Unearthing Prehistory: The Archaeology of Northeastern Luzon, Philippine Islands 9781407300559, 9781407330808

Table of contents :
PRELIMS.pdf
1613 verso.pdf
John and Erica Hedges Ltd.
British Archaeological Reports
Mijares-prelims.pdf
Acknowledgment
Abstract
Table of Content
Introduction 1
References 83
Appendices 95
List of Figures
Figure 1.1 Satellite generated map of the Philippine Islands
Figure 1.2 Philippine climate map (Wernstedt and Spencer 196
Figure 1.3 An Agta hunter with his bows and arrows 6
Figure 8.23 Spindle whorl: [a] from Fengpitou (Chang 1969),
List of Tables
Mijares-Chapters.pdf
Mijares-Appendix A and B.pdf
Mijares-Appendix C.pdf
Plate C1
Plate C2
Plate C3
Plate C4
Plate C5
Plate C6
Plate C7
Plate C8
Plate C9
Mijares-Appendix D-E-F.pdf
Mijares--Appendix G.pdf
Appendix G Ceramic Petrographic Plates
Plate G8. Eme black pottery
Plate G 10. Atulu modern pot
Appendix-H and I.pdf
Front Cover
Title Page
Copyright
Acknowledgment and Abstract
Table of Contents
List of Figures
List of Tables
CHAPTER 1 Introduction
CHAPTER 2 The Archaeology of Southeast Asia during the Late Pleistocene to mid-Holocene periods
CHAPTER 3 The Archaeology of Northeastern Luzon
CHAPTER 4 The Archaeological excavation of the Peñablanca Cave Sites
CHAPTER 5 Understanding site formation in Peñablanca Cave Sites
CHAPTER 6 The excavated bone and plant remains
CHAPTER 7 The Unchanging Flakes?
CHAPTER 8 Earthenware Ceramics and non-lithic Materials: Their Relevance to Interaction
CHAPTER 9 The Prehistory of Northeastern Luzon
References
APPENDICES

Citation preview

BAR S1613 2007  MIJARES   UNEARTHING PREHISTORY

Unearthing Prehistory The Archaeology of Northeastern Luzon, Philippine Islands

Armand Salvador B Mijares

BAR International Series 1613 9 781407 300559

B A R

2007

Unearthing Prehistory

Unearthing Prehistory The Archaeology of Northeastern Luzon, Philippine Islands

Armand Salvador B Mijares

BAR International Series 1613 2007

Published in 2016 by BAR Publishing, Oxford BAR International Series 1613 Unearthing Prehistory © A S B Mijares and the Publisher 2007 The author's moral rights under the 1988 UK Copyright, Designs and Patents Act are hereby expressly asserted. All rights reserved. No part of this work may be copied, reproduced, stored, sold, distributed, scanned, saved in any form of digital format or transmitted in any form digitally, without the written permission of the Publisher. ISBN 9781407300559 paperback ISBN 9781407330808 e-format DOI https://doi.org/10.30861/9781407300559 A catalogue record for this book is available from the British Library

BAR Publishing is the trading name of British Archaeological Reports (Oxford) Ltd. British Archaeological Reports was first incorporated in 1974 to publish the BAR Series, International and British. In 1992 Hadrian Books Ltd became part of the BAR group. This volume was originally published by John and Erica Hedges Ltd. in conjunction with British Archaeological Reports (Oxford) Ltd / Hadrian Books Ltd, the Series principal publisher, in 2007. This present volume is published by BAR Publishing, 2016.

BAR PUBLISHING BAR titles are available from:

E MAIL P HONE F AX

BAR Publishing 122 Banbury Rd, Oxford, OX2 7BP, UK [email protected] +44 (0)1865 310431 +44 (0)1865 316916 www.barpublishing.com

Kay love and understanding, and my wife Aileen May Paguntalan-Mijares inspiring support. I thank them all.

Acknowledgment

Funding for the fieldwork was provided by the Archaeology and Natural History Department, Research School of Pacific and Asian Studies, Australian National University. Funding for the radiocarbon dates was provided by the Centre for Archaeological Research, ANU.

This research has benefited from the support of many people. First of all I would like to express my deepest gratitude to Peter Bellwood, Atholl Anderson, Mathew Spriggs, Glenn Summerhayes, and Johan Kamminga. I am also grateful for the contributions of specialists who conducted different material analyses: Victor Paz and Jane Carlos for conducting the macrobotanical analysis, Jeff Parr for his phytolith analysis, Angel Bautista and Carmencita Mariano for their identification of the faunal remains, Judith Cameron for her analysis of the spindle whorl, and Helen Lewis for her patience in teaching me methods of soil micromorphology analysis.

Abstract Northern Luzon is an important area for understanding and reconstructing the prehistory of the Southeast Asian region. From archaeological work undertaken in the 1970s, we can see the potential of the area in contributing to our understanding of the peopling of the Philippine Islands, from the Pleistocene foragers to the migration of the early Austronesians.

I also would like to extend my gratitude for the support of the National Museum of the Philippines and in particular, Director Corazon Alvina, Wilfredo P. Ronquillo, Eusebio Dizon, Cecilio Salcedo, and Maharlika Cuevas. The National Museum personnel who assisted in the excavations; Clyde Jago-on, Alexandra De Leon, Nida Cuevas, Pablo Pagulayan and Domeng Pagulayan. Eduardo Bersamira and Antonio Peñalosa drew the Scientific Illustrations.

This study attempts to synthesize past and current archaeological research in the area, as well as to present new findings from archaeological excavations in the Peñablanca caves. The excavations of Callao, Dalan Serkot, and Eme caves provide fresh data for reconstructing the transition from the Preceramic into the ceramic periods. Several analytical approaches are used to reconstruct past culture and subsistence strategies. Analysis of cultural materials includes lithic and ceramic analysis. To reconstruct past diets and environments, specialists have conducted a suite of analyses, such as phytolith analysis, macrobotanical analysis and faunal identification. A soil micromorphology analysis has been conducted in order to understand cave depositional histories and estimate degrees of post depositional disturbance.

Governor Edgardo Lara and his staff at the Cagayan Provincial Government, particularly Carol Tobias-Pobre, and the DENRPAMB-Peñablanca under the supervision of Tito Mangantulao facilitated my research in their province. For the Lallo shell midden research area, Hsiao Chun Hung and Ray Santiago allowed me to observe the excavation at Nagsabaran site; Amalia de la Torre helped me with the documents relating to these sites; Hidefume Ogawa allowed me to use his rim form illustrations; and Kazuhiko Tanaka discussed the Cagayan pottery sequence.

The recent excavations in the Peñablanca caves have provided the earliest dated evidence of human occupation in Luzon, at c. 25,000 BP. Evidence from faunal identification, macrobotanical, and phytolith remains show broad-spectrum subsistence strategies employed by these early foragers. The lithic analysis shows some changes from Late Pleistocene into early Holocene technology.

The Archaeological Studies Program, University of the Philippines provided the Lithic Laboratory microscopes used in my analysis. Fernando Siringan of UP NIGS and Rose Berdin of RSES-ANU discussed sea-level change with me. At the Earth Science Department, ANU, Andrew Christy commented on the volcanic ash, and Tony Phimphisane facilitated the production of the thin sections. New Initiatives Grant funding for processing the thin sections was provided by the Centre for Archaeological Research at ANU. Frank Brink and Sally Stowe of the Electron Microscope Unit (EMU) Research School of Biological Sciences, ANU, facilitated and trained me in the use of the Jeol 6400 Scanning Electron Microscope.

Interaction between the foragers of the Peñablanca cave sites and the early Austronesian farmers of the Cagayan Valley was established by at least 3500 years ago. Farmers exchanged earthenware pottery, clay earrings, spindle whorls, and shell beads with foragers, possibly for forest products. This exchange, however, did not present evidence to include cerealbased foods such as rice. The botanical evidence from the cave sites shows a heavy reliance on wild and arboreal food sources. This study, therefore, proposes a general culture history of northern Luzon from the late Pleistocene to the mid-Holocene period.

I am also grateful for the encouragement and support of academic staff and students at ANH-ANU, in particular Janelle Stevenson, Andy Fairbairn, Jean Kennedy, Jack Golson, Geoff Hope, Sue O’Connor, Gill Atkins, Matiu Prebble, Vince Kewibu, and Amanda Kennedy of CAR-ANU I thank my colleagues at Anthropology Watch, Abe Padilla, Miks Guia, Doris Bacalzo, Aileen Paguntalan-Mijares, Bunny Soriano, and Perla Espiel for their support particularly in mapping technology. My colleagues and friends who supported the excavation, Danny Galang, Jun Obille, Arnulfo Dado, Eliza Valtos, Jane Carlos, Jack Medrana, Lee Neri, Michelle Eusebio, and Darly de Leon. Mike Bourke engaged me in discussions in his house on 16 Dianella, O’Connor. Joy Belmonte, my colleague and friend, helped me deal with my visa issues. My parents Angel and Milagros Mijares and others in the Mijares clan provided moral support; my daughters Kim and

i

Table of Content 1

2.

3.

4.

5.

6.

7.

Introduction

1

1.1 The Foraging Era

2

1.2 The Austronesian Debate

6

1.3 The Research Objectives

10

1.4 General Methodology

11

1.5 Significance of the Study

11

1.6 Thesis Outline

11

The archaeology of Southeast Asia during the Late Pleistocene-Holocene Periods

12

2.1 Mainland Southeast Asia

12

2.2 Island Southeast Asia

16

2.3 Discussion

21

The archaeology of north-eastern Luzon

24

3.1 The Geology and Geomorphology of North-eastern Luzon

24

3.2 The Cabalwanian Industry and the Fossil Record

25

3.3 The Flake Tool Industry

26

3.4 The Period of Ceramics and Shell Midden Formation

28

The archaeological excavation of the Peñablanca Cave Sites

33

4.1 Introduction

33

4.2 Methodology

37

4.3 Eme Cave

37

4.4 Callao Cave

37

4.5 Dalan Serkot Cave

38

4.6 Discussion

39

Understanding site formation in Peñablanca Caves

40

5.1 Geomorphology of Caves

40

5.2 The Soil Micromorphology Approach

40

5.3 The Archaeological Stratigraphy

41

5.4 Soil Micromorphology Analysis

42

5.5 Peñablanca Cave Site Formation

46

The excavated bones and plant remains

47

6.1 Faunal identifications

47

6.2 Macrobotanical Remains

48

6.3 Microbotanical Remains

49

6.4 Discussion

53

The Unchanging flakes?

54

7.1 Flake Analysis

55

7.2 Flake analysis in the Philippines

58

7.3 Methodology

59

7.4 Technological Analysis

60

7.5 Usewear Analysis

64

ii

8.

9.

7.6 Discussion

67

Earthenware ceramic and non-lithic materials: their relevance to Interactions

68

8.1 The Peñablanca Pottery Tradition

68

8.2 Methodology

68

8.3 Ceramic Analysis

70

8.4 Other Ceramic Materials

76

8.5 Shell Beads

77

8.6 Discussion

77

The Prehistory of north-eastern Luzon

78

9.1 The Preceramic hunters and gatherers

78

9.2 The coming of the Austronesians

79

References

83

Appendices

95

A. The Excavation Plates

96

B. Descriptions of Soil Micromorphology: Thin-sections

99

C. Soil Micromorphology Plates

103

D. Flora and Fauna Plates

109

E. Flake Analysis Plates

112

F.

123

Descriptions of Petrographic Thin Sections

G. Ceramic Petrographic Plates

127

H. Experts’ Reports

129

I.

138

Glossary of Soil Micromorphology Terms

List of Figures Figure 1.1 Satellite generated map of the Philippine Islands Figure 1.2 Philippine climate map (Wernstedt and Spencer 1967) Figure 1.3 An Agta hunter with his bows and arrows Figure 1.4 The Spread of Austronesian languages (Bellwood 2005) Figure 2.1 Selected archaeological sites in Southeast Asia Figure 2.2 The Hoabinhian lithic industry (Bellwood 1979, after P.I Boriskovskly) Figure 2.3 Maros points (Glover 1977) Figure 2.4 Blade points and Levallois-like flakes from LBS (Glover 1981) Figure 2.5 Cord-marked and comb-incised sherds from Fengpitou (Chang 1969) Figure 3.1 Geological map of Luzon Figure 3.2 Mountain ranges of Luzon Figure 3.3 Satellite Generated map of northeast Luzon (Google Earth 2005) Figure 3.4 Cagayan Valley geomorphological structure (Vondra et al. 1981) Figure 3.5 Major archaeological areas in north-eastern Luzon Figure 3.6 Cabalwanian pebble tools (Fox 1978) Figure 3.7 Minori Cave flake tools (Mijares 2002) Figure 3.8 Neolithic adzes based on Beyer’s (1947) typology Figure 3.9 Major shell midden sites in north-eastern Luzon Figure 3.10 Magapit pottery designs (Photo courtesy of Bellwood and Yoji Aoyogi) Figure 3.11 Dimolit house plan (Peterson 1974) Figure 4.1 Topographic map of the Peñablanca area (NAMRIA Map) Figure 4.2. Locations of the archaeological sites excavated Figure 4.3. Plan of the Eme Cave excavation Figure 4.4 Plan of the Callao Cave excavation

iii

1 2 6 7 12 14 18 18 20 24 24 24 25 26 26 27 28 28 29 31 33 33 36 37

Figure 4.5 Plan of the Dalan Serkot excavation lay out Figure 5.1 Eme Cave stratigraphy: 1 and 2 are oriented samples Figure 5.2 Dalan Serkot stratigraphy: 1 and 2 are oriented samples Figure 5.3 Callao Cave stratigraphy: 1 to 5 are oriented samples Figure 6.1 Identified faunal remains Figure 6.2 Comparative volume of shells in 4m2square in Peñablanca cave sites Figure 6. 3 SEM of Eme parenchyma tissues (courtesy of Paz and Carlos) Figure 6. 4 Comparison of Peñablanca parenchyma tissues with Colocasia cell size Figure 6.5 Comparison of Peñablanca parenchyma tissues with Dioscorea Alata cell size Figure 6.6 Comparison of Peñablanca parenchyma tissues with I. Batatas cell size Figure 6.7 Comparison of Peñablanca parenchyma tissues with Manihot cell size Figure 6 8 SEM of Callao charred wood (Courtesy of Paz and Carlos) Figure 6. 9 SEM Dalan Serkot [a] parenchymatous tissue and [b] charred wood (Paz and Carlos) Figure 6.10 Possible Moraceae phytolith (Courtesy of Jeff Parr) Figure 6. 11 Poaceae phytoliths: [a] block type, [b] long type and [c] prickle type (Jeff Parr) Figure 6.12 Poaceae phytoliths: [a] bilobate type, [b] blocky and articulated, [c] bulliform, and [d] saddle type (Courtesy of Jeff Parr) Figure 6.13 [a] Bambusoid and [b] Cyperaceae (Courtesy of Jeff Parr) Figure 6.14 Phytolith identical to Metroxylon sagu (Courtesy of Jeff Parr) Figure 6.15 Distribution of starch-bearing palm genera for sago making (from Ruddle et a.l 1978) Figure 6.16 Phytolith resembling those of Cocos nucifera L (Courtesy of Jeff Parr) Figure 6.17 [a] Cyperaceae and [b] Poaceae phytoliths (Courtesy of Jeff Parr) Figure 6.18 [a] S.brachycladum, [b] Bambusoid, [c] Myrtaceae [c] starch 6µm – 9µm [d] starch polarized light (from Jeff Parr) Figure 7.1 Flake terminations; [a] feather, [b] hinge, [c] step, and [d] plunging (Cotterel and Kamminga 1979) Figure 7.2 Olympus BXFM (left) and Olympus SZX-9 (right) microscopes Figure 7.3 Comparative mean values of flake measurements Figure 7.4 Eme ceramic layer flakes Figure 7.5 Eme preceramic layer flakes Figure 7.6 Callao ceramic layer flakes Figure 7.7 Callao preceramic layer flakes Figure 7.8 Dalan Serkot ceramic layer flakes Figure 7.8 Dalan Serkot preceramic layer flakes Figure 7.10 Comparative materials used in making flakes Figure 7.11 Percentage of surface cortex observed Figure 7.12 Types of flake terminations Figure 7.13 Types of working edges Figure 7.14 Types of edge damage scar terminations Figure 7.15 Comparative flake edge rounding Figure 7.16 Types of polish on internal flake surfaces Figure 7.17 Types of polish on external flake surfaces Figure 7.18 Directions of striations Figure 7.19 Comparative hand kinematics Figure 7.20 Contact materials Figure 8.1 Location map of pottery sites in Cagayan Valley Figure 8.2 Red-slipped pottery rim forms from Irigayen (Ogawa 2002a) Figure 8.3 Black pottery forms from Conciso (Ogawa 2002b) Figure 8.4 Callao Cave red-slipped pottery: 1) incised and punctuate design, 2) part of ring foot with perforations, 3) carinated lid, 4) part of a ring foot Figure 8.5 Decorated sherds from Dalan Serkot: 1 and 2 small incised design along the carination, 2 and 4 grass impressed design, 3 notch design, `and 5 incised design below the neck

iv

38 43 43 43 48 48 49 50 50 50 50 49 49 51 52 52 52 52 52 53 53 53 64 56 60 61 61 61 62 62 62 63 63 64 64 65 65 65 65 66 66 67 68 69 69 70 70

Figure 8.6 Dalan Serkot black pottery rims with notched design Figure 8.7 Body sherd colours by site and weight (grams) Figure 8.8 Comparative surface finishing of rim sherds per vessel type Figure 8.9 Mean rim diameter by site and vessel type Figure 8.10 Comparative mean diameter per surface colour Figure 8.11 Mean rim sherd thicknesses, by site per vessel type Figure 8.12 Eme black pottery Figure 8.13 Eme brown pottery Figure 8.14 Callao red-slipped pottery Figure 8.15 Callao black pottery Figure 8.16 Callao brown pottery Figure 8.17 Dalan Serkot red-slipped pottery Figure 8.18 Dalan Serkot black pottery rims Figure 8.19 Dalan Serkot brown pottery Figure 8.20 Dalan Serkot large brown rims Figure 8.21 Comparative size frequencies for red-slipped sherd inclusions Figure 8.22 Comparative size frequencies for black pottery sherd inclusions Figure 8.23 Spindle whorl: [a] from Fengpitou (Chang 1969), [b] from Callao Cave Figure 8.44 Earthenware ling ling o earrings: [a] and [b] from Nagsabaran Site, [c] and [d] from Callao Cave List of Tables Table 3.1 Selected radiocarbon dates from archaeological sites in the Cagayan Valley Table 4.1 Radiocarbon dates from the Peñablanca cave excavation Table 4.2 Summary of materials found in past archaeological excavation in Peñablanca cave sites Table 4.3 Radiocarbon Dates from the 2003 excavations Table 5.1 Summaries of Soil Micromorphology Features Table 7.1 Number of flakes measured Table 7.2 Comparative percentage of blade-like flakes among collected flakes per site and period Table 8.1 Descriptions of ceramic matrix fabric Table 9.1 Major features of each excavated site during the preceramic period Table 9.2 Major features of each excavated site during the ceramic period Table 9.3 A summary of northern Luzon culture history

v

70 71 71 71 72 72 72 72 73 73 73 73 74 74 74 76 76 76 77

30 35 35 39 44 60 61 75 81 81 82

vi

CHAPTER 1 Introduction The Philippine Islands lie between 5° and 20° latitude in the northern hemisphere (Figure 1.1). The Philippine Sea to the east, the China Sea to the west, and the Celebes Sea to the south surround the islands. Except for Palawan and parts of Mindoro and Zamboanga Peninsula, the Philippine islands are a product of Tertiary period tectonic events forming volcanic island arcs (Corby 1951:79-80). Luzon, in particular, was formed during the subduction of the Philippine plate under the Asian Continental plate (Mathisen and Vondra 1983). “Active vulcanism, strong seismic activity, and considerable isostatic imbalance characterize the island arc system of the western Pacific of which the Philippines is a part” (Wernstedt and Spencer 1967:9).

mainland. The months of April until early June are the hotter season (Wernstedt and Spencer 1967).

The geomorphological formations of the Philippines consist primarily of mountain ranges with accompanying hills, plateaus, basins, and riverine flood plains. In Luzon, some of the major rivers are the Cagayan and Abra Rivers in the north, and the Agno, Tarlac, and Pampanga Rivers in the south. The Agusan River, the longest river in the Philippines, is located in Mindanao. The other major rivers in Mindanao are the Pulangi, Rio Grande de Mindanao, Maridagao, and Tagoloan. There are also a few lakes and swamps, such Candaba Swamp in Central Luzon, and the swamps associated with the Agusan and Pulangi rivers in Mindanao. The climate of the Philippines is fundamentally maritime and tropical with high humidity and heat (Figure 1.2). But there are regional variations, with areas in the highlands having relatively cooler temperatures. There are four conditions that affect regional climate variability: 1. The peripheral position off the Asiatic mainland places the Philippines within range of continental air masses, storms, and atmospheric pressure conditions. 2. The archipelagic and oceanic nature of the Philippines insures a considerable maritime modification of the elements of land climate. 3. The Philippines lies astride the main tracks of the westward-moving violent tropical cyclonic disturbances (typhoons and hurricanes) that affect the western Pacific, and is affected by these. 4. Most of the larger islands have relatively high mountains areas that themselves play extremely important roles in climatic modification by producing lower temperature regimes, blocking air masses, causing rain shadows, and forcing orographic lifting and consequent adiabatic cooling (Wernstedt and Spencer 1967: 40).

