of operation (sche­

matic).

Fig. 6. The hoekwant (schematic).

Fig. 7. Ledger-line for catching sturgeon (schematic).

13

hook hang vertically, then we see that the hook-shaft hangs more or less as an extension of the line. In such a case the line must be attached to the extremity of the hook-shaft. If we now place the hook with its point against a hard object (the roof of the mouth of the fish) and pull on the line, we see that the hook takes up a different position. This position is characterized by an imaginary line running from the hook-point to the point of attachment of the line to the hook-shaft. We call this the line of pull (A-B). A second imaginary line is formed by the longitudinal axis of the hook-point (the line C-D). The two imaginary lines A-B and C-D form an angle with each other. This angle is called the angle of grip (fig. 5a). To obtain optimal penetration, the direction of the hook-point C-D should coincide with the line of pull A-B (fig. 5b). With a fish-hook of this kind, 'hooking' a fish is not very easy. The hook opening (the distance between the hook-point and the hook­ shaft measured perpendicular to the hook­ shaft) is small. On the other hand, the hook does meet the requirement that the chance of the fish escaping is slight. The hook 'holds' well. . If the hook-point is bent away slightly from the hook-shaft, then hooking a fish will be easier. The hook opening is then bigger (fig. 5c). On the other hand 'holding' the fish will be a problem, as the line of pull is a different one to that where the point of the hook becomes embedded. Then there is a chance that the fisherman will pull the· hook out of the mouth ofthe fish. To prevent this the hook­ point is provided with a barb. A good fish-hook will therefore have to be a compromise between one that 'hooks' well and one that 'holds' well. Borne, cited by Lekholm (1951) estimated an optimal 'angle of grip' of 10°. Present-day steel hooks with a barb have angles of grip of 10° to 30°. With steel hooks the 'angle of grip' can be modified by making the hook-shaft longer or shorter, by slightly bending the hook-point or by altering the curvature. A relationship evidently exists between the characteristic features of a particular steel hook and the species of fish that can be caught with it. A combination of data, based on experience, of Biscoff, cited by Lekholm (1951), and data of Lekholm (1951) provides us with the information on

14

D.C.BRINKHUIZEN

Table 1. Relationship between the size and hook opening of steelfish-hooks and the species of fish that can be caught with them Fish species Estimated maximum length of the fish (cm) Bleak Roach Rudd Whitefish (houting) Bream Tench Perch Eel Pike Cod Salmon Tunny

20 40 45 50 80 70 50 100 130 150 150 300

Estimated maximum weight of the fish (kg)

Totallength Hook of the hook opening (mm)

(mm)

2

up to 10 15 18-22 25

3-5 4-5 ? 5-9

9 8 3.5 4.5 28 40 36 300

22-28 22-28 28-30 28-35 30-80 30-70 up to 110 100-160

5-6 6 5-8 ? 7-23 ? 7-23

0.04 1 1.5

steel hooks presented in table 1. From the table it follows that small fish cannot be caught with large hooks. In general we can say that the fisherman uses small hooks and thin lines to catch smaller kinds of fish and large hooks and thick lines to catch larger fish. At the present time steel hooks are used almost exclusively. This applies to fishing activities on the European mainland, that are for the most part carried out by amateurs as recreation. In commercial fishery at sea, lines with hooks are sometimes used because fishing . with nets may not be possible for various reasons. Until recently, in the Netherlands, fishermen both of inland waters and at sea worked with ledger-lines. These were called beug or hoek­ want in I)utch:They consisted of- very long lines with at certain intervals shorter transverse lines, that each carried one baited iron hook. At the same time floats and weights were attached (fig. 6). The hoekwant was cast in such a way that it came to lie more or less on the bottom. In inland waters eels were caught in this way. At sea the species mainly caught in this way were cod and haddock. Along the coasts of the Black Sea, the Caspian Sea and in the rivers that debouch into them fishermen use ledger-lines to catch sturgeon. These ledger-lines are comparable to the hoekwant. There are two important

differences however in the way in which the ledger-lines are set up. In the hook-bend of the non-baited hooks extra side-lines are af­ fixed, each with a float. Consequently the side­ lines take up a vertical position in the water (fig. 7). The main line lies on the bottom. The distance between two successive hooks is 20­ 30 cm. The sturgeon, which swims close to the bottom in its journey upstream, has to pass through the side-lines and thus may get caught on a hook (Mohr, 1952).

2.2.6.1. Primitive fish-hooks Sirelius (1934) gives ethnological descriptions of some fish-hooks made of wood and bone that were used by the Finns, Woguls and Ostjaks. On the basis of the different forms in use Sirelius was able to establish a series showing different stages of development (fig. 8). In hook a he sees a double-pointed wooden gorge with a pointed side-piece. The line is attached at the point where the main shaft and the side-piece meet. Hook b already has the shape of a true fish-hook. However, the line is still attached at the point where the main shaft and the side-piece meet (in the hook-bend). Hook c has the same shape as hook b, except that the line is attached in the middle of the hook-shaft. At this point a notch has been made on the inner side of the hook­ shaft. (Also hook b has a notch here, though it is not clear why). The final result of the shifting of the point of attachment of the line is to be seen in hook d. The hooks illustrated all have a total length of 8-9 cm. They are all cut from the tough wood of juniper, heather, honeysuckle and (sometimes) birch. At the same time we see that the hook-bend is wound around with cord. The fisherman knows from experience that the hooks made out of a single piece of wood (or bone) break easily in the hook-bend. Therefore he makes composite hooks, that consist of two separate components, a hook­ shaft and a hook-point. If a fish is hooked that is too heavy, then it is possible that the hook-point will break off. The fisherman then still has the hook-shaft. He only has to make a new hook-point. The Woguls and the Ostjaks used the fish­ hooks exdusivel~ as passively-openlting fish­ ing gear in tlie· first Place for catching pike,

Recent and pre- and protohistoric fishing gear

also for large perch. The hook, baited with a small fish, was put in the water on a line with a large float. The tip of the line was made of a strip of leather and the main line of split roots of pine (fig. 9). 2.3. Passively-operating fishing gear In the preceding section we have already become acquainted with a passively-operating kind offishing gear, the Tedger-=1ine.·With this fishing gear, however, only one individual can be caught per hook. Moreover the hooks have to be baited again every time. This is not very effective and other kinds of fishing gear have been developed that enable fishermen to catch many fish at the same time. These kinds of passively-operating fishing gear are: the weir, the fish-surround, the wickerwork fish-basket and the fishing net. 2.3.1. The weir The weir is a kind of barrier or obstruction that leads fish to a spot where they can be easily caught and taken out of the wate'r. The kind of construction material used depends on the sediment of the river-bed, lake bed or sea bottom (whether consisting of rock, gravel, sand or clay) and the construction materials available in the local environment. Thus weirs ma,y be made of stones, felled trees, posts, sticks, bunches of twigs or woven screens of laths, bamboo, etc. In stagnant or sluggishly flowing water the weir is always used in combination with a more elaborate kind of passively-operating fishing gear. On the other hand the weir itself can function as a kind of passively-operating fish­ ing gear in places where there is a strong current or where there are great diurnal dif­ ferences in water level. In the former case the fisherman will have to use a splashing stick (2.4.1.). Sirelius (1906) described how the Ostjaks and the Woguls were still using the weir in its most primitive form at the beginning of this century (fig. 10). A few hundred metres upstream from the weir a fisherman uses a splashing stick to drive the fish in the direction of the weir. The fish that have collected up

15

against the weir are taken out of the water by a second fisherman with the aid of a dip­ net or stick with a hook. Fish regularly caught in this way include pike, perch and members ofthe carp family. This method of fishing must be carried out by at least two people, because the fish only stay close to the weir when someone continually splashes about further upstream. This is therefore not a true kind of passively-operating fishing gear. Sirelius also described the passively-opera­ ting weir in its most simple form as used by these same ethnic groups. This is called laltam in those parts. A wooden weir is built trans­ verse to the direction of the stream flow. At the same time, slightly further upstream from this weir two barriers are built, that point diagonally towards the weir. In this way a funnel construction (throat) is the result, with a chamber behind itserving to collect the fish (fig. 11). The fish are driven into the chamber with the aid of a splashing stick, the throat. opening is closed with a post or screen and the fisherman takes the fish out of the water. using a dip-net' or stick with a hook. This ~ method of fishing is carried out by one person. Another primitive kind of passively-opera­ ting weir is to be found in the Departeinent du Var (France). In a swiftly flowing stream a V-shaped weir of stones is built. The apex of the weir points downstream. In the apex itself an opening is left, in which is placed a bunch of twigs weighted down with stones (fig. 12). The fish present upstream of the weir are driven towards the weir by means of splashing the water-surface with a splashing stick. At the weir itself the increased strength of the current forces the fish into the bunch of twigs where they become caught up and killed. The bunch of twigs is now lifted out of the water and the fish caught up in it are removed. 2.3.1.1. The development of the weir It is possible that natural barriers could have served as an example for the construction of an artificial weir. Natural barriers are to be found in places where the water level fluctu­ ates. This occurs e.g. in places where rivers overflow their banks or along coastlines with a large difference between the low and high tide marks (e.g. the coast of Bretagne). As the water rises the fish come close to the coast

D.C.BRINKHUIZEN

16

fig.8

fig.8

fig.l0

fig. 11

fig.13 fig.8

fig.15

fig.14

Recent and pre- and protohistoric fishing gear

to search for food, and as the water goes down they swim back again. In places where there is a natural barrier, however, there is a chance that some fish will not be able to swim back in time. They stay behind in pools or become stranded and thus become an easy prey for animals or for man. Obviously the fisherman will make use of such situations, and will extend natural bar­ riers with large stones and pieces of wood so as to construct effective passively-operating fishing gear. In places along the sea coast where no natural barriers occur, wooden weirs are built that are V-shaped or in the form of an arc. With weirs shaped like this the fish are prevented from escaping as the water goes down. We find examples of such weirs on the coasts of Thailand, India and Madagascar. They also occur in Western Europe, namely in France along the coast of the Atlantic Ocean (fig. 13), and in Germany on the Baltic coast near Schleswig. Here such a weir is known as a gaard (Sirelius, 1906).

2.3.2. The fish-surround (German: Fischzaun) Obviously in wide rivers and lakes there is no point in building a weir or liiltiim. The collection chamber would be far too big and it would be impossible to take all the fish out of it. Therefore the chamber is made smaller so as to form a 'fish-surround', which is more or less round in shape. The two sides of the weir both project into the chamber (fig. 14). As the fish will remain in the chamber for some time, they will try to find all possible hiding places. The use of lath screens can solve

Fig. 8. Development of the fish-hook based on examples

from Finland and Russia (after Sirelius, 1934).

Fig. 9. Wooden fish-hook with line and float (Ostjaks)

(after Sirelius, 1934).

Fig. 10. The weir. S direction of flow of the river.

Fig. 11. The liiltiim (after Sirelius, 1906).

Fig. 12. V-shaped stone weir with bunch of twigs (after

Sirelius, 1906).

Fig. 13. Weir on the coast of the Atlantic Ocean in France

(after Sirelius, 1906).

Fig. 14. Primitive fish-surround with weir construction

(Woguls) (after Sirelius, 1906).

Fig. 15. Method of binding screens and wickerwork fish­

traps , using the Zwirnbindung.

17

this problem, as the laths are bound together so tightly that no escape is possible. At the same time the fishing gear can quickly be lifted out of the water in the event of a storm or floods . Thus Finnish fishermen build their fish­ surrounds out of lath screens. These screens consist of pieces of pinewood, that are shar­ pened at one end. The pieces of wood are bound together with birch twigs or straw using the Zwirnbindung (fig. 15). Cord is also made out of strips of lime or willow bast twined together. When the fish-surround and the weir are to be put in position the screens are driven into the bottom of the lake or river. Then the fisherman drives in heavier posts so as to give extra support to the screens.

2.3.2.1. The development of the fish-surround In contrast to the passively-operating weir, the fish-surround is also suitable for use in stag­ nant and sluggishly flowing water. It can catch fish swimming both upstream and down­ stream. Combinations of the fish-surround with the elements of a weir and the develop­ ment of the fish-surround itself have resulted in very complicated fishing gear. A few exam­ ples, taken from Sirelius (1906), illustrate this (fig. 16). a. When the water in the middle of a river or lake is too deep to build a fish-surround, the fisherman builds the fish-surround close to the . bank or shore (fig. 16a). The throat opens towards the bank. Immediately in front of the throat, along its midline, a weir element is placed at right angles to the bank. At the same time the bank itself serves as a weir. With this construction the fish can come into the surround from two sides. b. If the fisherman expects the fish to come from a certain direction, then he places the fish-surround with the throat opening in this direction (fig. 16b). Here too two weir elements and the bank lead the fish into the fish­ surround. c. Small, shallow rivers and lakes can be exploited by placing across their entire width a combination of a number of fish-surrounds and weirs (fig. 16c). The throat openings of the fish-surrounds face in opposite directions alternately, so that fish can become trapped from both sides. d. In the previous example each fish-sur­

18

D.C.BRINKHUIZEN

round has to be emptied individually by the fisherman. This has led to the joining together of the fish-surrounds to form one single fish­ surround. To facilitate emptying the fish-sur­ round, the fisherman can build a small col­ lection chamber on to the fish-surround. This collection chamber is also fitted with a throat (fig. 16d). e. An interesting variation is the spiral fish­ surround of Anjala (Finland) (fig. 16e). In this structure, which is made out of lath-screens, duck not infrequently get trapped. On account of the spiral shape they can no longer find the way out. Also the screens are placed so close together that the birds can no longer fly upwards. The fish-surrounds with their weirs are also often placed in rows, one behind the other (fig. 17). From historical sources and recent data it is evident tl)at the fish-surround occurs widely in Asia and Eastern Europe (Russia, Estonia, Poland and Hungary). In Western Europe on the other hand it is only found in ,Sweden, Finland (except for the region where the Lapps live), Yugoslavia and Italy (in the Po delta). On the basis of this distribution and the etymology of this fishing apparatus Sirelius (1906) comes to the conclusion that the fish­ surround came to Northern Europe from a south-eastern region of origin, via river-cour­ ses and lakes. It is a matter of some doubt, however, whether this distribution pattern is correct. In the past, in shallow places along the Dutch coast a kind of fishing gear was used that on account of its shape is strongly reminiscent of the fish-surround. This kind of fishing gear, known as a kom (literally: bowl), consisted of a long weir element that led the fish through a throat-shaped opening into a semicircular collection chamber. The underside of this collection chamber was closed off with a horizontal net. The vertical part was also made of network. The whole contraption was sus­ pended by means of ropes from vertical posts. 2.3.2.2. The use and duration of the fish­ surround (and weirs) The fish-surround is built as early as possible in the spring. This is done preferably at the onset of the thaw, when the ice can still be trodden. It is easier to build a fish-surround

from the ice than from a boat, and at the same time fish can be caught earlier. There is a risk, however, that the fish-surround may be de­ stroyed by ice-flows. As far as the catch is concerned the best results are obtained in spring and in the autumn. At the end of the autumn the fish­ surround is taken out of the water, dried and put away. The cord used to bind the laths together has often rotted away by this time. The laths are often in sufficiently good con­ dition to last another year, but many fishermen make a new fish-surround and use the old laths as firewood. 1 In Finland the fish-surround is used to catch pike, perch, bream, roach, cru­ dan carp, burbot, (whitefishes) and eel. The way in which the fisherman takes the fish out of the fish-surround is described vividly by Sirelius (1906) in the following passage: Der fischzaunfang wird in Finland nirgends als berufs­ fischerei getrieben. In den meisten fallen beschriinkt er sich auf die beschaffung der zukost fUr den hausbedarf. Es ist daher natiirlich, dass in voller arbeitskraft stehende leute ihre zeit nicht auf diese fangart verwenden, wenn sie auch bei der schwierigs­ ten aufgabe, der errichtung des fischzauns, mit hand anlegen. Das visitieren wird gem den alten miinnem und kriiftigen knaben, wenn solche im hause sind, iibertragen. In fischreichen zeiten wird es mitunter sogar zweimal am tage vorgenommen, wird aber die beute geringer, beschriinkt man auch das visitieren allmiihlich, bis es zuletzt ganz eingestellt wird. Ge­ wohnlich sieht man friih am morgen nach den fang­ geriiten. Der betreffende mann nimmt einen fisch­ zaunhamen und einen rindenranzen oder korb mit, in welch em er die fische nach hause transportiert. Gem versieht er sich, wenn er die wahl hat, mit einem kleinen boot, bestenfalls mit einem nachen und wrickt es zu seinem fanggeriit. Am ziel angelangt, steuert er sein fahrzeug zuerst an das grasufer und scheucht die fische durch trampen gegen den fischzaun zu-. Dann rudert er an die leitwand des fischzauns und kommt, indem er beiderseits im wasser start, schliess­ lich zur einkehlung des fanggeriits. Diese verschliesst er vermittelst mitgebrachter stangen oder ruder und treibt die fische aus dem vorhof in die kammern. Nachdem er auch die einkehlungen der letzteren versperrt hat, nimmt er schliesslich den hamen zur hand. Er driickt diesen am einen ende der kammer auf den grund und hebt ihn erst am anderen ende empor. So kommen aile fische auf einmal herauf. Nachdem er eine kammer oder einen fischzaun un­ tersucht hat, nimmt er die verschliisse der einkeh­ lungen weg und begiebt sich nach den anderen kam­ mem oder fischziiunen, urn bei ihnen sein glUck zu versuchen. Ausser fischen steigen als beute manchmal krebse auf, die vermutlich ein in dem fischzaun

Recent and pre- and protohistoric fishing gear verendeter fisch in die gefangenschaft gelockt hat. Es ist auch nicht selten, dass sich eine ente, gewohnlich ein von fischen 1ebender taucher, in den fischzaun verirrt. Durch die einkehlung kann sie sich nicht zuriickfinden, und die kammer ist gewohnlich so eng, dass sie beim versuch zu fliegen mit den fliigeln an die wande schlagt. Durch mannichfache misslungene anstrengungen ermiidet, muss sie notgedrungen das schicksal der fische und krebse teilen.

2.3.3. The wickerwork fish-trap This kind of fishing gear generally consists of a long funnel made of wickers plaited together, of willow or some other kind of wood that is similarly pliant. The wide opening of this funnel where the fish enter is called the mouth. This is often formed by a hoop. This word implies that the mouth is circular. However, there are also fish-traps with oval, semicircu­ lar, square or rectangular mouth openings. The external wall of the fish-trap is referred to here as the outer casing. The narrow, hindmost part of the outer casing is called the tail. Two main types of wicker fish-traps can be distinguished, namely the fish-trap without throat (German: Trompetenreuse) and the fish­ trap with throat (Sirelius, 1906) (fig. 18). The throat (Dutch: inkeling, inkel, enkel) consists of a small wicker funnel that is inserted in the big funnel.

2.3.3.1. The fish-trap without throat or trumpet fish-trap This is a funnel-shaped wickerwork fish-trap which does not have a special throat. The trumpet fish-trap is mostly used in combina­ tion with a weir in places where the water current is very strong. It is placed in the water with its wide opening facing against the cur­ rent. The fish that enter the trap are forced by the water pressure into the narrow part (the tail) of the funnel, where they become stuck fast. Consequently the fish are unable to turn around and swim away. Sometimes there is a so-called fish-hole in the tail of the trumpet fish-trap. The fish that have been caught can be removed from the trap through this hole. When the fish-trap is in use the fish­ hole is seale9 off with a wooden plug, or a bunch of twigs, or the wickers of the outer casing are tied together at the tail end with a piece of cord. In principle most kinds of

19

fish can be caught with the trumpet fish-trap. Fish-traps of this kind are placed in swiftly flowing streams and rivers, and in places where there is temporarily a strong water current e.g. as a result of heavy rainfall, or a thaw, or tidal action. In England a kind of trumpet fish-trap is actually used for catching flatfish. Figure 19 shows a trumpet fish-trap from Finland. The total length of this trap is 1.6 m. The first hoop is made of the thin stem of a fir tree, the outer casing is made out of fir branches split lengthwise and the cord used for binding consists of fir roots. An interesting kind of trumpet fish-trap is known from the estuary of the Severn, that opens out into the Bristol Channel between England and Wales (fig. 20). This so-called salmon putcher is 1.7 m long and the largest qpening measures 53 em in diameter. The salmon putchers are put in position, with their largest opening facing downstream, in rows and fixed above and below on to wooden constructions. In the' estuary' there is a big difference in water level between low and high tide. Moreover, the water that rises with the incoming tide is very turbid. Consequently the salmon swimming upstream do not see the fish-trap, and become caught inside it. In a state of 'panic' the fish tries to escape through the narrow opening. In the attempt the fish jerks itself so forcefully that it becomes stuck fast, and cannot be released by even the strongest current of the ebb tide. In the salmon putcher only one fish can be caught at a time. This is in contrast to the normal trumpet fish-traps, which can catch several fish at a time.

2.3.3.1.1. The development of the trumpet fish­ trap. Sirelius (1906) presumes that this kind of fish-trap is derived from the bunch of twigs weighted with stones, which is placed in the opening of the V-shaped weir (2.3.1.). Evi­ dently it is not very easy to lift a bunch of twigs, containing fish, out of the water. There would thus have been an incentive to develop a light-weight and portable kind of fishing gear, with which the catch could be lifted out of the water all at once. Thus Sirelius sees the prototype of the trumpet fish-trap in a kind of fishing gear used in the Departement du Var (France). Here, in addition to the V­ shaped weir with a bunch of twigs, a V-shaped

20

D.C.BRINKHUIZEN

fig.16

fig.16

fig.16

fig.16 fig.16

fig.16

fig.16

Recent and pre- and protohistoric fishing gear

weir with two bundles of reed is used. These bundles of reeds, that are held together by strips of lime bast, are joined together at one end. When the fishing gear is in use there is a bunch of twigs between the two bundles. A further development of the trumpet fish­ trap would be the type of fishing gear that comes from the Departement d'Aude (France) (fig. 21). With this kind of gear the hindmost part, where the fish become trapped, is already closed in the form of a funnel. The foremost part is still open, however.

2.3.3.2. The fish-trap with throat In sluggishly flowing and stagnant waters the trumpet fish-trap cannot be used. Because the water pressure is too low, the fish can swim freely in and out of the trap. To prevent this the fisherman makes use of the fish-trap with throat. 2.3.3.2.1. The weaving technique used for fish­ traps with throat. The fish-traps with throat can be made in two different ways, as regards weaving technique. On the basis of the. weaving technique used we can distinguish between the so-called light fish-trap and the closely woven fish-trap (German: Korbreuse). The former owes its name to the fact that the wickerwork has an open structure. As a result of this much light can penetrate the fish-trap, in contrast to the closely woven fish-trap. With the light fish-trap the outer casing and the throat are formed mainly of wickers that run in the longitudinal direction of the trap. These wick­ ers run more or less parallel to one another. The distance between adjacent wickers is usual­ ly 0.5 tot 1.5 cm. 2 The wickers are connected

Fig. 16. Various forms of fish-surround (after Sirelius,

1906).

Fig. 17. Fish-surrounds in Finland (after Sirelius, 1906).

Fig. 18. The fish-trap (schematic):

a. fish-trap without internal funnel; b. fish-trap with

internal funnel.

Fig. 19. Trumpet fish-trap with semicircular mouth

opening from Finland (after Sirelius, 1906).

Fig. 20. The salmon putcher (after Sirelius, 1906).

Fig. 21. Trumpet fish-trap from France (after Sirelius,

1906).

Fig. 22. Eel-trap from Lunow (Germany) (after Peesch,

1966).

21

to one another at regular intervals by means of cord. This binding is done using the Zwirn­ bindung technique. The cord forms the so­ called cross-strips. The mutual distance be­ tween these cross-strips is usually 6 to 10 cm. They can run around the fish-trap in the form of circular or spiral bands. Figure 22 gives an example of a light fish-trap with two throats, that is used for catching eels. It is made out of laths of pine wood. The cross­ strips are made out of pine roots. These have been divested of their bast layer and tanned in boiling soda water with oak- and beech­ bark (Peesch, 1966). The closely woven fish-trap also has wickers running in the longitudinal direction of the fish-trap. The distance between adjacent wick­ ers is greater than in the light fish-trap, however, for the outer casing is largely formed by the cross-strips (fig. 23). These cross-strips are also made out of wickers. They are woven close to one another by means of the LeilJ­ wandbindung techniqu~ (Vogt" 1937). Sirelius (1906) finds these closely woven fish-traps only among the Magyars and related people~. :rhey are used for catching very small kinds of fish. Thus in the north of Yugoslavia small cypri­ nids are caught with a closely woven fish-trap (fig. 24). To lure the fish inside the trap, the throat opening is rubbed over with bread.

2.3.3.2.2. Weaving methods for fish-traps and the kinds of wood used. The fish-traps can be woven out of different kinds of wood. In Finland they are often made out of willow wickers. For this purpose wickers are used that are one year old, that are gathered in the autumn or at the end of the winter. Here fish­ traps are also made out of the wood of fir, pine and juniper. For cross-strips are used: roots of fir, pine and juniper, strips of bast of birch, willow and lime or sometimes flax. Closely woven fish-traps are also sometimes made out of wood shavings. In the Netherlands the fish-traps known as the prikkorf(lamprey­ basket), the aalkorf and the aalkubbe (both eel-baskets) were woven exclusively out of willow wands. These were also wickers of one year of age. Usually the basket-maker used unpeeled wickers for the hoop of the mouth opening. He then used peeled ('white') wickers for the outer casing of the aalkorf that was used mainly in salt and brackish water (a so- '

22

D.C.BRINKHUIZEN

fig. 23

fig.25

fig.25

fig.26

Recent and pre- and protohistoric fishing gear

called white basket). Unpeeled ('green') wic­ kers were used to weave the so-called dark basket (Dorleijn, 1977) So far as I have been able to ascertain all recent Dutch fish-traps are light fish-traps. The terms prikkorj and aalkorj do not mean that we are concerned with closely woven fish-traps. The basket-maker can begin weaving a light fish-trap in different ways. The basket-makers of Lunow (Germany) begin by weaving the tail end. The throat is woven separately and is later inserted in the outer casing (Peesch, 1966). Other basket-makers begin by weaving the first hoop (Van Doorn, 1971; Michelsen, 1952) or by drawing together into a circle the ends of the first hoop, if it is made out of one twig (Sirelius, 1906). The throat is woven on to this first hoop. The basket-maker then weaves the wickers of the outer casing on to the first hoop. In this way a fish-trap is produced, in which the outer casing completely encloses the throat (fig. 25a). It is also possible to weave the outer-casing wickers on to a cross­ strip of the throat. Then a fish-trap is produced in which the throat projects slightly out of the outer casing (fig. 25b). A transitional form between the fish-trap with a built-in throat and the fish-trap with the throat sticking out of the outer casing is to be seen in figure 26. The outer-casing wickers are woven on to the first hoop. At the level of the middle of the throat the outer casing becomes narrower, and behind this level it becomes wider again. To strengthen the wickerwork of the outer casing extra hoops are sometimes fitted. To give the fish-trap its funnel shape, the basket-maker weaves the cross-strips not around every outer­ casing wicker, but around every pair of these wickers. He does this at regular intervals. Thus towards the tail the outer-casing wickers come to lie increasingly closer together. The basket­ maker can also make the outer casing (and

Fig. 23. Method of weaving fish-traps, using the Lein­

wandbindung.

Fig. 24. Closely woven fish-trap from Batrovci (Yugo­

slavia). Photo: B.A.I., Groningen.

Fig. 25. Placing of the throat: a) fish-trap with built­

in throat (Finland); b) fish-trap with throat sticking out

of the outer casing (Russia) (after Sirelius, 1906).

Fig. 26. Fish-trap with built-in throat and narrowed

outer-casing (Sweden) (after Sirelius, 1906).

23

the throat) narrower by cutting away regularly an outer-casing wicker. A combination of these two methods is also known to occur. The opening through which the fisherman removes the fish from the trap is often made by the basket-maker at the end of the tail. This fish-hole is closed off when the trap is in use by means of a wooden plug, a woven lid or a bunch of straw; alternatively the outer­ casing wickers may be tied together with a tough binding, that can be undone. There are also fish-traps that have a fish-hole in the outer casing. The opening is sealed off with a woven or solid wooden flap. One of the many fish­ traps of the Syrjans has no true fish-hole. In this particular case the outer casing is not woven concentrically around the throat, and the space that is thus created between them is used as a fish-hole. When this fish-trap is in use the pole to which it is attached covers this opening. The dimensions of the fish-traps (particu­ larly the light fish-traps) are variable. In Ger­ many (Lunow) the eel-baskets are 120 and 150 cm long and the lamprey baskets 100 cm. There are also so-called Wehrriisen (fish-traps that are used in combination with a weir). These measure 5-6 m in length and have a mouth opening of 1-1.5 m (Peesch, 1966). For a number of Finnish fish-traps Sirelius (1906) mentions lengths of between lOO and 130 cm and mouth openings between lOO and 130 cm and mouth openings between 35 and 60 cm. The Dutch fish-traps also varied in size. The prikkorj was 50-lO0 cm long, and measured 20-30 cm in diameter. This fish-basket had a cylindrical outer casing, and two throats (the so-called hoedje and petje - the little hat and the little cap). The basket was used for catching river lampreys and smelt. So far I have not been able to find any measurements for the eel-baskets. Van Doorn (1971) mentions only the use of the eel-basket with two throats and two wings that were woven on to the first hoop (fig. 27), the welie (a large eel-basket), the vleugelkorj (literally: winged basket, an eel­ basket with one weir element), the waas (an eel-basket especially suitable for use in strong currents), a Friese korj (literally: Frisian bas­ ket, a short basket with two wings) and the kachelkorj (literally: stove basket, an eel-bas­ ket without wings, in the shape of a pot-bellied stove).

24

D.C.BRINKHUIZEN

fig.25

fig.25

Recent and pre- and protohistoric fishing gear

The Dutch aalkubbe (a special kind of eel­ basket or eel-pot) is a bell-shaped trap mea­ suring 50-60 cm in length. At the rear opening of the throat a silk net (called an inkeltje) is attached, that is fastened to the tail with a piece of cord (called an enkelsnoertje). The fish-hole is situated at the end of the tail and is closed off by means of a wooden plug. The space between the peeled outer-casing wickers measures 0.5 cm and the space between the cross-strips 4.5 cm. These cross-strips are wo­ ven in a spiral around the outer casing. The diameter of the rear opening of the throat measures 5 cm (fig. 28). 2.3.3.2.3. Fishing with a trap and the kinds of fish caught in it. To fish effectively with fish­ traps (and other fishing gear), the fisherman must have a broad knowledge of the habits of the different kinds of fish. This means that he must know his fishing waters (with regard to depth, current, substratum, etc.). In addi­ tion insight into weather conditions and fa­ miliarity with his equipment are important. On the basis of this knowledge he sets out his fishing gear. 3 The larger fish-traps are placed in combi­ nation with a weir. These can be very large permanent weirs. The fish-trap is then lowered into the weir and lifted out of it with the aid of a vertical gate construction (fig. 29). In the Netherlands this kind of fishing is known as weervisserij (literally: weir fishing). It involves the use of a V-shaped weir, almost 1 km long, with a fish-trap made of yarn to catch herring and anchovy and the zalmsteek (a row of fish-traps) to catch salmon. In addition fish-traps are used in combi­ nation with loose (transportable) weir ele­ ments. The fish-trap is then often anchored to a post, that has a central position in front of the mouth opening. Also the tail can be attached to another post with a piece of cord. Sometimes the tail is weighted with a stone. Fish-traps may also be put down in the water completely weighted with stones.

Fig. 27. Eel-basket with two throats and two wings, from the Netherlands (after De Groot & Schaap, 1973). Fig. 28. Aalkubbe (wooden plug missing) from the Netherlands. Photo: C.F.D., Groningen.

25

In the Netherlands, in the big rivers a number oflamprey-baskets or eel-baskets were placed next to one another in the water on a steel wire or thick rope (a so-called weel or kubbenUjn) (Van Doorn, 1971). In the former Zuiderzee many eels were caught with the aalkubbe. The position of this kubbe in the water is quite different to that of other fish-traps, i. e. it is vertical instead of horizon­ tal. The baited kubbe is weighted with stones and is hung on a cord from a stick stuck at a slight angle into the bottom. This is done in such a way that the mouth hoop just touches the bottom. The position of the kubbe on the ground can be altered by placing the stick at a steeper or less steep angle accordingly. Du­ ring cold weather the eel moves down deeper and then the hoop has to rest on the bottom. During warm weather the kubbe may be lifted up a little higher (Dorleijn, 1977). The wicker fish-traps are not very durable. According to Sirelius (1906) the Finnish fish­ traps are worn out after a few years. According to Dorleijn (1977) the aalkubbe that was used in the Zuiderzee lasted for two years. In principle almost any kind of fish can be caught with a fish-trap. For technical reasons this is not done, however, and for many kinds of fish no special fish-traps are made. Such fish are caught with nets or lines. So far I have not found any special fish-traps in West­ ern Europe for catching the larger kinds of fish such as sturgeon and catfish. This can be simply explained: large (predatory) fish are rare compared to other, smaller kinds of fish. A special fish-trap for large fish would provide too low a yield. Moreover smaller kinds of fish can easily swim in and out of such a fish­ trap. We thus find fish-traps for catching those kinds of fish that are present in large numbers in the waters concerned. These are nearly always kinds of fish that are not very big compared to sturgeon and catfish. When the fisherman uses a fish-trap in combination with a weir or a fish-trap with wings, he does not need to use any bait. The aalkubbe and the prikkorf are baited, however. In the Nether­ lands the freshwater fisherman used the fish­ trap to catch the following kinds of fish: eel, perch, members of the carp family, pike, ruffe, smelt and river lampreys. When a migrant species of fish was temporarily present in freshwater (e.g. salmon, sea trout, houting,

26

D.C.BRINKHUIZEN

fig.29

fig.25

fig. 3D

fig.33

fig.32

fig.34

Recent and pre- and protohistoric fishing gear

twaite shed and allis shed), these were fished too. In brackish/salt water herring and an­ chovy were caught. The herring fishery with fish-traps on the Frisian Wad (the tidal shal­ lows) yielded a supplementary catch including flounder, gurnard, sole, codling and ray (Van der Molen, 1976).

2.3.3.2.4. The development of the fish-trap with throat. On the one hand Sirelius (1906) regards the fish-trap with throat as a further devel­ opment of the fish-surround, on the other hand he regards it as a further development of the trumpet fish-trap. It is evident that tish-surrounds placed in sandy and muddy bottoms are not completely escape-proof. Fish can dig them selves into the substratum and in this way escape under the barrier. Because there is no roof present, birds of prey visit the fish-surround. Also strong sunlight can have a damaging effect on the fish. In the event of sudden flooding the fish-surround may disappear under the water thus enabling the fish to escape. More­ over, there are obvious difficulties involved in emptying a fish-surround of its catch, on account of its necessary situation in deep water. A transitional form between the fish­ surround and the fish-trap with throat is seen by Sirelius (1906) in a fish-surround from Southeast Asia (fig. 30). This fishing gear has a base, and the top ends of the fencing are tied together. The contraption is placed ver­ tically in the water, however. Figure 31 shows a fish-trap with throat that is used by the Syrjans. This trap can be regarded as a further developed form of trumpet fish-trap, since it has an especially long and narrow tail. 4 The trap is made out of willow wickers. The cross­ strips are roots of fir. The mouth-opening is square. In addition to the above-mentioned fish-

Fig. 29. Weir with two fish-trap openings (Syrjans) (after

Sirelius, 1906).

Fig. 30. Fish-surround from Siam (after Sirelius, 1906).

Fig. 31. Fish-trap with a long narrow tail (Syrjans) (after

Sirelius, 1906).

Fig. 32. The square net.

Fig. 33. The cast-net (after Zivkovic, 1956).

Fig. 34. The drift-net.

27

trap the Syrjans also use a trap that consists of a loose throat and a loose outer casing. This throat is inserted in the outer casing. This kind of fishing gear can also be regarded as a transitional form between the trumpet fish­ trap and the fish-trap with fixed throat. The most developed fish-traps are those with an outer casing and throat made entirely out of cording (2.3.4.4.).

2.3.4. The fishing net The final stage in the development of the fishing gear is the fishing net. This consists entirely of cords that are knotted together in such a way that a meshwork is produced with diamond-shaped openings. The cording of recent fishing nets is made out of hemp, cotton, nylon or other synthetic fibres. Sirelius (1906) mentions that in Finland also willow bast was used formerly. Fishing with nets is in almost all respects more advantageous than fishing with lines, weirs and wickerwork fish-traps. The fishing net is less conspicuous, so the fish are caught more easily in it. Once fish are trapped in a net they become confused in the meshwork and are hardly able to escape. Moreover fishing nets are easy to handle and take up little space when they are stored away. On the basis of their shape fishing nets can be divided into four groups, namely the square net, the round net, the rectangular net and the bag-shaped net.