Figure 1.1 Satellite generated map of the Philippine Islands (Google Earth 2005) The amount of precipitation also varies in the different part of the archipelago, with the island of Mindanao receiving moderate year-round rains. Sheltered areas such as the Cagayan Valley, the Central Plain of Luzon, and the Ilocos coast in Luzon, and some central Visayan islands receive light precipitation (Wernstedt and Spencer 1967). Topography and amount of precipitation affect the type of vegetation. Wernstedt and Spencer (1967:83) believe that dense tropical forests covered the Philippine islands before human arrival. The current vegetation includes humid rainforest, tropical grassland (savanna), and open pine forest in areas with elevation above 1300 to 2000 meters (Figure 1.2). The native forest is composed primarily of Dipterocarp trees, such as lauan (S. negrensis and Pentacme contorta), yakal (Shorea spp) and narra (P. indicus). Philippine wild grasses include cogon (Imperata cylindrica and I. exaltata), and sedges. Wernstedt and Spencer (1967) opine that large grassland areas were due to repeated burning after the introduction of farming. Most of the wild animals of the islands are believed to have migrated from Borneo into Palawan and Mindanao, and the species diversity decreases going north to Luzon. “It would appear that the greater part of the land-animal and nonmigratory bird populations reached the Philippines via Borneo only during Pleistocene time, and for many of the forms, their recent distributions represent the early patterns of natural colonization” (Wernstedt and Spencer 1967:106). According to Groves (1985:46), Luzon and Mindanao have many endemic species. However the non-endemic mammal species are shared with Sundaland. The possibility of a land bridge connecting the main body of the Philippines to the

Precipitation in the Philippines is mainly due to tropical cyclones, or the presence of the Intertropical Convergence Zone (ITC). The ITC is the zone where the southern and northern hemisphere air masses converge, rise and move from a position south of the equator in January towards Taiwan in July. The Island of Mindanao in the south receives precipitation in January, but Luzon in the north is more affected in July, which is also the typhoon season. From December until early March, the northeast monsoon affects the islands, bringing cool and dry air from the Asian 1

1970). Recent direct dating of the skullcap fossil using uranium series provides a revised age determination of 16,500 +/-2,000 BP (Dizon 2003:65). The Tabon Cave lithic assemblage is primarily flake tools made from chert. Fox (1978:64) observed that there was no basic change in the method of manufacture of the flaked artefacts during the entire inhabitation of Tabon Cave. Recent re-excavation of Tabon Cave (Detroit et al. 2004, Dizon 2003, Dizon et al. 2002) has resulted in the recovery of twelve new human bone fragments. A tibia fragment has been dated to 47,000 +/- 11-10,000 BP using uranium-series analysis (Dizon 2003:65). This joint Franco-Filipino endeavour has also resulted in further physical analysis of the human remains. The initial assessment of metrical dimensions could suggest a presence of two distinct groups of people: The associations of specimens exhibiting respectively large and small dimensions points to the presence of at least two distinct Homo sapiens morphologies in the Tabon Cave during the Upper Pleistocene times. It could be the result of a taphonomic association of two culturally distinct groups of Homo sapiens (biologically and/or chronologically) or only reflect the presence of a single group characterized by high sexual dimorphism (Detroit et al.2004:711)

Asian land mass has been refuted by Heaney (1985:141). He, however, did not refute the connection of the Palawan chain to Borneo and the Asian landmass during the middle Pleistocene. The land animals include deer (Cervus sp.), pig (Sus philippinensis), palm civets (Paradoxurus spp), cats (Felix spp), and monkeys (Macaca irus). There were also large placental mammals that are now extinct, such as Elephas, Stegodon, and Rhinoceros sp. found in Luzon (Koenigswald 1956, de Vos and Bautista 2003). Crocodiles are thought to have been widely distributed in the past, but now have a restricted area. The tamaraw (Bubalus mindorensis), a wild water buffalo, is restricted to Mindoro Island. Squirrels, tree shrews, and lemurs are found in some Visayas islands and Mindanao.

Palawan Island is part of the former Sundaland continent, which is now partly submerged as a result of sea level rise after the last glaciation. While some archaeologists up to the 1980s believed that there was a land bridge from China through Taiwan into Luzon (Beyer 1948, Fox 1971, 1978, 1979), the faunal and geological evidence indicates that no such land bridge existed (Bellwood et al. 2003). Modern humans might have used sea craft to cross the narrow channels from Sundaland to Palawan. Initial arrival from Taiwan in an earlier era is far less likely as Bellwood and Glover suggested: Dates for initial human occupation in the Philippines, a process that almost certainly occurred from Borneo rather than from Taiwan, fall generally within the past 30,000 years and the former presence here of Homo erectus seems very unlikely (Bellwood and Glover 2004:14) According to Atholl Anderson (2000) building a raft requires simple technology. “ A rough-edge stone knife and some ability in using stripped bamboo as cordage are the only technological requirements for building a seaworthy and durable craft capable of carrying a group of people and their baggage” (Anderson 2000:15).

Figure 1.2 Philippine climate Map (Wernstedt and Spencer 1967)

The Foraging Era The peopling of the Philippine Islands during the mid- to late Pleistocene period is still largely untraced. There have been earlier claims of precursors of modern humans living in Luzon during the mid-Pleistocene (Fox 1978, 1979), and these early hominids were allegedly responsible for the extinction of megafauna (Fox and Peralta 1974:101). But the associations of pebble and flake tools with fossil fauna are still an issue of debate. With no human fossil remains recovered for this period, these claims of very early settlement are at best speculative.

Archaeologists mostly agree that the ancestors of the Negritos were the initial human population in the region. The Negritos include the Agta, Aeta, Ati, Mamanwa, and Batak who currently inhabit forestlands in Northern Luzon, Palawan, and Mindanao in the Philippines, plus the phenotypically related Semang of Peninsular Malaysia and the Andaman Islanders (Bellwood 1979, 1992, 1997, Beyer 1948, Omoto 1981, 1985, Solheim 1981). According to Bellwood (1992:74) “the Negritos are thus of great significance in Southeast Asia: they seem to represent the modern members of an AustraloMelanesian population which may once have occupied much of the region…” Omoto’s (1985:129) genetic study of the Negritos led him to postulate that the Negritos “evolved in the upper Pleistocene times, probably during the 20,000 to 30,000

Tabon Cave on Lipuun Point, Palawan, excavated by Robert Fox (1970:24) in the early 1960s, has yielded the oldest human remains found in the Philippines so far. Initially, the Tabon mandibles and skullcap were dated to 23,200 ±1000 BP (UCLA 699) on charcoal from associated strata (Fox 2

Previous Models Concerning Ancient Philippine Subsistence

years ago, in the tropical rain-forest of Sundaland and developed phenotypic peculiarities through genetic adaptation.” Fix (2002) is cautious about inferring ancestry of present Negritos from genetic or morphological analysis and suggests a direct skeletal analysis. “In particular, the theory that the entire region was originally occupied by peoples of Negrito physical type could be tested with skeletal evidence (insofar as “Negrito” traits can be identified skeletally)”(Fix 2002:194). Unfortunately, there are too few skeletal remains that date back to 20,000-30,000 years ago from the Philippine region to shed much light on this issue.

In reconstructing Paleolithic subsistence economy in the Philippines, a number of models have been proposed. For Northern Luzon in particular, two previous models were presented in PhD theses by Peterson (1974), and Thiel (1980). Peterson, working with a system model in which the interrelationships between humans and their environments changed through time hypothesized that there was an increase in population at the end of Pleistocene, resulting in an overkill of large mammals (Peterson 1974:186). The extinction of the megafauna led to an expansion into marginal areas and an adoption of broad-spectrum subsistence. These stresses in food procurement were aggravated by the postglacial rise in sea level, submerging vast tracts of optimal resource areas. Peterson believed that: Northern Luzon has produced evidence of a narrow spectrum Pleistocene hunting and collecting subsistence pattern, a broadspectrum post-Pleistocene hunting and collecting subsistence mode, and a postPleistocene subsistence style based on seasonal reaping of wild or early domesticated species of grain plants (Peterson 1974:191).

Subsistence Strategies Understanding and reconstructing the subsistence strategies of late Pleistocene hunter-gatherers in the Philippines has been difficult due to poor preservation conditions. Most sites are caves, more often than not merely frequentation sites and thus selective in their archaeological records. Archaeological data from most open sites are surface collections of lithic implements (Fox and Peralta 1974, Junker 2002a, Pawlik and Ronquillo 2003). Kuhn and Stiner (2001) suggested that modern hunter-gatherer subsistence strategies were already in place globally by the Early Upper Paleolithic (EUP), stating that “modern hunter-gatherer adaptations in diet and technology were globally established by the LUP (20,000 BP), perhaps even by the EUP (c. 45,000 BP)” (Kuhn and Stiner 2001:127).

He related these changes in subsistence strategy to supposed changes in tool assemblages: 1) The early development of the chopper/chopping tool kit reflected the hunting and butchering of large game animals. 2) The Late Pleistocene and post-Pleistocene sites with utilized flake tools reflected broad-spectrum exploitation of small game animals, with no large game animals utilized (ibid. 195).

Hunter-gatherer, or foraging subsistence, is “derived from non-domesticated resources, species not actively managed by themselves or by other human beings” (Winterhalder 2001:12). Winterhalder (2001:12) further states “foragers are those peoples who gain their livelihood fully or predominantly by some combination of gathering, collecting, hunting, fishing, trapping or scavenging the resources available in the plant and animal communities around them”. Anthropologists and archaeologists largely rely on extant hunter-gatherer group behavior in order to reconstruct Palaeolithic subsistence strategies. Though we can learn from ethnographic analogy, caution must be used in interpreting recent and modern behavior as representative of behavior in the past (Morrison 2002).

Peterson assumed that there was a correlation between the size of a tool kit and the size of the fauna hunted: small flake tools for small animals, and chopper/chopping tools for megafauna. I do not agree with this correlation for a number of reasons. As was stated above, the associations of the pebble tools with extinct megafauna are questionable since both are found in redeposited terrace sediments. Second, chopper/chopping tools can also be used for disarticulating medium to small mammals as well cutting wood or bamboo.

There is also a dispute concerning the ability of tropical forests to support hunter-gatherer groups over long periods of time. Tropical forest may have a high biomass that can support animals (especially arboreal mammals), but it has few plant foods easily available for human consumption (Griffin 1984). Even the availability of wild yams or other wild plants that humans can collect and consume may not be adequate to support a large group of foragers (Headland 1986). Junker (2002b: 209-210) infers that: Tropical forest hunter and gatherers, faced with an environment with high species diversity, but low biomass and patchy (and somewhat unpredictable) resources, generally have a high level of residential mobility (i.e., frequent moves), depend very little on the long-term storage of resources, and have few resources that can be targeted for intensive exploitation through logistical foraging.

Working further within a similar model, Barbara Thiel (1980) tried to reconstruct the subsistence strategy of early hunters and gatherers and the changes toward agriculture. Like Peterson (1974), Thiel viewed population growth as a major factor for change in subsistence strategy. “The model states that there was population pressure, which was brought about by environmental and cultural factors, and that this produced a stress situation in various groups of people” (Thiel 1980: 38). These pressures would have resulted either in the exploitation of new resources and a broader-spectrum subsistence pattern or, alternatively, a concentration on a particular resource resulting in a narrow-spectrum subsistence pattern. Thiel hypothesized that a generalized hunter-gatherer or a broad-spectrum subsistence pattern was in place during the late Pleistocene (c. 20,000 BP). A change followed after 18,000 BP toward either unspecialized hunter-gatherer that was more area- or locality restricted, or specialized huntergatherer that was more resource-specific. Hunter-gatherer, according to Thiel (1980: 161), was augmented by cultivation 3

of plants and horticulture by 16,000 to 12,000 BP, and subsequently augmented by agriculture by 8000 BP.

It is suggested that, while the upland lithic assemblages may reflect the production of perishable extractive tools used in daily hunting and collecting pursuits, the formalized tools and retouched components of the lowland lithic assemblage may represent more specialized emphasis on the production activities associated with seasonal exchange relations with the lowland agriculturists (Junker 2002a: 237).

Thiel, in her model, assumed that hunter-gatherer groups in Southeast Asia in general, and northern Luzon in particular, independently adopted domestication of plants as an additional resource. The shift, according to her, was due to population and resource pressure. Unfortunately, there are no archaeological data concerning plant domestication in the Philippines during the late Pleistocene or even the early Holocene (Bellwood and Glover 2005:5). Current evidence suggests that migrating early farmers (see section on Austronesian debate) brought domesticated plants and animals into the Philippines. There are only a few places in the world that saw purely indigenous local evolution from foraging to agricultural subsistence (Bellwood 1995).

These lowland sites, according to Junker, were used in processing meat and hides, or manufacturing items such as bows and arrows or fishing gear. These processed or manufactured items were exchanged with nearby agriculturist settlements. Though this hypothesis is plausible, it is also possible that the upland sites were used as lithic workshops and the tools produced were used in the lowland sites. As Junker described, the upland lithic assemblages are composed of cores and unutilized flakes. If the upland sites were used as sites to produced perishable tools, then use-wear analysis of the flakes would indicate this.

Headland (1986), working among the Agta of northeastern Luzon, discussed why these foraging groups never became farmers. He also debunked an earlier isolationist model that viewed Negritos as pure hunters and gatherers up to the time of Spanish conquest, living in relative isolation even after the arrival of Austronesian-speaking lowland farming populations (Headland 1986, Headland and Reid 1989). In his Interaction Model, Headland posited that the Negritos had already developed mutualistic relations with the lowland farmers by at least 2000 years ago. He further proposed that “the Agta were not living as independent Pleistocene-type hunter and gatherers during late prehistoric times” (Headland 1986:178) and they were familiar with plant cultivation. Although they have known plant cultivation and could probably have assisted lowland farmers in agricultural work, they have remained largely hunters and gatherers to the present day. “The archaeological record, then, suggests that rice-farming populations and Negrito hunters were living within a day’s walk of each other in northeastern Luzon for at least the last 3000 years” (Headland and Reid 1989:46).

Junker also identified a number of problems for huntergatherer archaeological research in the Philippines. These include a widespread presence of chronologically undiagnostic lithic assemblages, the ephemeral nature of settlements, and the inferred domination of technology by perishable materials (Junker 2002a: 238). Although there have been a number of Paleolithic and Neolithic sites excavated in the Philippines, especially in Palawan and Northern Luzon, the transition between the two cultural periods is still not fully understood. These problems are real but can still be addressed. There are a number of well-stratified cave sites that could reveal chronological relations between different sets of archaeological materials. There are a number of approaches that an archaeologist can now use to reconstruct prehistoric environments and human diet, such as macrobotanical and phytolith analysis, and isotopic analysis of human teeth. These approaches will be used in this study.

Working in the Central Philippines on the issue of preHispanic interaction between hunter-gatherers and agriculturists, Laura Junker agrees with the Headland interaction model: …many Southeast Asian foragers incorporated trade interactions with adjacent agriculture-oriented populations, and the specialized production activities associated with these symbiotic exchanges, into their diverse economic “repertoires” at a very early date (Junker 2002b:161)

Foraging in a Tropical Forest Environment A crucial debate in reconstructing prehistoric hunter and gatherer subsistence strategy concerns the issue of whether inland tropical forest resources are sufficient to sustain human subsistence. Headland (1986, Headland and Reid 1989) and later Bailey et al (1989:73) questioned the possibility of hunters and gatherers subsisting in tropical rain forest without support from agriculturists. They noted problems with continued access to carbohydrate resources, which could be patchy, scarce, and seasonal, and wild forest animals that are lacking in fat (Bailey et al. 1989, Bailey and Headland 1991, Headland 1986, Headland and Bailey 1991, Headland and Reid 1989). However, there are ethnographic accounts pointing to the occurrence of seasonal hunting of animals such as wild pigs that are rich in fat (Griffin 1984, 1985, Griffin and Estioko-Griffin 1978, Mudar 1985).

In her Tanjay, Negros Island, research, Junker (Junker 2002a) used settlement pattern analysis to reconstruct aspects of hunter-gatherer mobility. She distinguishes between upland and lowland sites in terms of lithic assemblages, the upland tending to have denser archaeological deposits than the latter: Upland sites primarily yielded cores and small, unretouched, unutilized flakes, while lowland sites contained a significant proportion of utilized flakes, as well as large flakes and pebbles that had been retouched into morphologically distinct “tools” such as scrapers, burins, notched tools, and knives (Junker 2002a: 235).

Headland reasoned that the availability of wild yams in the rainforest is not sufficient to sustain the carbohydrate needs of hunter and gatherers. He (1986:179) argued “such wild foods are so scarce in rainforest environments that they could not support human foraging populations unless they supplemented their diet by part time cultivation and/or trade with neighboring farmers.” Bailey et al. (1989:61) in addition

Junker viewed these assemblage differences as signifying different settlement components in a pattern of seasonal mobility: 4

to raising the scarcity of carbohydrates also pointed to the protein and fat requirements of a hunting group. The rainforest animals they hunted are distributed in patchy areas and are lean and lacking in calorie-rich fat. Both Headland and Bailey (1991) think that extant hunters and gatherers survive in interior rainforest because of resources derived from agriculturists. They also sometimes practice a limited form of farming, and now inhabit forest that has been substantially modified by clearing, burning, and swidden farming to allow cultigens or exotic plants to grow.

The debate on the probability of foragers living independently in rainforests continues. We must also remember that environments have radically changed since the Last Glacial Maximum. Rainforests were much less extensive then and foragers probably more widespread. Bailey and Headland (1991:297) are correct to state that: Archaeology promises the definitive test of the hypothesis if it can uncover a human site for which there is unequivocal evidence that it was occupied at a time prior to plant and animal domestication, and if that site was in tropical rainforest at the time of occupation or use.

In a special edition of Human Ecology (1991), a number of arguments were raised against Headland and Bailey’s hypothesis. Brosius (1991) and Endicott and Bellwood (1991) argue for independent inland Hoabinhian and ethnographic foragers in Peninsular Malaysia. Brosius contends that though foragers in tropical rainforests exploit diverse resources, some existing foragers focus on a limited range of plants and animals, to the extent that they are managing these resources. An example of such management of forest resources is the utilization of sago (Eugeissona utilis) by Penan foragers in Borneo (Brosius 1991:142). Brosius also raises the important point that extant foragers now live in interior areas, though before they might have been living in the alluvial and coastal areas now inhabited by agriculturists. These alluvial plains are rich in resources, including fish and shellfish. Therefore, current observations of extant interior foragers might not be sufficient to reconstruct past subsistence strategies in full.

Ethnoarchaeology of Hunter-gatherers: The Agta As earlier stated, the extant hunters and gatherers of the Sierra Madre are believed to be hybridized descendants of Philippine foragers of the late Pleistocene. The Agta groups of the Sierra Madre have been studied by anthropologists and archaeologists wishing to gain insights into subsistence strategies in tropical rainforest environments (Griffin 1984, 1985, Griffin and Estioko-Griffin 1978, Headland 1986, Headland and Reid 1989, Peterson 1978, Peterson and Peterson 1977, Peterson 1974). This section will discuss briefly the ethnographic accounts of these Agta groups. If the assumption of descent from the late prehistoric forager populations of Luzon is correct, then the Agta lifestyle at contact with the ethnographic record should give us some insights into subsistence strategies and group interactions, relevant for interpreting the later prehistoric record of the Peñablanca Caves.

The Batek of Peninsular Malaysia were also observed to have survived with minimal or no support from agriculturist neighbors (Endicott and Bellwood 1991). Endicott and Bellwood (1991:154) observe that Batek have access to wild yams (Dioscorea sp.), honey, game, and fish throughout the year. They also challenge the view of Headland that yams are always scarce and seasonal. Dioscorea are available every month, although the quantities might vary. They also cite the archaeological record that spans up to 10,000 years with evidence of pre-agricultural occupation in the interior Malay Peninsula.

The Sierra Madre Agta are primarily a hunter-gatherer population but have incorporated, in different degrees, some forms of horticulture. Different Agta groups practice a combination of this dual-subsistence strategy based on the locations of their base camps. Griffin (1985:350) categorized them as semi-settled and unsettled. Agta who are semi-settled and reside near lowland farmers tend to do more horticulture and may work as farm labourers. These Agta have small swidden farms on the hillslopes, planted with corn, root crops, and upland rice. Griffin (1985:350) proposes that this knowledge of horticulture among the Agtas could have some antiquity. Agta males are also normally engaged in special relationships, each with a non-Agta Filipino farmer (Peterson 1978, Peterson and Peterson 1977). This relationship usually involves the Agta hunting wild animals so that they can exchange wild meat for agricultural products. The non-Agta Filipino farmer partners may also ask the Agta to help with their farming activities.

Latinis (2000) and Kealhofer (2003) propose that subsistence in lowland rainforest must have been based on an arboreal subsistence economy: An arboreal-based subsistence economy is a subsistence economy whose practitioners meet a majority of their dietary, nutritional, and economic needs through the exploitation of arboreal resources including resources other than trees that are located in or proximate to a forest environment (Latinis 2000:43).

Unsettled Agtas live in more remote areas and have minimal contacts with non-Agta Filipino farmers. The Ebukid Agta, for example, live in the forested interior of the Sierra Madre and maintain a traditional hunter-gatherer lifestyle. They are not involved in any form of cultivation or trade (Peterson and Peterson 1977:553).

Arboreal-based economies would require some management of activities that might include clearing and low-level burning to enhance edible plant densities. Forest plants such as sago, root crops, and nuts could have been easily managed.

Agta hunt primarily wild pig (Sus philippinensis) and deer (Cervus philippinus), but may supplement these species with monkeys, civet cats, and snakes. Fishing and shellfish collection are carried out along the rivers. The bow-and-arrow is traditional hunting equipment, and different types of arrows based on size, weight, and types of head are used for different types of prey (Figure 1.3). A heavy arrow is used for hunting pig and deer, and a lighter arrow for monkeys (Griffin and

The Malay Peninsula rainforest witnessed disturbance and burning during the early Holocene (Kealhofer 2003) that may relate to increased forest management. The change in mainland Southeast Asia from flake tools during the late Pleistocene to pebble tools during the early Holocene might actually indicate adaptation to rainforest conditions.

5

Estioko-Griffin 1978:39). The Agta also use animal traps infrequently.

such interaction would have ranged from initially intimate contact, through full bilingualism, to eventual shift.

Ethnographic accounts of Agta hunting practice show their deep knowledge of their environment (Griffin 1984, 1985, Griffin and Estioko-Griffin 1978, Mudar 1985). Hunting pigs, though physically demanding, is best done during the rainy season, when the animals are fatter due to the availability of better forage. Mudar (1985:80) observed that a greater number of pigs were killed during the rainy season. Agta hunting skills relate to their knowledge of terrain and animal feeding behavior (Allen 1985:54). They recognize four vegetation zones, including coastal strandline, coastal forest, lowland forest, and montane forest. They hunt pigs in the montane forest during the rainy season and in the lowland forest during the dry season. Allen (1985:60) also noted the extensive Agta knowledge of the plants that animals prefer as forage. For instance, Agta would hide and wait when hunting near fruiting trees that normally attract feeding pigs.

The Austronesian Debate H. Otley Beyer (1948) hypothesized that a Negrito population initially peopled the Philippines, which was followed three successive groups, Indonesian A, Indonesian B, and Malay. This multiple-wave migration theory is no longer widely accepted, and this section will review the different current theories of Austronesian migration into the Philippines, with their associated archaeological assemblages, including agriculture.

The Meta Narrative The spread of agricultural subsistence and associated material culture through Southeast Asia has been a source of contentious debate among proponents of sometimesconflicting models. Peter Bellwood articulated the current dominant model, in his “Out of Taiwan model” (Bellwood 1979, 1992, 1995, 1997, 2001, 2002, 2004a, 2004b, 2005). This model proposes a migration of Austronesian-speaking people from Taiwan, who carried Austronesian languages and knowledge of agriculture into the Southeast Asian islands, previously inhabited to the west of New Guinea by huntergatherer groups. The model is based on a combination of archaeological, linguistic and genetic evidence, and Diamond and Bellwood applied it to the Bantu dispersal in southern Africa, and the colonization of the uninhabited islands of the Pacific by the Austronesians. As Diamond and Bellwood (2003:598) stated “the simplest form of the basic hypothesis that prehistoric agriculture dispersed hand-in-hand with human genes and languages is that farmers and their culture replace neighbouring hunter-gatherers and the latter’s culture.” However, they did not apply this model to Island Southeast Asia (Bellwood and Diamond 2005:504). The Austronesian Languages The Austronesian language family covers a very wide area (Figure 1.4) and includes an estimated 1200 languages (Pawley 2002:251). The Austronesian Diaspora took place in various stages between 5,000 and 1,000 years ago, and spread from a homeland in the region of southern China and Taiwan to places as far away as Madagascar, Hawaii, and Easter Island (Bellwood 2004b:25).