2.3.4.1. The square net This net is suspended by its corners from two flexible sticks placed crosswise. At the point where the sticks cross a cord is fastened, and the other end of this cord is attached to the end of a long stick. This kind of fishing gear is called a square net (Dutch: kruisnet or toteheT) (fig. 32). The net has a closely spaced meshwork, and is let down horizontally into the water, until it comes to lie on the bottom. The fisherman may place a bait on the net, but this is not essential. When the water is turbid, the fisherman raises the net haphazard. If the water is clear, then he actively goes in search of fish. Then, whenever a fish swims above the net, the net is lifted up. The di­ mensions of square nets vary. Those that are operated by hand measure c.4 m 2 • If the

28

D.C.BRINKHUIZEN

fisherman is fishing from a boat, he uses a bigger net (e.g. 25 m2). Such a net would be mechanically operated (Van Doorn, 1971). 2.3.4.2. The round net This circular fishing net has a diameter of about 4 m. The periphery of the network is strengthened with a thick rope, to which lead weights are attached. At the centre-point of the net there is a small opening, in which a cow's horn or a ring made of horn is fixed to protect the network. Passing through this ring are a number of ropes (Dutch: pezen), that are attached to the periphery of the net at regular intervals (fig. 33). This kind of fishing gear is known as a cast-net. The fish­ erman stands in a boat or wades into the water and throws this net as far outspread as possible on to the surface of the water. On account of the lead weights the net sinks rapidly to the bottom and covers the fish there present. By pulling up the ropes that run through the central hole the fisherman is able to trap the fish inside the net, which is then pulled out of the water. Van Doorn (1971) states that the cast-net is derived from the stulpmand. This is a basket without a base. With this device the fisherman stands motionless in shallow water and waits until a fish or a shoal of fish comes along. As soon as the prey comes within reach, he quickly thrusts the basket down over the fish and can easily remove the fish from the basket. Barthel (1977) mentions the find of a stulp­ mand dating from Roman times in Oberdorla. 2.3.4.3. The rectangular net This fishing net is a rectangular network, of which the two long sides are attached to strong ropes (the top and bottom ropes). The dimen­ sions of this kind of fishing gear are very variable, and depend on the waters that are to be fished with it. Smaller specimens measure e.g. 10 m in length and 1 m in height; others may be as much as several hundreds of metres long and ten or more metres high. The rec­ tangular net is always placed vertically in the water. This can be done in two ways: either the net is suspended on vertical posts that are driven into the lake, river or sea bottom (a standing net) or the net is used as a so-called drift-net. In the latter case, to the upper rope are attached floats made of wood, cork or

synthetic material and to the lower rope weights made of stone, lead or some other heavy material. s It is clearly obvious that rectangular nets are derived from the weir. Figure 34 shows the simplest form of drift­ net, that is still widely used. From this basic kind of drift-net other kinds have been devel­ oped that are more effective. I will not go into this matter in any further detail. An important point is that the use of a boat is nearly always necessary for fishing with rectangular nets. 2.3.4.4. The bag-shaped net This kind of fishing gear consists of a network funnel that is held open by means of a number of wooden or metal hoops. In this funnel a number of constrictions are present. To lead the fish towards the trap one or two network wings are attached to the first hoop (fig. 35). These network fish-traps are attached to posts and placed in shallow water in lakes, rivers and along the coast. The wickerwork fish-trap stood as a model for the development of the bag-shaped net. This in turn gave rise to the development of the drag-net or trawl. Seeing that this kind of fishery involves such a high degree of development of the fishing boat and fishing techniques, however, I will not go into any further detail on this matter. 2.4. Auxiliary fishing equipment

In the actual process of catching fish only a few items of auxiliary equipment are used. These are: the splashing stick (Dutch: plons­ stok, German: Trampe), the gaff (German: Schlaghaken) and the scoop-net or dip-net (Dutch: haam or schepnet). At the same time one could regard the boat and the live-box as auxiliary fishing equipment. In view of the fact that terms like scoop-net and boat need no further description, I shall not devote any attention to them here. 2.4.1. The splashing stick This is a long wooden stick with a small flat wooden disc at one end. There are also splash­ ing sticks that have a hollow wooden disc, a hollowed-out gnarled piece of wood or a

Recent and pre- and protohistoric fishing gear

piece of leather (Sirelius, 1906). To alarm the fish and to drive them towards the fishing gear noise is made in the water. The fisherman can do this by merely hitting the water surface with an ordinary stick. It is evident that the splashing stick results in a more rapid and effective flight of the fish, however, as it produces a noise of greater resonance. A species of fish that does not flee from the noise, but rather is attracted by it, is the European catfish. The predatory fish is only active at night. In order to make the fish active in the daytime, the fisherman makes use of the splashing stick. In Yugoslavia a wooden or iron stick c. 40 cm long is used for this purpose. At the end of this a hollowed-out disc, c. 5 cm in diameter, is attached (fig. 36). With this splashing stick the water is struck obliquely from above. This is done very re­ gularly (c. 20 strokes per minute), and in this way a resonant singing noise is produced, that carries a long way. According to Ristic (1977) this sound imitates the noise that is made when a large catfish snaps up its prey. Impelled by their insatiable appetite the catfish swim to­ wards the source of the noise. Here they are caught with a strong line and large baited hook, which the fisherman moves up and down in the water. As it is difficult to haul in large catfish from the riverbank or shore, this kind of fishing is only done from a boat.

2.4.2. The gaff The gaff consists of a long wooden stick with a large iron hook at one end. The hook is sometimes barbed. It is attached to the stick by means of thread (fig. 37). With the aid of the gaff the fisherman removes from the water the larger fish that have swum into the pas­ sively-operating fishing gear. I have also men­ tioned the gaff as an actively operated kind of fishing gear under 2.2.1. Among the people living around Lake Ladoga a similar gaff is used for catching seals (Sirelius, 1934). In the Netherlands, those who fish in rivers are familiar with specific types of gaff, namely the steurhaak and zalmhaak (literally: sturgeon hook and salmon hook). The sturgeon hook is a large crescent-shaped hook that is thrust into a sturgeon when it threatens to escape from the net. The salmon hook is used in

29

catching salmon and is usually slightly smaller than the sturgeon hook (Van Doorn, 1971). Van der Molen (1976) mentions the use of a gaff in the hoekwant fishery. The fishery mu­ seum in Moddergat (Friesland) possesses a finely decorated iron hook that was used in bringing in large cod.

2.4.3. The live-box This is a square or rectangular wooden con­ tainer with a lid. In the bottom and in the lower part of the sides small holes are present, so that water can flow in and out of the container. The live-box provides a means of keeping alive fish that have been caught, and also of transporting them alive. In the Netherlands large wickerwork trailing baskets were used for transporting river lam­ preys from Arnhem to Vlaardingen. These baskets were secured alongside the fishing boat (Lobregt & Van Os, 1977). A fixed live-box in a fishing boat is known in Dutch as a beun or bun. In Yugoslavia wooden live-boxes are used that are in the shape of small boats c. 1 m long. These are pulled along by means of a rope attached to the fishing boat. During excavations of Early Medieval Do­ restad (Wijk bij Duurstede), a rectangular oaken frame was found in the bed of the Rhine, dating from Carolingian times. This frame had a bottom made of a wickerwork of willow wands, while in the frame there were holes that had been bored into it, in which remains of the side-walls were still present. The object has been interpreted as the remains of a live­ box (Van Es, 1974; Casparie & Swarts, 1978). 3. ARCHAEOLOGICAL FINDS OF FISHING GEAR In the article 'The development of fishing in prehistoric Europe' (Clark, 1948), detailed information is given about the ways in which prehistoric man in Europe exploited the na­ tural fish populations as a source of food, and about the kinds of fish that were certainly or probably caught. After this survey had been published, in the field of archaeology hardly any more attention was devoted to this subject. Recent research has shown, however, that

30

D. C.BRINKHUIZEN

fig.25

fig.25

fig. 37

fig.38

Recent and pre- and protohistoric fishing gear

some assumptions made in that article were premature, and therefore stand in need of readjustment. 3.1. The method of approach In looking for references to pre- and proto­ historic fishing gear in the archaeological li­ terature I have had to restrict myself in many ways. For example, I have not attempted to make an inventory of the non-independently operating types of fishing gear. That would be a very time-consuming task, as there are many scattered references to such finds in the literature. Rather, I have limited myself to a few brief remarks on fish-spear prongs and gorges, and also to the description of six bone fish-hooks from the Netherlands, in conside­ ration of their effectiveness from a theoretical viewpoint. With regard to the independently operating' fishing gear I have not made any special effort to look for references to archaeological finds of weirs, fish-surrounds and fishing nets. Finds of weirs and fishing nets are well known. On the other hand, so far I have not come' across any finds of fish-surrounds. Emphasis is laid, however, on the inventorization of Mesolithic and Neolithic wickerwork fish-traps and frag­ ments thereof from the Netherlands, Northern Germany, Denmark and Sweden. In addition, with regard to these finds attention is devoted to the plaiting method used and the degree of regularity in the wickerwork, as far as can be ascertained. Also in a number of cases a reconstruction is attempted of the fish-trap or fragment in order to gain some idea of the possible type. Figure 38 indicates the sites of the various

Fig. 35. Network fish-trap with two wings.

Fig. 36. Splashing stick for activating catfish (Yugoslavia)

(after Zivkovic, 1956).

Fig. 37. The gaff (after Sirelius, 1934).

Fig. 38. The Netherlands at the present time, showing

the findspots of the various types of pre- and proto­

historic fishing gear: 1. Rotterdam-Europoort; 2. Vlaar­

dingen; 3. Rotterdam-Bergschenhoek; 4. Molenaars­

graaf; 5. Olst; 6. Spoolde; 7. Noordoostpolder; 8. terpen

area; 9. Emmerer.fscheidenveen; 10. Utrecht; 11. Dore­

stad (Wijk bij Duurstede).

31

types of pre- and protohistoric fishing gear from the Netherlands, as mentioned in the text. 3.2. The fish-spear In prehistoric assemblages since the Upper Palaeolithic, objects evidently occur that strong­ ly resemble the loose prongs of recent fish­ spears. These are long (c. 15 cm) narrow points that have a number of small barbs on one long edge. They are made out of bone or antler. and are called leister prongs or barbed points. These barbed points are found especially in those countries around the southern part of the North Sea, the North Sea itself and coun­ tries around the Baltic Sea, as isolated finds in old lake-bottom sediments. Barbed points have also been found in excavations of Me­ solithic (Maglemose) settlements in Denmark. In .these settlements fish remains were also found. Among these remains those of pike evidently predominate, or at least occur re­ gularly (Degerb.0l, 1945). Also the find of a pike skeleton: in the clos'e vicinity of a barbed point (lndreko, 1948) makes it probable that the barbed points were for catching fish. On the basis of this evidence Clark (1948) stated: 'Evidently, from the rarity of remains of other species, we have to do with a spe­ cialized pike fishery. In summer this was mainly carried on by spears as the fish lay quiet in still, shallow water'. That this state­ ment is too one-sided and may lead to wrong conclusions is clear if we look at the Late Preboreal settlement of Star Carr. Partly on the basis of the absence of pike remains (no fish remains were found at all) Clark (1954) concludes that the settlement cannot have been inhabited in the summer. However, besides the fact that pike can be caught just as well in winter, Wheeler (1978a) comes to the conclu­ sion, on the basis of biological and palaeogeo­ graphical factors, that pike could not have been present at the time in the lake next to which Star Carr was situated. The data that have been obtained from the material of the Danish Maglemose settlements are not appli­ cable a priori to contemporary settlements elsewhere. Many barbed points were indeed present at Star Carr. Despite the absence of fish remains, in my opinion these barbed points could certainly have been used for

32

D.C. BRINKHUIZEN

fig.39

fig. 40

f ig. 41

Recent and pre- and protohistoric fishing gear

catching fish, namely fish of the salmon family. On the basis of the biology and the distribution of salmonids it is well possible that there existed a standing popUlation of arctic charr or a migrant population of salmon and sea­ trout in the Late Preboreal lake. However, for physiological and chemical reasons the chan­ ces of remains of salmonids becoming fossi­ lized are very slight (Lepiksaar, 1975; Wheeler, 1978b). Also cultural factors can be respon­ sible for decreasing the likelihood of preser­ vation of salmonid remains. For example, it is known that certain North American Indian tribes prepare the salmon they catch by split­ ting them lengthwise, drying them and then pounding them into fish-meal (Casteel, 1976). It will thus be clear that salmonid fish pro­ cessed in this way will leave no or hardly any remains. In Neolithic and later periods it is evident that the finds of barbed points decrease. This may be, on the one hand, the result of the development of better fishing gear and, on the other hand, due to the transition from hunting, fishing and food-collecting to agriculture and stockbreeding. Fish-spears remain in use, how­ ever. The Greeks and Romans used the fish­ spear, as is evident from frequently occurring illustrations in pictorial art. This does not mean that for the Greeks and Romans this was the most important kind of fishing gear. The fish-spear is rather the kind of tackle used by those who fish for pleasure, and as such emphasizes the element of sport. 3.2.1. Barbed points from the Netherlands Also from the Netherlands a number of barbed po,ints are known. One of these comes from Emmererfscheidenveen (fig. 39). On the basis of the data of Clark (1936), the object is dated by Louwe Kooijmans (1970/1971) in the Pre-

Fig. 39. Barbed point from Emmererfscheidenveen.

Photo: R.M.O., Leiden.

Fig. 40. Double-pointed bone rods from the Dutch terpen

area. Photo: C.F.D., Groningen.

Fig. 41. The relationship between length and width of

bone fish-hooksJrom Southern Sweden (after Lekholm,

1951) and the Netherlands.

33

boreal or Early Boreal. About 10 barbed points have been found as a result of sand­ dredging operations at Europoort, Rotterdam. They probably come from a deposit that is dated in the very beginning of the Boreal (Louwe Kooijmans, 1970/1971, 1976). Some of these barbed points are very small (c. 5 cm) in comparison with barbed points from other sites. Despite their small size, that these objects functioned as fish-spear prongs is not impossible in my opinion. 3.3. The gorge Bone objects interpreted as gorges are known from as early a context as Upper Palaeolithic sites in France (Sirelius, 1934; Clark, 1948). Reinerth (1926) illustrates two objects from a Swiss Neolithic lake-shore settlement which he interprets as gorges. The larger specimen measures 11.2 cm in length and 1.2 cm in diameter; the smaller one is 3.9 cm long and 0.4 cm in diameter. Barthel (1977) mentions a number of bone gorges dating from Neoli­ thic, Roman and Early Medieval times from Thiiringen. The author also mentions that the constriction in the middle of the gorge is not necessary. This became evident to him as a result of experiments with gorges. 3.3.1. Gorges in the Netherlands In the Netherlands many bone artifacts resem­ bling gorges have been found in the terpen of Friesland and Groningen (fig. 40). Most of them have not been dated, however. Similar objects are also known from the Early Medi­ eval town of Dorestad. Some Dutch authors think that these objects were used in spinning, while others leave the question open as to what function they may have had. Roes (1963) describes a number of these objects yet she does not attribute any function to them, but says 'not all the rods need have served the same purpose for there are many different types among them'. In my opinion it is possible that a number of the double-pointed short bone rods that have been found in the Netherlands were used as gorges. In the middle of some of these rods a number of notches or a small hole can be

34

D.C.BRINKHUIZEN

seen. These notches could serve as a distin­ guishing mark for the fisherman, while at the same time they would prevent the line from slipping off the rod. 3.4. The fish-hook The oldest fish-hooks that we find in North­ western Europe date from the Boreal-Meso­ lithic. These fish-hooks measure up to several centimetres in length and have no barb. They are made out ofa single piece of bone or antler. Bone fish-hooks with a barb and combination hooks with separate hook-shaft and hook­ point appear in Northwestern Europe at the end of the Neolithic. Soon afterwards metal fish-hooks make their appearance (approxi­ mately in the Late Bronze Age) (Clark, 1948). Nevertheless bone fish-hooks both with and without a barb continue to remain in use (see e.g. So~berg, 1909). Until recently fisherfolk who had little or no contact with technoJo­ gica!ly advanced, cultures still used bone and wooden fish-hooks. Consequently it is obvious that an isolated find of a bone or metal fish­ hook is difficult to date. Lekholm (1951) studied 40 bone fish-hooks from Southern Sweden. Of these 29 had been found in freshwater deposits and 11 in marine deposits. None of the hooks were dated. From an examination of the complete specimens it was evident that the hooks from the freshwater deposits were generally smaller than those from the marine deposits (fig. 41). Also the hook opening was evidently bigger with in­ creasing length of the hook-shaft. In addition the grip angles were measured. These varied from -25° to 65°, though most of the hooks had a grip angle of between 30° and 40°. This range corresponds fairly well with the grip angles of modern steel hooks.

3.4.1. Bone fish-hooks from the Netherlands; description Louwe Kooijmans (1974) mentions six bone fish-hooks from the Netherlands. Three of these were found together with a tool made of red-deer antler and a few flint artefacts in grave II at Molenaarsgraaf in the Ablasser­ waard. The grave dates from the Bell Beaker

period. These three hooks I shall refer to from here on as hook a, hook b and hook c. In addition to these, three isolated finds of fish­ hooks are known, to which I shall refer as hooks d, e and f. Hook a is a bone hook (maximum length 4.3 cm) without a barb (fig. 42a). At the top of the hook-shaft there is a small thickening or knobble. Below this knobble a groove appears to be present, that runs around the hook-shaft. In the hook-bend one can still see part of the perforation that determined the shape of the hook-bend. Hook b is a bone hook (3.1 cm long) without a barb (fig. 42b). there IS no knobble present to prevent the line from sliding off. However, the diameter of the hook-shaft does increase towards the top. With this hook, too, part of the perforation is visible. Hook c is a bone hook (2.6 cm long) without a barb (fig. 42c). No provision has been made to prevent the line from slipping off. With this hook, too, part ofthe perforation is still visible in the hook-bend. Hook d measures 7.6 cm in length and is made out of antler (fig. 42d).6 It was found in the River IJssel at Olst. This is a hook with a barb. Around the hook-shaft there runs only a single groove, and not several grooves as mentioned in the description given by Louwe Kooijmans (1974). Hook e measures 9.2 cm in length and is made out of antler (fig. 42e).1 It was found at some time around 1950 in the Zwolse Vaart in the Noordoostpolder (Van der Heide, 1972). The hook has a barb. Just below the top of the hook-shaft there are three notches present on both the inner and outer side. These notches have been cut in such a way that they alternate with respect to one another. This implies that the line was attached to the hook-shaft spirally. The cross-sections of the hook-shaft, the hook­ bend and the hook-point are more or less round. Hook f measures 11.7 cm in length and is made out of antler (fig. 421).8 It was found in 1948 on the Frisian side of the Noordoost­ polder (Boeles, 1951). The hook has a barb and the top of the hook-shaft consists of a large rectangular knobble. In cross-section the hook-shaft is rectangular and the hook-point round. The inner side of the hook-bend is more or less flat, while the outer side is rounded.

35

Recent and pre- and protohistoric fishing gear

About 1.5 cm away from the knobble of the hook, on one side of the hook-shaft a shallow transverse groove is present, measuring about 2 cm in breadth.

3.4.2. Measurements The various measurements and grip angles of the hooks a-f are given in table 2. They are taken according to figure 43. 9 Two different grip angles were determined. Grip angle A is the angle between the ima­ ginary line through the axis of the hook-point and the imaginary line from the hook-point to where in my'opinion the line was attached, namely below the knobble or around the groove situated nearest the top. I have also assumed that the line was attached to the outer side of the hook-shaft, facing away from the hook-point. Grip angle B is the angle between the imagmary line through the axis of the hook­ point and the imaginary line from the hook­ point to that part of the top of the hook-shaft situated closest to the hook-point (Lekkolm, 1951).

3.4.3. Conclusions On the basis of the data presented in table 2, that have been included to some extent in figure 41, we can draw a few conclusions. On the basis of their measurements the hooks a, band c fit well in the group of freshwater fish-hooks from Sweden. This is in agreement with their place of origin, which was situated in a former freshwater tidal delta. Hook d could have been used for fishing in either freshwater or the sea. The findspot suggests rather that it was used in freshwater. On the basis of their measurements it seems that hooks e and f were intended for use in the sea. On ethological and anatomical grounds, e.g. the way in which fish of a particular species bite, the position of the mouth of fish of a particular species, etc., the data of table 2 can also be compared' with those of table 1. It would then appear that hooks a-d could only be used for catching pike, salmon and Euro­ pean catfish. This last-mentioned species,

Table 2, The measurements and grip angles of the fish-hooks a-f

Hook a Hookb Hook c Hookd Hooke Hookf

Angle of grip

C)

(mm)

Hook opening (mm)

A

B

17 15 14 25 43 57

13 10 11 16 24 28

57 48 63 55 46 42

45 33 62 32 36 28

Length

Width

(mm) 43 31 26 76 92 117

which does not appear in table 1, is fished for with a hook that is of the same size as or larger than those used for catching pike and salmon. Remains of large European cat­ ,fish have been shown to occur regularly in pre- and protohistoric settlements in the catch­ ment area of the ,Maas, Rhine and IJssel (Brinkhuizen, 1979a). The hooks e and f could have been used for catching cod and other large species of marine fish. On the basis of the size of its grip angles it would seem impossible to catch a fish with hook c: the fish would very readily get hooked, but could just as easily escape. Moreover no provision has been made to prevent the line from slipping off the hook-shaft. This could mean that the hook was specially made to serve as a grave gift for the dead, and that it was never used for catching fish. It is difficult to assign a date to the undated hooks d, e and f. Metz (1975) dates hook d in the Mesolithic. On the basis of the presence of a barb we should date the hook, according to Clark (1948), in the Late Neolithic or thereafter. Hook e shows traces of working on the inner side of the hook-bend and the hook-shaft. Clason (pers. comm.) deduces from the nature of these traces that the hook was made with the aid of flint tools. On the basis of this a dating in the Late Neolithic or Early Bronze Age would seem justifiable. Traces of inha­ bitation dating from this period have been found in the region concerned. Although the exact findspot of the hook is not known, we know that the local aquatic environment was a freshwater one formerly. In that case the hook must have been used for catching fresh­ water fish, quite possibly European catfish.

D. C. BRINKHUIZEN

36

fig. 42

fig.43

fig. 44 fig.43

fig.43

fig.43

fig. 47

Recent and pre- and protohistoric fishing gear

Van der Heide (1955) dates hook f, on the basis of isolated finds in the vicinity, in the Late Neolithic or in the Early Bronze Age. Although this dating cannot be rejected, it is my impression that the hook dates from a younger period. This is suggested by its shape and the way in which the hook-bend and hook­ shaft are neatly finished off. The size of the hook indicates that it was used for sea fishing. On the basis of the assumption that in the Netherlands the first seaworthy fishing boats were built only in Roman times, the hook can perhaps better be dated between the beginning of the Christian Era and the Early Middle Ages. Seeing that the environment in the vicinity of the findspot was a freshwater or brackish one, the hook must have been lost from a fishing boat on its way to or from the sea. Because of its size, it is also possible that the hook was attached to a stick and used as a gaff. 3.5. The weir If the water conditions were suitable,. prehis­ toric fisherfolk would certainly have made use of weirs. However, archaeological finds of weirs or fragments thereof are rare. There are various reasons to account for this (3.6). 3.5.1. Archaeological finds offish-weirs outside the Netherlands Clark (1948) mentions the find from KyrksHitt (Finland) of a row of vertical posts, about 13.5 m long, with horizontally lying branches. This

Fig. 42. Fish-hooks made of bone or antler, from the

Netherlands: a, band c. Molenaarsgraaf (after Louwe

Kooijmans, 1974); d. Olst (photo: R.M.O., Leiden); e,

f. Noordoostpolder (photo: C.F.D., Groningen).

Fig. 43. Measurements of the fish-hook. a length; b width;

c hook opening; A, B grip angles.

Fig. 44. Reconstruction of the enclosure for catching

sturgeon at Vlaardingen (after Boddeke, 1971).

Fig. 45. Layer 2 of the fish-trap fragment from Nid10se

(after Becker, 1941).

Fig. 46. Fish-trap fragment from Magleby Long (after

Becker, 1941).

Fig. 47. Median cross-section of the reconstructed fish­

trap fragment from Fjellenstrup.

37

framework must almost certainly have been part of a weir. In H0yland (Western Norway) a similar weir was found in the channel of a silted-up stream, that once conn;!cted two small lakes. The sediment in which the find was present was dated by means of pollen analysis to the period of transition from Stone Age to Bronze Age in Norway. Becker (1941) describes a fragment of a wattle screen from Svinninge Velje (Holbaek district, Sjaelland, Denmark). The find was made in a marine deposit, dating from the Atlantic (zone VIIb according to a palynological investigation by Troels-Smith). The fragment, which measured 65 x 50 cm, was made out of parallel wickers connected to one another by means of twisted strips according to the Zwirnhindung techni­ que. The wickers were lime twigs and the strips were of lime bast. In the immediate vicinity of the fragment some 20 or so pointed stakes were found. These stakes, which measured 3­ 5 cm in diameter, had been driven into the bottom in a row. Becker interprets the whole construction as a weir, that led the fish towardS a fish-trap. In my opinion this is not necessarily so. The find-spot once formed part of -the Lammefjord and with the tidal movement of water in this locality the weir could have functioned as an independently operating kind of fishing gear (a so-called gaard: 2.3.1.1.). Becker (1941) also describes the find of a wickerwork fish-trap from a peat-bog at Mag­ leby Long (Som district, · Sjaelland, Den­ mark). At a distance of 2-3 m from the fish­ trap a fragment of a screen was found. This did not have any cross-strips, but the wickers were most probably clamped between obli­ quely standing sticks. It is clear that we are concerned here with a fish-trap with a weir. 3.5.2. Fish-weirs in the Netherlands Up to the present time no finds have been made of clearly recognizable weirs. During the excavation of the Neolithic settlement in Vlaar­ dingen, in the tidal creek that ran alongside the settlement many wooden posts were found. Boddeke (1971) interprets these as the remains of a weir for catching sturgeon. When the water was rising the sturgeon would first pass a V-shaped weir with a throat opening and then come up against a solid barrier of posts.

38

D.C. BRINKHUIZEN

Before the tide turned the throat opening was sealed off, so that the sturgeon could no longer escape down to the sea. At low water the sturgeon could then be simply removed from the enclosure (fig. 44). In this method we see the principle of the liiltiim (2.3.1.). During the excavation remains were also found of a tiet with a mesh-width of 24 cm. Among these remains a dermal scute of stur­ geon was present. Van Iterson Scholten (1977) is of the opinion that this net was used to seal off the throat opening of the weir. The net from Vlaardingen could also be a net for transporting heavy objects, like a sturgeon. To my mind the use of the liiltiim in Vlaardingen is a reasonable supposition. Detailed exami­ nation of the excavation ground-plans, that so far have only partly been published, may provide more information on this matter. In a photograph published by Van Iterson Schol­ ten (1977) a row of posts is visible that is strongly reminiscent of a weir. In Spoolde, during the investigation of a site where objects made of antler had been found, a gully was cut into which probably dates from the Bell Beaker period. At the . bottom of the gully filling a number of roughly hewn wooden posts were present (Van der Heide, 1962). It is possible that these posts were used for the attachment of fish-traps or nets; alternatively they may have formed part of a weir. At Molenaarsgraaf in the Ablasserwaard, the excavation was carried out of a settlement dating from the period of transition from the Neolithic to the Bronze Age (Early Bell Beaker/Barbed Wire Beaker). In the stream channel that ran alongside the settlement, a number of pointed wooden posts were found. Some of these were still standing vertically, others were leaning at an angle as a result of the pressure of water flowing through the channel. The excavation plan of the stream channel does not show any systematic distri­ bution of the posts. Louwe Kooijmans (1974) presumes that nets or fish-traps were attached to the posts. Alternatively these posts could be regarded as the remains of a weir. In the Netherlands weirs were used for fishing in the Middle Ages. An 11th century copy of the list of property belonging to the church ofSt. Martin in Utrecht mentions seven fish weirs in Muiden (Blok, 1974). Whether

we are concerned here with the use of a weir as an independently operating kind of fishing gear either by itself or in combination with a fish-trap is not clear. The latter alternative seems to me more likely. 3.6. The wickerwork fish-trap The find of a wickerwork fish-trap or a frag­ ment of such a trap can be regarded as a very rare find indeed. This becomes clear if we look at the conditions under which a wickerwork fish-trap has to function. Fish-traps are set in shallow water in places where fish are abundant. In such places the water is eutrophic and thus rich in oxygen. Fish-traps are subject to rapid oxidation. Moreover it is evident that e.g. the wickerwork of an eel-basket disinte­ grates as a result of the exertions of the eels trapped inside in their attempt to escape. After a couple of years the trap is worn out and is then thrown away or burnt. In view of the local conditions (a eutrophic environment), the chances of the trap ending up in an environment with good preservation condi­ tions are extremely small. This occurs only if the trap becomes rapidly covered by a sediment or if the oxygen content of the water suddenly decreases drastically. Also it is evi­ dent that a fish-trap or fragment thereof that has ended up in an environment with good preservation conditions becomes as soft as butter in the course of time. Before the trap is recognized as such, it has therefore often been partly dug away. Nevertheless a number of fish-traps or fragments of traps have been found and recovered. Also in a number of cases fish remains have been found in the immediate vicinity, that may originate from fish that were caught with the aid of these traps. 3.6.1. Archaeological finds of wickerwork fish­ traps from Denmark

Fish-trap fragment, found in 1905 in the Amosen at Nidl0se, Holbaek district, Sjael­ land. This find is described by Becker (1941). The fragment dates from the Early Atlantic (zone VIla according to the zonation of Knut Jessen). It measures 52 by 38 cm and consists of two layers of wickerwork. The wickers

Recent and pre- and protohistoric fishing gear

measure 0.5-0.7 cm in thickness and are made of birch twigs. The space between the wickers measures at most 0.9 cm. The cross-strips are made of split pine roots. The distance between the cross-strips, that hold the wickers together by means of the Zwirnbindung, measures 4­ 8.5 cm. In the wickerwork of one layer the wickers run almost parallel. In the other layer, how­ ever, the wickers converge towards one another and a few wickers have been cut away with some degree of regularity (fig. 45: see arrow). On the basis of this difference Becker concludes that layer 2 is part of the throat. Layer 1 is then part of the outer casing at the level of the throat. The find is thus part of a light fish-trap with throat. Fish-trap fragment that was found in 1940 in the Magleby Long, Som district, Sjaelland. This find is described by Becker (1941). With some reservation the fragment can be dated to the beginning of the Subboreal (zone VIn according to the zonation of Knut Jessen). The fragment consists of a flattened funnel (fig. 46). The wide end of this measures 18 cm; the width at the narrow end is 5.5 cm. The total length of the fragment is 34.5 cm. The wickers measure 0.4-0.6 cm in thickness and are made of lime twigs. These are attached to one another with twisted lime-bast cross­ strips by means of a Zwirnbindung. To some extent the cross-strips run in a spiral around the funnel. The distance between cross-strips is 3.5 cm at most. The top of the cone is closed off by a double cross-strip. The funnel has been made narrower by cutting away wickers. In the funnel a stone c. 7 cm in diameter was found. At some distance from the fish-trap fragment the remains of a weir were found (3.5.1.). Becker interprets the find as the tail­ end of a fish-trap, which was placed in the water weighted with stones. It is not possible to ascertain whether a throat structure was present. In making a reconstruction of the tail­ end, I noticed that the wide end of the funnel measured c. 11 cm in diameter. The other end of the funnel is tied together so as to make it shut. Therefore the tail-end must have been very narrow. It is thus well possible that the fragment is the tail-end of a trumpet fish-trap. It could also be a transition form between a trumpet fish-trap and a fish-trap with throat, i.e. the long narrow fish-trap with throat.

39

Fish-trap fragment, found in 1940 during turfcutting operations in Fjellenstrup near Gilleleje, Sjaelland. Before the fish-trap was recognized as such, 1-1.5 m of it had already been dug away. The find is described by Becker (1943) and has been dated by Iversen, on the basis of pollen analysis, in the Middle Atlantic. From the geological investigation it became evident that the fish-trap lay in the bank zone of a former small island. This small island was situated in the narrow fjord that connected the S0borg Lake with the Kattegat in the Early Atlantic. The excavated fragment has a total length of 2.95 m. The maximum width is 90 cm. The width at the end of the tail is 20 cm. The wickers of the wickerwork measure 0.3­ 0.6 cm in thickness and are peeled. The dis­ tance between adjacent wickers is at most 0.7 cm. The twisted cross-strips are made of twigs, split lengthwise. The space between the cross­ strips measures 5.5-6.5 cm. To give extra strength to the outer casing there are some­ times two or three cross-strips close together. The entire structure is plaited together by means of the Zwirnbindung method. The wick­ ers of the tail-end are tied together with a piece of cord, which could be undone. In the tail a small stone is present. From that part of the fish-trap that had been dug away a few remains were retrieved. These consisted of two layers of wickerwork, which differed from each other. In one layer the wickers run parallel to one another (the wick­ ers of the outer casing); in the other layer the wickers converge towards each other (the throat wickers). Also peat was found contai­ ning three layers of wickerwork. On the basis of this Becker interprets the find as a fish­ trap with throat measuring c. 4 m in length. If we estimate the diameter of the wide part of the excavated fragment, this comes to c. 60 cm. The tail of the funnel is bound together to make it shut. From the reconstruction it is therefore clear that the fish-trap was very narrow (fig. 47). In my opinion we are here also concerned with the transition form be­ tween a trumpet fish-trap and a fish-trap with throat, i. e. the long narrow fish-trap with throat. Fish-trap fragment, found in 1968 during the excavation of a coastal settlement dating from Early Atlantic times at Villingebaek, Sjaelland. Kapel (1969) mentions the find, but

40

D.C.BRINKHUIZEN

fig. 50

fig. 50

fig . 51

fig. 50

fig.52

Recent and pre- and protohistoric fishing gear

gives no description of it. He interprets the fragment as being the throat of a fish-trap. From the photos accompanying his article it can be deduced that the wickers are most probably peeled. They converge towards one another and are joined together with cross­ strips according to the Zwirnhindung techni­ que. Here too the wickerwork has been made to become narrower by the cutting away of wickers. The distance between the cross-strips becomes smaller towards the narrow end. At least two layers of wickerwork are visible. Whether the fragment is the throat or the tail of a fish-trap will only become clear when the entire structure has been finally prepared. In the course of the excavation fish remains were found. These come from spurdog, cod and pike (identification U.M0hl). Two fish-traps and fish-trap fragments, found in 1946 in Lille Knapstrup, Holbaek district, Sjaelland. The find was made in the course of turf-digging operations in a peat­ bog and has only been published in a short communication in a Danish journal (Becker, 1946). The find consisted of two almost com­ plete fish-traps with throat. These lay c. 2.5 m apart. Moreover about 6 m away from these fish-traps the fragments of five or six other traps were found. Together with one of the fragments pieces of cord were found, as well as stones around which cord had been wound. It is clear that this fish-trap was formerly anchored with the aid of stones. The find is provisionally estimated to date from the At­ lantic period. Clark (1965) illustrates the two almost com­ plete specimens from Lille Knapstrup. The findspot that he mentions is however incorrect. Figure 48a shows a short, light fish-trap with throat. The trap has two hoops, namely one where the wickers of the outer casing meet

Fig. 48. The two complete fish-traps from Lille Knap­

strup, Holbaek, Sjaelland (after Clark, 1965).

Fig. 49. Fish-trap fragment from Jonstorp (after Pe­

tersson & Olausson, 1952).

Fig. 50. Fish-trap from the Schliisbeker Moor (after

Sprockhoff, 1953).

Fig. 51. Closely woven fish-trap from Oberdorla (after

Barthel, 1977).

Fig. 52. Fish-trap C from Rotterdam-Bergschenhoek

(after Sarfatij, 1978).

41

those of the throat, and one halfway down the outer casing behind the throat. Figure 48b shows a light fish-trap with throat too. This specimen also has a hoop halfway down the outer casing. The wickers of the outer casing are tied together at the end of the tail with a piece of cord, which could be loosened. The tail is fairly wide. This implies too that the specimen is not a trumpet fish-trap. Fish-trap fragments, found at Koster Vig (to be more correct: Sprove), Damsholte, Praes­ tm district, M0n. The find consists of 12 frag­ ments of one or more fish-traps. No further information is available, as the fragments were found in the course of turf-digging ope­ .rations. They are kept in the National Museum in Copenhagen where their classification num­ ber is A 39570-81. 10 Mathiassen (1948) mentions this find, about which no more pUblications have appeared as far as I am aware. Mathiassen (1948) mentions the find of a fish-trap at Svinnige Velje. In his publication he refers to Becker (1941). The latter author, however, interprets the find as a fragment of a weir (3.5.1.).