Figure 1.3 Agta hunters with their bows and arrows (Photo courtesy of Richard Guzman) Agta will butcher an animal at the kill site and later smokedries the meat in their hunting camps (Griffin and EstiokoGriffin 1978:39). Some bones may be left in the hunting camps to ease the weight for the long trek back to the base camp.

The term Austronesian (from the Greek “of the southern islands”) was first introduced by W. Schmidt in the 1900s to replace the term Malayo-Polynesian that was widely used at the time for the whole family. Schmidt, and later HeineGeldern, inferred that the homeland of the Austronesian language family was somewhere in mainland Southeast Asia. Dyen, on the other hand, proposed a Melanesian origin on the basis of lexicostatistical evidence (Dyen 1965, Heine-Geldern 1945). Robert Blust’s work on the reconstruction of proto Austronesian (PAn) and the sub-grouping structure within the family, based on shared innovations, particularly in verbal prefixes and pronouns for the Extra-Formosan (MalayoPolynesian) languages, pointed clearly to Taiwan as the most likely candidate for the Austronesian homeland (Blust 1976, 1985). Another linguist, Andrew Pawley supports this and states: PAn was spoken in Taiwan, which is both the area of the highest concentration of firstorder subgroups and the site of the earliest

Agta supplement their diet with wild tubers (Dioscorea, Stenomeris) and other starchy foods like sago (from the trunk of Caryota cummingii), and some fruits like Diplodiscus sp. (Allen 1985, Griffin and Estioko-Griffin 1978). Agta also collect and chew betel nuts (Areca) with lime. Linguistically, all extant Agta languages are Austronesian, belonging to the Malayo-Polynesian (Extra-Formosan) subgroup, as do all other Philippine languages. The Agta might have spoken different languages prior to the arrival of Austronesians, but if so, they have undergone language shift as a result of contact with farming groups (Reid 1994). Headland (1986:177) proposes “cyclic interaction between prehistoric Agta and farming populations.” Linguistically, 6

Neolithic cultures in Island Southeast Asia (Pawley 2002:268).

Polynesian (CEMP) subgroup, while other movements to the west founded the many Western Malayo-Polynesian (WMP) subgroups. The WMP category includes languages in the Philippines, Borneo, Sumatra, Vietnam and Malay Peninsula. The break up of CEMP occurred when groups moved into Irian Jaya to found the Eastern Malayo Polynesian (EMP) subgroup, the main member of which is Oceanic. The other CEMP subgroup is Central Malayo Polynesian (CMP), which includes languages in the Lesser Sundas and southern Moluccas.

According to Blust’s linguistic reconstruction, PAn broke up into ten branches. Nine remained confined to Taiwan, while the tenth became the Malayo-Polynesian (MP) subgroup that spread out of Taiwan to cover the remainder of Austronesia, from Madagascar in the west to Easter Island in the East (Blust 1985, see also Pawley 2002). The MP dialect chain eventually disintegrated, with one group moving to the Moluccas to found the Central and Eastern Malayo-

Figure 1.4 The Spread of the Austronesian language family (Bellwood 2005) Bellwood suggests that there would be four different situations of farmer and hunter-gatherer contact and interaction during the spreads of prehistoric agricultural societies, these being:

Subsistence The spread of Austronesian occurred in tandem with the spread of agriculture and domestic animals. “After 5,000 years ago, the world of the hunter-gatherer economy across Southeast Asia and on the western edge of the Pacific was confronted rather suddenly by another economic system, that of the first agricultural population” (Bellwood 2004b:21). There are only a few areas in the world where there has occurred an independent origin of plant domestication. In East Asia, the main area for this development was in central China, where cereals (rice and millets) were first domesticated.

homeland/starburst zones of agricultural origin, spread zones of rapid and coherent farming dispersal, friction zones where human or environmental factors restricted or induced strong interaction during farming dispersal, and the beyond farming zones where farmers continued into other, nonagricultural adaptations (Bellwood 2002:22, italics in the original).

In East Asia, the foci of interest as far as agricultural origins are concerned lie in the middle and lower basins of the Yellow (Huanghe) and Yangzi rivers, and the several smaller river basins, especially the Huai, that lie between them (Bellwood 2005:111).

In addition, in Island Southeast Asia there are propositions that rice might have been replaced by other cultigens such as yams (Dioscorea alata) and taro during the early movement of farming communities into Island Southeast Asia (Paz 2004). This possibility has been recognized by Bellwood (2005:139), who stated that “this subtropical cereal faded from the economic repertoire as populations moved towards eastern Indonesia.”

From China, agricultural economies spread into Island Southeast Asia via Taiwan, and Mainland Southeast Asia via Vietnam and the major inland rivers. Rice and foxtail millet cultivation appeared in Taiwan by at least 5000 years ago (Bellwood 2004b:29). Gua Sireh in Sarawak has AMS-dated rice at about 4200 years ago (Bellwood et al. 1992, Paz 2002). Austronesian farming communities spread through Island Southeast Asia and later into island Oceania (except for most of New Guinea, which had an independent development of agriculture), passing through areas west of the Solomon Islands that were formerly inhabited by hunter-gatherer groups.

Archaeological evidence In the 1964, K.C. Chang initially proposed a cultural link between Taiwan and the Southeastern Tradition of China, hinting that it could be important for the origins of the Austronesians (termed ‘Malayo-Polynesians’ at that time): It is both interesting and instructive to note that the modern aborigines of the island of Formosa, whose languages are related to the Malayo-Polynesian family and who exhibit many characteristic features of the Pacific 7

cultures, can be demonstrated to be, at least in great part, the direct descendants of prehistoric inhabitants on the island whose culture remains were a part of the Southeastern Tradition (Chang 1964:375)

and rice in the Cagayan Valley by at least 3700 BP (Snow et al. 1986:8). A number of layers stratified beneath the mainly Iron Age shell middens in Cagayan Valley, as well as in the cave sites in the foothills of the Sierra Madre, have produced red-slipped pottery in the earlier layers and polished black incised and impressed pottery in the younger shell midden layers above (Thiel 1988, Tsang, Santiago, and Hung 2001). These sites also contain spindle whorls, clay lingling o ear ornaments, shell and stone beads, and polished stone adzes (Aoyagi, Ogawa, and Tanaka 1997).

In Chang’s terms, the Southestern Tradition was the Neolithic culture in the southeastern Chinese coastal area that immediately followed the Lungshanoid. It was characterised by tanged and stepped adzes, stone bark cloth beaters, geometric impressed pottery, stone knives, rice cultivation with fishing and mollusc collection, and probably a maritime navigation technology.

Genetic evidence The relationship between language and biological populations has recently gained prominence (Bellwood and SanchesMazas 2005, Sagart, Blench, and Sanches-Mazas 2005). Correlations can never be absolute, since a population can adopt a new language without changing biologically. As noted above, the Agta have lost their original languages and adopted Austronesian ones.

Shutler and Marck presented a similar view in 1975. They proposed that a farming group moved from Southern China into Taiwan, then into the Philippines. This group then diverged into two groups, with one moving to western Indonesia and the other into eastern Indonesia (Shutler and Marck 1975).

Nevertheless, a number of genetic markers have been used to understand relationships between populations, such as those in the mitochondrial and Y genomes (Underhill 2005), RH and GM serological markers (Poloni et al. 2005), and HLA haplotypes (Sanches-Mazas et al. 2005). The details of each genetic analysis and the statistical analyses will not be discussed here, and only the major findings relating to Austronesian dispersal will be noted. The opposing interpretation by Oppenheimer (2002, 2004) based on genetic data will be discussed in the next section.

It was Peter Bellwood (1978) who popularised and further developed the archaeological evidence for the Taiwan origin of the Austronesians. Proto-Austronesian society, with a subsistence economy based on agriculture, fishing and probably some hunting, developed in Taiwan following migrations from Southern China before c. 5000 BP. Taiwan prehistory is discussed in the next chapter, but a brief overview is warranted here. The Dabenkeng (TPK) is the oldest Neolithic culture in Taiwan, represented by cord marked and incised pottery associated with rice cultivation, spindle whorls, stone adzes, stone barkcloth beaters, slate points and stone net sinkers (Chang 1969, Tsang 2005). The TPK is dated to c. 5000 BP, and developed into regional cultures such as the Xuntangpu and the following Yuanshan in the north, Niumatou in the west, Niuchouzi in the south, and Fushan followed by Beinan in the east (Hung 2004). Chang (1969) referred to the post-TPK phase in southwestern Taiwan as ‘Lungshanoid’, a term no longer in general use, and described the Fengbitou assemblage in southwestern Taiwan as including red-cord marked, painted and black pottery with tripods and perforated ring feet, clay spindle whorls, ground slate reaping knives and projectile points. Bellwood (1997:215) believed that the Yuanshan culture of northern Taiwan could have made a significant contribution to the spread of Neolithic cultures into the Philippines, with its assemblages of pottery with ring feet, occasional examples of incised and punctuate decoration, and red slipping. Bellwood (pers.comm. 2005) now regards the central and southeastern coast of Taiwan as a more likely homeland. Hung (2005) observes that red-slipped pottery became widespread in Fushan sites by c. 4000 BP, and this region of eastern Taiwan could have been the source region for Luzon’s red-slipped pottery.

Using serological markers (RH and GH), Poloni et al (2005) have attempted to see correlations between linguistic groups in terms of genetic affinity. Using a number of statistical approaches such as multivariate and ANOVA analysis, their results show that: the correlation analyses between linguistic and genetic distances carried out in this study argue in favour of hypothesis 2, that is for a common origin of the population of Southeast Asian linguistic families (AustroAsiatic, Tai-Kadai, Hmong-Mien, and Austronesian) and a separate origin of SinoTibetans (Poloni et al 2005:268). They also observe that the Austronesian group shows a low within and between group diversity that might suggest a recent origin from a homogenous ancestral population. Using HLA genetic markers, two further studies show the relationships between aboriginal Taiwan people and other Austronesian groups. Lin et al (2005) conducted an analysis of Taiwan aboriginal groups and observed a high degree of diversity between them, despite there being a high degree of homogeneity within each group. They argued that Taiwan’s indigenous peoples have a closer genetic relationship to southern Asian populations. This is especially true for insular Southeast Asian populations (Indonesia, Philippines), in whom many high-frequency HLA alleles specific for Taiwan’s indigenous peoples are also seen (Lin et al 2005:242).

Migrating early Austronesians from Taiwan to the southern islands thus carried a Neolithic suite of artifacts and crops. This included plain and red-slipped pottery sometimes with perforated ring feet. The complex also included stone adzes, spindle whorls, barkcloth beaters, and stone and shell ornaments (Bellwood 2005). Dentate-stamped and incised pottery designs followed an early plainware phase (Bellwood and Dizon 2005).

Sanches-Mazas et al (2005:290) have used HLA genetic markers, applied in a regional context. They conclude that their investigations are congruent with a Taiwanese homeland of Austronesians, specifically along the east coast of Taiwan.

Luzon is believed to have been colonized by Austronesians by c. 4000 BP. There is evidence for both red-slipped pottery 8

They further argued that Austronesian populations experienced rapid differentiation once they left Taiwan Island. This differentiation may have been due to genetic drift that can be seen in the high genetic variation between groups, but low genetic variation within them.

Modified Models and Differing Austronesian Origins and Dispersals Modified Models

Views

Opposing Views to the Out of Taiwan Model Wilhelm Solheim II (1988) has a different view from Bellwood (1997). He believes that a maritime-based culture developed in Mindanao and northeastern Indonesia and expanded northward to Taiwan by 5000 BC. He also believes that edge-ground stone tools and shell adzes were brought northwards from Indonesia.

about

Solheim objected to the hypothesized southward movement from Taiwan to Luzon due to the strong north-flowing Kuroshio Current, a view supported by Anderson (forthcoming). Anderson also thinks that it would have been easier to sail northwards from Luzon to Batanes than the other way. However, a recent study by the University of the Philippines Marine Science Institute has identified a southflowing current east of the Kuroshio Current (Isorena 2004:81). This current flows between 123° and 125° east longitude and could have been used to sail from Taiwan to eastern Luzon.

The Austronesian expansion that Bellwood proposed has been confronted with differing views, some being modifications of the model, others putting forth stronger disagreements. Matthew Spriggs (2003) agrees with Bellwood that a population movement occurred from South China to Taiwan and into Northern Luzon. However, he is cautious concerning the process of neolithization beyond Luzon, stating: we need to look at new models of language and cultural history for the region which do not rely on a simple ‘demographicsubsistence model’ of mass migration of Austronesian speakers from Taiwan to Tonga during the second and the beginning of the first millennium BC (Spriggs 2003:65).

Another opponent of a Taiwan origin is William Meacham, who believes in a convergent development of Austronesian languages and archaeological cultures across the whole of Island Southeast Asia, with no specific region of origin (Meacham 1988). He argues that the archaeological cultures of Island Southeast Asia are unrelated to the archaeological cultures of Taiwan, and that in turn the cultures of Taiwan are unrelated to those of southern China. But recent archaeological evidence (Hung 2005, Tsang et al 2001) shows strong affinity between the Neolithic cultures of both the Batanes Islands and the Cagayan Valley of Luzon and those of eastern Taiwan, as discussed above.

One specific concern with Bellwood’s Out of Taiwan Model was to explain early evidence (c. 4200 BP) for rice in pottery in the site of Gua Sireh in Sarawak, associated with carved, cord wrapped and basket-wrapped paddle-impressed pottery (Bellwood 1997, Paz 2002). A modification to the model is favoured by Daud Tanudirjo (2004), who noted (after Tsang 1992) that the initial Guoye phase of the TPK in Taiwan was succeeded, c. 4700 BP, by the Suo-kang phase, a locally derived culture with fine cord-marked, painted and basketimpressed pottery associated with rice production (Tanudirjo 2004). He suggested that a group of people might have moved southwards from Taiwan to the western coast of Luzon and Palawan, reaching Borneo at c. 4500 BP or earlier (Tanudirjo 2004:88). A separate dispersal occurred a little later to the east, involving a movement of red-slipped pottery from Taiwan to the Philippines and Indonesia.

Arguing for an origin for Austronesians somewhere in Wallacea, Stephen Oppenheimer (Oppenheimer 2004, Oppenheimer and Richards 2002) uses genetic evidence to contend that there was no biological link between Formosans and Filipinos on the one hand, and with Polynesians on the other. Using data on globin genes, mtDNA and the Y chromosome, he argues for a late Pleistocene genetic origin east of the Wallace Line for the most common Polynesian mtDNA motif. Oppenheimer and Richards (2002: 293) argued that the “motif probably originated long before farmers of Taiwanese origin could have arrived in the Moluccas.”

Bellwood (1997:237-238) initially suggested that the Gua Sireh pottery had a Malay Peninsular origin since it is different from the red-slipped pottery found in the Philippines and Sabah. But Bellwood (2001) and Bellwood and Dizon (2005) now propose a double movement from Taiwan.

Oppenheimer has also used Y chromosome data to support his views of non-relationship between Formosans and Polynesians. The M119 Y haplotypes are generally rare in Polynesia, but M122 is common in French Polynesia and Tonga, a circumstance that Oppenheimer relates to recent Chinese migration. But M122 is actually found widely in Mainland China, Vietnam, Philippines, Sulawesi, southern Borneo, Java and Sumatra, and amongst the Amis of Taiwan. M119 is also common in Taiwan and the Philippines (Oppenheimer and Richards 2002). Oppenheimer (2002:295) asserted that there is no genetic evidence to support that modern Polynesians share a recent common ancestor with modern aboriginal Taiwanese. He argues that there is ancient genetic evidence of a spread from Wallacea into lowland and coastal Oceania.

Victor Paz (2002:275) views the passage through Luzon as a “friction zone” (in Bellwood’s terms), rather than the “spread zone” hypothesised by Bellwood (2002), owing to the limited evidence for rice/cereal cultivation during the Neolithic in the Philippines and Indonesia (Paz 2002). Atholl Anderson (2005) has also proposed two successive dispersals. He calls them Neolithic I and II respectively: Neolithic I, if it can be called that, may be presented by expansion from South China through Thailand and Vietnam then through Malaya to Borneo and Sulawesi, if not more widely, of basket and cord-marked ceramics among other types….Neolithic II can be propose as the characteristically red-slipped pottery expansion in the Philippines and elsewhere in the eastern islands of Southeast Asia, and later into Malaya and Vietnam (Anderson 2005:37).

But it is quite possible to read the distributional data for the M119 and M122 haplotypes as strong evidence for Austronesian (male) migrations from Taiwan via the Philippines into Indonesia and Polynesia. 9

Bellwood (2005:207) has addressed the genetic issues that Oppenheimer raised as follows: 1. If populations with a Polynesian-like and Asian-derived morphology had spread into eastern Indonesia long before the period of Austronesian dispersal, then the haplotype could have been acquired from them much later by the Austronesian populations ancestral to Polynesians. This is possible in theory, and would allow for a greater level of accommodation between the two opposing reconstructions. The main problem is that there is no skeletal evidence to indicate the presence of such a Polynesian-like population in eastern Indonesia prior to the Holocene, although this may simply reflect lack of data. 2. The molecular clock calculation for the origin of the Polynesian motif may be wrong, given that there is now voluminous literature of debate and disagreement about this kind of chronological calculation. If wrong, then perhaps the mutation could have occurred during Austronesian dispersal after 4,000 years ago in eastern Indonesia. In other words, is the place of origin of the Polynesian motif correct, but the calculated date of its mutation wrong? 3. Could early Austronesians have picked up the Polynesian motif from an indigenous source in eastern Indonesia, initially through only one or two female individuals, whose descendants soon thereafter, as a result of considerable fecundity in bottleneck situations, raised the frequency of its occurrence in the ancestral Polynesian (Lapita) population? This option seems possible, but is only necessary if a Pleistocene date for the origin of the Polynesian motif is supported by future calculations.

between New Guinea and the islands to the west, even before the Austronesian arrived. These exchange networks might have dispersed bananas (Eumusa section) from New Guinea and betel nut (Areca) from Southeast Asia. Denham proposes the abandonment of regional history as represented by the Austronesian model and suggests that we “instead open ourselves to the contextual understanding of the past, indicated here with respect to the roots of agriculture and arboriculture” (Denham 2004:616). Szabó and O’ Connor (2004) questioned the reliance on Austronesian linguistic reconstruction over the archaeological record. They discuss the case of shell beads that are considered as an item in the Austronesian package. Szabó’s analysis points to different manufacturing techniques in producing shell beads from sites in Palawan, Maluku and East Timor. “This certainly problematizes the use of a monolithic category of ‘shell beads’ as a marker for migrating Austronesian-speaking populations” (Szabó and O'Connor 2004:624). While there is no doubt that sub-regional data and local variations are the building blocks in constructing regional Prehistory, I disagree with Terrell and Denham that there is no need for a big picture reconstruction or meta-narratives. Humans have a propensity for adventure and will always interact with other people and nature. In doing this, humans leave traces across vast areas of the earth’s surface that archaeologists can follow and reconstruct from. In doing this, meta-narratives play an important role. Scale is significant in these debates. We cannot only concern ourselves with our own area of research; we must go beyond. As Szabó and O’ Connor (2004:624) suggest, there is much variation on the scale of the local region. But this variety can become irrelevant when examined from the perspective of larger scale meta-narratives.

Research Objectives The research area is in the karstic limestone formation of Peñablanca, Cagayan, northeastern Philippines. The area is home to a number of caves, most of which contain archaeological materials. Previous excavations by Thiel (1988), Ronquillo (1988) and Mijares (2002), demonstrated the widespread existence of two successive cultural units. In general, the lower layers contain stone artefacts, animal bones and riverine shells, while the higher layers have the same kind of assemblages, but with the addition of earthenware sherds. This presence of two successive cultural deposits makes the Peñablanca caves pertinent in any investigation of the changes that occurred between these two periods. Unlike the caves, the open sites in the Cagayan Valley contain only Palaeolithic or Neolithic assemblages separately, and not together in associated cultural horizons.

Murray Cox (2005) supports the views of Bellwood (2005). He recalculated the data used by Oppenheimer and Richards (2002) to infer the ‘most recent common ancestor’ or TMRCA for the mtDNA Polynesian Motif. Oppenheimer and Richards’ (2002) calculation of the TMRCA suggested c. 17,000 years ago in Eastern Indonesia. The recalculation presents a new value and questions the statistical reproducibility of Oppenheimer and Richards’ calculation (Cox 2005:186). Cox’s recalculated TMRCA of the Polynesian motif is consistent with a rapid expansion of the early Austronesian-speaking people from southern Taiwan via eastern Indonesia and into the Oceanic islands during the mid-Holocene.

The following are the objectives that this research intends to address 1. To reconstruct the Late Pleistocene to Middle Holocene prehistory of northeastern Luzon, from the Upper Paleolithic (or Preceramic) period to the Neolithic (or ceramic period). This will entail a detailed description of each period’s assemblages, and corresponding environmental data.

Objecting to the ‘simplicity’ of the ‘Out of Taiwan model’, a recent World Archaeology volume (2004) raised issues concerning the alleged affinity between Oceanic people and Austronesian-speaking populations in Southeast Asia. Denham (2004) and Terrell (2004, see also Terrell et al. 2001) both suggest that early Holocene interaction occurred

2. To examine if the Cagayan Valley archaeological record reveals a replacement of people, or simply exchange involving Neolithic cultural materials, in 10

the relationships between hunter-gatherers and farming communities during the past 4000 years.

no use-wear traces were found, the flake was identified as unutilised.

3. To understand the relationships between caves and open sites in terms of their cultural assemblages, chronologies, and patterns of usage by human groups.

Significance of the Study This study is an attempt to synthesize the vast quantity of archaeological material and data produced from Northern Luzon. Archaeological research in Northern Luzon has been an on going process since the 1950s, when fossil evidences were first discovered.

General Methodology To characterise the changes from the upper Paleolithic to the Neolithic in the cave sites in Peñablanca, controlled stratigraphic excavation of new cave sites was necessary. Three caves were excavated, namely Eme, Callao and Dalan Serkot.Understanding site formation processes is of the utmost importance.

This study attempts to understand the transition from the Late Upper Palaeolithic culture (or preceramic period) into the introduction of pottery and farming during the Neolithic period. The study also attempts to understand the interactions between the lowland communities of the Cagayan Valley and the cave-using groups from the foothills of the Sierra Madre. The application of different methods of analysis in illuminating the archaeological past is also a major contribution of the study. The artifactual analyses were used in reconstructing the functions of the cultural materials, and possible changes in the technology through time. The ecofactual analyses (flora, fauna and micromorphology) were used in reconstructing past environments and human diet.

This research used a number of approaches to reconstruct the archaeological record of the Peñablanca cave sites. In order to establish the chronology of the area, conventional and AMS radiocarbon dating were employed, with funds from the Centre for Archaeological Research at the Australian National University. The radiocarbon dates were presented both in uncalibrated date (uncal.) and calibrated (cal.) date range using OXCAL calibration software. In order to understand site formation processes in the caves, soil micromorphology was studied from impregnated samples in thin-section. According to Courty and colleagues (Courty et al.1985:5), soil micromorphology can be used to infer depositional and post-depositional archaeological sedimentation processes, as well as to interpret archaeological ‘features’ that can add a new breadth in interpreting an archaeological site. This approach, combined with recovery of archaeobotanical remains, can provide in-depth insight into Palaeolithic and Neolithic subsistence strategies. Identification of macro and microbotanical remains provides information on habitat vegetation and any cultigens used.

The Book Outline The next Chapter (Chapter 2) discusses selected archaeological sites in Mainland and Island Southeast Asia that date to the Late Pleistocene and the Early to Middle Holocene. The aim here is to identify patterns in human behaviour, as reflected by the archaeological records of the different sites in each particular time period. Chapter 3 discusses the relevant archaeological literature for northeastern Luzon in terms of the two recognised major cultural assemblages of the region; Palaeolithic (or Preceramic), and Neolithic (or Ceramic). Chapter 4 presents the results of my excavations in Eme, Callao and Dalan Serkot caves and relates the new data to previous archaeological results from other cave sites excavated previously in the Peñablanca area.

Faunal and shell midden deposits were also analysed. Comparisons of richness of the taxa for each cultural horizon reveal factors of diet breadth and choice. Analysis of material remains such as ceramics and lithics was also conducted. Analysis of ceramic vessel form and manufacture entailed both descriptive study and measurements of rim forms and recording of surface treatment. A sample of earthenware sherds was subjected to petrographic analysis to determine their mineral composition. The form, surface treatment and petrographic characteristics of sherds excavated in the caves were then compared with sherds from open sites in the Cagayan Valley.

The results of the excavations are further analysed in Chapter 4, where the stratigraphic descriptions and the results of soil micromorphological analysis are presented. Different formation processes are identified and assessed in order to determine the degrees of post depositional disturbance in the various sites studied. These data are then used to form a contextual backdrop for the archaeological analyses. Chapter 5 discusses the results of the various analyses by specialists who have assisted me during the course of the project. These cover faunal and botanical materials, and the results are then used in reconstructing past environments and human diets. The lithic implements recovered from the excavation are described and analysed in Chapter 7. I assess using technological and use-wear analysis if there was any change in lithic technology through time, especially from the Preceramic into the Ceramic, and within the different cave sequences. Non-lithic materials, particularly sherds, are analysed in Chapter 8. Vessel forms and the petrographic characteristics of sherds from cave sites are compared with those from open sites on the Cagayan Valley alluvium, in order to understand group interactions during the Neolithic period. Finally, Chapter 9 synthesises all the results recovered in terms of their significance for reconstructing the prehistory of northern Luzon.

The lithic analyses use a number of approaches. First, individual lithic analysis was conducted on all newly recovered artefacts. This included metrical measurements, and the recording of attributes such as completeness, platform classes, extent of cortex, shape of working edge, and type of flake termination. The data were sorted according to cultural periods (pre-ceramic and ceramic), and compared across periods and sites. Use-wear analysis occurred in two stages. First was lowpowered microscopy (6-60x) to identify use-scar terminations and edge-rounding (Kamminga 1982, Odell and OdellVereecken 1980). The second stage was the use of 100-500x microscopic magnifications to identify striations and polishwear forms (Mijares 2001, 2002, Vaughan 1985) Combinations of use-wear attributes identify probable contact materials, which can be characterised as either soft or hard. If 11

CHAPTER 2 The Archaeology of Southeast Asia during the Late Pleistocene to mid-Holocene periods the focus on flakes at Nguom towards increasing use of pebble tools.

This chapter selectively reviews and discusses for comparative purposes some of the relevant archaeological sites in Mainland and Island Southeast Asia. These sites are dated within the Upper Pleistocene to mid- Holocene periods (Figure 2.1), extending into the Neolithic. Some of the major sites in Luzon will be discuss in the next chapter and some sites in the Philippine Islands will be discussed further below, in context. This chapter is aimed at situating the prehistory of Northern Luzon within the larger regional archaeological context during the time period of concern.

Hoabinhian Industry Xom Trai cave contained a classic Hoabinhian industry, composed of mostly unifacially flaked pebble tools such as sumatraliths (Figure 2.2 a), short axes (Figure 2.2 d), and an edge-ground form that became popular during the succeeding Bacsonian industry. At Xom Trai, the radiocarbon dates for the Hoabinhian are between 18,000 and 17,000 BP (Ha Van Tan 1997:35). The Bacsonian (Figure 2.2 f) is believed to be the later phase of the Hoabinhian industry and is dated to c. 11,000 to 7000 BP. As stated above, it is characterised by the presence of edge-ground axes and cord or vine impressed pottery.

This survey of archaeological sites in Southeast Asia reveals trends in the material culture and subsistence strategies used by modern humans during the Upper Pleistocene to midHolocene periods. The archaeological record of the different major regions will be discussed first. A discussion will then follow, which attempts to present the different developments that occurred in each major region. This discussion will be divided into three time periods: Upper Pleistocene (40,000 to 18,000 BP) Terminal Pleistocene-Early Holocene (18,000 to 6000 BP), and mid-Holocene (6000-2000 BP).

An Epi-Palaeolithic site called Bau Du in Quang Nam Da Nang province is the only known Hoabinhian site along the coast of Vietnam (Ha Van Tan 1997:37). The site is in a sand dune and contains shell midden and five human burials in sitting and flexed positions. Faunal remains include deer, serow, monkey and rhinoceros, associated with pebble tools including sumatraliths, short axes, pointed tools, choppers, bifacially flaked tools, and pestles and mortars. The Bau Du site has radiocarbon dates of 5030±60 uncal. BP or 59205640 cal. BP (Bln-3040) and 4510± 50 uncal. BP or 53205030 cal. BP (Bln-3260) (Ha Van Tan 1997:37). The assemblage, according to Ha Van Tan (1997:37), is similar to the Hoabinhian industries of Malaysia in the importance of bifacial tools, and has no pottery or edge-ground tools.

Mainland Southeast Asia Vietnam Vietnam has a number of sites that extend back into the Pleistocene period. As Reynolds (1993:8) stated, “this country already possesses the richest Upper Pleistocene archaeology in South-East Asia and such richness of itself will create a more detailed picture of variability when compared to elsewhere.” Vietnam has a number of lithic industries that some archaeologists view in a chronological sequence (Ha Van Tan 1997). These are the Nguomian, Sonvian, Hoabinhian and Bacsonian.

Two roughly contemporary coastal sites with Bacsonian affinities are Quynh Van and Da But (Nguyen, Pham, and Tong 2004, Ha Van Tan 1997). The latter is a shell mound in Thanh Hoa Province containing thick coarse pottery, flaked pebble tools and edge-ground tools. The pottery is basket marked, has round bases and straight or flaring rims. The earlier phase in Da But has flaked pebble tools and edgeground axes, while the later phase has a predominance of edge-ground axes with pounders and mortars. Stone and clay net sinkers are also found, attesting to fishing subsistence. Da But has a radiocarbon date range from 6500 to 6000 BP (Nguyen Viet 2005:91). Quynh Van, on the other hand, is dated to about 5000 BP and has flakes, pestles and pounders with thick pottery.

Nguomian Industry The Nguom rockshelter in Bac Thai province, Northern Vietnam, is the site type for the Nguomian flake industry. The lower strata of Nguom contain primarily flakes made from quartzite and rhyolite materials and a few pebble core tools. Most of the flakes are amorphous in shape and some have retouching on their edges. This industry at Nguom is dated to c. 27,000 BP (Reynolds 1993:8), and Ha Van Tan (1997:38) argues it is contemporaneous with the flake industry of Lang Rongrien in Thailand and the lowest cultural level of Bailian Cave in Southern China.

Late Neolithic Moving into the Late Neolithic, Phung Nguyen is a large site in the Red River Valley that has some of the oldest evidence for rice cultivation in Vietnam, and possibly animal domesticates such as cattle and pigs, dating to c. 4000 BP (Higham 2004:47). The pottery has incised designs in parallel bands filled in with punctations, forming meandering shapes. Higham (2004:48) has suggested that this pottery design is related to a similar shape from Yunnan Neolithic sites in China. Other cultural materials at Phung Nguyen include shouldered stone adzes, whetstones, polished stone points, clay spindle whorls, stone rings mostly made from white nephrite, stone beads and clay net sinkers. Phung Nguyen represents the establishment of fully a Neolithic mixed economy, possibly derived from Southern China.

Sonvian Industry Sonvian industry is believed to be the immediate precursor of the Hoabinhian industry and is dated between 23,000 and 13,000 BP. According to Ha Van Tan (1997: 37) the Sonvian is characterised by pebble tools flaked only on the edges with the natural cortex preserved on both faces. The Tham Khuong rockshelter assemblage in Lai Chau Province contains a Sonvian assemblage and is initially dated to 28,130 ± 2000 uncal. BP (Bln-1408) and 33,150 ± 2300 uncal. BP (Bln1412) on land snails (Reynolds 1993:9). However, Ha Van Tan (1997:37) doubts the accuracy of these dates and prefers the more recent shellfish date of 15,800 ±150 uncal. BP (HCMV-03/93). The Tham Khuong has pebble tools, choppers and picks, and this industry indicates a shift from 12

1. Dabenkeng

13. Lena Hara

25. Khok Phanom Di 26.Spirit Cave

2. Beinan/Fushan

14 Liang Bua

3. Nanguanli 4. Penghu

15. Ulu Leang/ Leang Burung 16. Tingkayu

28. Tham Khuong

5. Fengpitou

17. Niah

29. Quynh Van

6. Batanes

18. Gua Sireh

30. Da But

7 Cagayan Valley

19. Song Keplek

31. Son Vi

27. Buon Triet

8. Ille Cave

20. Pondok Silabe

32 Xom Trai

9. Tabon/Sa’ gung

21 Kota Tampan/ Gua Gunung Runtuh

33. Nguom

10. Leang Tuwo Mane’e/ Leang Sarru

22. Gua Cha/ Gua Peraling/ Gua Chawas

34. Phung Nguyen

11. Tanjung Pinang/ Daeo Cave 2

23. Lang Rongrien

35. Pengtoushan

12. Halmahera Sites

24Ban Kao/ 36. Kequitou Lang Kamnan Figure 2.1 Selected archaeological sites in Southeast Asia

13

bifaces, choppers, a short axe made from chert, utilized and unutilised flakes, whetstones and cores. Most of these lithic implements were made from quartzite. The upper layers of Lang Rongrien contained four burials, two of which were secondary. Associated grave goods include pottery and a polished stone adze. The burial pottery is pedestalled with carinations and and flaring rims regarded by Anderson (1990) as similar to Ban Kao pottery (below). Pottery and stone implements were also recovered in nonburial contexts.

Lang Kamnan Cave Lang Kamnan Cave is located in Kanchanaburi, western Thailand (Shoocongdej 2000). The site was excavated by Rasmi Shoocongdej of Silpakorn University and contains a long sequence of ephemeral human habitation. The lowest layer is of late Pleistocene date, between 27,110 ± 500 uncal. BP (GX-20072) and 10,030±110 uncal. BP (GX-20066) on land snails (Shoocongdej 2000:22). The materials recovered are lithic implements, animal bones and shellfish. The preHoabinhian lithic assemblage consists of utilized and waste cores and flake stone, hammers and grinding stones. Animal bones include flying squirrel, porcupine, bamboo rat, turtle, cervids and bovids. Canarium nutshells were also recovered.

Figure 2.2 Hoabinhian lithic industry (Bellwood 1979, after P.I. Boriskovsky) The Late Neolithic sites of Vietnam reveal increasing diversity and include a number of regional variants, such as the archaeological cultures Ha Giang, Mai Pai and Ha Long in the north, and the Bau Tro, Bien Ho, Buon Triet and Lung Leng cultures in central and southern Vietnam (Nguyen, Pham, and Tong 2004).

The early Holocene assemblage is the same as that in the previous deposit. This period is dated between 7990 ±100 uncal. BP (GX-20069) and 7740±140 uncal. BP (OAEP 1192), both dates on land snails. No radiocarbon date is reported for the first appearance of pottery at the site, but Shoocongdej (2000) estimated that it appeared about 17701300 BC based on the dates for black burnished pottery from the Ban Kao site (below).

Generally, later Neolithic pottery has more diverse forms, some with pedestals and foot rings. The decoration is also more diverse, with cord marking, comb-incision, red slipping and punctation all occurring. Added to the lithic inventory are stone adzes, both shouldered and untanged, barkcloth beaters, pestles and shouldered hoes.

Shoocongdej (2000:28) characterised the Lang Kamnan lithic assemblage as an expedient technology based on the production of “amorphous, unpatterned sizes and shapes of tools.” She also stated that Lang Kamnan Cave was used as a temporary campsite during wet seasons, where stone-tool making had occurred.

Thailand Lang Rongrien rockshelter Lang Rongrien rockshelter in Krabi, southwestern Thailand, is a well-stratified site that has yielded a deep-time depth of human occupation (Anderson 1990, 1997). Excavated by Douglas Anderson (1990) since 1974, the site has at least three cultural horizons. The oldest assemblage is composed of chert flakes, pebble tools, bone and antler objects. The lithic implements include retouched and utilized flakes, end and side scrapers, and a few choppers and core bifaces. The lithic implements were also associated with a number of hearths. This horizon has been dated on charcoal to between 37,000 ± 1780 uncal. BP (SI-6819) and 27,110 ± 615 uncal. BP (SI6816) (Anderson 1990:21). With the emphasis on flakes rather than pebble tools it is similar to Nguomian of Vietnam.

Spirit and Banyan Caves Chester Gorman’s archaeological work in the uplands of Thailand included excavation of a number of cave sites (Gorman 1971, 1972). The two most important of there were Spirit Cave Spirit Cave and Banyan Valley Cave. In Spirit Cave, Gorman recognized two cultural horizons. They both contained stone implements such as unifacial discoids, flakes and grinding stones, but the upper horizon also had the addition of cord-marked pottery, some with resin coating, and stone adzes. Subsistence evidence in the lower cultural horizon included shellfish gathered from the Khong stream below the cave, and bones of deer, civet, macaque, squirrel, suids, snake and bamboo rat. Plant remains included seeds of Canarium, gourds, almonds and legumes. Cultural Horizon 1 was dated by Gorman (1971) to between 11,690 ± 560 uncal. BP or 15,650-12,350 cal. BP (FSU-315) and 9455±360 uncal. BP or 11,950-9550 cal. BP (GaK 1845). Cultural Horizon 2 was dated 8776± 290 uncal. BP or 10,650-9050 cal. BP (FSU- 318) to 8142± 390 uncal. BP or 10,150-8150 cal. BP (FSU 314) (Gorman 1971:303). However, the pottery now dated much younger than Gorman thought (Lampert et al 2003).

The middle cultural unit in Lang Rongrien is a Hoabinhian assemblage dating to 9655 ± 90 uncal. BP or 11,250-10,650 cal BP (SI-6817) and 7765 ± 65 uncal. BP or 8710-8390 cal. BP (SI-6213), both dates on charcoal (Anderson 1990:21). This thick deposit contains stone artefacts, faunal remains, and a variety of both freshwater and marine shell species. The faunal remains are dominated by different species of deer such as sambar (Cervus unicolor), barking deer (Muntiacus muntjak), hog deer (Axis Porcinus), and Cervus eldi. There are also pig (Sus sp.), cattle (Bos sp.) and jungle fowl (Gallus sp.). The Hoabinhian lithics include discoid and elongated

The upper layer which contained the pottery was dated by Gorman to 7622± 300 uncal. BP or 9350-7750 cal. BP (FSU 14

The cave site has a basal date of 13,600 ± 120 uncal. BP (Beta-38338) on a freshwater shell. The ‘Perak Man’ remains are bracketed between dates of 9460 ± 90 uncal. BP (Beta37818) and 10,120 ± 110 uncal. BP (Beta-38394), also on freshwater shell (Saidin and Tija 1994:16). Perak Man was an adult male in a crouched position, interred in a shell midden. Physical studies reveal an Australo-Melanesian affinity (Majid et al. 1994). The lithic implements recovered from the cave include cores, anvils, pebble tools, hammerstones, and large flake tools and debitage (Majid et al. 1994:151). Majid et al (1994:166) believe the cave was also used as a tool-making site owing to the quantity of lithic debitage. Pebble tools were the main items in the assemblage, made from quartzite. They were further classified as unifacial and bifacial pebble tool, choppers, and perimeter-flaked tools. Faunal remains included monkey, pigs and reptiles, with abundant riverine shellfish (Brotia costula/spinosa).

317) from associated charcoal. AMS dating of the resin coating on two sherds (Lampert et al. 2003:128) has given a result of 3042 ± 37 uncal. BP or 3360-3160 cal. BP (OxA10271) and 2995 ± 40 uncal. BP or 3340-3060 cal. BP (OxA10272). This makes the original date by Gorman no longer tenable and places the date of pottery in Spirit Cave close to c. 3000 BP. Banyan Valley Cave had a similar archaeological assemblage, but with the addition of rice husks and a tanged stone projectile point in the upper layer.

Khok Phanom Di and Ban Kao Two important Neolithic open (non-cave) sites in Thailand are Khok Phanom Di (Higham 1989, 2004, Higham and Thosarat 1994), and Ban Kao (Sørensen 1972). Khok Phanom Di in central Thailand is also a habitation and burial site dating to 2000 to 1400 BC. This large site, about 200 meters in diameter, is located 22 km north of the Gulf of Thailand. Excavated by Higham and Thosarat (1994) in 1985, the site yielded 104 mostly extended burials with grave goods such as shell beads, shell bracelets, stone adzes and pottery. The pottery is black burnished with incised decoration. Faunal remains include pig, dog, macaque and deer. There is a high quantity of fish bones, shellfish, crabs and turtles. Rice husks were retrieved from the occupation layers and were also observed as inclusions in or impression on pots.

Ulu Kelantan Sites Adi Haji Taha has excavated a number of sites in Ulu Kelantan, two of them for his PhD Thesis at ANU. These are Gua Peraling and Gua Chawas (Adi Taha 2000). Gua Peraling is south of the Perias River and has four cultural phases. The lowest cultural deposit contains dense accumulation of cobble flaking debris and implements. Some of the implements are bifacially flaked. Faunal remains include deer, rodents and freshwater shellfish. This layer is dated to 11930 ± 100 uncal. BP (ANU 9902) and 11,770 ± 90 (ANU 9903) from freshwater shell (Adi Taha 2000:124 and 136). Adi Taha (2000:124) observed that radiocarbon determination on freshwater shells excavated from two Hoabinhian sites in Kelantan, West Malaysia, tended to be much older than charcoal dates from the same layer (see discussion in Chapter 4 on the problem of dating shells).

Ban Kao, located 200 km west of Khok Phanom Di, is another burial and occupation site, and is the type-site for the Ban Kao culture of Sørensen (1972), dating to 2000 to 1300 BC. The site is in Kanchanaburi Province in central Thailand. The 44 extended burials contained grave goods such as untanged stone adzes, shell beads, shell bracelets, barbed spear points and pottery. The pottery vessels have tripod and pedestal supports, with cord-marked surfaces. Radiocarbon dates for Ban Kao are 3720 ± 140 uncal. BP or 4500-3650 cal. BP (K838) and 3310 ± 140 uncal. BP or 3900-3200 cal. BP (K842) (Gorman 1971:303).

The next cultural deposit is dated to 9730 ± 90 (ANU 9901), also on freshwater shell. A flexed primary burial of a female was excavated in this layer and dated to 9590± 100 uncal. BP (ANU9910) from freshwater shell. The date of this burial is similar to that of the Perak Man (above). Both burials were dated from freshwater shell and hence problematic in terms of chronology.

Peninsular Malaysia Kota Tampan The Tampan Paleolithic site in Perak, West Malaysia, has been beset by interpretive problems since its discovery by H. Collings in 1938. The site contains stone tools overlaid by volcanic ash, and has problems in terms of lithic typology and in understanding the site formation processes (Harrisson 1975). A re-excavation of the site to address these problems was undertaken in 1987 by a team headed by Zuraina Majid (1990). Majid found evidence of a lithic workshop, which included quartzite chunks used as cores, anvils, hammerstones, flakes and debitage. Flakes and flake tools predominated. A layer of volcanic ash then buried the Kota Tampan workshop that was located near a palaeo-lake. This ash was initially dated to 31,000 ± 3000 uncal. BP by the fission-track method, and was traced to the eruption of Mount Toba in Sumatra (Reynolds 1993:3). The eruptive history of Mt. Toba has been a subject of investigation by Chesner et al (1991) and they were able to correlate the Malaysian ash to the Youngest Toba eruption but to a much older date. Using 40 Ar/39Ar dating, the Younger Toba eruption was dated to 74 ± 2 ka (Chesner et al. 1991:202). This makes the Kota Tampan site much older than originally projected.

The third Cultural Horizon is dated to 6910± 250 uncal. BP or 8350-7250 cal. BP (ANU 9910) on charcoal (Adi Taha 2000:136). This layer has an extensive ash deposit associated with turtle, large fish and shellfish. Large edge-ground tools were also present. The ceramic period follows and is dated to 5770 ± 210 uncal. BP or 7250-6050 cal. BP (ANU 9906) on charcoal. The pottery is plain, cord-marked and red-slipped. Some of the unrestricted vessels are footed and carinated like those from Gua Cha, where they are dated to only 3000 BP. Gua Chawas has three cultural periods of interest to this period. The lower Hoabinhian layer is dated to 10,770 ± 90 uncal. BP (ANU 9938) on freshwater shell (Adi Taha 2000:134). The succeeding Hoabinhian layer is dated to 4390 ± 90 uncal. BP or 5300-4830 cal. BP (ANU 9462) on charcoal, while the upper layer with pottery dated to 1874 ± 70 uncal. BP or 1990-1610 cal. BP (ANU 9914), also on charcoal. Gua Cha in Ulu Kelantan (Adi Taha 1983, 1993) was first excavated by Sieveking in 1954, and in 1979 by Adi Haji Taha and Peter Bellwood. The site contains a thick alluvial deposit with bifacial Hoabinhian pebble tools and human burials. The Hoabinhian layers also contain many animal

Gua Gunung Runtuh Zuraina Majid (1994) also excavated Gua Gunung Runtuh cave in Perak, which contains human remains associated with a Hoabinhian industry (Majid 1994, Saidin and Tija 1994). 15

1976 (Majid 1982), and by a large team of specialists headed by Graeme Barker starting in 2000 (Barker 2005, Barker et al 2002,). In reconstructing Harrisson’s excavations, particularly the stratigraphic context of the ‘deep skull’, Barker’s (2002 et al.:157) team has produced two new dates for this skull of 42,600 ± 670 uncal. BP (Niah-310) and 41,800 ± 620 uncal. BP (Niah-311) on charcoal, using the ABOX-SC technique. These new dates essentially confirm the previous Upper Pleistocene date obtained by the Harrissons.

bones, predominantly pigs (Sus scrofa and Sus barbatus), with deer, monkeys, gibbons, squirrels, rats, rhinos and cattle. The middle of the Hoabinhian deposit is dated to 6280 ± 250 uncal. BP or 7750-6550 cal. BP (ANU-2218), on charcoal (Adi Taha 1983:51). The upper layer of Gua Cha contains Neolithic burials with pottery, beaked adzes, a bark cloth beater, shell beads and shell spoons, dating to 3020 ± 270 uncal. BP or 3950-2450 cal BP (ANU-2217) (Adi Taha 1983:51). The pottery vessels have flat or round bases with cord marking (Bellwood 1979) and are related to pottery in southern Thailand, rather than Malaysian Borneo.