3.6.2. Archaeological finds of wickerwork fish­ traps from Sweden Fish-trap fragment, found in 1952 in Jonstorp, parish of Jonstorp, Schonen. The find is des­ cribed by Petersson & Olausson (1952). The fragment was found, when a well was being dug, at a depth of 6.3 m below the surface in a marine sandy gyttja, that dates from the Atlantic (probably an older phase of zone Vs according to the zonation of Tage Nilsson). The fragment measures maximally 10 by 20 cm and consists of two layers of wickerwork. In one layer the wickers converge towards one another and wickers have been cut away with some degree of regularity (fig. 49: see arrow). In the wickerwork of the other layer the wickers run parallel and with a certain degree of regularity two adjacent wickers are enclosed within a single loop of a cross-strip (a 'double wicker'). The working scheme followed in this layer is: 1 double wicker, 2 single wickers, 1 double wicker, 3 single wickers, 1 double wicker, 2 single wickers, 1 double wicker, 3

42

D.C. BRINKHUIZEN

single wickers, etc. On the basis of these differences the authors conclude that one layer is a fragment of the throat structure and the other part of the outer casing. The wickers used are unpeeled birch twigs. As a result of the pressure exerted by the overlying sediment they have been deformed into wickers 0.1 cm thick and 0.3-0.4 cm wide. The cross-strips consist of peeled willow twigs split lengthwise. The space between the wickers is at most 0.7 cm. The distance between the cross-strips, that hold the wickers together by means of the Zwirnbindung, is 3.5-5 cm. Also a double cross­ strip is present, consisting of two closely ad­ jacent single cross-strips (fig. 49: the cross­ strips on the extreme right). We also know of one species of fish that was caught with this fish-trap. Between the two layers of wick­ erwork some fish remains were present, namely those of a c. 45 cm long cod (identification O. Nybelin/J. Lepiksaar). Fish-trap fragments, found together in 1951 in the course of turf-digging activities in the Nebbe mosse, parish of Ostra Vemmerl6v, Schonen. The find is described by Stjernquist, Nilsson & Nybelin (1953). It dates from the Late Atlantic (the younger part of zone Vs according to the zonation of Tage Nilsson). The fish-trap is made of willow twigs split lengthwise. The connecting elements consist of twisted cross-strips of a different species of willow that hold the structure together by means of the Zwirnbindung technique. The space between the wickers measures c. 0.6 cm and between the cross-strips c. 4 cm. It cannot be ascertained from which part of the fish­ trap the fragments originate. That the frag­ ments do indeed come from a fish-trap and not a weir is clear from a fragment, in which a narrowing of the wickerwork is visible. Petttrsson & Olausson (1952) also mention the following unpublished fish-trap fragments from Sweden: Fish-trap fragment, found in 1946 in the Nebbe mosse, parish of Ostra Vemmerl6v, Schonen. Fish-trap fragment, found in 1947 in Hal­ jarp, parish of Tofta, Schonen. Fish-trap fragment, found in 1951 in the Store mosse, S6sdala, parish of Norra Mellby, Schonen.

3.6.3. Archaeological finds of wickerwork fish­ traps from Northern Germany Fish-traps, found in Lake Dummer. Petersson & Olausson (1952) mention this find, with reference to the publication of Reinerth (1939). In the latter a broad description is given of the results of the excavation of a settlement of the Funnel-Beaker Culture (Dummer-Lem­ bruch). There is mention of finds made of fish remains, from pike and perch, but not of any fish-traps. On the other hand, Durr (1962), who was excavation technician for Reinerth, does indeed mention two fish-traps in com­ bination with two weirs for this site. In contrast to Reinerth's publication, Durr states that the settlement was occupied twice, namely in the Neolithic and in the Iron Age. He sees in­ dications of inhabitation during the Iron Age in the presence of bones of domesticated horses. Also Reinerth mentions that bones were found of both domesticated and wild horses. As neither author gives any measu­ rements of these bones, we are unable to verify whether or not the horses were indeed domes­ ticated. In addition Durr points out that an Iron Age pot was found. This pot is currently regarded as a Late Havelte pot of the Funnel­ Beaker Culture (Bakker & Van der Waals, 1969). Another indication of inhabitation du­ ring the Iron Age is given by the find of a pointed stake from the vicinity of the weir. This stake is said to have been cut by a metal axe. However, in my opinion the find of a single stake cut with a metal axe gives no decisive proof of inhabitation in the Iron Age. Durr sees the most convincing argument for multi-period inhabitation in the presence of the fish-traps. If the Neolithic settlement is contemporaneous with the fish-traps, then it follows from the location of the fish-traps and weirs that these stood on dry land. This means that the fish-traps cannot have been contem­ poraneous with the Neolithic settlement. They must therefore be older or younger than the settlement. A dating in the Iron Age is possible, but on the basis of the above-mentioned ar­ guments not imperative. Fish-trap fragment, found in the Priestermoor near Duvensee, Kreis Herzogtum Lauen­ burg. This fragment is said to date from the Mesolithic (Schwabedissen, 1949). No further details of this find have been published.

Recent and pre- and protohistoric fishing gear

Fish-trap, found in 1952 during turf-digging operations in the Schliisbeker Moor, Gem. Moorsee, Kreis PIon. Sprockhoff (1953) gives a brief description of this find and an outline sketch of its position in situ (fig. 50). From a palynological investigation it is evident that the fish-trap dates from the Early Atlantic. The fish-trap, which was originally fastened to posts, was found as a 10-12 cm thick layer of wickerwork. This wickerwork consisted of willow wickers c. 0.5 cm thick, which had been carefully attached to one another by means of cross-strips made of bast. The distance between the wickers measured just 0.5 cm and that between the cross-strips about 5 cm. At the same time the author states that the funnel structure, which was at least 2 m long and had a mouth opening about 1.5 m across, can clearly be distinguished from the tube-shaped end of the fish-trap which was almost 2.5 m long. From the outline sketch of the fish-trap in situ it is not clear to me where Sprockhoff made these observations. On the basis of the drawing I myself am inclined to regard this find as a light fish-trap with or without a throat, which was fixed to posts, in combi­ nation with two weir sections (wickerwork screens). A closer study of this find will provide more details. In addition Sprockhoff mentions the find of a second fish-trap, which was found 12.5 m away from the first one, lying undis­ turbed in the peat. The remains of a third specimen were found that same year elsewhere in the peat. Also in Southeastern Germany wickerwork fish-traps and fragments thereof have been found. Thus in the peat-bog near Oberdorla (Thiiringen) many specimens came to light in the course of excavations. These dated from Roman times (Barthel, 1977). They had all been plaited from willow twigs. Among the specimens found there was only one light fish­ trap, which had no throat. All the other fish­ traps (about 20 specimens) were fish-traps with throat. The wickerwork of these fish-traps had been made according to the Leinwandbindung technique (fig. 51). In a number of these fish­ traps stones were still present, that had served as weights. One fish-trap still contained fish remains, namely of a pike.

43

3.6.4. Archaeological finds of wickerwork fish­ traps/rom the Netherlands Fish-traps, found in 1978 at Rotterdam-Berg­ schenhoek during the excavation of a small 'extraction camp', situated on a small 'peat island' measuring 10 m 2 in area and only 30­ 35 cm in thickness. The group of finds dates from the Late Atlantic (5415 ± 60 BP: GrN 7764). The finds consist of a few sherds, three pieces of flint, a few bone tools, faunal remains and worked wood, including the fish-traps. Three almost complete fish-traps were found as well as a fragment of a fourth. Together with a few large planks this fragment formed the basal part of one of the layers added to the peat island to make it higher. The three other fish-traps lay in the surrounding clay. In addition, scattered among the find-bearing layers were a few fragments of wickerwork, that may have come from the fish-traps. Louwe Kooijmans, who carried out the ex­ cavation, presents a brief report on this group of finds in Sarfatij (1978).11 No special study of the fish-traps has appeared as yet. Evidently all of them were light fish-traps. Their wick­ erwork thus consists of wickers held in place by cross-strips by means of the Zwirnbindung. From one of the fish-traps it can be deduced that the cross-strips were plaited around the wickers as separate 'hoops'. At the spot where the 'hoop' closes, the cross-strip passes diag­ onally out of line so as to commence the next 'hoop'. The three complete fish-traps are about 1.50 m long. Two of them are provided with a throat. The third one is throatless, however. This implies that it is a trumpet fish-trap. If we look at one of the fish-traps with throat (fig. 52), however, then we see the remains of the two hCJopslt coiitained:These-hoops have been broken or came loose as a result of the tension within the wickerwork. When the hoops sprung loose part of the outer casing came loose with them, though not the throat. This could indicate that the throat was faste­ ned to the outer casing in only a few places. It is also possible that the throat was a separate element which was inserted into the outer casing. If this theory is correct, then the fish­ trap without throat could be a trumpet fish­

44

D.C.BRINKHUIZEN

trap or a separate outer casing. Further study of the fish-traps and their attempted recon­ struction will probably provide the right ans­ wer. 12 During the excavation sieving was carried out, and many fish remains were thus found. These come from pike, perch, ruff, eel, mem­ bers of the carp family and European catfish. Fish-trap fragments, found at Vlaardingen during the archaeological investigation (1959­ 1961) of the settlement of the Vlaardingen Culture (4330 ± 60 BP: GrN 2303; 4330 ± 100 BP: GrN 2487). The find is described by Van Herson Scholten (1977). It consists of two fragments, namely a hoop with remains of wickerwork and a small fragment of wicker­ work, that may come from the former. The wickerwork consists of slightly converging wickers, held together with cross-strips by means of the Zwirnbindung. The cross-strips are made of bands of twisted bast and measure 0.3 cm in diameter. The distance between the wickers is at most 0\7 cm. About the hoop the author says: 'The weel fragment consisted of an outer hoop of two or three pieces of wood, between which the side bars were clam­ ped. The outer hoop was wound around with strips of bast 8 mm wide. In situ, the outer hoop measured 56 x 68 cm and the circum­ ference was 193.7 cm, which, after reconstruc­ tion, resulted in a mouth diameter of 62 cm' and further on 'on account of its shape this fragment is probably the internal funnel of a weel, but in the absence of other fragments no further reconstruction can be attempted'. If this interpretation is correct, then some­ where on the fragment there should still be remains present of the area of attachment between the wickerwork of the outer casing and that of the throat. The most obvious place for this is the hoop. Here no remains are present however. It is also possible that the attachment area of the outer casing lay in that central part of the throat that has been de­ stroyed. To find out whether this may have been the case, it is necessary to make a reconstruction. For this purpose we have to consider the technique that was applied to make the wickerwork narrower. This narrow­ ing begins from the fourth cross-strip. Here we see the so-called double wickers, like those of the fish-trap fragment from Jonstorp. The plaiting system that was followed is: 1 double

wicker, 2 single wickers, 1 double wicker, 2 single wickers, etc. These double wickers are coloured black in figure 53. The double wick­ ers run as far as the eighth cross-strip, after which they can be traced no further. The eighth cross-strip lies more or less concentrically within the hoop. The distance between the two is about 15 cm. The number of wickers that are plaited on to the hoop amounts to c. 200 and the number of wickers of the eighth cross­ strip to c. 150 (the double wicker is counted as one wicker). The ratio is thus 4:3. The estimated diameter of the hoop is about 62 cm. From this diameter and the estimated ratio it follows that the diameter of the eighth cross­ strip must have been about 46.5 cm. On the basis of this data a reconstruction can now be made. From this reconstruction it follows that it must have been possible for the outer casing to be plaited to the throat. The fragment could thus be part of a short fish-trap with throat (fig. 54a). As we do not know how the narrowing continued after the eighth cross­ strip, there still remain three other possibilities. If the fragment did not become any narrower, then it would have been a trumpet fish-trap (fig. 54b). Alternatively the fragment may be an isolated throat, that was placed inside a separate outer casing (fig. 54c). The reverse is also possible: the fragment may be an isolated outer casing, inside which a separate throat was inserted (fig. 54d). The construction of the hoop is strange. Generally, for a hoop a single long thick wicker was used. The hoop of this fragment consists of two or three pieces of wood. In my opinion this indicates that the hoop became weakened or broken in the course of use. By the addition of elements to provide extra strength this problem was overcome. This also explains the fact that the wickers in the NW-quadrant of the hoop are clamped between the hoop and the strengthening elements. Elsewhere around the hoop this is not the case. During the excavation of the settlement fish remains were found, which could be identified as remains of sturgeon, thin-lipped grey mullet and pike (Brinkhuizen, 1979b). Sturgeon was certainly not caught with this fish-trap (3.5.2.). Fish-trap found in 1971 during the exca­ vation of a former bed of the river Rhine in Utrecht (Van Regteren Altena & Sarfatij, 1973). The find dates from the second half

Recent and pre- and protohistoric fishing gear

45

Fig. 56. Type classification of fish~traps.

of the first century A.D. (1880 ± 35 BP: GrN 6633). It was fitted with a throat and measured 95 cm in length. The wickerwork of willow twigs had been made according to the Lein­ wandbindunf( technique. Between the wicker­ work ot the throat and outer casmg a number of stones and a clay net sinker which had served as weights were present (fig. 55). The fish-trap strongly resembles the closely woven fish-traps from Oberdorla. Two fish-traps found in 1975 during the excavation of the Early Medieval harbour complex in the Kromme Rhine riverbed at Dorestad (Wijk bij Duurstede). Both fish-traps were approximately 3.5 m long and plaited according to the Zwirnbindung technique. One fish-trap was fitted with a throat. Due to the difficult excavation conditions (rising ground­ water) this observation could not be made on the second one. The two fish-traps will be published in another article in the near future. 3.6.5. A few remarks concerning the prehistoric wickerwork fish-traps

3.6.5.1. The typology If we wish to ascribe a wickerwork fish-trap to a particular type, then first of all a typology must be established. A sentence such as 'a fish­ trap made of wickers and cross-strips of the type well known from South Scandinavia' (Glasbergen et al., 1966) is of little value as so far no typology has been established for fish-traps. To arrive at a typology, we must first return to the trumpet fish-trap. Sirelius (1906) proposes that this fish-trap developed via some intermediate forms into the fish-trap with fixed throat. 13 On the basis of some

intermediate forms known to us from ethno­ graphical studies we are confronted with the typological series given in figure 56. On the basis of this typology we can now classify the prehistoric fish-traps (table 3). As the finds usually consist of fragments, for any particular case we cannot determine the exact type. Thus in the table for each fish-trap the different possibilities are indicated. The fish­ traps are listed not according to site location, but in order of age. From table 3 a few conclusions can be drawn. The Lille Knapstrup fish-traps (type C2 or D) date in all probability from the Atlantic period. The Magleby Long fish-trap (type A, B or Cl) dates from the beginning of the Subboreal. From this it follows that on the Danish island of Sjaelland a more primitive type was in use at a later time than a more highly developed type. If we now consider the fish-traps found in Northwest Europe as one group, then it is clear that the Lille Knapstrup fish-traps are typologically comparable with the Bergschenhoek fish-traps C and F. Also the latter are some 500 years older and as regards type more highly devel­ oped than the Magleby Long tish-trap.14 Thls implies that we cannot link the typological series to the archaeological time-scale that has been established for Northwestern Europe. In other words: the typology is only descriptive and has no chronological significance. This can be explained from the ethnological data. In Finland, France and England, up until recently the trumpet fish-trap (type A) and the fish-trap with fixed throat (type D) were both in use at the same time. Which type of fish-trap was used depended on the speed of

46

D .C. BRINKHUIZEN

fig. 55

fig. 55

[

fig. 54

fig. 57

47

Recent and pre- and protohistoric fishing gear

flow of the water. The Woguls, Ostjaks and Syrjans eJen had the types A, Cl, C2 Clnd D. On the basis of this evidence it is reasonable to assume that in prehistoric times too the various types of fish-trap in any particular region could have been used alongside one another at the same time.

3.6.5.2. The origin of the wickerwork fish­ traps The Nidl0se fish-trap, which dates from the Early Atlantic, already possesses a loose/fixed throat and the wickerwork is regularly plaited. This high degree of precision indicates that the fish-traps had undergone development already before the Atlantic. Where this de­ velopment took place and when it began cannot be said at present. In principle the area of origin could have been any locality where fish were present and from which the necessary construction materials could have been obtai- , ned. As regards the Upper Palaeolithic rein­ deer hunters, who lived in Northwest Euro­ pe near the periphery of the ice sheet, we can almost certainly say that they were not familiar with wickerwork fish-traps. In the enyiron­ ment in which they lived the necessary con­ struction materials were lacking at that time. It is true that fish lived in the rivers and lakes that were present near the edge of the ice­ sheet, namely cold-resistant freshwater and migratory fish. These may have been caught using e.g. the fish-spear and perhaps nets made of strips of leather knotted together. In the course of the climatic amelioration and the accompanying changes in flora and fauna the reindeer hunters were gradually forced to switch over from hunting reindeer to hunting wild boar, red deer, roe deer, elk, aurochs and birds. In addition their means of subsistence included fishing and fruit-collecting. It is ob­ vious to assume that fishing was done in the

Fig. 53. Vlaardingen, cutting 17, in situ drawing of the

weel by S. Hoek, I.P.P. (after Van Iterson Scholten, 1977).

The double wickers are shown in black.

Fig. 54. The four possibilities for the reconstruction of

the fish-trap from Vlaardingen.

Fig. 55. Closely woven fish-trap from Utrecht. Photo:

Foto Dienst Gemeente Utrecht. '

Fig. 57. Drawmg of the unfolded' net I from the Nebbe

mosse (after Stjern'quist et al., 1953).

Table 3. Classification of the prehistoric wickerwork fish-traps fuund in ~orthwest ~urope Fish-tr::ps

Type A rl Cl C2 D Dating Atlantic • • • • Early Early Atlantic Atlantic • • • • Early Middle Atlantic • • Middle Atlantic • • • • Late Atlantic ••

Nidl¢lse Villingebaek Schliisbeker Moor Fjellenstrup Jonstorp Bergschenhoek fish-trap C Bergschenhoek fish-trap F Bergschenhoek fish-trap B • Lille Knapstrup Lille Knapstrup Magleby Long Vlaardingen

• • • ••

Late Atlantic • • Late Atlantic • Atlantic • • Atlantic Early Sub boreal • • • •• •• Early Subboreal • ••

long familiar way, namely using the fish-spear and fishing nets. Finds of fish-spear barbs and parts of drift-nets (3.2 and 3.7) are known from the Early Mesolithic of Northwest Europe. The. first wickerwork fish-traps found here date, from the Middle Mesolithic. They are, how­ ever, already in an advanced stage of de\iel­ opment. There are two possible explanations for this. If the development of the fish-trap did take place in Northwest Europe, then this occurred fairly rapidly. The alternative expla­ nation is that the fish-trap developed elsewhere and was introduced into Denmark at a far­ advanced stage of development. This devel­ opment could have taken place e.g. in the Late Palaeolithic in the interior or on the coasts of Southern Europe. The construction mate­ rials were available there and the fish fauna was presumably the same as that present today. Which ofthe two explanations is correct cannot be ascertained, also on account of the lack of finds. From Rosenlund (1976) it is evident that in Denmark the number of fish species that could be identified among the bone material from Early Atlantic settlements shows a sharp increase with respect to those from older settlements. This could indicate the introduction of a more effective means of catching fish, but it is more likely that this increase in fish species is due to the fact that the colonisation of the Danish waters had almost been completed at the Boreal-Atlantic boundary.