A later Pleistocene series of pit features has been dated by Barker (2000 et al:159, 2005) to between 19,650 ± 90 uncal. BP (OxA-11550) and 8630 ± 45 uncal. BP (OxA-11549) on charcoal materials. These pits contain mammal, reptile, bird and fish bones, as well as charred plant remains.

Adi Taha (2000:275) has observed that, ‘the Hoabinhian lithic assemblage occurs right throughout the cultural horizons including the Neolithic phase.’ He surmised that either the Hoabinhians continued to live along with the Neolithic people, or they adapted Neolithic subsistence but continued to use Hoabinhian stone tools.

The stone artefacts during the pre-Neolithic period are mostly flakes, with a few pebble tools made from coarse-grained stone. The lithic technology is described by Majid (1982:90) as a process of smashing pebbles and selecting suitable flakes. Bone tools include spatulae and points. There are a few edgeground tools, estimated to date between 20,000 and 10,000 years ago (Majid 1982).

Island Southeast Asia Malaysian Borneo Gua Sireh Cave Gua Sireh is a cave located in the Gunung Nambi limestone massif southeast of Kuching. Harrisson and Solheim first excavated the cave in 1959 and Zuraina Majid in 1977 and Ipoi and Bellwood in 1989 carried out later research. Solheim (pers. comm.2005) recently (2004) went back to Kuching to work on the materials he co-excavated with Harrisson in 1959. Gua Sireh has pottery with carved, cord-wrapped or basket-wrapped paddle-impressed surfaces (Bellwood 1997: 237). The site has an AMS radiocarbon determination of 3850± 260 uncal. BP or 5050-3550 cal BP (CAMS-725) on a rice-grain inclusion in a pottery sherd and 3990± 230 uncal. BP (ANU 7049) on charcoal (Ipoi and Bellwood 1991:391). This evidence of early rice in the area was initially contentious, but has been confirmed by macroscopic examination of the pottery bearing sediment that contains rice phytoliths (Beavitt et al. 1996:31). Bellwood (1997: 237) initially stated his opinion that the early date for pottery and rice in Gua Sireh might reflect an arrival of Austroasiaticspeaking populations from Mainland Southeast Asia. His view now is that the cord-marked pottery of Niah and Gua Sireh is very similar to the cord-marked Middle Neolithic pottery of Taiwan, and hence related to Austronesian migration at about 4500 BP (Bellwood and Dizon 2005).

Niah Cave has an enormous assemblage of biological remains. The identification of the fauna dated to during and after the period of the deep skull includes the now-extinct giant pangolin (Manis palaeojavanica), bearded pig, orangutan, macaque, mouse deer, and wild cattle. Pre-Neolithic plants identified from phytoliths, starch grains and macrobotanical remains include charred parenchymatous tissue of yams, breadfruit, legumes, and tubers of the carrot family (cf. Apiaceae) and charred nuts (Barker 2005, Barker et al 2002:159). The Neolithic cultural assemblage at Niah was found within a cemetery area, and was associated with a variety of burial practices. Pottery forms are mainly globular in shape with cord-marked and carved paddle-impressed designs, as well as the painted and incised ware known as “Three-Colour Ware”, and a range of double-spouted vessels. Some of the carved paddle-impressed pottery is of recent date, similar to Kupang pottery in Brunei and ethnographic Iban pottery (Bellwood and Omar 1980, Freeman 1955). The Three-Colour Ware and cord-marked pottery are likely to predate 2000 years ago. Quadrangular adzes were also found in this period. The Neolithic botanical remains include charred parenchymatous tissues of wild taro and yams and charred fruit and nuts. This led Barker (2005:101) to infer that the Neolithic foragers who buried their dead at Niah were in contact with farmers but did not engage in farming themselves. The Niah Neolithic period is dated to c. 1400 BC onwards but may be as early as c. 2500 BC, like Gua Sireh.

Gua Sireh also has a preceramic assemblage of quartz flakes and shellfish dating to 21,630 ± 80 uncal. BP (ANU 7048) on freshwater shell (Ipoi and Bellwood 1991:391).

Niah Cave The West Mouth of the Great Cave at Niah in Sarawak was excavated by Tom Harrisson between 1954 and 1968. The cave was used as a habitation and burial site from the Late Pleistocene to mid-Holocene, spanning a period of 40,000 years. A human cranium known as the ‘deep skull’ was dated by associated charcoal to 40,000 years ago, and identified as morphological grounds by Brothwell (1960) as AustraloMelanesian and resembling Tasmanians. The inner chamber of the West Mouth also contains numerous Neolithic burials.

Tingkayu In Eastern Sabah, the Tingkayu stone industry presents a unique (for this region) Late Pleistocene stone technology. The site is located close to what was inferred to be a former lake dammed by a lava flow about 28,000 years ago (Bellwood 1988, 1997). The Tingkayu lithic industry includes a fully bifacial lithic technology. Slabs of laminated grey chert were reduced bifacially to produce lanceolate knives and other forms. Use-wear analysis on these knives suggests a cutting function, but some may also have been used as spear points. This bifacial lithic technology is rare in Southeast Asia. Bellwood (1997:179) proposes ‘that this tradition was

Though Harrisson and others published a number of reports about Niah Cave, no final report was ever produced and the stratigraphic contexts of the finds have always been contentious. The site was re-excavated by Zuriana Majid in 16

preserved on, a Wallacean island refuge—in the same way that Flores was a refuge for Stegodon, the only other larger mammal on the island during the Pleistocene.

developed locally, perhaps to meet a specific need in this rather unusual lacustrine environment.’

Indonesia and East Timor Pondok Silabe and Song Keplek Caves

Uai Bobo 2 Cave

Java and Sumatra have been the focus of archaeological research by Truman Simanjuntak, on both Pre-Neolithic and Neolithic assemblages. Two particular cave sites will be cited here; Pondok Silabe Cave in southern Sumatra and Song Keplek in Gunung Sewu in eastern Java (Simanjuntak and Asikin 2004, Simanjuntak and Forestier 2004). Pondok Silabe cave is located in the limestone formation in Baturaja, Sumatra. Simunjuntak and Forestier (2004) identified two cultural deposits, a lower Pre-Neolithic and an upper Neolithic. The Pre-Neolithic contains flake tools with faunal remains. The Neolithic deposits also contain flake tools, with the addition of pottery, globular in shape some with cordmarked, net- and mat-impressed designs. The fauna include pig, deer, monkey, orangutan, civet, tortoise and porcupine. Simunjuntak noted that while no ground stone adzes were recovered in the cave excavation, they were common finds in nearby open sites. He suggested that ‘the adzes as tools were not known to these Neolithic cave dwellers, but were introduced later in the region in the open land occupation’ (Simanjuntak and Forestier 2004:113). This is similar to the Peñablaca cave sites in Luzon as will be discussed in the next chapter. The upper layer of the Neolithic deposit was dated to 1875 ± 47 uncal. BP or 1930-1700 cal. BP (WK-11248) (Simanjuntak and Forestier 2004:111).

In East Timor, excavated by Ian Glover (1981), contains unretouched flakes associated with rodents, reptiles and shellfish, dated 14,000 to 9000 uncal. BP. At about 5000 BP a different assemblage appeared with the introduction of redslipped pottery, tanged points, shell adzes, shell fishhooks and shell beads. Pig, dogs and later, goat, were also introduced.

Lena Hara Cave An archaeological project in East Timor was conducted in 2000 to 2002 with the objective of understanding the initial colonization of the area during the Late Pleistocene period (O’Connor et al 2002, O’ Connor and Veth 2005). O’ Connor and colleagues (O’Connor et al 2002) conducted test excavations in Lene Hara Cave, previously excavated by a Portuguese team directed by Antonio de Almeida (Almeida and Zybszweski 1967). Lena Hara is a solution cave in raised limestone about 1 km from the current coastline. During the 2000 test excavation, the team observed at least two phases of occupation. The bulk of the deposit contains flaked chert artefacts, shells and animal bones. Radiocarbon determinations for the lower deposit are between 30,110± 320 uncal BP (ANU 11398) and 34,650± 630 uncal BP (ANU 11418), both on Strombus luhuanus shel (O’Connor et al 2002:48). Late Neolithic materials directly overlie this Pleistocene deposit. Pottery was recovered from the top 25 cm, and associated with continuing stone artefacts, shells and animal bones.

Song Keplek is located in eastern Gunung Sewu and contains three cultural deposits. The lowest is late Pleistocene, dated to 24,000 to 12,000 BP (Simanjuntak and Asikin 2004). Flake tools were recovered with faunal remains of large mammals such as cervids, bovids and elephants. The Pre-Neolithic, or Keplek, period is dated between 12,000 and 4000 BP (Simanjuntak and Asikin 2004:16). A high volume of lithics was recovered, which includes retouched and unretouched flakes. The retouched flakes were classified as scrapers, denticulates, borers, knives, arrowheads and points. Bone tools such as spatulas and points were made from cervid, bovid and Sus bones. Primary and secondary human interments were also placed in the caves during this period. The Neolithic period in Keplek is dated c. 4000-2000 BP, and contains earthenware sherds as well as unfinished ground stone adzes (Simanjuntak and Asikin 2004:14).

During the 2002 test excavation, fishhooks made from marine gastrapod shell (Trocus niloticus) were recovered and two were directly dated (O’ Connor and Veth 2005). One is dated to 9741± 60 uncal BP (NZA 17000), another to 6890± 50 uncal BP (OZG 894). O’ Connor and Veth (2005:255) suggest that these early-Holocene fishhooks antedate the arrival of Austronesians in Timor by 5000 years, and that this technology may have its origin in Wallacea.

Ulu Leang, and Leang Burung 1 and 2 A number of cave and rock shelter sites that contain Late Pleistocene deposits have been excavated in South Sulawesi (Bellwood 1979, Glover 1977) Three important caves in south Sulawesi have also been excavated; these are Ulu Leang, and Leang Burung 1 and 2. Leang Burung 1 was excavated by Mulvaney and Soejono (1970), and LB2 and Ulu Leang by Glover (1977). All contain a Toalian lithic industry dating after 6000 years ago (Bellwood 1997). This unique lithic industry includes flake and blade tools and geometric microliths. Glover dates the pre-Toalian basal cultural horizon of Ulu Leang from 10,000 to 8000 uncal. BP, and contains core and steep edge scrapers. In the middle of the sequence, long backed blades and bone points were recovered, dating after 6000 uncal. BP (Glover 1977:51). By 5000 uncal. BP, geometric microliths and Maros points were introduced (Figure 2.3), the later being hollow-based points sometime serrated and presumably used as arrowheads or spearheads (Bellwood 1997, Mulvaney and Soejono 1972). A radiocarbon determination on charcoal for the Ulu Leang ceramic period provided a date of 3550 ±130 uncal. BP or 4250-3450 cal. BP (PRL-230) (Spriggs 2003:70).

Liang Bua In the Nusa Tenggara chain, a recent controversial finding has been the discovery of bones of Homo floresiensis in Liang Bua, Flores (Brown et al 2004, Morwood et al 2004). This small statured extinct hominin (about 1.5 meters in height) may have existed until as recently as at least 18,000 years ago. Associated animal bones include remains of Komodo dragon and extinct Stegodon. Stone implements as described by the excavators include macroblades and microblades, ‘burins’ and perforators. Whether these stone implements were all made and used by this hominin remains to be proven. Morwood et al (2004: 1090) state that: Liang Bua provides evidence for distinct hominins descended from ancestral H. erectus population that survived until at least 18 kyr, overlapping significantly in time with H. sapiens. We interpret H. floresiensis as a relict lineage that reached, and was then 17

7782), both on marine shell. This preceramic deposit contains pitted stones, unretouched flakes and flaked pebbles tools. Above it, the ceramic deposit has incised earthenware sherds as well as human burials dating to 2090 ± 180 uncal. BP (ANU 8439) on human bone. At Daeo cave 2 a similar cultural assemblage was excavated. Incised pottery similar to that of Tanjung Pinang pottery was found on the ground surface. The deposit contains a mainly preceramic assemblage of stone flakes and manuports, ochre, bone points and cooking stones. Faunal remains consist of cuscus (Phalanger), fish bones and shellfish. This assemblage is dated to between 5530± 70 uncal. BP or 6470-6170 cal. BP (ANU 9452) and 13,930 ± 140 uncal. BP (ANU 9450) on charcoal and marine shell, respectively (Bellwood et al. 1998:339).

Leang Burung 2 was excavated by Glover’s team in 1975 and contains an upper Palaeolithic assemblage dated to 19,000 ± 250 uncal. BP (BM-1492) and 27,645 ± 200 uncal. BP (GRN8292), both determinations on freshwater shells (Glover 1981:16). The bulk of the stone artefacts are flakes, with some exhibiting edge gloss, and waste flakes. There are a few distinct retouched flakes and points that show a relatively advanced stone technology, that Glover has identified as Levallois (or Levallois-like) (Figure 2.4).

Golo Cave Golo Cave is located in Gebe Island, just east of Halmahera. This is a cave site with a good stratigraphic sequence. The lowest preceramic assemblage has been divided into three phases. The oldest phase is dated to 32,210±320 uncal. BP (Wk-4629) on marine shell and contains flaked lithics, volcanic cooking stones and a relatively high volume of marine shells. The middle phase also contains flaked stone artefacts, and volcanic cooking stones, but has a lesser volume of marine shell. This phase has a number of shell adzes made from Tridacna gigas and Hippopus. These adzes were initially estimated to date between 13,000 and 8000 BP (Bellwood et al 1998, Tanudirjo 2001), but more recent AMS dating of the shell indicates that all are probably of Holocene age (Bellwood pers. comm. 2005). The upper phase is dated to 7400± 110 uncal. BP (ANU-9449) on marine shell (Bellwood et al. 1998:339). This upper phase is a rich shell midden deposit with bone points, flake stones, volcanic cooking stones and ochre. Identified faunal remains were cuscus, wallaby, fish bones, birds and reptiles. Finally, the ceramic period postdates 3200 ±180 uncal. BP or 3900-2900 cal. BP (ANU 9448) on charcoal. The pottery was associated with shell adzes made from Cassis shell.

Figure 2.3 Maros points (Glover 1977)

Siti Nafisah Cave On southern Halmahera, Siti Nafisah cave has an assemblage devoid of any lithic tools. The preceramic deposit is mainly a shell midden with a few bone points, dated between 5120± 100 uncal. BP (ANU7789) and 3410± 70 uncal. BP (ANU 7786), both determination on estuarine shell (Bellwood et al. 1998:338). The ceramic horizon is dated to 2540±70 uncal. BP (ANU 7785) on estuarine shell. This horizon has redslipped and incised earthenware associated with fragments of ground stone adzes.

Figure 2.4 Blade points and Levallois-like flakes from LB2 (Glover 1981) Glover suggested (1981:29) that the Levallois character of the technology was an independent local invention. The upper and disturbed layers of Leang Burung 2 also contain small earthenware sherds with paddle-stamped decoration

Northern Moluccas

Bellwood et al (1998:360) has posited that humans were capable of sea crossing by c. 30,000 years ago, though only on rare occasions. The stone implements from these preceramic sites in the Moluccas were made from local materials. There was no evidence of trading or exchange of raw materials resources. Bellwood and his colleagues (1998:64) have also discussed the red-slipped pottery found at the site of Uattamdi on Kayao Island and related cultural materials as evidence for the colonization of these islands by Austronesians.

Bellwood directed a number of archaeological excavations in the northern Moluccas from 1990-1996. These included the islands of Morotai, Gebe and Halmahera (Bellwood et al. 1998). The geographic location of these islands is important since they lie between the southern Philippines, Sulawesi and New Guinea.

Tanjung Pinang and Daeo Cave 2 In southern Morotai Island two sites were excavated: Tanjung Pinang and Daeo Cave 2. Tanjung Pinang has two depositional layers; the upper archaeological layer which has a basal date of 10,000 BP, and the lower archaeologically sterile shell deposits dating to 37,510± 650 uncal. BP (ANU 7783) on marine shell (Bellwood et al. 1998:338). The upper layer has a preceramic deposit, dating between 3390± 70 uncal. BP (ANU-7778) and 8860± 100 uncal. BP (ANU

Leang Sarru and Leang Tuwo Mane’e In northern Sulawesi, the most pertinent rockshelter sites are Leang Sarru and Leang Tuwo Mane’e (Bellwood 1979, 1997, Glover 1977, Tanudirjo 2001). Leang Sarru on Salebabu Island (Talaud) was excavated by Tanudirjo (2001, 2005). 18

a thick shell-midden deposit with a few weathered earthenware sherds that might be intrusive (Szabó, Swete Kelly, and Penalosa 2004).

The lowest layer has a Pre-Neolithic assemblage of flakes and blade-like flakes made from chert, and basaltic hammer stones. The stone implements were found associated with shellfish, dated to 29,590 ±630 uncal. BP (ANU-10498) on Turbo shell (Tanudirjo 2001:264). The next layer is dated between 9,750 ±90 uncal. BP (ANU-10203) and 18,880± 140 uncal. BP (ANU-10960), again on Turbo shell Tanudirjo (2001:264). The site was used as a lithic workshop during this second period. Pottery with red-slipped, net- and matimpressed, and incised designs was recovered in the upper layer of the site. Tanudirjo (2001) dates the pottery to the early metal age, similar to the neighboring jar burial cave of Leang Buidane.

Balobok Cave on Tawi Tawi Island, southwestern Philippines, is another controversial site. Spoehr first excavated it in 1969 (Spoer 1973). A re-excavation was conducted in 1992, and three cultural components were reported (Ronquillo et al. 1993). An earlier occupation believed to be by hunting and gathering group, contained flake tools without pottery. This layer is dated to 8760 ± 130 uncal. BP and 8000 ± 110 uncal. BP1. The middle occupation is dated to 7290 ± 120 uncal. BP and contained pottery sherds, an adze preform of Tridacna gigas, a stone adze and flake tools. The later occupation is dated to 5140 ± 100 uncal. BP (Ronquillo et al 1993:8). This layer also includes pottery, shell and stone adzes, with the addition of a bronze-socketed adze. The early date for pottery from the middle occupation layer is contentious. Spriggs (2000, 2003) believes that site has been considerably disturbed and with claims for pottery and bronze several thousand of years earlier than other sites, making this date unreliable.

Leang Tuwo Mane’e was excavated by Peter Bellwood in 1974 and by Tanudirjo in 1996. It contains a lower preceramic assemblage of blade-like flakes made from grey chert, with a few prismatic cores (Bellwood 1997). The lithic implements were recovered with marine shells. Tanudirjo (2001:229) dated this preceramic assemblage to 4560 ± 70 uncal. BP (ANU-12011) on marine shell. The Neolithic pottery at the site is plain and red-slipped vessels dating to 3690 ± 70 uncal. BP (ANU-10209) on marine shell (Tanudirjo 2001:229). Bellwood’s excavation provided a radiocarbon date of 4030 ± 80 uncal. BP or 42504250 cal. BP (ANU 1515) for charcoal in the Neolithic layer. The flakes and blade-like flakes were associated with the newer Neolithic elements (Tanudirjo 2001).

Batanes Islands The Batanes Islands lies between the islands of Taiwan in the north and Luzon in the south. This is also the area where the South China and Philippine Seas meet, creating sometimes turbulent sea, especially in the Bashi and Balintang channels. Previous archaeological surveys and excavations in Batanes (AFAP 2002, Bellwood et al 2003, Dizon 2000, Koomoto 1983, Mijares and Jagoon 2001) have revealed no Palaeolithic or preceramic cultural remains. This negative evidence suggests that these islands were not inhabited until the Neolithic. The Japanese survey in 1982 (Koomoto 1983) in Batan Island identified an important Neolithic period assemblage at Sunget. This site has been re-investigated by the Asian Fore-Arc Project (Bellwood et al 2003), and contains a Neolithic assemblage, with Taiwan nephrite and slate, buried under a thick volcanic deposit. Sunget contains red-slipped pottery of shapes similar to Beinan in Taiwan (with vertical handles, high ring feet, and high everted rims), some with stamped circle designs, and pottery biconical spindle whorls. Two AMS dates on food residues in excavated sherds are 2910 ± 190 uncal. BP or 3650-2450 cal. BP (ANU 11817) and 2915 ± 49 uncal. BP or 3220-2920 cal. BP (Wk-14640) (Bellwood et al 2003:145).

Other regions of the Philippines Palawan Island In southern Philippines, two important sites on Palawan Island are Sa’gung rockshelter (Kress 2004) near Tabon Cave, and Ille Cave in northern Palawan. Kress (2004) excavated Sa’gung rockshelter in 1969 but the report has only recently been published. Eleven burials were excavated in this site, dated by Kress to the Neolithic, although no radiocarbon dates have been reported. Grave goods included edge-ground axes, a crocodile-tooth necklace, shell pendants, and a plug of lime probably connected with betel chewing. Except for lumps of unfired clay, no pottery was recovered from the excavation. The Ille Cave archaeological excavation is an ongoing project headed by Wilhelm Solheim II and Victor Paz. The excavation of the site started in 1998, and has continued every year since. The excavation is concentrated in the east mouth of the cave, on an extensive earth platform. The stratigraphy of the site has been complicated by disturbances cutting into previous burials. The lowest dated layer (Layer 5) has dates of 8770 ± 260 uncal. BP or 10,550-9050 cal. BP (ANU 11867) and 8170± 170 uncal. BP or 9500-8600 cal. BP (ANU 11870). In the 2001 excavation, when these charcoal samples were collected, this level contained no cultural material. Succeeding excavations in 2002-2005 have produced an assemblage of Paleolithic stone implements and ‘fossilised’ human remains from a disturbed context (Paz pers. comm. 2005).