48

. D.C.BRINKHUIZEN

3.6.5.3. The plaiting technique and the materials used All the Mesolithic and Neolithic fish-traps found in Denmark, Sweden, Northern Ger­ many and the Netherlands are light fish-traps. The wickerwork of these fish-traps was with­ out exception plaited using the Zwirnhindung technique. The material used for the wicker­ work varies. For the wickers of the outer casing and the throat, twigs of lime, willow and birch were taken. The cross-strips were made of twisted bands of willow and lime bast, longi­ tudinally split willow wickers and pine roots. All these materials are the same as those which were used until recently in Finland. 3.6.5.4. The catch In view of the skill with which the prehistoric fish-traps were plaited, it is certain that these were by no means inferior, in terms of effi­ ciency, to their recent counterparts. On the analogy of recent wickerwork fish-traps some authors regard the prehistoric ones as intended for catching eels. That this is by no means necessarily so is shown by the Jonstorp fish­ trap, which was used in a marine environment. In this fish-trap remains of cod were present. The fish remains that were found with the Bergschenhoek fish-traps do not only come from eels, but also from various other species of fish. This is not unusual, for as a general rule a fish-trap is not selective as to its catch. Almost all fishes and crustaceans that are smaller in diameter than the diameter of the rearmost throat opening can be trapped in­ side. 15 Once inside the trap generally only those fish which are smaller in diameter than the space between the wickers of the outer casing can escape. One fish species that could have been caught with prehistoric traps is the herring. Clark (1948) states: A point to emphasize is that, apart from a few rib­ bones from Solager in Denmark, remains of herring are conspicuously absent from the prehistoric settle­ ments of Europe. The herring fishery is notoriously subject to great local fluctuations, but such can hardly be held to account for the absence of herring-bones from so many regions over so lengthy a period. The absence of a fish, which to-day is caught in larger numbers than any other, from early settlements in the whole region from the Baltic to the west coast of Norway and northern Britain, and extending from

the Stone Age to the Early Iron Age, can only mean that the fishery was not developed during the pre­ historic period.

However, in putting forward this theory the author has disregarded the possibility that certain fish species may be considerably un­ derrepresented in the faunal remains. For this is evidently the case with species of fish having a skeleton rich in fat. These species include: fish belonging to the salmon family, to the herring family, to the mackerel family and to flatfish (Lepiksaar, 1975). That herring were indeed caught is shown by the excavated herring remains from a Middle Neolithic settle­ ment at Korsnas, Grodinge, Sodermanland, Sweden. From a refuse pit, measuring 2 m x 0.6-0.7 m and at least 0.3 m in depth, an earth sample weighing 2,375 grammes was taken. This sample was found to contain, in addition to several hundred remains of pike, perch, eel and whitefish, at least 21,000 re­ mains of herrings (Aaris-S"Hensen, 1978). 3.7. The fishing net The oldest fishing net known to us dates from the Early Boreal. The remains of this net were found in 1913 at Korpilahti in the neighbour­ hood of Antrea. This place lies to the west of Lake Ladoga in what was formerly Finnish territory. The find is mentioned by i.a. Clark (1948, 1965, 1975), Stjernquist et al. (1953), Welinder (1969) and Van Iterson Scholten (1977). Eighteen flat, more or less oval floats of pine bark were found, each perforated at one end. Below several of these floats remains of the net were still present. The cord from which this net had been made consisted of two strips of willow bast twisted together. Close by these floats lay 31 unworked fist­ sized stones. The find is interpreted as the remains of a seine,16 that must_have measured 27-30 m in length and 1.3-1.5 m in height. In addition to this fishing net Van Iterson Scholten (1977) also mentions the find of a Mesolithic fishing net from Vuoksenranta. Ayrapaa (1950) shows a photo of this find. However, this is the same as the illustration given by Clark (1975) of the net from Kor­ pilahti. Also from the texts and references of the two authors it is clear that one and the

Recent and pre- and protohistoric fishing gear

same fishing net is concerned here. Thus one Mesolithic fishing net is known from Finland and this is mentioned in the literature as the fishing net from Korpilahti, Vuoksenranta or Antrea. Another find of part of a drift-net is a disc­ shaped float of pine bark from the Bare mosse (Sweden). From a pollen-analytical investiga­ tion it has been established that this float dates from an early stage of the Boreal (Welinder, 1969). Thus it is contemporary with the drift­ net from Antrea. Indreko (1948) describes the find of a frag­ ment of a flat oval float of pine bark from Siivertsi (Estonia). An unworked fist-sized stone with cord remains from the same region is interpreted as a weight for a fishing net. The finds have been dated by pollen analysis to the Boreal!Atlantic transition. From Northwest Europe net fragments da­ ting from younger periods are also known. Schwabedissen (1957, 1957/1958) mentions the find of an almost complete fishing net from Rude 2 (Forstermoor, Kr. Schleswig), a settle­ ment of the Erteb011e/Ellerbek period. The author says that the net consists of triangular and diamond-shaped meshwork, the. cord being knotted with so-called Fischer- oder Weberknoten.17 From a peat-bog near Ordrup (Denmark) a fragment of an unknotted net is known, dating from the Neolithic (Clark, 1965). Dating from the same period are a number of net fragments from the Nebbe mosse, parish of Ostra Vemmerlov, Schonen, Sweden. This find has been excellently des­ cribed by Stjernquist, Nilsson & Nybelin (1953). The' fragments were found to have come from three different nets. It was possible to Istudy two of these (fragments I and II) in detail. Thus the material used could be iden­ tified (lime bast), and the mesh-width (net I: 4-5 cm; net II: 5-6 cm) and the type of knot (net I: Pfahlbau knot; net II: another type) could be ascertained. From the position of the knots in the two fragments it was evident that net I had been made in the round and net II not. In addition net I showed some par­ ticular features from which it was clear that this had been part of the tail of a bag-shaped net (fig. 57). In view of the small size of the remaining fragment (c. 30 x 50 cm), however, the overall shape of the net cannot be ascer­ tained. On the' basis of comparison of data

49

provided by these nets with data from recent Swedish lime-bast fishing nets it seems not unlikely to the authors that the recent ones are directly descended from their prehistoric predecessors. As to the kinds of fish caught by net I, among the netting a number of bones of tench were found (estimated length of the fish: 50 cm), as well as a caudal vertebra of a pike (estimated length of the fish: 50 cm) and scales of a roach. 4. SUMMARY AND CONCLUSIONS In the first part of this article a number of fishing methods and types of fishing gear known from ethnographical sources are des­ cribed. Three principal fishing methods can be distinguished: a. Fishing without the use of any fishing gear; such as catching fish with one's bare hands, stunning fish with a stick, catching fish by making the water turbid, and poisoning. The last two mentioned methods are more or less unselective with regard to th~ fish species ' and their sizes, and fish are caught en masse. b. Fishing by means of actively operSlt~d fishing gear. The types of gear include the gaff, the rod with a bob, the stick with a noose, the fish-spear, the line with a gorge and the line with a fish-hook. When the fisherman is fishing with one of these types, he searches for the fish, or lures it towards himself, and tries to catch it. The method can be time­ consuming and is more or less selective with regard to the species and the size of the fish. Usually one fish is caught at a time. Deter­ mining factors are e.g. the clearness of the water, the size of the gorge/fish-hook, the strength of the line and the kind of bait. c. Fishing by means of passively-operating fishing gear. The kinds of fishing gear are: the weir, the fish-surround, the wickerwork fish­ trap and the fishing net. When the fisherman uses these he only has to collect the fish which have become trapped. The weir itself can function in places where there is a strong current or great diurnal differences in water level. In stagnant or slowly flowing water it is always used in combination with the fish-surround, fish-trap or fishing net. These combinations have resulted in very large and complicated complexes of fishing gear.

50

D.C.BRINKHUIZEN

The duration of these complexes depends on the construction materials used. Those parts made of wickers, roots or strips of bast (the screens, collection chambers, fish-traps and fishing nets) are worn out after a couple of years. Heavier vertical wooden posts to which these contraptions are fastened last longer. In regard to the wickerwork tlsh-traps two main types can be distinguished: the trap without internal funnel, or trumpet fish-trap, and the trap with internal funnel. Which of the two is used depends on the speed of flow of the water. A further division of the fish-traps with internal funnel can be made on the basis of the weaving technique used. In regard to this we distinguish the light fish-trap, which is woven according to the Zwirnbindung, and the closely woven fish-trap, which is woven ac­ cording to the Leinwandbindung. The latter is mainly used for catching very small kinds of fish. In general a fish-trap or fishing net is not selective as to the catch. However, the plaiting technique used, the diameter of the rearmost throat opening, the mesh-size of the net, the kind of bait and the position of the fishing gear in the water can influence the fish species and sizes caught. The most highly developed fishing gear is the fishing net, and on the basis of its shape four groups can be distinguished: the square net, the round net, the rectangular net and the bag-shaped net. In the actual process of catching fish, the fisherman sometimes uses a few items of auxiliary fishing equipment. These are: the splashing stick, the gaff, the scoop-net, the live­ box and the boat. In the second part of this article a few general comments are made on the pre- and protohistoric actively operated fishing gear, the bone and antler fish-spear prongs, gorges and fish-hooks. Six fish-hooks from the Neth­ erlands are described and considered from a theoretical viewpoint with regard to their effectiveness and the possible fish species that could have been caught with them. In regard to the passively-operating fishing gear, some comments are made on the finds of prehistoric weirs and fishing nets. Finds of fish-surrounds are conspicuously absent in the

archaeological record. Emphasis is laid on the pre- and protohistoric wickerwork fish-traps from the Netherlands, Northern Germany, Denmark and Sweden. On the basis of the ethnographically known types of wickerwork fish-traps a simple type classification for the prehistoric wickerwork fish-traps has been made. The few Mesolithic and Neolithic wickerwork fish-traps found fit into this typology. However, from the datings of these traps it is clear that the established typology is only descriptive and has no chro­ nological significance. The most striking conclusion concerning the prehistoric wickerwork fish-traps found in Northwestern Europe is that their overall shape, the plaiting technique and the materials used were the same as those which have been used until recently. In view of this, it is certain that the prehistoric wickerwork fish-traps were very efficient and could have trapped nearly all the fish species which were living at that time. It is also clear that they were used both in a freshwater and in a marine environment. 5. ACKNOWLEDGEMENTS For the realization of this article I am grateful to a number of people. In the first place to Dr. A.T. Clason and Dr. W. Prummel who read through the text and commented upon it; furthermore to Professor L.P. Louwe Kooij­ mans who provided me with useful informa­ tion on fish-traps; to K. Vlierman (Rijksdienst voor de IJsselmeerpolders) for information concerning the two fish-hooks from the N oord­ oostpolder; to F.H. Kuiper and Son for placing the aalkubbe at my disposal; to 1.M. Smit (B.A.I.) for executing the drawings; to Mrs. H. Klaassens and Mrs. E. Rondaan-Veger (B.A.I.) for typing the manuscript; and to Mrs. S.M. van Gelder-Ottway M.Sc. for the trans­ lation into English. Postscript

The present article was concluded at the end of September 1979. Since that date no fewer than two bone fish-hooks and more than 200 barbed points or fragments thereof have been

Recent and pre- and protohistoric fishing gear

recovered from the Rotterdam-Europoort sand dredging works (pers. comm. L. Verhart, R.M.O.). Also a fish-trap possibly dating from 200-100 B.c. was found during the excavation of terp 100 in the Uitgeesterbroekpolder near Uitgeest (pers. comm. A. van Gijn, I.P.P.). 6. NOTES 1. In the Netherlands, however, it is known that the screen of a zalmsteek (= a row of posts with hurdles in between that salmon swim up against, before being driven into fish-traps) rotted away already halfway through the summer and then had to be replaced by a new one (Van Doorn, 1971). 2. This distance is dependent on the size of fish that the fisherman wants to catch in his trap. 3. Fish-traps can also be used for fishing in winter. The Syrjans cut a long trench in the ice, in which a weir of young evergreen trees is constructed. At the place of the fish-trap an even bigger hole is made. When fishing is in progress the fisherman must ensure that this hole' does not freeze over. An already existing weir can also be used in this way (Sirelius, 1906). 4. Fish-traps may also be provided with a long tail because the fisherman expects a lot of fish and does not wish to pick up his catch immediately. 5. Until recently the Finns used flat oval' wooden floats. As weights for nets they used i.a. oblong stones wrapped in birch bast (Sirelius, 1906). 6. Collection Rijksmuseum van Oudheden, R.M.O. d 1928/12.1. 7. Collection Rijksmuseum voor de IJsselmeerpol­ ders, without inventory number. The hook is made out of red-deer antler (pers. comm. Dr. A.T. Clason and Dr. W. Prummel, B.AJ.). 8. Collection Fries Museum, F.M. 240/1. Boeles (1951) says that the hook is made of (whale?) bone. According to Van der Heide (1955) the hook is made of whalebone. However, in a more recent publication (Van der Heide, 1972) the whalebone has changed into 'probably antler of elk'. On the basis of closer inves­ tigation it is evident that the hook is made of antler. The raw material was most probably the antler of an elk although a large antler of red deer cannot be excluded (pers. comm. Dr. A.T. Clason and Dr. W. Prummel, B.A.I.). 9. The length measurements of the hooks a-d are taken from Louwe Kooijmans (1974). Their width, hook ope­ ning and grip angles were estimated from illustrations. 10. This information was kindly given to me by P.O. Nielsen, assistant curator of the First Department of the National Museum in Copenhagen. 11. The find is erroneously mentioned in the Archaeo­ logische krQniek van Zuid-Holland for the year 1977. 12. In particular the diameter of the hindmost throat opening must be noted. This opening determines the size of fishes able .to swim inside. 13. Also Sirelius (1906) does not exclude the possi­

51

bility that the fish-trap with throat is derived from the fish-surround. 14. Becker (1941) ascribes the Magleby Long fish-trap to the Funnel-Beaker Culture. 15. Only the aalkubbe (2.3.3.2.3.) that was baited with herring roe, worms, etc. caught exclusively eels. The diameter of the rearmost throat opening (5 cm) and the deviating position of the trap in the water were deter­ mining factors for this. 16. Seine = a drift-net, one end of which is taken out from the shore and subsequently hauled in again on land so that the net encompasses a wide arc. The net is hauled in from both ends, so that the fish enclosed within the net are driven ashore in a heap. 17. There are at least five different known types of knots (Van Iterson Scholten, 1977).

7. REFERENCES Aaris-S0rensen, K., 1978. Knoglematerialetfra den mel­ lemneolitiske boplads ved Korsnas, .Grodinge socken, Sodermanland ( Riksantikvarieambetet och Statens Historiska Museer Rapport 8). Stockholm. Ayrapaa, A., 1950. Die altesten steinzeitlichen Funde aus Finland. Acta Archaeologica 21, pp. 1-43. Bakk~r, J.A. & J.D. van det Waals, 1969. Denekamp'­ Angelslo. Cremations, collared flasks and a corded ware sherd in Dutch final TRB contexts . .In,: G. Daniel & P. Kjaerum (eds.), Megalithic graves and ritual. Papers presented at the III Atlantic Collo­ quium, Moesgard 1969. K0benhavn, pp. 17-50. Barthel, H.J., 1977. Die Germanische Binnenfischerei im Gebiet des See- und Moorheiligtums von Ober­ dorla. Alt-Thiiringen 14, pp. 148-185. Becker, c.J., 1941. Fund af ruser fra Danmarks sten­ alder. Aarbnger for Nordisk Oldkyndighed og Historie, pp. 131-149. Becker, c.J., 1943. Et 6000-aarigt fiskeredskab. Fra det gam Ie Gilleleje, pp. 70-87. Becker, c.J., 1946. Stenalderens ruser var konstrueret som nutidens. Salomonsen-Leksikon- Tidsskrift 10, pp. 1118-1120. Blok, D.P., 1974. De Franken in Nederland. Bussum. Boeles, P.C.J.A., 1951. Friesland tot de elfde eeuw. 2nd ed. 's-Gravenhage. Boddeke, R., 1971. Vissen & vissen. Amsterdam. Brinkhuizen, D.C., 1979a. On the finds of European catfish (Siluris glanis L.) in the Netherlands. In: M. Kubasiewicz (ed.), Archaeozoology 1. Proceedings of the IIIrd International Archaeozoological Conference held 23-26th April 1978. Szczecin, pp. 256-261. Brinkhuizen, D.C., 1979b. Preliminary notes on fish remains from archaeological sites in the Nether­ lands. Palaeohistoria 21, pp. 83-90. Casparie, W.A. & J.E.J. Swarts, 1978. Hout. Spiegel Historiael13, pp. 267-281. Casteel, R.W., 1976. Fish remains in archaeology and paleo-environmental studies. New York. Clark, J.G.D., 1936. The Mesolithic settlement of North­ ern Europe. Cambridge.

52

D.C.BRINKHUIZEN

Clark, J.G.D., 1948. The development of fishing in prehistoric Europe. The Antiquaries Journal 28, pp. 45-85. Clark, J.G.D., 1954. Excavations at Star Carr. Cam­ bridge. Clark, J.G.D., 1965. Prehistoric Europe, the economic basis. 2nd ed. London. Clark, J.G.D., 1975. The earlier stone age settlement of Scandinavia. Cambridge. Degerb01, M., 1945. Subfossile fisk fra kvartaertiden i Danmark. Vidensk. Medd. fra Dansk Naturh. Poren. 108, pp. 103-160. Doorn, T.H. van, 1971. Terminologie van de riviervissers in Nederland. Assen. Dorleijn, P., 1977. Geen moed vist ook. Bussum. Durr, H.W.A., 1962. Zur Ausgrabung an der Hunte bei Dummerlohausen 1938/39. Unpubl. man. Es, W.A van, 1974. Vis uit Dorestad voor mijnheer Calkoen. Westerheem 23, pp. 89-94. Forde, D.C., 1934. Habitat, economy and society. London. Glasbergen, W. et al., 1966. De Neolithische nederzet­ tingen te Vlaardingen (Z.H.). In: W. Glasbergen & W. Groenman-van Waateringe (eds.), In het voet­ spoor van A.E. van Giffen. Amsterdam, pp. 41-65. Groot, S.J. de & S. Schaap, 1973. De Nederlandse visserij rond 1900. Amsterdam. Heide, G.D. van d'er, 1955. Archaeological investigations on new land. Antiquity and Survival 2, pp. 93-122. Heide,G.D. van der, 1962. Een bijzondere bodemvondst in Overijssel. Verslagen en Mededelingen ~an de Vereniging tot beoefening van Overijsselsch Regt en Geschiedenis 77, pp. 21-35. Heide, G.D. van der, 1972. Graven naar het verleden. Utrecht. Indreko, R., 1948. Die mittlere Steinzeit in Estland (= Kungl. Vitt. Hist. och Antik. Akad. Handl.). Stock­ holm. Iterson Scholten, P.R. van, 1977. Rope and fishing tackle. In: B.L. van Beek, R. W. Brandt & W. Groenman­ van Wateringe (eds.), Ex Horreo. Amsterdam, pp. 135-143. Kapel, H., 1969. En boplads fra tidlig- atlantisk tid ved Villingebaek. Pra Nationalmuseets Arbejdsmark, pp. 85-94. Lekholm, C.G., 1951. A technical study of some Scanian bone fish-hooks. Meddelanden fran Lunds Univer­ sitets Historiska Museum, pp. 123-143. Lepiksaar, J., 1975. Uber die Tierknochenfunde aus den mitteralterlichen Siedlungen Sudschwedens. In: A.T. Clason (ed.), Archaeozoological studies. Papers of the Archaeozoological Conference 1974, Groningen. Amsterdam, pp. 230-239. Lobregt, P. & J. van Os, 1977. De laatste riviervissers. Zutphen. Louwe Kooijmans, L.P., 1970/1971. Mesolithic bone and antler implements from the North Sea and from the Netherlands. Berichten van de Rijksdienst voor het Oudheidkundig Bodemonderzoek 20-21, pp. 27­ 73. Louwe Kooijmans, L.P., 1974. The Rhine-Meuse delta. Pour studies on its prehistoric occupation and Holo­ cene geology. Leiden.