Another site on Batan, also buried under volcanic sediments, is Naidi. This also contains red-slipped pottery, but has different rim forms from the Sunget, and appears to be slightly younger. The oldest radiocarbon date for Naidi site is 2620± 30 uncal. BP (ANU-11708) (Bellwood et al 2003:145). Itbayat Island is the largest of the Batanes Islands, and is mainly raised coral around two eroded ancient volcanoes. The coast consists of high and steep cliffs, and landing is quite difficult. Recent archaeological research on the island by the AFAP (Bellwood and Dizon 2005) has identified a number of sites. Mijares and Jagoon (2001) first reported Torongan Cave during their 2001 survey. Excavation in 2004 and 2005 revealed red-slipped and plain pottery in a layer of buried in washed topsoil at 40-65 cm, with a series of AMS dates of food residues in sherds, and marine shells, which indicate

Layer 4 of Ille contains flake tools and shellfish dating to between 6020± 330 uncal. BP or 7650-6150 cal. BP (ANU 11872) and 8580 ± 200 uncal. BP or 10,250-9050 cal. BP (ANU 11868(B2) both on charcoal (Szabó, Swete Kelly, and Penalosa 2004:220). Layer 3 is dated between 5200±210 uncal. BP or 6450-5450 cal. BP (ANU 11866) and 5720± uncal. BP or 7350-5850 cal. BP (ANU 11869(B1)). Layer 3 is

1

Note: all the radiocarbon dates for Bolobok are on marine shell and conducted by the Gakushuin Laboratory, the authors have not published the laboratory numbers. 19

stone hoes and adzes, reaping knives and projectile points (Chang 1969). The lower shell mound in Fengpitou is dated between 3722± 80 uncal. BP (Y-1580) and 3192 ±80 uncal. BP (Y-1578), both on marine shell (Spriggs 2003:66).

initial occupation by at least 4000 BP. An AMS date on a marine shell is 3352±35 uncal. BP (Wk-14641) while an AMS date on food residue has a date of 3320± 40 uncal. BP or 3640-3460 cal BP (Wk-14642) (Bellwood and Dizon 2005:14). New dates obtained for Torongan in 2005 are 3860± 70 uncal. BP or 4450-4080 cal. BP (OZH-771) on food residue in a pot and 3880±40 uncal. BP (OZH-772) on a marine shell (Bellwood pers. comm. 2005). Anaro is a hilltop defensible site of ijang type (fortified settlement, as described by Dampier in 1687) that contains red-slipped pottery with stamped circle designs, and plentiful objects of slate and nephrite, both imported from Taiwan. The jade includes discs, rings and pendants, all discarded during the manufacture of ornaments, as well as two drilled out cores that appear to have been discarded during the making of ling ling-o-type earrings similar to those recovered from the Tabon Caves, Niah Cave, and sites in southern Vietnam. Sourcing analysis (Iizuka, Hung and Bellwood 2005) of the Itbayat jade shows that it was derived from the Fengtian source region in eastern Taiwan. Anaro has AMS dates of 2770±50 uncal. BP or 2970-2760 cal. BP (OZH774), 1876±41 uncal. BP or cal 1900-1710 cal. BP (ANU-14643) and 1360± 39 uncal. BP or 1350-1220 cal. BP (Wk-14645) all on food residues in sherds (Bellwood pers. comm. 2005).

Figure 2.5 Cord-marked and comb-incised sherds from Fengpitou (Chang 1969)

Formosa Formosa, today known as Taiwan, lies east of southern China. It is home to fourteen recognised indigenous tribal groups that speak Austronesian languages. A Changpinian preceramic assemblage consisting of pebble and flake tools has been found on the eastern and southern coasts, dated to between 15,000 and 5000 years ago (Lien 1995, Tsang 1995). It is believed that this late Pleistocene culture was derived from South China when Taiwan was joined by a land bridge to the Asian continent.

In the survey and excavation of sites in the Penghu Islands in the Taiwan Strait, Tsang (1986) found the Kuo-yeh site, dated between 3600 and 4160 uncal. BP to be part of the TPK culture of mainland Taiwan (Tsang 1992:98). Also in southwestern Taiwan, two contiguous TPK sites were recently excavated under the supervision of Tsang (2005). The Nanguanli and Nanguanli East sites are located within the coastal floodplain in Southern Tainan County. Both contained charred rice and foxtail millet grains, with faunal remains of fish, deer, pigs and dogs. Associated material culture includes polished adzes, slate arrowheads and notched net sinkers. The pottery contains jars and bowls with cordmarked, painted and incised decoration. The Nanguanli sites are dated between 3000 and 2500 BC (Tsang 2005:71).

Tapenkeng Culture Taiwan’s Neolithic is currently divided into early, middle and late periods. The oldest known Neolithic culture in Taiwan was identified by K.C. Chang as the Tapenkeng (Dabenkeng) culture (Chang 1964, 1969, 1995). The Tapenkeng (TPK) is characterized by pottery that has impressed cord marks (Figure 2.5), comb and single incision, with some painted and red-slipped sherds. The TPK is also the period when the earliest stone adzes, normally with quadrangular cross sections, appeared in the archaeological record. Other stone implements include pecked pebbles, barkcloth beaters and slate point. The TPK is the period of the initial use of Penghu basalt and Fengtian jade in making stone adzes. The TPK is dated between 5500 and 4500 BP.

Middle Neolithic Generally, the middle Neolithic, or the fine corded-ware culture, replaced the TPK by 4500 BP and lasted until 3500 BP. This middle Neolithic, which KC Chang (1969) originally termed the Lungshanoid, is composed of four regional cultures—Xuntangpu in the north, Niumatou in the west, Niuchouzi including Fengpitou in the south and Fushan in the east (Hung 2004, 2005). The early phase of the Middle Neolithic is characterised by fine cord-marked pottery, but the cord marking decreases in quantity on the east coast with the dominance, by 4000 BP, of the red-slipped pottery tradition recently excavated by Hsiao-chun Hung (pers. comm.2004) at the site of Chaoliaqiao, near Fushan and just north of the modern city of Taitung.

The TPK was initially based on the excavation of two sites in the late 1960s, Tapenkeng in northern and Fengpitou in southwestern Taiwan (Chang 1969). Fengpitou is partially a shell mound in which Chang (1969) distinguished a number of cultural horizons. The lowest is the corded ware culture typical of the TPK period. The later cultural horizons were assigned to the Lungshanoid culture by Chang and comprised the lower fine and sandy red-ware settlements and to the latter upper-shell mound occupation. After the TPK, Fengpitou contained red cord-marked and painted pottery, with a few black pots. Pottery forms included unrestricted vessels with tripods or perforated ring feet, and restricted vessels. The black pottery has incised designs and began to appear during the sandy red-ware settlement period (c. 1900-1400 BC). Other cultural materials recovered were clay spindle whorls,

In the Penghu Islands, a phase contemporaneous with the mainland middle Neolithic also existed, called Suo-kang (Tsang 1995:192). The Suo-kang site contains fine cordmarked pottery and rice impressions, mostly reddish-brown in colour and associated with chipped and polished stone hoes, net sinkers, stone adzes, stone knives, stone anvils and bone fish hooks. 20

Kealhofer’s (2003) reconstruction of the Malay Peninsula environment during the late Pleistocene points to mosaics of savannah and woodlands. The hunters and gatherers of the Upper Pleistocene were involved in broad-spectrum subsistence, exploiting different terrestrial, riverine and marine ecological zones. The range of animals recovered in archaeological sites shows that these prehistoric humans were trapping animals from the forest canopy, including birds, squirrels and macaques. Terrestrial animals hunted included pigs, deer, bovids, rhinoceros, crocodiles and elephants. Shellfish, either from rivers or the sea, were a regular source of protein. Most sites contain at least some shell midden. There is also evidence, as in Niah that people exploited a number of plant foods, such as wild yams, legumes, seed, nuts and fruits. As Barker et al (2002:160) suggest: The implication of the Niah evidence is that Homo sapiens groups who colonised southeast Asia were able to do so because they had developed ways of living in and off the variety of landscapes they encountered, including tropical rainforest, rather than being restricted to coastal ecologies

Late Neolithic The Late Neolithic (3500-2000 BP) of Taiwan has predominantly plain pottery and is represented by regional cultures such as the Yuanshan and Zhishanyan in the north (Taipei), Dahu in the southwest, Beinan (or Peinan) in the east, and Yingpu in the middle west (Hung 2004). Yuanshan sites are located in the Taipei basin with red-slipped pottery with some incised, net-impressed, circle-impressed and brushpainted designs. Stone tools include shouldered adzes, stepped adzes, hoes and points. The Yingpu site, on the west coast, has a characteristic black ware that emerged around the late second millennium BC. Yingpu site has a radiocarbon determination between 2970 ±80 uncal. B.P or 3360-2920 cal. BP (Y-1630) and 2250 ±60 uncal. B.P 2360-2110 cal. BP (Y1632) both on charcoal (Sung 1965:148). In eastern Taiwan, a large site was excavated at Beinan, which has Neolithic burials and remains of a settlement (Lien 1995). The site is on a terrace of the Beinan River, immediately north of Taitung city, and is dated between 3500 and 2800 BP (Lien 1995:357). Beinan contained a large Neolithic village with about fifty house structures built with stone walls and pavements, and has yielded over a thousand burials, many in slab graves of slate. Pottery is plain, sometimes red-slipped, and many forms have vertical handles and high ring feet.

Terminal Pleistocene to Early Holocene The period between 18,000 and 6000 years ago saw the continuation of a broad-spectrum subsistence strategy by hunters and gatherers (Gorman 1971), who were exploiting different ecological zones and different sizes of prey. This can be seen from the faunal remains that also show an opportunistic hunting strategy. As the sea level rose after the last glacial maximum and as Sundaland became submerged, some of these hunter-gatherers might have chosen to go inland and into the rainforest while some remained along the rich coastal areas. Latinis (2000) has suggested that huntergatherers began to incorporate arboreal subsistence during this period. ‘Arboreal-based economies were perhaps more “sophisticated” by the late Pleistocene than is commonly assumed, as suggested by the practice of inter-island translocation of plant and animal resources’ (Latinis 2000:50).

Discussion Upper Pleistocene The archaeological record of the Pre-LGM Upper Pleistocene period, from 40,000 to 28,000 years ago in Southeast Asia, is highly dependent on cave sites. Movius’ chopper/chopping tool industry of the mid to late Pleistocene has been placed in serious uncertainty through archaeological research during the past three decades (Movius 1944, 1948, Reynolds 1993). The previous view of Movius, favouring a predominance of pebble tools, cannot be substantiated in all circumstances. Several Pleistocene archaeological sites, such as the lower layers of Long Rongrien, Lang Kamnan, Nguom, Niah, Kota Tampan, Tingkayu, Lena Hara, Leang Sarru, Leang Burung 2, Golo, Keplek and Tabon, have flake tools as the main lithic component, with a few pebble tools. As Anderson (1990:69) argued: ‘Upper Pleistocene archaeological sites that are securely dated are characterized by flake tools.’

There are different patterns between mainland and island Southeast Asia in terms of lithic technology. On the mainland, different pebble stone technologies developed, starting with the Vietnamese Sonvian industry, followed by the Hoabinhian lithic industry as seen in the caves of Xom Trai, Gua Gunung Runtuh, the middle layer of Lang Rongrien, and the lower layers of Spirit Cave, Gua Peraling, Gua Chawas and Gua Cha. The earliest lithic technology in Formosa, the Changpinian, is also considered to be of Hoabinhian affinity (Tsang 1995). Though the excavator of Sa’gung rock shelter in Palawan (Kress 2004) did not state so, my initial observation of the ground axes from this site indicates that they are similar to the Bacsonian edge-ground tradition of northern Vietnam.

The flake industry during this period is generally made using simple percussion with minimal modification. There are a few sites that developed more formal tool forms. Leang Burung 2 has Levallois preparation of striking platforms and Tingkayu has lanceolate bifaces. Since these forms do not occur elsewhere in Southeast Asia, Glover (1981) for Leang Burung 2 and Bellwood (1997) for Tingkayu think that these lithic technologies are local developments, probably adaptive strategies in specific environments.

Ha Van Tan (1997) attributed the change from flake tools to pebble tools as an adaptation to the changing climate. According to Kealhofer (2003:77), after the last glacial maximum the sea level rose, submerging much of low-lying Sunda and Sahul. The rainforest environment expanded and there was a need for the use of pebble tools, probably to manage the forest through limited clearing.

Although simple unmodified flakes were the most common lithic artefacts, there are occasional retouched flakes and/or blade-like flakes. Some flakes were modified to suit a particular function, such as high angle scrapers, tanged points and knives. Ian Glover (1977, 1981) stated that the Late Pleistocene in eastern Indonesia (or Island Southeast Asia generally) witnessed the development of relatively advanced stone-working techniques, in stylistically diverse traditions.

A different story can be inferred from the lithic industry of the islands, where flake tool assemblages persist, with relatively very few pebble tools. Sites such as Tabon, Niah, 21

Ulu Leang, Tanjung Pinang, Daeo 2 Cave, the lower layers of Keplek and Pondok Silabe, are still dominated by flake-tool assemblages. The Cabalwanian industry in the Cagayan Valley (though the dating is still problematic, but most probably Terminal Pleistocene) that was once thought to be mainly a pebble-tool industry, has actually more flake tools than pebble tools (Wasson and Cochrane 1979).

materials continuing to come into Itbayat from Taiwan until after AD 500. Further to the south, plain red-slipped pottery heralds the beginning of the Neolithic in. eastern Indonesia and East Timor. There seem to be two major pottery traditions that spread into Island Southeast Asia. The first is characterised by the cordmarked and paddle-impressed pottery in Niah and Gua Sireh in western Borneo. The second is characterised by red-slipped and often stamped pottery, as can be seen in the Philippines and Eastern Indonesia. As Bellwood (2005:139) stated:

The Mid-Holocene Southeast Asians would experience dramatic changes in their subsistence strategies and their cultural materials during the middle Holocene period. Agriculture and domestication of animals were flourishing in Central China, particularly in the Yellow and Yangzi basins, by 9000 BP (Bellwood 2005). Sites such as Cishan and Peiligang in the Yellow River Basin, dating to 8500 BP, have evidence for millet cultivation and domestication of pigs, dogs, and chicken. Pottery is pedestalled or on tripods, with cord-marked and incised decoration. In the Yangzi Basin, Pengtoushan has evidence for rice cultivation as early as 9000 BP, associated with cordmarked and sometimes red-slipped pottery, reaping knives and simple flakes.

Exactly how this Sundaland series of paddleimpressed assemblages relates to the redslipped pottery tradition to the east is unknown, but a homeland for both amongst populations located in Taiwan and the Philippines at about 2500/2000 BC seems likely. Stone technology in most agricultural communities is more concentrated in the production of ground and well polished stone adzes and axes. There is lesser concern with the production of flake tools, except for reaping knives. The adzes are sometime tanged or shouldered to facilitate hafting with a wooden handle.

This Neolithic culture of Central China expanded into different areas. One path was to the southeast, reaching the areas along the coastline and forming communities in Fujian and Guangdong, who were later to settle in Taiwan Island. Taiwan’s Tapenkeng culture, characterised by cord-marked and incised pottery, was derived from southern China. Chang (1969) had earlier proposed that the Austronesian inhabitants of Taiwan probably descended from the earliest TPK people, via the Lungshanoid and Yuan-shan cultures. In his 1995 paper, Chang describe the similarities in material culture, particularly between the Fujian sites of Xitou, Kequitou and Fuguodun, and the Tapenkeng of Taiwan.

Cave sites such Keplek and Pondok Silabe in Indonesia, Bolobok in Tawi Tawi, Tabon in Palawan, and the Peñablanca caves witness a continuation of flake tools into the period of pottery usage, sometimes without stone adzes (except for Keplek which has adzes). Simanjuntak and Forestier (2004:109) believe that such patterns reveal a continuity from the Pre-Neolithic to the Neolithic: ‘Cave findings show a cultural continuity between the Pre-Neolithic and the Neolithic indicating that the early Neolithic developed from the Pre-Neolithic with the introduction of new elements from the outside.’ I think this ‘cultural continuity’ could probably indicate that hunter-gatherers descended from the local Preneolithic or preceramic period continued to inhabit these cave sites. They continued to utilise simple flakes, but were acquiring pottery and other items from farming communities through exchange.

Tsang (1995:213) has pointed out that the ‘Tapenkeng Culture in the Taiwan area are actually very similar to those found in the sand dunes sites on the Fukien and Kwantung coasts.’ Fujian and Guangdong Neolithic sites are characterised by cord-marked, basket-marked, shell-edge impressed and painted pottery, with adze and axes and chipped pebble tools dating to c. 7000 BP.

In mainland Southeast Asia and Peninsular Malaysia we can see an expansion of agriculture, particularly rice cultivation, even though hunters and gatherers still persist today in areas either remote or less suitable for cultivation. The same occurs in the Philippines. Whether the Hoabinhian population directly participated in these new subsistence strategies or whether new migrants were actually settling is an issue of much debate (Bellwood 2005). Higham (2004:460) has observed ‘all attempts to trace a local Southeast Asian transition from hunting and gathering to farming societies have been fruitless.’ Yet Nguyen Xuan Hien (1998) believes that Vietnam is a major region for local domestication of rice. She says, ‘the territory of ancient Vietnam was situated in the region where rice became domesticated and … the Vietnamese were one of the first rice cultivators’ (Nguyen 1998:34). Unfortunately, there are no sites with radiocarbon dates for rice in Vietnam that parallel those from China.

The TPK culture is thus believed to have belonged to the early Austronesian population whose descendants eventually dispersed throughout Island Southeast Asia. The TPK evolved into different regional cultures during the middle Neolithic of Taiwan, and the red-slipped pottery that became dominant in eastern Taiwan sites after 4500 BP is believed to reflect the source of the early Malayo-Polynesian migrants to Luzon (Bellwood 2005, Hung 2005). I think that the Fengpitou culture of western Taiwan with its black pottery could also have contributed to this Austronesian culture out of Taiwan, as we will see in the discussion in the next chapter. Island Southeast Asia also became a recipient of expanding farming communities. Niah and Gua Sireh both have cordmarked pottery and the latter has evidence for rice cultivation as early as 4200 BP. The Batanes Islands and the Cagayan Valley have pottery that is initially red-slipped, later with incised and circle- or dentate-stamped designs. Bellwood and Dizon (2005) believe that the Batanes Islands, which lie at the crossroad between Taiwan and Luzon, have evidence for early migration of Austronesians at about 4500 BP. This contact and exchange between Batanes and Taiwan continued well after the initial migration, with slate and nephrite

Most sites in Vietnam and Thailand are habitation sites, burial sites, or both. The occupation of Vietnam sites of Quynh Van and Da But seem to have continued from Hoabinhian predecessors into the Neolithic without evidence of population shift. Da But has evidence for pottery by 6500 BP 22

that is either cord-marked or basket-impressed. However, these sites do not have evidence for agriculture and might still have been foraging communities. Phung Nguyen, on the other hand, could represent a cultural tradition that has an external origin, with its rice cultivation and highly decorated pottery. Sites in Thailand and Malaysia, such as Ban Kao open site, and the upper layers of caves such as Spirit Cave, Lang Rongien, Lang Kamnan and Gua Cha, witness the advent of these changes with the introduction of pottery and polished stone adzes. Pottery in these sites has cord-marked surfaces. Of particular interest is the pottery of Ban Kao, which has cord-marked decoration with tripods or pedestals, like much Malay Peninsula pottery (Bellwood 1993). Sørensen (1972) points out the similarity of this Ban Kao pottery with the ‘Lungshanoid’ period pottery of Southern China, and concluded that it originated from this region.

migrants from the north, and were probably the ancestors of the Austroasiatic-speaking populations who still inhabit many regions in Thailand as a non-Thai minority (Higham 1989). A similar phenomenon occurred in Luzon, particularly in the alluvial plains of the Cagayan Valley. Here, Austronesian migrants from Taiwan started to settle around 4000 BP (Bellwood 1997). They brought with them a suite of cultural materials and economic systems. Alluvial plain sites in Luzon, as we will see in the next chapter, are mostly of Neolithic date and reveal no clear continuity from the preceramic sites closer to the edges of the valley. Most preceramic sites in the Cagayan Valley are on old abandoned alluvial platforms and are not associated stratigraphically with more recent pottery. The same phenomenon can be observed in mainland Southeast Asia, where Neolithic sites tend to be on the active alluvial terraces.

Ban Kao site in Thailand has strong evidence for early farming communities. Khok Phanom Di, on the other hand, seems to have been initially a fishing and shellfish-gathering site in a rich estuarine environment. The burials in these sites are rich in grave goods, such as shell ornaments, pottery, barbed spear points and stone adzes.

In both mainland and island Southeast Asia, and in Luzon, hunter and gatherer communities in decreasing numbers survive today on the forested fringes of agricultural territories. They can sometimes live independently of the farming communities and maintain trading relationship with them, so long as competition for land and resources is not too overpowering (Headland 1986, Headland and Reid 1989, Peterson and Peterson 1977). Through such interaction, foraging communities were able to acquire pottery and other items, as well as carbohydrates-based food, such as rice, from their farmer neighbours.

Luzon and Southeast Asia The archaeological record of Southeast Asia reflects a mosaic of human adaptation to a rich and diverse region that contains many different ecological zones—alluvial plains, karst environments, rainforest and coastlines. Modern humans in this region were already capable of crossing sea by at least 40,000 BP, by which time they had populated nonlandbridged islands such as Luzon, Talaud and the northern Moluccas. The lithic technology across both mainland and island Southeast Asia (including Luzon) during the period 25,000 to 10,000 BP, as we have seen from the discussion above, was based primarily on flake tools. The colonisation of these islands was probably related more to Late Pleistocene hunter-gatherers, rather than to the predominantly Holocene ‘Hoabinhians’. As discussed above, there was a divergence between the mainland and island lithic technologies during the terminal Pleistocene and early Holocene periods, with the former becoming increasingly dominated by pebble tools (Hoabinhian), while the latter continued to produce flake tools. I agree with Kealhofer (2003) and with Latinis (2000), that the advent of ‘Hoabinhian’ stone technology could have been related to forest management, perhaps for arboriculture. Although Hoabinhian lithic technology was widespread on mainland Southeast Asia (including Peninsular Malaysia) at around 9000 to 6000 BP, it never reached Luzon or elsewhere in Island Southeast Asia beyond northwest Sumatra except for the Changpinian of Taiwan which is derived from South China. In Luzon, except for the Cabalwanian industry, that has both pebble tools and flakes, there are generally few pebble tools recovered as relative to flake tools. During the mid-Holocene, Southeast Asia experienced the advent of new technologies and subsistence strategies. Farming systems were introduced, carried in part by spreading groups of people with their ancestral languages. Here, we can see parallel developments in Neolithic cultures in mainland Southeast Asia, in particular Thailand, with contemporary island regions. Agricultural communities started to spread over the alluvial plains of Thailand by about 4500-4000 BP. Some of these farming communities were 23

CHAPTER 3 The Archaeology of Northeastern Luzon during the westward subduction of the Philippine Plate beneath the Asian continental block. Second, the eastward subduction of the South China Sea floor beneath the Philippine plate resulted in the formation of the Cordillera Central in the late Oligocene. An inter-arc basin (Cagayan Valley) was then formed between the now-inactive Sierra Madre and the active Cordillera Central, and filled with deep marine sediments during the Oligocene-Miocene periods (Mathisen and Vondra 1983:371).