Louwe Kooijmans, L.P., 1976. Local developments in a borderland. Oudheidkundige Mededelingen uit het Rijksmuseum van Oudheden te Leiden 57, pp. 227­ 297. Mathiassen, T., 1948. Danske Oldsager I Aeldre Sten­ alder.K0.benhavn. Metz, W.H., 1975. Pre- en protohistorie in Veluws museum 'Nairac' te Barneveld. Barneveld. Michelsen, P., 1952. Risflettede aIekuber, Fra National­ museets Arbejdsmark, pp. 91-100. Mohr, E., 1952. Der Star. Leipzig. Molen, S.J. van der, 1976. Vissers van wad en gat. Leeuwarden. Muus, B.J. & P. Dahlstf0m, 1968. Zoetwatervissengids. Translated and edited by R. Boddeke. Amsterdam/ Brussel. Peesch, A, 1966. Holzgeriit in seinen Urformen. Berlin. Petersson, M. & E. Olausson, 1952. Eine mesolitische Fischreuse aus Jonstorp, Schonen. Meddelandenfran Lunds Universitets Historiska Museum, pp. 141-157. Regteren Altena, H.H. van & H. Sarfatij, 1973. Waar­ nemingen in de bouwput 'van V. & D. Achter Clarenburg. Maandblad Oud Utrecht 46, pp. 68-70. Reinerth, H., 1926. Die jiingere Steinzeit der Schweiz. Augsburg. Reinerth, H., 1939. Ein Dorf der Grossteingraberleute. Germanen-Erbe 4, pp. 226-242. Ristic, M., 1977. Ribe i ribolov u slatkim vodama. Beo­ grad. Roes, A, 1963. Bone and antler objects from the Frisian terp-mounds. Haarlem. Rosenlund, K., 1976. Catalogue of subfossil Danish vertebrates, fishes. K0benhavn. Sarfatij, H., 1978. Archeologische kroniek van Zuid­ Holland over 1977. Holland 10, pp. 297-312. Schwabedissen, H., 1949. Die Bedeutung der Moorar­ chaologie fUr die Urgeschichtsforschung. Offa 8, pp. 46-74. Schwabedissen, H., 1957. Ausgrabung mesolithisch/ neolithischer Wohnplatze im Satruper Moor, Kr. Schleswig. Germania 35, pp. 371-373. Schwabedissen, H., 1957/1958. Die Ausgrabungen im Satruper Moor. Offa 16, pp. 5-28. Sirelius, U.T., 1906. Uber die Sperrfischerei bei den Finnisch-Ugrischen Valkern. Helsingfors. Sirelius, U.T., 1934. Die Volkskultur Pinnlands I Jagd und Pischerei. Berlin/Leipzig. Solberg, 0., 1909. EisenzeitJunde aus Ostfinmarken (= Videnskabs-Selskabets Skrifter II. Hist.-Filos. Klasse 7). Christiania. Sprockhoff, E., 1953. Fischreuse aus dem Schlusbeker Moor. Germania 31, pp. 229-230. Stjernquist, B., T. Nilsson & O. Nybelin, 1953. Some stone age fishing tackle from Scania. Meddelanden fran Lunds Universitets Historiska Museum, pp. 123­ 148. Vogt, E., 1937. Geflechte und Gewebe der Steinzeit. Basel. Welinder, S., 1969. Ett barkf16te fran Bare mosse. . Fornviinnen 64, pp. 37-38. ­ Wheeler, A., 1978a. Why were there no fish remains at Star Carr? Journal of Archaeological Science 5, pp.85-89.

Recent and pre- and protohistoric fishing gear Wheeler, A., 1978b. Problems of identification and interpretation of archaeological fish remains. In: D .R. Brothwell, K.D. Thomas & 1. Clutton-Brock (eds.), Research problems in Zooarchaeology (= Oc­ casional publication). London, pp. 69-75 . Zivkovic, 0.,1956. Ribolov i ribarske sprave na Tamisu. Rad Vojvodjanskih Muzeja 5, pp. 194-204.

S3

8. KEYWORDS The Netherlands, Germany, Denmark, Sweden, Europe, prehistory, protohistory, recent, ethnography, ·archaeology, .zoology, fishing technique, fishing gear, fish-trap, type classification.

THE MOLLUSCS OF THE DWELLING MOUND GOMOLAVA, YUGOSLAVIA

AN ENVIRONMENTAL INVESTIGATION ON AND NEAR GOMOLAVA

Jan Willem Eggink CONTENTS 1. INTRODUCTION 2. THE HISTORY OF MOLLUSC RESEARCH IN ARCHAEOLOGY 3. MATERIAL AND METHODS 4. TABLES AND DIAGRAMS 5. DISCUSSION 6. CONCLUSIONS 7. ACKNOWLEDGEMENTS 8. REFERENCES 9. KEYWORDS

55

56

J.W.EGGINK

1.INTRODUCTION* Gomolava, in Yugoslavia, is a dwelling mound on the left bank of the river Sava near Hrtkovci, a small village about 60 km north­ west from Belgrade (fig. 1). The mound shows eight periods of habitation, the oldest belon­ ging to the Vinca culture that, with the aid of C14, may be fixed at c. 4000-3800 B.C. (Clason, 1977). The most recent traces, inclu­ ding a burial-ground, date from the early Middle Ages (table 1). After trial excavations during the fifties, it was decided in 1970 to excavate the tell, as systematically as possible, in its entirety. The location of Gomolava along the outside bend of the river Sava, has largely contributed to this decision. For, owing to erosion, portions of the tell disappear into the river every year. Its original surface area has no doubt been pigger than what now remains (about 230 x 45 metres) (fig. 2). In the earliest periods of its habitation, the tell may possibly have been situated at some distance from the Sava. It is more than 5 metres high. The excavations that still continue are being carried o~t by the Vojvodjanski Muzej of Novi Sad and the Universities of Belgrade and Novi Sad, under the direction of Dr. B. Brukner, Dr. B. Jovanovic and Dr. N. Tacic. Staff-members of the Biologisch-Archaeolo­ gisch Instituut at Groningen take part in the Gomolava-excavations in pursuance of a co­ operation between this institute and the Voj­ vodjanski Muzej, as provided for in the Yugo­ slav-Dutch Cultural Treaty. The mollusc material, which is the subject matter of this research, was excavated in 1976, 1977 and 1980. Material excavated in 1980 was collected by myself in co-operation with Drs. D.C. Brinkhuizen of the Biologisch-Archae­ ologisch Instituut at Groningen. Object of the present research was to exa­ mine whether conclusions may be drawn from mollusc material found in dwelling mounds. It was mainly focused on a reconstruction of the environment of the tell Gomolava and,

* Tables 4, 5, 7-13 and the Appendix have been reproduced as microfiches (I:AI-B6) in an -eiwelope attached to the rear of this volume.

Fig. I. The geographic location of Gomolava, Yugo­ slavia.

under supervision of Dr. A.T. Clason, carried out as an optional part of a university course in biology, following a staff-membership of the Archaeozoological Department of the Bio­ logisch-Archaeologisch Instituut at Gronin­ gen. 2. THE HISTORY OF MOLLUSC RESEARCH IN ARCHAEOLOGY* As a hunter and a food-gatherer man has always affected his surroundings more or less. But when the domestication of animals began and farms came into being where crops were grown, his influence increased considerably. Forests were cut and the resulting clearings used for various purposes. Investigations of the environment. of such farms is therefore most interesting and data derived from it may afford some insight into the way in which these early dwellers lived and how they used their land. Following the development of pollen ana­ lysis it became possible to recognize the re­

*

In part after Evans (1972: p. 3 ff.).

Molluscs of the dwelling mound Gomolava

57

Fig, 2. The dwelling mound Gomolava. Schematic lay-out.

gional vegetation of certain areas. At first the local structure of the vegetation could only be determined by macroscopic plant fossils. Now, with the use of three-dimensional pollen diagrams, local aspects, such as big clearings in woods, can also be distinguished (Turner, 1975). As pollen grains can not be preserved quite so well in a basic as in an acid environment, one was, when examining settlements in cal­ careous areas, forced to look for other means. Therefore, the generally large numbers of molluscs in these areas seemed, together with the local information they may divulge about their earlier environment, eminently suitable for further study. In the beginning of this century, much research on mollusc material oflate prehistoric periods in the calcareous countryside of South­ east England was done by A.S. Kennard and B.B. Woodward. B.W. Sparks introduced for the first time a classification of mollusc species in ecologi­ cally related groups. He made use of histo­ grams in which, on a similar basis as in pollen analysis, the occurrence of these groups is shown in percentages (Sparks, 1961). He spent

a great part of his investigation on molluscs contained in fresh-water deposits from the Pleistocene, in which period climatic factors may also play an important role in the dif­ fusion of snails (Sparks, 1964). On the basis of the same techniques, M.P. Kerney and J.G. Evans studied Late-Glacial deposits in South-east England (Kerney, 1963). Outside England the investigation of snail material from Pleistocenic deposits was carried out mainly in Central Europe. Calcareous loess deposits in Czechoslovakia were tho­ roughly examined by Lozek. Successive cold and warm periods could be defined by means of the mollusc material found (Lozek, 1964). In a great many places shells from recent archaeological excavations have been collec­ ted, for instance the sediment in ditches, post­ holes, pits, deposits on slopes at the foot of walls and banks, dump-hills, etc. Mollusc investigation of dwelling mounds in order to reconstruct the local environment has not yet taken place (Evans, 1978a). The construction and interpretation of mol­ lusc diagrams is to a great extent hampered by lack of information concerning the exact recent distribution of a great number of mol­

58

J.W.EGGINK

Fig. 3. Gomolava 1977 and 1980.

luscs. Only when more quantitative data about recent snail and mussel associations would become known, more questions about earlier environments may be answered by means of mollusc diagrams.' 3. MATERIAL AND METHODS 3.1. Material The investigation is directed at shells of animals belonging to the class of Gastropoda and the class of Bivalvia, both Mollusca. The material collected is twofold. Firstly the shells sampled by myself in 1980 (3.2.1.1.), and secondly the material from pits excavated in 1976 and 1977 (3.2.1.2.). In addition to this, mollusc material sam­ pled at Gomolava by Dr. B.S. Ottaway of the University of Edinburgh (5.2.3.) will be con­ sidered summarily. 3.2. Methods

3.2.1. Sampling 3.2.1.1. Material sampled in 1980 In August 1980 I visited Gomolava during the excavation campaign. Object of my visit was to take a successive series of samples. About half a meter north-east of a trench, resulting from an earlier excavation in 1977 by Ottaway, and fixed on the tell by the co­ ordinates S44 (Jovanovic, 1965), a column of soil with a surface of 40 x 40 cm and a depth

of c. 5 m was cut away (figs. 3, 4). A possible disturbance within this column, caused by infiltration of recent snail material, may be ignored. The upper layer of the column (c. 10 cm) was removed because of its disturbed character. The column was cut away in layers of 10 cm each, whereby the layer with a depth of o to 10 cm was numbered sample 1, from 10 to 20 cm sample 2 and so on. The samples were sifted in their entirety, so that the quantity of soil sifted amounted to 161itres per sample. Mesh size of the sieves amounted to one and a half and half a millimetre, the latter being the smallest that Kerney (1963), Sparks (1964) and Evans (1972) also used. Only odd-numbered samples were chosen for analysis. Dating of each single layer was done partly by means of pottery that was found, partly by the use of data collected at the investiga­ tions by Ottaway in 1977. Some observations concerning the stratigra­ phy: - At a depth of 3.00 to 3.10 m the so-called 'Eneolithic humus' formed in a period when the mound was uninhabited, was found. In this layer practically no remains of pottery were met with. - The last two samples (47 and 49) contained no pottery fragments.

3.2.1.2. Material from pits, excavated in 1976 and 1977 In the study of molluscs a few pits were included that were found next to the remains of houses dating from the Vinca period. The width of these houses was c. 7 m, and their length sometimes over 20 m. The walls were probably a wattle and daub construction. The floors may also have been made of loam, quarried in the direct vicinity of the house, whereby pits remained. Loam was also remo­ ved from the· banks of the Sava, and possibly from other sources. The pits next to the houses were used for the disposal of garbage: most of the bones from these pits are broken and show carving marks (Clason, 1979). In these garbage pits layers are found consisting of shells of the genus Unio, as well as of Helix pomatia. The measurements of the pits vary. At a width of 1 to 3 m, the length may be between

Molluscs of the dwelling mound Gomolava

59

Fig. 4. The column of soil dug up in 1980.

1 and sometimes more than 7 m. Their depth is never more than two and a half metres. The pits, excavated in 1976 and 1977, were identified by means of letters. In 1976 pit A, B, C, D and H were excavated; in 1977 pit K, L, M, 0, P, Rand X. All were excavated in blocks I-VI (fig. 2). Fully 100 litres of soil from each pit were sifted. How much precisely can not be ascer­ tained. Mesh sizes of the sieves were 3 and one and a half millimetres.

3.2.2. Identification On the identification of subfossil and recent Yugoslav shell material no standard-work is available . A great support at the identification was Lozek's study (1964), in which also the regional spread of molluscs in Czechoslovakia is recorded. Moreover, the following textbooks that are based on recent material, were consulted: Adam (1960), Ehrmann (1937), Ellis (1969),

60

J.w'EGGINK

Table 1. Habitation periods of the dwelling mound Gomolava (after Clason, 1979) Gomolava I Gomolava V Horizon of Bosut group a) Older dwelling horizon of Vinca group (Basarabi complex) b) Younger dwelling horizon of Gomolava VI Vinca group a) Older dwelling horizon of Gomolava II La Tene settlement a) Horizon with Lengyel pottery b) Younger dwelling horizon b) Horizon with prototiszapolgar of La Tene settlement and Tiszapolgar pottery c) La Tene-Early Roman Gomolava III dwelling horizon Gomolava VII a) Horizon of pits of Baden group Dwelling horizon of Roman­ provincial settlement b) Dwelling horizon of Kostolac group Gomolava VIII c) Horizon with Vucedol Horizon of Middle Age pottery settlement and necropolis Gomolava IV a) Horizon Omoljica-Vatin group b) Dwelling horizon of group BelegiS I c) Dwelling horizon of group BelegiS II

Gittenberger et al. (1970), Janssen & de Vogel (1965), Janus (?), Riedel (1980), Kerney & Cameron (1979) and its slightly more extensive translation by Gittenberger (1980). For the identifications use was made of recent mollusc material collected in 1980 in the surroundings of Gomolava (see Appen­ dix). On some identifications Gittenberger, Back­ huys and Ripken were personally consulted. A few notes on some of the identifications: Landsnails. SuccinialOxyloma. Separation of species of Succinia and Oxyloma is, particularly where juvenile material is involved, extremely diffi­ cult. From a few adult specimens it may be inferred that Succinia is, within the scope of this investigation, only represented by Succinia oblonga and S.putris, which are both included in recently collected material (Appendix). Spe­ cimens otOxy/oma couldnotbe identified to within the species. In view of these problems of identification, and of the fact that S.oblonga, S.putris and O. sp come within the same ecological group (3.2.5.), all specimens belonging to these ge­

nera were classified in group SuccinialOxy­ loma. Granaria Jrumentum Jrumentum. Identifica­ tion to within the sub-species Jrumentum was done by Dr. E. Gittenberger. Clausiliidae. To this group belong apices of sinistrally coiled shells, that could not be more closely identified. Limacidae. Small internally asymmetric shells of slugs were, at the investigations in 1980, found on the half millimetre sieve. Nei­ ther in 1980, nor in 1977 were these shells found on the one and a half millimetre sieve. Closer identification to within the species is impos­ sible. Water-molluscs. Viviparus sp. Material of this group consists of apices of young specimens of the genus Viviparus. Most probably these are of the species acerosus (Bourguignat,' 1862). Recently this species has been found along the Sava (see Appendix). Opercula. At both investigations a great number of opercula were found in the samples. These are mainly belonging to Bithynia ten­ taculata and only occasionally to B.leachi.

3.2.3. Counting To count fragments of shells I applied the following rules: Identifiable apertures (f.i. of Chondrula tri­ dens) are added to the total number of intact shells of the species or genus in question. Apices of identifiable shells (f.i. of Clau­ siliidae) are added to the species, genus or family in question. However, if of one group the apices and apertures are singly identifiable, only one of both is counted. Thus, of Chon­ drula tridens and Granaria Jrumentum only the apertures are counted. Like Sparks (1961) and Evans et al. (l978b) I count the valve of a bivalve's shell for one individual. Fragments of shells of Unio and Pisidium may, when more than half of the hinge is discernable, be counted as individuals. At none of the investigations the opercula of Bithynia are included in the calculations, as its counts are open to discussion and might show an over-representation of the species (Sparks, 1964).

Molluscs of the dwelling mound Gomo/ava

61

Table 2. Species and families found at the 1980 and pit investigations. land

water

Pomatias elegans (Muller 1774) Succinea oblonga Draparnaud 1801 Succinea putris (Linnaeus 1758) Oxyloma sp. Cochlicopa lubrica (Muller 1774) Cochlicopa lubricella (Porro 1838) Truncatellina cylindrica (Ferussac 1807) Orcula dolium (Draparnaud 1801) Granaria [rumen tum [rumen tum (Draparnaud 180 I) Pupilla muscorum (Linnaeus 1758) Vallonia excentrica Sterki 1892 Vallonia pulchella (Muller 1774) Chondrula tridens (Muller 1774) Ena montana (Draparnaud 1801) Punctum pygmaeum (Draparnaud 1801) Vitrea contracta (Westerlund 1871) Aegopinella ressmanni (Westerlund 1883) Oxychilus inopinatus (Ulicny 1887) Daudebardia brevipes (Draparnaud 1805) Cochlodina laminata (Montagu 1803) Clausilia pumila Pfeiffer 1828 fam. Clausiliidae Bradybaena [ruticum (Muller 1774) Euomphalia strigella (Draparnaud 1801) Per[oratella rubiginosa (Schmidt 1853) Monacha cartusiana (Muller 1774) Cepaea vindobonensis (Ferussac 1821) Helix pomatia Linnaeus 1758 fam. Helicidae fam. Limacidae Cecilioides acicula (Muller 1774) Cecilioidesjani (De Betta & Martinati 1855)

Theodoxus danubialis (C. Pfeiffer 1828) Theodoxus transversalis (C. Pfeiffer 1828) Viviparus sp. Valvata piscinalis (Muller 1774) Valvata macrostoma March 1864 Valvata cristata Muller 1774 Lithoglyphus naticoides (c. Pfeiffer 1828) Bithynia leachi (Sheppard 1823) Bithynia tentaculata (Linnaeus 1758) Lymnea stagnalis (Linnaeus 1758) Lymnea truncatula (Muller 1774) Planorbis planorbis (Linnaeus 1758) Planorbis corneus (Linnaeus 1758) Anisus leucostomus (Millet 1813) Anisus septemgyratus Rossmassler 1835 Anisus vortex (Linnaeus 1758) Bathyomphalus contortus (Linnaeus 1758) Gyraulus albus (Muller 1774) Armiger crista (Linnaeus 1758) Segmentina nitida (Muller 1774) Fagotia acicularis (Ferussac 1823) Amjimelania holandri a[ra (Ross massIer 1839) Unio crassus crassus Philipsson 1788 Unio pictorum (Linnaeus 1758) Unio tumidus Philipsson 1788 Pisidium sp.