The Geology and Geomorphology of the Cagayan Valley Northern Luzon (Figure 3.1) is composed of a large basin called the Cagayan Valley, surrounded by mountain ranges, notably the Sierra Madre in the east, the Cordillera Central in the west, and the Caraballo Mountains in the south (Figures 3.2 and 3.3). The Palaeogene Sierra Madre is a rugged mountain range with peaks of up to 1800 m (Wernstedt and Spencer 1967:19). It is composed of intermediate andesitic rocks, metavolcanics and metasediments, as are the Caraballo Mountains. The western Cordillera Central is composed of intermediate to mafic plutonic masses, with bedded basalts and graywackes (Durkee and Pederson 1961:140).

During the Pliocene there was change in the Cagayan Valley sedimentation regime from shallow marine to fluviatile. The Cordillera continued its uplift during the late Pliocene to early Pleistocene, forming asymmetrical anticlines and synclines in the sediments of the Cagayan basin (Vondra et al. 1981, Wasson and Cochrane 1979).

Figure 3.3 Satellite generated map of northeastern Luzon (Google Earth 2005) There are four major geomorphological structures (Figure 3.4) within the confines of the Cagayan Valley: the Tabuk Plateau, the anticlinal belt, the riverine alluvium, and the homoclinal belt (Vondra et al. 1981, Wasson and Cochrane 1979). The Pleistocene Tabuk plateau (Kalinga foothills) is an alluvial fan with terraces along the eastern flank of the Cordillera Central. The plateau is composed of pyroclastic flows and fluvial conglomerates, with elevations ranging between 30 and 170 m above sea level (Vondra et al. 1981:280).

Figure 3.1 Geological map of Luzon

The Cagayan anticlinal belt is a result of folding. The anticlines appear as elongated hills, while valleys occupy the synclines. These anticlinal belts expose megafaunal fossil bones of Elephas, Stegodon and Rhinoceros sp., particularly in the Cabalwan anticline. The Awidon Mesa Formation, of welded tuffs and dacitic tuffaceous sediments (Wasson and Cochrane 1979:10), forms the western flank of the Enrile-Liwan Anticline (with in the Cagayan Valley Anticlinal Belt) and overlies the Ilagan Formation described below. Figure 3.2 Mountain ranges of Luzon

Down the eastern side of the Cagayan valley and along the western foothills of the Sierra Madre lies the homoclinal belt. This comprises the Ilagan Formation and the Callao Formation. The Plio-Pleistocene Ilagan Formation is

Northern Luzon was formed during the Mesozoic as part of a volcanic island arc. First, the Sierra Madre was uplifted 24

began to accumulate in the northernmost part of the Cagayan Valley.

composed of sandstones, siltstones and mudstones, overlain by an upper pyroclastic flow deposit. The tilted homoclinal features are a result of tectonic activity that occurred after the Ilagan Formation was deposited, forming a ridge with a 10°-dip slope. The underlying Callao Limestone formation consists of reef carbonates, 540 m thick, formed during the late Miocene and early Pliocene (Mathisen 1980, Durkee and Pederson 1961). The Pinacanauan de Tuguegarao River divides the Callao Limestone formation into northern and southern sections, at the Bagaba Gorge.

The geographic circumscription of the Cagayan Valley gives it a different climate from other areas in the Philippines. The Sierra Madre partially shields the valley from typhoons, which are frequent in northern Luzon. The annual rainfall ranges from 150 to 200 cm per year in the valley, but may reach up to 250 cm per year in the Sierra Madre (Wernstedt and Spencer 1967:316). Vondra et al (1981:301) suggest that ‘a tree savannah environment probably predominated in Cagayan Valley during the Pleistocene while monsoon forest grew along the mountain front and streams.’ Adams and Faure (1997) have attempted to reconstruct vegetation maps of the world since the Last Glacial Maximum (LGM). In their reconstruction they postulate that Luzon would have had a monsoon or dry deciduous forest at 18,000 radiocarbon years ago, with a change to evergreen or semi-evergreen forest of humid tropical type by 8000 radiocarbon years ago (Adams and Faure 1997:639). The Sierra Madre and contiguous areas are currently vegetated by dipterocarp forest.

The Cagayan Valley occupies an extensive syncline that is now filled, for the most part, with deep alluvium deposited by the Cagayan River and its various tributaries, principally the Chico, Ilagan and Magat rivers. The Cagayan Valley contains approximately 3,125 sq km of deep Tertiary and Quaternary alluvial fill (Wernstedt and Spencer 1967:17).

The Cabalwanian Industry and the Fossil Record As early as 1947, H. Otley Beyer reported the existence of a fossil bed that contained rhinoceros teeth and large mammals dating to the mid-Pleistocene in the Cagayan Valley (Beyer 1947). He postulated that the antiquity of humans was as early as the mid-Pleistocene, with cobble tools that were perhaps associated with Stegodon and Rhinoceros sp. (Beyer 1948). The fossil fauna from Luzon was initially associated with the fauna from China by von Koenigswald (1956), and was believed to have crossed via Taiwan to Luzon over a land bridge sometime during the Pleistocene. On another field visit to the Cagayan Valley, von Koenigswald collected from the surface about a dozen pebble tools that he later named the ‘Cabalwanian Culture’ (Koenigswald 1958). Figure 3.4 Cagayan Valley geomorphological structure (Vondra et al 1981)

In 1971, the National Museum under the direction of Robert Fox started the Early Man Project in search of precursor of modern humans in the Cagayan Valley. The survey resulted in the identification of at least sixty-four sites that contained flake and pebble tools and fossils of extirpated large mammals (Fox 1971, 1978, 1979, Fox and Peralta 1974).

The present Cagayan riverbank in the lower valley is about 10 meters above present mean sea level (pmsl). Research has been conducted to reconstruct sea level changes during the Holocene in a broad region of the Philippines that includes Ilocos Norte, Palawan, Samar and Bohol (Berdin, Siringan, and Maeda 2003, Maeda et al 2004). Relative sealevel changes have been established using observations of tidal notches etched into the limestone, combined with radiocarbon and Thorium/Uranium dates for coral and oyster shells associated with them. This research was able to identify a sequence of sea-level changes for the Philippines.

Three sites in the Liwan area (Figure 3.5) in Kalinga Province, on the western side of Cagayan Valley, were excavated. These sites are located on the Awidon Mesa Formation that contained the megafaunal fossils found by von Koenigswald. According to Shutler and Mathisen (1979:111): The Awidon Mesa Formation is the term used to described Mid-Pleistocene (?) tuffaceous sediments derived from a dacitic source and characterised by the presence of bypyramidal quartz, euhedra of hornblende and sodic feldspar.

In the Ilocos study area, the reconstructed sea level at about 7000 uncal. BP was about 3 m above pmsl, and decreased to 2.3 m above pmsl by 6170 BP. By 5073 BP, sea level was still 0.9 m above pmsl (Maeda et al 2004:20). At the time of higher mid-Holocene sea level, the lower Cagayan River, which discharges into the Philippine Sea, would have flowed at least a meter higher than its current level. It was only after c. 2000 BP when sea level dropped to the present level that the surviving shell middens of Batissa childreni and Corbicula fluminea (Ogawa 2000b)

A number of pebble tools were collected from the surface of the Awidon Mesa Formation. At Espinosa, the Museum team excavated a number of Elephas bones seemingly associated with one pebble-cobble tool, according to Fox and Peralta (1974:126). In the nearby Madrigal site, an 25

cultural stagnation (Movius 1944). This assessment was contradicted by Hutterer (1977) and Reynolds (1993), who associated the simplicity of the stone tools with the working of plant resources such as rattan and bamboo. In support of this argument, Mijares (2001,2002) posited that the simplicity of the tools was due to both to the working of plant resources and to the nature of the volcanic raw materials.

tools after a previous dominance of flake tools (Reynolds 1993). But in Island Southeast Asia, flake tools remained the prevalent type of lithic assemblage. Flake tools dominated lithic assemblages throughout Philippine prehistory, and even the Cabalwanian industry is dominated by them (Fox and Peralta, 1974:135). Fox (1979), using changes in lithic assemblage, believed that a new group of people reached the Philippines about 10,000 to 13,000 years ago, bringing with them a microlithic tool tradition found in Palawan and Luzon. The Tabon Cave assemblage in Palawan is primarily flake tools, made from chert (Fox 1970), but is known to date to between 47,000 and 16,000 years ago (Dizon 2003).

An expedient lithic technology thus persisted in Minori Cave in Northern Luzon (Philippines), as in the rest of the Southeast Asian region. The expediency of the industries is, however, not due to cultural stagnation, but is an appropriate cultural adaptation by prehistoric people to their environment and its resources (Mijares 2003:74).

After archaeological work on the fossil-bearing open sites in the Liwan area, the National Museum research expanded across the Cagayan Valley towards the western foothills of the Sierra Madre. In 1976-77, archaeological exploration of caves sites in the Callao limestone formation in Peñablanca, Cagayan Province, was conducted in search of Palaeolithic sites. The extensive explorations resulted in the identification of forty-three caves and rock shelters containing archaeological materials on their surfaces (Ronquillo and Santiago 1977). A number of caves in the Callao limestone formation have since been excavated, especially between 1977 and 1982, then subsequently in 1999 and finally in 2003-4 for this PhD research. These caves contain primarily flake tools, although a few pebble tools have also been recovered. The flake tools are generally amorphous in shape and of chert, chalcedony, basalt and andesite stone types (Figure 3.7). Flake tools from Musang Cave (Thiel 1990b). Rabel Cave (Ronquillo 1981), Laurente Cave (Henson 1978), and Minori Cave (Mijares 2001, 2002) have been previously analysed (Figure 3.7), and will be discussed in Chapter 7.

This issue was again revived by Pawlik (2004:10), when he stated that: the idea of Palaeolithic bamboo tools remains to be seen….But there is at present no archaeological evidence for the use of bamboo as a raw material for tools nor have archaeologists presented convincing arguments for the former existence of bamboo tools able to replace heavy duty tools like hand-axes and cleavers. Neither have they conducted comparative experiments with bamboo tools to compare them with hand-axes and other stone tools. Bamboo, being organic, is difficult to recover in tropical archaeological contexts unless in charred form. Ethnoarchaeological research by Toth and his colleagues (Toth et al 1992) among the highlanders of New Guinea shows the versatility and strength of bamboo: we saw men butchering pigs, for example, with razor sharp pieces of split bamboo. One only resharpens such a tool by simply tearing off a strip of bamboo with the thumbnail or teeth. The serviceability of bamboo has been invoked to explain the simplicity of stone tools in East Asian sites that were inhabited by protohumans for many hundreds of thousands of years (Toth et al 1992:92) Johan Kamminga’s experimental work among the Agta of Sierra Madre (Davenport 2003, Davenport and Kamminga 2002) also demonstrated the versatility of bamboo for lightduty cutting. The experiment shows flake tools utilised in a number of activities, particularly in manufacturing wooden implements from bamboo, rattan and reed. Use-wear analysis by Mijares (2002, 2003) on experimental flake tools used on bamboo and rattan shows a high degree of correlation with archaeological flake tools. The flake tools from Minori Cave might have been used on bamboo and rattan.

Figure 3.7 Minori Cave flake tools (Mijares 2002)

The view of Pawlik (2004) is a return to the Movius Line Model (Movius 1944, 1948) that implied the dominance of pebble tools in early industries. There is no doubt that pebble tools, such as choppers and chopping tools were used by prehistoric humans in Southeast Asia. But they were probably for making tools made from organic materials and other implements needed to successfully exploit the tropical environment.

The lithics from the Peñablanca Caves have been recovered from both lower preceramic horizons as well as upper ceramic horizons. Technological analyses of the lithics generally show that there was no identifiable change between these two periods (Mijares 2002). The reason for this unchanging flake stone technology over such a long time in many parts of Southeast Asia is a subject of ongoing debate. Previous archaeologists attributed it to inferred 27

from layer 5 are reported to be red-slipped, and are from carinated restricted vessels with everted rims, and unrestricted vessels with flat and rounded lip profiles. However, a review of the archaeological inventory record for Catugan indicates that some sherds of black pottery were still being recovered from a depth of 120-149 cm, or within layer 5. On the eastern side of the Cagayan River, Takeji Toizumi and Ame Garong excavated the Conciso shell midden site in 1996 and 2000. Conciso was excavated to 200 cm below present ground surface, with the base of the upper shell midden deposit 160 cm below surface (Garong 2002). The shell midden layers contain primarily black pottery, with both restricted and unrestricted vessels. The rim forms of the Conciso black pottery have been analysed by Hidefume Ogawa (2000), whose typology was also used in the ceramic analysis in this study (see Chapter 8). The shell midden overlies silty clay sediment bearing red-slipped pottery. The upper layer of the Conciso shell midden is dated to 1220±25 uncal. BP or 1190-1060 cal. BP (NUTA21854), and the lower shell midden deposit to 1240±25 uncal. BP or 1270-1060 cal. BP (NUTA2-1851) (Ogawa 2002a:219). Both radiocarbon determinations are from animal bone.

Figure 3.10 Magapit pottery designs (photos courtesy of Peter Bellwood and Yoji Aoyagi) The Bangag site lies just across the river from Magapit and was initially explored by Aoyagi and Soriano in 1977. Kasuhiko Tanaka excavated the site in 1996, excavating a 1.5 x 1.5 sq m of almost 3 m in depth. Tanaka (1999, 2002) identified twelve layers in Bangag; layers 1 to 11 contain shells, and layers 3 to 5 are pure shell deposits. Layer 12, the lowest layer, is a silty clay deposit without shell. The shell midden primarily contains black and brown pottery, some with short incised lines along the rims or necks, and comb-incised designs that are either straight or zigzag. There are unrestricted and restricted vessels. Two AMS radiocarbon date were taken from organic materials sealed in sherds of the black pottery (Mihara et al 2004:410), the results being dates of 2170± 30 uncal. BP or 2310-2060 cal. BP (NUTA2-5367) and 2290 ±30 uncal. BP or 2360-2300 cal. BP (NUTA2-5368).

Another shell midden on the eastern side of the Cagayan River is Irigayen, excavated by Amalia de la Torre in several field seasons between 1995 and 1997. Irigayen has a clear stratigraphy, with layer 1 being the plow zone, layer 2 the shell midden, and layer 3 the silty clay (de la Torre 2000). The site was excavated to 2 m below the surface. Human skeletons in extended postures were excavated within the shell midden and were associated with black pottery and glass beads.

Layer 12 at the base of the site contains red-slipped pottery, although a few black sherds were also recovered. Unfortunately, Tanaka excavated only about 10 cm of this layer.

Layer 3, silty clay sediment with no shells, has rich cultural remains. These include complete and broken polished adzes, fragments of jade ornaments, clay lingling o earrings and red-slipped unrestricted vessels. De la Torre (2000:115) observes that in Layer 2, or the shell midden layer, only black pottery occurred, while in Layer 3, only red-slipped sherds occurred with associated shells.

Tanaka (2000, 2003) also excavated in the village of San Lorenzo, just northwest of Bangag, in 1997. The San Lorenzo shell midden has a well-stratified deposit with five distinct layers. Layer 1 is topsoil with fragmented shells. Layers 2 to 4 are alternating layers of shell with black and brown soil layers. The midden contains both jar burials and pit burials (Tanaka 2000). The pottery from layers 2 to 4 consists of restricted vessels with black, brown and a few red-slipped surfaces. The rims are everted, with both rounded and flat lips. Layer 5, which is a soil deposit, has both unrestricted and restricted vessels.

When I reviewed the archaeological inventory record for Irigayen, I determined that some sherds of black pottery still occurred in layer 3. In a personal communication with de la Torre (2005), she confirmed that the red-slipped pottery existed in association with some black pottery in this silty clay layer. The Irigayen shell midden is dated to 1490±35 uncal. BP or 1520-1300 cal. BP (O-W8700), while the silty clay layer is 3165±25 uncal. BP or 3470-3340 cal. BP (NUTA2-913), both determinations on charcoal (Ogawa 2002a:219). Hidefume Ogawa (2002c) has formulated a typology for the rim forms of the Irigayen red-slipped pottery that were used in the pottery description found in Chapter 8.

Tanaka (1998) excavated the Catugan shell midden site in 1996, which is also on the western terrace of the Cagayan River. Catugan is a well-stratified site consisting of five depositional layers. The upper layer 1 is the present cultivated soil. Layers 2 and 4 are both shell deposits with some soil, separated by layer 3, a soil layer with a few fragmented shells. Layer 5, on the other hand, is a loamy soil layer. The site was excavated to a depth of 310 cm below ground surface, with the shell midden being about 110 cm thick. Tanaka excavated human skeletal material within the shell midden in layer 2.

The joint Filipino-Taiwanese archaeological excavation of Nagsabaran was conducted in 2000 and 2001 (Tsang and Santiago 2001, Tsang, Santiago, and Hung 2001), and Hsiao-Chun Hung re-excavated the site in 2004. Nagsabaran is near the west bank of the Cagayan River, just northwest of the San Lorenzo site. The site was excavated to 300 cm below the surface and has nineteen stratigraphic units. Layers 1 to 14 are shell midden, while layers 15 to 19 are silty clay. Layer 18 contains a great deal of pottery, ,

Tanaka segregated the pottery of Catugan by layers; with layers 2 to 4 containing restricted and unrestricted vessels with brown and black surfaces. Some of the black sherds of layer 4 have short line incisions on the rims, while the brown pottery has some paddle impression and combincision (Tanaka 1998:159). All the earthenware sherds 29

30

On shell (Thiel 1989), red-slipped and black pottery

Layer 2 (Tanaka and Orogo 2000), red-slipped pottery Layer 3 (Tanaka and Orogo 2000), red-slipped pottery Layer 5 (Peterson 1974), red-slipped pottery Layer 5 (Peterson 1974), red-slipped pottery

Magapit

Pamittan Pamittan Dimolit Dimolit

3390±100 3810±200 3280±110 5120±220

3680 ±110

Gak-17967 Gak-17968 Gak-2939 Gak-2938

Gak-7048

unknown OOHM19 OOHM16 NUTA2-5367 NUTA2-5368 O-W8700 NUTA2-914 NUTA2-913 GX-26797 GX-26806 GX-38379 GX-28381 N5396 N5397

AMS date on rice husk in a red-slipped sherd (Snow et al. 1986) Upper shell layer,on animal bone (Ogawa 2002), black pottery. Lower shell layer, on animal bone (Ogawa 2002), black pottery AMS date on organic material sealed in black pottery (Mihara et al 2004) AMS date on organic material sealed in black pottery (Mihara et al 2004) Shell layer, Charcoal (Ogawa 2002), black pottery Silty clay layer, Charcoal (Ogawa 2002), red-slipped and black pottery Silty clay layer, Charcoal (Ogawa 2002), red-slipped and black pottery Upper shell layer, Charcoal (Tsang et al 2001), black pottery Lower shell layer, Charcoal (Tsang et al 2001), black pottery Silty clay layer, charcoal (Tsang et al 2001), red-slipped and black pottery Silty clay layer, Charcoal (Tsang et al 2001), red-slipped and black pottery Red-slipped pottery (Aoyagi et al 1997) Red-slipped pottery (Aoyagi et al 1997)

Andarayan Conciso Conciso Bangag I Bangag I Irigayen Irigayen Irigayen Nagsabaran Nagsabaran Nagsabaran Nagsabaran Magapit Magapit

3400 ±125 1220± 25 1240± 25 2170 ±30 2290 ±30 1490±35 3025±20 3165±25 1470±50 2150±150 3050±70 3390±130 2800±140 2760±125

Lab No.

Table 3.1 Selected Radiocarbon Dates from archaeological sites in the Cagayan Valley Site Context (source) Date BP

3890-3390 BP 4850-3650 BP 3850-3250 BP 6450-5250 BP

Oxcal 2 sigma calibration range 4000-3350 BP 1190-1060 BP 2160-2000 BP 2310-2060 BP 2360-2300 BP 1520-1300 BP 3470-3340 BP 3470-3340 BP 1520-1280 BP 2500-1700 BP 3400-3000 BP 4000-3350 BP 3350-2700 BP 3350-2700 BP

together with polished adzes and the skull of a water buffalo (Tsang, Santiago, and Hung 2001). Human burial pits cut through the shell midden deposit. The shell midden layers contain black, brown and a few red-slipped pottery sherds of both everted rimmed restricted and unrestricted vessels. Some have comb-incised decoration of straight or zigzag lines, and circle stamped designs. Lingling o earrings and glass beads were also recovered from these shell midden layers. The lower silty clay deposit has red-slipped pottery, some with incised linear patterns and stamped dentate designs (Tsang and Santiago 2001). Clay lingling o earrings, shell bracelets, nephrite stone and polished adzes were recovered from these layers. Hung (2005) observed two cultural layers, lower alluvial silt with red-slipped pottery, and upper shell midden with unslipped black pottery. Figure 3.11 Dimolit house plans (Peterson 1974)

During the recent re-excavation of Nagsabaran in November 2004, I observed the stratigraphic context of the recovered cultural remains. Some black pottery was associated with the red-slipped pottery in the lower silty clay level, as at Irigayen. The black pottery becomes dominant in the shell middenwherein the number of red-slipped sherd dramatically declines. Clay spindle whorls and three clay lingling o earrings with black surfaces were also recovered in the lower silty clay level.

Peterson (1974a) observed two cultural horizons. The upper contained red-slipped pottery, a broken polished adze and whetstone, and clay net weights. The lower cultural horizon had two complete house forms based on the excavated post holes (Peterson 1974a). There were flake tools made from jasper and basalt, red-slipped pottery, jadeite beads and sandstone mortars. Some of the vessels had ring feet with circular perforations. Peterson (1974b:34) reported radiocarbon dates of 3280±110 uncal. BP or 3850-3250 cal. BP (Gak- 2939), 3900±140 uncal. BP or 4850-3900 cal. BP (Gak-2937) and 5120±220 uncal. BP or 6450-5250 cal. BP (Gak-2938), all on charcoal. The oldest date of 5120 uncal. BP is problematic, considering that it is considerably older than the other two dates mentioned and from the same single-phase house floor.

The shell midden deposit at Nagsabaran has conventional radiocarbon determination on charcoal to 1470±50 uncal. BP or 1520-1280 cal. BP (GX-26797) for the upper part, and 2150 ±150 uncal. BP or 2500-1700 cal. BP (GX-26806) for the lower part. The silty clay level is dated to 3390±130 uncal. BP or 4000-3350 cal. BP (GX-28381) on charcoal (Tsang and Santiago 2001:12).

Ogawa (2000b, 2002a, 2002b, 2002c) has hypothesized that there were four chronological phases of pottery development in Cagayan Valley prehistory. The first had red-slipped pottery, some with stamped decoration as found at Magapit. The second had undecorated red-slipped pottery, as that stratified below the shell midden levels in many of the alluvial open sites such as Irigayen. The third had black pottery with incised decoration, of the type excavated from the lower shell midden layers at Bangag and Catugan. The final phase had undecorated black pottery, found in most upper shell middens.