3.2.4. Frequency of land and water-molluscs

When calculating the frequency of land and water-molluscs per layer or per pit no indi­ viduals of the genus Cecilioides were consi­ dered. Cecilioides acicula can burrow to depths of c. 2 m (Evans, 1972: p. 201). It is, like C.jani, as such of minor importance for stratigraphic investigation. In the tell Gomolava C.acicula was found alive to a depth of c. 70 cm. 3.2.5. Ecological groups If all species would be singly represented in

a mollusc diagram, its interpretation would become virtually impossible. In order to get a clearer view a classification in ecological groups is being used. Water-molluscs are grou­ ped according to the system adopted by Sparks (1961); land-molluscs according to Evans (1972: p. 194 ff.). For the present study Lozek's classification (1964; 1965) was used as well. Molluscs that could not be identified to within the species, were not classified in ecological groups (Helicidae, Clausiliidae, Viviparus sp, Lymnea sp, Segmentina sp, Pisidium sp). The

exception is fam. Limacidae which, in accor­ dance with Evans (1972), are classified into the group 'Catholic land species'. The following ecological groups are being distinguished: Land:woodland species,

catholic species,

open country species,

marsh species.

Water:slum species,

catholic species,

ditch species,

moving water species.

When classifying the various species in eco­ logical groups, it is assumed that the present distribution is identical to what it was in the past. However, the interference of man may have caused a change in the distribution of certain species (Sparks, 1964). Concurrently with the classifications men­ tioned above, data on the distribution of molluscs derived from Kerney & Cameron (1979), Riedel (1980), Janssen & de Vogel (1965), Janus (?), Boycott (1936), Gittenberger et al. (1970) and Jaeckel (1954) were used.

62

J.W.EGGINK

Table 3. Classification in ecological groups ofthe species found at the present investigation, as well as of the species of live mollusc material collected round Gomolava in 1980 (see Appendix). land

water

A:woodland species B:catholic species C:open country species D:marsh species

E:slum species F:catholic species G:ditch species H:moving water species

A: woodland species Vitrea contracta Aegopinella ressmanni Daudebardia brevipes Carychium tridentatum (Risso 1826) Ena montana Acanthinula aculeata (Muller 1774) Punctum pygmaeum Perforatella incarnata (Muller 1774) Helix pomatia Bradybaena fruticum Cochlodina laminata Clausilia pumila Orcula dolium

E: slum species Lymnea truncatula Anisus leucostomus

B: catholic species land Euomphalia strigella Cochlicopa lubrica Cochlicopa lubricella fam. Limacidae Cepaea nemoralis (Linnaeus 1758) Cepaea vindobonensis Pomatias elegans C: Open country species Vallonia excentrica Vallonia pulchella Pupilla muscorum Truncatellina cylindrica Monacha cartusiana Chondrula tridens Granaria frumentum frumentum Helicella obvia (Menke 1828) Oxychilus inopinatus D: marsh species Succinea oblonga

Succinea putris

Oxyloma sp.

Vertigo antivertigo (Draparnaud 1801)

Zonitoides nitidus (Muller 1774}

Perforatella rubiginosa

F: catholic species water Armiger crista Bathyomphalus contortus Gyraulus albus Bithynia leachi Bithynia tentaculata G: ditch species Valvata cristata Planorbis planorbis Anisus septemgyratus Anisus vortex Segmentina nitida Planorbis corneus H: moving water species Valvata piscinalis Lithoglyphus naticoides Unio crassus crassus Unio pictorum Unio tumidus Valvata macrostoma Fagotia acicularis Amfamelania holandri afra Theodoxus transversalis Theodoxus danubialis Viviparus acerosus water-molluscs, not classified

Viviparus sp.

Lymnea sp.

Segmentina sp.

Pisidium sp.

land-snails, not classified fam. Clausiliidae fam. Helicidae

Recent distribution data drawn from mol­ lusc material collected in the vicinity of Go­ molava (1980) were used as well (Appendix). The species found at the mollusc investi­ gation (Gomolava 1980 and pit-exploration) are listed in table 2. For the classification in ecological groups of these species and of the species found recently in Gomolava, see table 3. 3.2.6. The mollusc diagram In a mollusc diagram the number of molluscs per ecological group or per family or genus

may be presented in terms of either absolute or relative abundance. Both methods of pre­ sentation have their pros and cons. One disadvantage of the use of the latter is that, in case of an increase in absolute numbers of one species, another species which in absolute numbers remains constant, shows a decrease in percentage on a relative scale. Another disadvantage of this method would be that, when the total numbers are small, the percentages might be statistically unreli­ able. Kerney (1963) gives four reasons why, in his analysis, histograms are best constructed

Molluscs of the dwelling mound Gomolava

in terms of absolute abundance. He presents the results of his investigation at Dover Hill in terms of both absolute and relative abun­ dance. Evans (l978a) points out that particularly the relative differences from period to period are important for the investigation of the earlier environment. When considering the number of molluscs, found in dwelling mounds, the rate of speed at which the level of the mound was raised as a result of man's depositing all sorts of material on it, plays an important role. It will, during periods of intensive habitation, rise faster than when a few people choose it for their dwelling place. The deposits may contain land as well as water-molluscs. Owing to the varying speeds at which the mound rose, the total number of molluscs per sample (and thus per period) may vary to a great extent. To compare the ecological groups as to their abundance layer by layer, it seems to be best to consider their relative numbers. The ab­ solute numbers may vary to an extent that makes them hardly suitable for comparison. As the size of the samples taken at the ex­ ploration of the pits is unknown, the absolute number of molluscs per sample cannot be used for drawing a comparison between the pits. After considering the foregoing arguments, presentation in terms of relative abundance has been chosen for this investigation. The construction of a mollusc diagram is comparable to that of a pollen diagram. In the case of this investigation, the vertical axis represents the depth from the top of the tell (measured after removal of the disturbed top­ layer). At the investigation of the pits, this axis represents the identification of the various pits (figs. 6 and 9). The horizontal aXIS shows the percentage of the ecological groups as well as of not-classified groups. The percentages of the groups of land-molluscs at the left side of the diagram, are calculated as percentages of the total number ofland-molluscs. A similar calculation was separately made in respect of water-molluscs. Alongside the vertical axis of the diagram Gomolava 1980, a sketch of the profile is giv:en, i.e. the north-eastern profile with a width of 40 cm (fig. 3). For Gomolava 1980 dating of the layers is, with an occasional note, given to the right of the land-water frequency.

63

In figure 6, the frequency of the genus Cecilioides is expressed as a percentage of the total number of land-molluscs, contained in a particular sample. 4. TABLES AND DIAGRAMS The results of the investigation of 1980 are to be found in tables 4 to 8 inclusive. The first two tables show the species found and their numbers per sample: table 4 land-mol­ luscs, table 5 water-molluscs. A + sign means that the species in question is represented in a sample. Owing to deficient preservation it is, however, impossible to distinguish indivi­ duals separately. In the calculations a + sign was counted as I individual. Table 6 gives the abundance, absolute and relative, of land and water-molluscs per sample. The uppermost rows indicate the combined results of either of the sieves used. Under it the results are split up in material found on the half millimetre and one and a half millimetre sieves,~n;spec­ tively. The absolute and relative numbers of the various ecological groups are plotted in tables 7 and 8. Tables 9 to 12 inclusive show the results of the pit investigation. The absolute numbers of the collected species are shown in tables 9 and 10. In table 11 the various groups are listed in absolute, in table 12 in relative terms. The mollusc diagrams are plotted in figure 6 (Gomolava, 1980) and figure 9 (the pits). Figure 5 is a graphic presentation of the land­ water frequency per layer, concerning the investigation of 1980. The results are split up in material found on the half millimetre and one and a half millimetre sieves, respectively. The total result of both sieves (one and a half plus half millimetre) is shown in between. 5. DISCUSSION 5.1. General aspects of the mollusc investigation Before considering the tables and diagrams concerning the various investigations, it will be wise first to discuss shortly some general aspects of mollusc investigation. The question will be raised of what may be the cause of

64

1.W.EGGINK

Table 6. Gomolava 1980. The absolute and relative number of land and water-molluscs per sample, specified in totals trapped on both sieves (V2 mm + 1V2 mm) and per sieve Sample number

o

Total

73 120516343834729576432863651 89 139595978374783452234128122

35791113151719212325272931333537394143454749

+ 3353

V2mm+ 1V2mm Water absolute Water % Land absolute Land %

24102121 1614 6 1725 11 57 28 35 35 20203349 19 30 21 30 39 34 66 78 69 39 73 9641 42 22 67 33 23 40 39 21 29 8 1654 100 8080675181707970616634223161

75 374271 356742419212 94 28 5463 71 91 95 95 93 91 73 23 6422 17 7 18 41 33 22 35 94 463729 9 5 5 7 9 27 77

2396 71% 957 29 %

V2mm Total Water absolute Water % Land absolute Land %

4 32224732 73 38 20445525742641 77 0 5 4161613 10 21422 650232830 o 16 18 34 50 18 26 10 3240 2468 88 68 39 4 27 1831 1660281830331924 31347 100 8482665082749068 60 76 32 12 3261

114454574357768432216112104 67333868343732405 199 85 25 5973 8492 96 95 94 92 76 24 4712 7 6 14 36 27 17 27 79 41 27 16 8 4 5 6 8 24 76

2877 2234 78 % 643 22 %

1V2mm Total Water absolute Water % Land absolute Land %

69 88 29 16 11 10 9 9 13 9 7 12 10 10 12 o 19 6 5 5 3 4 4 3 3 5 7 5 7 5 o 2221 31 45 30444423 33 71 58 50 7042 69 692311 6 7 5 5 10 6 2 5 5 3 7 100 7879695570565677 67 29 42 503058

18 17 18 13 9 3 72 53 17 5 8 15 28 47 83

476 162 34% 314 66 %

o o

the differences between the proportions in which the various species originally oc;.curred i.n living natl1fe, and those in which they are found in the sediment. The method of extrac­ ting molluscs from the sediment will be cri­ tically viewed, and the causes of fluctuations in numbers of molluscs from layer to layer examined.

5.1.1. Abundance of molluscs in nature The numbers in which snails can occur in a certain community depends on hereditary fac­ tors (f.i. the reproductive rate), and on en­ vironmental factors. The latter may be regi­ onal, such as climate and lime-content of the soil, or local, such as sloping country, exposure to sun, humidity, shade, etc. These local fac­ tors are generally of importance for archaeolo­ gical mollusc investigation. About the regional factor 'climate' but little can be concluded from mollusc investigation into the past 6000 years (Evans, 1972: p. 17).

5.1.2. Alterations owing to disturbance of the soil After death, snails will initially land on the

25 14 14 4 8 4 4 3 32292975 17 10 10 1 6871 71 25

17 13 76 4 24

15 10 67 5 33

20 14 70 6 30

ground, or stay in the ground when being there already. The proportional abundance of the various species immediately after death, which we equate with that within the living com­ munity may, however, be altered by all kinds of processes. Shells may be displaced horizon­ tally, for instance by moving water, as well as vertically. The latter displacement is of great importance for the investigation of dwelling mounds. The ground may be stirred up by worms and other animals - also man - , whereby recent and older material becomes mixed. Cameron (1978) reports on a recent inve­ stigation in which recent mollusc material, collected on the surface, appeared to be mixed with older shells. Mollusc diagrams plotted on the basis of recent material only on the one hand, and of mixed - recent and older ­ material on the other, showed appreciably different results. From this it may be conclu­ ded that, if recent and older shells cannot be sorted out, which is always the case with archaeological excavations, a correct interpre­ tation can be very difficult to achieve. At pollen analysis burrowing animals pre­ sent great problems as well. Thus burrowing bees impede the interpretation of pollen di­ agrams at the investigation of prehistoric sett­ lements, like Gomolava (Bottema, 1975).

Molluscs of the dwelling mound Gomolava

Fig. 5. The land-water frequencies of Gomolava 1980.

5.1.3. Weathering and fracture of shells Other factors that may alter the proportional abundance of the various species after death are weathering and fracture of shells. Gene­ rally, shells stay best preserved in calcareous regions. Fracture of material may occur in the sediment as well as be a result of the method of extraction of the shells from the sediment. However, whether the results shift in a certain direction as a result of such weathering and fracture. remains questionable. Among land as well as water-snails, some species are more vulnerable than other species, where their shells are concerned. When the original abundance of species is unknown, as always in archaeology, it is difficult to trace whether, and to what extent the results are shifted through weathering and fracture.

5.1.4. Mesh size of the sieves At the extraction of shells from the sediment, the mesh size of the sieves plays an important role. The smallest mesh size used indicates the order of magnitude of the smallest individuals from the material collected. Practically, no ad~lt snails will pass the half millimetre sieve

65

(Kerney & Cameron, 1980), so that a smallest mesh size of half a millimetre seems allright for optimal extraction of the smallest, iden­ tifiable, mollusc material from the sediment. In most cases, juvenile molluscs of less than half a millimetre are very hard to identify. Table 6 shows, that of 3353 identified mol­ luscs, 86 % (2877) were trapped by the half millimetre sieve and. 14 % by the one and a half millimetre sieve. When examining the land-water frequency in these totals, it appears that the percentages of land and water-mol­ luscs, trapped by the half millimetre sieve are 22 % and 78 % respectively; of these trapped by the one and a half millimetre 66 % and 34 %. These results make it clear that, when no smaller sieve than one and a half millimetre mesh size is used as happens at the pit in­ vestigation, a very great number of molluscs is not extracted from the sediment. The mol­ luscs that are washed through such a sieve · show a totally different land-water frequency in comparison 'Yith the material trapped by it. By using a half millimetre sieve next to f.i. a one and a halt millimetre sieve, the total number of extracted shells per volume of sediment increases. The percentages are, as a result of these greater numbers, statistically more reliable. Figure 5 shows that, sample by sample, the changes in land-water frequency of the half millimetre sieve and half millimetre plus one and a half millimetre sieves are practically identical. This similarity can be largely explai­ ned by the fact that the number of molluscs, trapped by the half millimetre sieve, amounts, during the whole investigation, to 86 % of the material collected in total (half millimetre plus one and a half millimetre). The differences per sample between the percentages of land and water-molluscs of the one and a half millimetre sieve and of the half millimetre plus one and a half millimetre sieves are much more pro­ nounced. From these results it may be concluded that percentages of land and water-molluscs per sample may vary with the use of a smallest sieve of half millimetre or one and a half millimetre. In many cases the percentage of shells of water-snails per sample will increase when both types of sieves are used, instead of using only the one and a half millimetre

66

J.W.EGGINK

one. Whether the numbers would show com­ parable differences, if the one or the other sieve were used at excavation on other sites, will only appear after more will have been publis­ hed about the method of extraction at mollusc investigations.

5.1.5. Fluctuations from sample to sample The relative numbers of various species, fa­ milies, ecological groups etc. may fluctuate from sample to sample. The cause could be the different distribution of certain species in time. The proportional abundance of the spe­ cies in the living community is reflected in the proportions immediately after death. These proportions may, as already mentioned, un­ dergo changes as a result of horizontal dis­ placement, stirring up of the soil, weathering and fracture. Import of material from elsewhere (in the present. investigation f.i. material from the surroundings of the tell) may change these proportions as well. To obtain statistically reliable percen~ages, the number of'molluscs extracted per sample should be as great as possible. Evans (1972: p. 83) demonstrates, by means of a test sample of 500 snails comprising 11 species, that a total number of 150 to 200 snails is sufficient to obtain the proportional abundance (5 % more or less) of the species in the original sample. As, at both investigations now under dis­ cussion, the total number of molluscs amoun­ ted often less than 100, it should be realized that some of the fluctuations in the diagrams may be caused by statistically unreliable per­ centages.

5.1.6. The build-up of the dwelling mound The build-up of a dwelling mound takes place through the import by man of material from the countryside to his dwelling place. When habitation is concentrated in the same place for a long time, dwelling mounds may reach a considerable height (sometimes over 15 m; Evans, 1978a). Generally, the material consists mostly of clay from the surroundings, impor­ ted for house-building. In many places of the tell Gomolava, re­ mainders of houses have been found, in most

cases consIsting of burnt clay from houses destroyed by fire. Remainders of houses that were not burned down are much more difficult to mark. In Gomolava, for house-building use was made of clayish soil dug away next to the houses. Accordingly, in the tell a great many pits dating from the Vinca and Baden period were found. It is expected that on the tell Gomolava mainly land-snails lived. Water-snails could only have lived in pits, partly filled with water, and possibly watertanks, basins, etc. As to the analysed layers of Gomolava 1980, no pits were apparent in the profile, so that it may be assumed that all water-molluscs were imported by man. Possible exceptions are Lymnea truncatula or Anisus leucostomus, which can be found in very moist surroun­ dings, such as puddles that are running dry. The imported land- and water-molluscs can be of different origin. Land-snails may have been brought up with building-material, bran­ ches for the construction of walls, grass, herbs and other vegetable food. Water-snails and fresh water mussels may have come along with loam from the river, or may originate from neighbouring marshes or other pools. When water was drawn for f.i. drinking and washing, water-molluscs and possibly land-molluscs, living alongside water ('marsh species') may have been carried off also. Great quantities of Unio sp and Helix po­ matia, which served as food for man, are found in layers in the pits. Pigs also eat them, but then the shells are crunched. As the shells from the pits are hardly ever crushed, they are considered to be human refuse. Probably, the pigs will have rooted for Unio and Helix themselves, just as they do nowadays. 5.2. Mollusc diagrams

5.2.1. Local and regional information When an attempt is made to obtain, by means of mollusc diagrams, data concerning the occurrene of land-snails that have not been imported from elsewhere, the information about the environment, derived from these data, is local. The local character is, however,

Fig. 6. The mollusc diagram of Gomolava 1980.

I::)

is' " ~

~

~

Cl ~

!:l..

~

;::

~

~

~.