Ogawa (2000a) also excavated Mabangog limestone cave, situated four kilometers east of Lal-lo, and recovered undecorated red-slipped pottery associated with flake tools. In the Municipality of Solana, two archaeological sites have been excavated. The first is Andarayan (Figure 3.5), excavated by Richard Shutler in 1978. This site yielded the first AMS radiocarbon date - 3400±125 uncal. BP or 4000-3350 cal. BPfor a rice husk inclusion in a sherd of red-slipped pottery (Snow et al 1986:5). The site also contained clay earrings, spindle whorls, chert flakes and fragments of polished stone adzes (Snow et al 1986). The other site is Pamittan, adjacent to Andarayan and excavated by Kazuhiko Tanaka and Alfredo Orogo in 1993. Cultural materials recovered included spindle whorls, clay ornaments, and fragments of polished stone adzes, chert flake tools and earthenware sherds. The pottery is mostly red-slipped, some with linear incision in triangular configurations with infilling punctate decoration (Tanaka and Orogo 2000). The radiocarbon dates, on charcoal, are 3390±100 uncal. BP or 3890-3390 cal. BP (Gak-17967) and 3810±200 uncal. BP or 4850-3650 cal. BP (Gak-17968) (Tanaka and Orogo 2000:124).

Kazuhiko Tanaka agrees with this general sequence of redslipped to black pottery. He stated that there was replacement of the people who made the red-slipped pottery by those who made the black pottery, and he believed that the former were Austronesian speakers, though he did not identify the latter (Tanaka 2002:508). He further suggested that there was a chronological gap between the red-slipped pottery occupation and the shell middens above, due to a high flooding phase of the Cagayan River (Tanaka 2004). According to Tanaka (2004:180), the pottery tradition of lower Layer 5 (red-slipped) is typologically discontinuous with that from upper Layer 4 (black pottery), with the red-slipped pottery disappearing before the occurrence of the black pottery.

On the eastern coast of Luzon, Warren Peterson in 1969 excavated the unique site of Dimolit, in Palanan Bay, Isabela Province, and recovered the remains of house post settings (Figure 3.11).

Tsang (2001) and Hung (2004) also agreed with the general chronology proposed by Ogawa, though Hung (pers. comm 2004) takes issues with Ogawa’s red-slipped decoration sequence of decorated followed by plain. Nagsabaran has shown that the decorated red-slipped pottery in fact occurs in the same stratigraphic context as the undecorated. 31

I do not agree with this general chronology for the Cagayan Valley pottery sequence. The recent excavation of Nagsabaran, (Hung pers. comm. 2004), and the archaeological inventory records for Bangag, Catugan and Irigayen, provide a different story. The red-slipped and black pottery styles occur together in the silty clay layers in these sites. Though the red-slipped pottery tends to be more frequent in this early stratigraphic context, black pottery is already present. This differs from Ogawa and Tanaka’s view that the black pottery only appears in the shell midden layer. It is unfortunate that most of their excavations were concentrated only in the shell midden layers, and rarely sampled into the non-midden layers below. The possibility that the black pottery found in the early non-midden layers was always redeposited from above is unlikely. The soil micromorphological analysis of Nagsabaran (Mijares forthcoming) reveals that there was minimal bioturbation, and there was also a 20 cm stratigraphic hiatus between the shell midden above and the red-slipped/black pottery assemblage below. I also do not agree with Tanaka’s (2004:181) view that there was a chronological gap or discontinuity between the redslipped and black pottery phases. The black pottery simply became dominant during the shell midden period, and red slipping declined in popularity. Hung (2004) has postulated that there is a link between the Middle Neolithic of Taiwan at c. 4000 BP and the Early Neolithic of Luzon, c. 3800 BP. Eastern Taiwan shows an increase in red-slipped pottery during this time, while in Luzon the earliest red-slipped pottery was already present, as at Andarayan (Snow and Shutler 1986). On the western side of Taiwan, the Fengpitou site has a pottery sequence that includes black pottery, clay lingling o earrings, and spindle whorls, in a shell midden deposit (Chang 1969). The earliest radiocarbon dated layer that contains black pottery at Fengpitou is 3722±80 uncal. BP (Y-1580), or 3839 (3632) 3443 cal. BP (Spriggs 2003:66). This shows that black pottery was already present in southwestern Taiwan prior to shell midden formation in the Cagayan Valley, at c.2000 BP. I believe that the red-slipped potteries with or without decoration, and associated black pottery, were already in use together at least 3600 years ago in the Cagayan Valley. The redslipped pottery tradition might have come from eastern Taiwan, the black pottery from western Taiwan. This Neolithic cultural material package would also have included clay lingling o earrings, spindle whorls, polished adzes, and shell and stone beads.

32

Pedro Pagulayan Caves, at 125 and 150 meters amsl elevations respectively. A 1 x 1 meter square was excavated to a depth of 40 cm below the surface in Alejandro Malanos cave. Earthenware sherds, mostly plain but some with incised designs were recovered. The sherds were associated with flake tools, animal bones and riverine shells. A thin ashy layer was identified as a possible hearth.

In 1978, Artemio Barbosa excavated Lattu –Lattuc cave, which is near Musang Cave. The cave is about 75 m amsl in elevation and 40 m above the river. There are 4 chambers in this cave and chambers A and D were excavated. In Chamber A, two 2 x 2 meter squares were excavated, and three 1 x 1 meter squares in chamber D (Barbosa 1979). Bedrock was reached about 1 m from the surface. Earthenware sherds were common and consistently recovered from the surface to the lowest excavated layer. They are brown, black and redslipped in colour, and vary in shape including jars, bowls and pots. A few chert and andesite flakes were also recovered. Animal bones (pig, deer, monkey and bat), riverine shells and land snails were also retrieved.

On the other hand, five 2 x 2 meter squares were excavated in Pedro Pagulayan Cave. The squares were excavated to a meter below the surface. Most of the pottery recovered had red-slipped surface finishing with no other decoration. The lithics comprised mostly of andesite flakes though there are a few chert flakes. Animal bones and riverine shells were also retrieved (Ronquillo 1977a).

Maharlika Cuevas’s team excavated Callao Cave in 1979 and 1980. Callao Cave is the biggest and longest cave in the area. With seven chambers, Callao has an elevation of 85 m amsl. The length is 366 m from the mouth to the innermost chamber, and the passages at 14 to 35 meters wide. The height of the cave ceiling varies from 10 to 45 m. A chapel was built in one of the outer chamber in the early 1970s. Cuevas’s team excavated a 4 x 4 meter square (see Figure 4.4) up to 7.5 m in depth (Cuevas 1980).

At the northern end of the Callao Limestone Formation, Barbara Thiel excavated two caves in 1976-1977: Arku and Musang (Thiel 1990a, 1990b). At an elevation of 250 meters amsl, Arku Cave is the highest cave excavated within the Peñablanca formation. Nine 2 x 2 meter squares were excavated. Thiel suggested that the cave was never used for habitation, but that it was primarily a burial area. Thiel estimated that at least fifty-seven individuals had been interred, some in thick-walled burial jars. A number of the bones had been burnt pointing to the possibility that these burial type included cremation.

Cuevas recognized two cultural horizons. The upper horizon is from the surface to 80 cm and contained of earthenware sherds, flake tools, a spindle whorl, beads, animal bones and shells. Earthenware sherds were recovered up to 50-cm depth. Flake tools of andesite, basalt, and chert were recovered up to 80 cm. There was intrusive human burial at 50-70-cm depth associated with Chinese monochrome pottery and metal implements.

Small vessels were also abundant, most red-slipped. There were also sherds of brown and black pottery, the latter highly polished. A number of ornaments were recovered that could have been burial goods. These include stone and shells beads, lingling o earrings of fired clay, stone and jade items, shell bracelets and shell pendants. Spindle whorls of conical and biconical shapes were also recovered. Bone points were also excavated, as well as one bark cloth beater.

The second cultural horizon that Cuevas identified is about 310-350 cm below the surface, and contains burnt animal teeth and bones, shells, and a high concentration of organic carbon. There is no associated artefactual material. For this preceremic horizon, Cuevas reported single carbon date of 5840 ± 140 uncal. BP or 7000-6300 cal BP (Gak-10529) date.

Arku Cave also produced a few stone flakes and two trapezoidal-sectioned ground-stone adzes. This is the only cave in the Peñablanca area where adzes and a barkcloth beater were found together. Thiel (1990a:259) suggested that the cave was used from 2000 BC until 75 BC (see table 4.1). Thiel also said that the cave might have been used by an agriculturalist population, based on the Neolithic assemblage that was recovered and the identification of lesions caused by porotic hyperostosis on one of the skulls, probably due to malaria-related thalassemia and sickle cell anemia. ‘Porotic hyperostosis is highly suggestive of agricultural subsistence but not conclusive’ (Thiel 1990a: 261).

Minori Cave was first excavated in 1980-81 by a team led by Fred Orogo (1982). It is the most northern excavated cave and has an elevation of 240 m amsl. Minori cave has four chambers, with chambers A and D being the archaeological areas. Seven 2 x2 meter squares were excavated in Chamber D and six in Chamber A. Chamber D of Minori Cave was re-excavated by me (Mijares 2002) in 1999 to recover more flakes for use-wear analysis. I excavated two 1 x 2 meter squares. Both the 1980 and the 1999 excavations observed two distinct depositional horizons, a lower with flake tools, animal bones and shells, and an upper with a similar assemblage with earthenware sherds. Some of this pottery has red-slipped surface finishing. The ceramic horizon is dated to 5190-5040 cal. BP (no lab number). Animal bones recovered from the excavation include deer, macaque, pig, monitor lizard, civet cat, snake, bird, and bony fish (De Vera 1983).

Musang Cave, on the other hand, has the lowest elevation of any excavated Peñablanca cave, 60 amsl and only 25 m above the river. Seven squares were excavated, two of 2 x2 meters and five of 1 x 1 meters (Thiel 1980). The excavation reached bedrock at 120 to 165 cm below the surface. Thiel also recognized two cultural horizons. The lower Cultural Horizon 1 contained flake tools, shells and animal bones. Analysis of the flakes is discussed in Chapter 7. The upper, Cultural Horizon 2, had the same assemblage with the addition of earthenware sherds and fired-clay lingling o earrings. More black than red-slipped pottery was recovered in Musang Cave. Thiel also observed that there were more shells in Cultural Level 1, but Cultural Level 2 had more animal bones. Cultural Level 1 is dated from 11,000 until 9000 uncal. BP. Cultural Level 2 has dates of 5000 and 4000 uncal. BP (Thiel 1990b:77).

A number of open sites were also identified during the 1977 explorations; mostly flake-tool workshops (Ronquillo and Santiago 1977). In Quibal Village, near Minori Cave, a grassy knoll about 100 m in elevation contained chert and andesite flakes and debitage and hammer stones. A similar open-site assemblage was also found in the Aggugaddan area, near Callao Resort. 34

35

4260± 380 4260 ±360 3690 ±310 3410 ±270 3130 ±220 2910± 230 5840 ±140 7830±170 4980 ±150 4110 ±130 9670 ±220 9390-±280 10,750 ±150 11,450 ±170 2460 ±80 2010 ±90 3040 ±130 2390 ±160 2740 ±120 4590 ±50

Rabel Cave Rabel Cave Rabel Cave Rabel Cave Rabel Cave Rabel Cave Callao Cave Laurente Cave Musang Cave Musang Cave Musang Cave Musang Cave Musang Cave Musang Cave Arku Cave Arku Cave Arku Cave Arku Cave Arku Cave Minori Cave

Rabel Laurente Alejandro Malanos Pedro Pagulayan Arku Musang Lattu-Lattuc Callao Minori

Cave Site

9 9 9 9 9

9 9 9

9 9

9

9

9 9 9 9

9

9 9 9

9

Adzes

Flakes

Flakes Faunal Bones

Ceramic Period

Preceramic Period Shells

9 9 9 9 9

9

9

9

Potte ry

9

9

Spindle Whorls

9 9

Clay earrings

Table 4.2 Summary of materials found in past archaeological excavation in Peñablanca cave sites

Charcoal, ceramic (Spriggs 2003) Charcoal, ceramic (Spriggs 2003) Charcoal, ceramic (Spriggs 2003) Charcoal, ceramic (Spriggs 2003) Charcoal, ceramic (Spriggs 2003) Charcoal, ceramic (Spriggs 2003) Charcoal, preceramic (Spriggs 2003) Charcoal, Preceramic (Spriggs 2003) Thiara scabra, ceramic (Thiel 1980) Thiara scabra, ceramic (Thiel 1980) Thiara scabra, preceramic (Thiel 1980) Thiara scabra, preceramic (Thiel 1980) Thiara scabra, preceramic (Thiel 1980) Thiara scabra, preceramic (Thiel 1980) Charcoal, ceramic (Thiel 1980) Charcoal, ceramic (Thiel 1980) Charcoal, ceramic (Thiel 1980) Charcoal. ceramic (Thiel 1980) Charcoal, ceramic (Thiel 1980) Charcoal, ceramic (Mijares 2002)

Date (BP)

Table 4.1 Radiocarbon dates from the Peñablanca cave excavation Site Material, Context, Source

9

9

Beads

2740-2350 BP 2200-1700 BP 3650-2750 BP 2800-2000 BP 3250-2450 BP 5190-5040 BP

5750-3650 BP 5750-3750 BP 4950-3250 BP 4550-2850 BP 4150-2650 BP 3650-2450 BP 7000-6350 BP 9250-8150 BP

9 9 9 9

9

9

9

Riverine Shells

Oxcal 2 sigma calibration range

9 9 9 9

9

9

9

Faunal bones

9

9 9

Human Burials

Gak-9932 Gak-9929 Gak-9933 Gak-9892 Gak-9893 Gak-9896 Gak-10529 Gak-7256 Gak-7044 Gak-7043 Gak-7045 Gak-7046 ISGS-497 ISGA-496 ISGS-495 Gak-7038 Gak-7041 Gak-7042 Gak-7040 Lab No not available

Lab. No.

conducted at the National Museum research station in Callao Resort.

In the southern area, near Dalan Serkot Cave, a chert source and workshop have been identified. This area is on the western side of the limestone formation and is about 110 m amsl. The exploration team recovered chert flakes, debitage and one andesite cobble tool.

Eme Cave The Eme Cave Complex lies about 2 km southwest of the National Museum Field Station at Callao Resort (Plate A1). It is composed of three caves under a good forest cover. The GPS coordinates are 17° 41.25’ N, 121° 49.71’ E. The cave complex was first discovered during the 1977 Archaeological Survey headed by Wilfredo P. Ronquillo and Rey Santiago (1977) of the National Museum. It lies at an elevation of 220 m above mean sea level.

Next I will discuss my own excavations carried in 2003, at Eme Callao and Dalan Serkot Caves. The methodology used is discussed first, followed by descriptions of the individual excavations.

Methodology My fieldwork was divided into three phases, each dedicated to the excavation of a particular cave site. Phase one was from August 4 to September 2, 2003, phase 2 from September 12 to October 16, 2003, and phase 3 from November 7 to December 4, 2003. The sites excavated were respectively Eme Cave, Callao Cave and Dalan Serkot Cave (Figure 4.2). This chapter will present the excavation results and archaeological contexts of the materials recovered. Stratigraphic descriptions will be presented in the next chapter. The archaeological field protocol was designed to guide the excavation and documentation processes. The protocol was discussed and reviewed by the different batch team members before the start of each phase.

Excavation Protocol Excavation was by arbitrary 5-cm spits modified to recognize natural-layer boundaries. Whenever a stratigraphic change was exposed, the remainder of the overlaying layer in the square was first removed before the new layer was excavated to avoid mixture of artefacts. Scraping was conducted within each spit to expose any soil features. A running number system was used for spits, starting from 1 at the top until the excavation was terminated. Loose sediment was removed using pointed and margin trowels, while archaeological picks, geological picks, and folding shovels were used on compacted and hardened sediments. All sediments removed were dry-sieved through a 2-mm screening table.

Figure 4.3. Plan of the Eme Cave excavation Two 2 x 1 squares were laid out as shown in Figure 4.3, and subsequently excavated. The northern square was designated SQ1 and the southern SQ 2. The Local Datum Point (surface) of SQ1 is 130 cm and of SQ2 is 133 cm below the datum point (DP). The DP was set on a limestone boulder against the north wall of the entrance. The excavation (Plate A2) revealed five natural layers (Plate A3). Layers 1 (Spit 1) and 2 (Spit 2) are fine ashy sediments, but the latter has small pebble inclusions. Materials recovered from these layers are mostly earthenware sherds and a few animal bone fragments.

Diagnostic finds were plotted three-dimensionally. Each flake, over 2 cm in any dimension was plotted, and then wrapped in plastic bubble sheet before being individually bagged. Earthenware sherds such as rims, footrings, bodyparts with designs, and sherds with food residue were also plotted three-dimensionally and individually bagged. The same was done for C14 -dating samples, such as charcoal. Shells, bones and teeth, body sherds and charcoal not for dating were collected per spit, to be later counted and weighed.

Layer 3 (Spits 3-10) is a compacted silty clay loam with earthenware sherds, stone flakes, lithic debris, land snails and riverine (Thiara sp.) shells, and animal bones, the latter including many small bat bones recovered during sieving. The sherds are mostly black in finish but there is also some plain brown pottery (Plate A4). About 340 lithic items were recovered, mostly from screening. We observed some deep vertical cracks in this layer due perhaps to periodic drying and shrinkage. Loose sediment from layers 1 and 2 had filled these cracks, and contains recent materials such as glass beads. Layer 3 has a radiocarbon determination of 1908 ± 74 uncal. BP or 2010-1690 cal. BP (Wk-14882) on charcoal from 48 cm below surface.

The Harris Matrix system was used during the excavation process to record relationships between layers and features as they were exposed (Harris 1989). A running number for each context (layer or feature) was used. The sediment pH was also tested for each stratigraphic unit. After the drawing and photographing of each trench profile, Kubiena tin box-oriented samples or an undisturbed sediment sample were collected at the boundaries between of natural layers. A spoonful of sediment for phytolith analysis was also collected at every 10-cm depth. Ten liters of sediments were collected per layer for flotation purposes. The flotation was

Layer 4 (Spits 11-18) is also a silty clay loam. No sherds were recovered. Flake tools and debitage of andesite, chert and basalt were the only cultural materials recovered. Substantially more lithic debitage was recovered from this 36

recovered (Figure 8.6a). This decoration is similar to some reported from the Magapit Hill site near the Cagayan River (Aoyagi et al. 1993). Human bones, stone flakes, riverine gastropods and land snail shells were collected. Four deer teeth, a wild boar tusk and nine other pig teeth were also found. Layer 4b is an intermediary layer in the northeast corner of SQ1. It is a distinctive reddish and loose sandy silt loam that may contain bat droppings. This layer was dated to 3335± 34 uncal. BP or 3650-3470 cal. BP (Wk-17010) by an AMS radiocarbon determination on charcoal.

layer than from layer 3 (1680 pieces as opposed to 340 for the same vertical thickness of deposit). Land snail and riverine shells, and teeth of deer (Cervus sp.) and pig (Sus sp.) were also recovered, as well as continuing large quantities of bat bones. The radiocarbon determination on charcoal from 67 cm below surface is 3569 ± 52 uncal. BP or 3990-3690 cal. BP (Wk-14883). Layer 5 (Spits 19-21) is again a silty clay loam, but no archaeological materials were recovered. Excavation stopped without reaching bedrock.

Callao Cave

Layer 5 (Spits 8-10) is a black sandy deposit devoid of cultural remains. However, it seems that this layer underwent some kind of burning event, which will be discussed in the next chapter. Though no artefacts were recovered in this layer. There is a reason to treat it as anthropogenic since the burning seems to have been due to human action.

Callao Cave (Plate A 5) has an elevation of 85 m amsl and coordinates of 17°42’ north latitude and 121° 49’ east longitude. The cave was first excavated in 1979-1980 by the team lead by Maharlika Cuevas (1980). In 2003, two contiguous 2 x 2 meter squares were opened next to the east wall of the cave entrance. SQ 1 is the northern square and SQ 2 the southern one (Figure 4.4). At 5 cm below the present ground surface a previous excavation was discovered, occupying the SW quadrant of SQ1 and the NE quadrant of SQ2. The back dirt was removed and the team discovered that this pit had been excavated to 132 cm below the surface. It was verified to be a treasure hunter’s pit, since most of the archaeological remains such as human bones, sherds and flake tools were not collected. The team used the walls of this pit as stratigraphic guide during the excavation. The cave deposits at Callao are generally undulating (Plate A6) so the excavation proceeded by removing natural layers in sequence, by 5-cm spits. Layer 1 (Spit 1) is a thin loose deposit with modern materials such as broken bottle glass, bottle caps and electrical wiring. Layer 2 (Spit 2) is more compact and contains Chinese glass beads, earthenware sherds, bones and lithic debris. Layer 3 (Spits 3-4) appears to be a late Neolithic deposit that contains shell beads (Plate A 7); clay lingling o earrings (Plate A8); brown, red-slipped and black earthenware sherds (Plate A9); flake tools; human bones (A 10) and teeth (Plate A11); bat bones, and riverine and land snail shells. Maharlika Cuevas’s excavation in 1980 recovered a spindle whorl (Plate A12) from this level. Thirty-eight shell beads were recovered from layer 3, from three to six mm in length and one to four mm in height. The perforation seems to have been drilled from both sides since they are hourglass-shaped, with diameters ranging from 1.5 to 2.5 mm.

Figure 4.4 Plan of the Callao Cave excavation

The human bones were disarticulated and fragmentary and might have been disturbed, even during prehistoric times. Two partial pelvic bones, a right and left, were recovered, belonging to two separate individuals. Visual assessment of the sciatic notches suggests that the right pelvic bone is female while the left one is male (Walker 2005:2).

Layer 6 (Spits 10-13) and Layer 7 (Spits14-20) are geogenic deposits devoid of cultural materials. Both are interbedded deposits of layers of cemented sand and very loose sand. Layer 7 has undergone an oxidation process, making it reddish in colour. A probable post-hole was observed in Spit 12 in the northeast quadrant of SQ1. It contained a few redslipped sherds and bat bones, and could have been dug from Layer 4.

Within this level a cemented deposit, probably due to calcium carbonate precipitation from the cave ceiling, occurred in the west central part of the two squares. This made excavation and recovery difficult.

In Layer 8 (Spit 21), chert flake tools were recovered (Plate A13). A probable hearth was also observed at the south end of SQ 1. Fragmentary burnt bones were recovered, but species identification was impossible. An AMS radiocarbon determination on charcoal for this layer is 25,968 ±373 uncal. BP (Wk14881).

Layer 4 (Spits 5-8) still contains pottery, but yielded no lingling o earrings or shell beads. Earthenware sherds with brown, red-slipped and black surface finishes continued to be recovered. A red-slipped carinated sherd with double parallel incisions enclosing a triangle with punctate in-filling was also 37

Layer 9 (Spits 23-29) was devoid of cultural remains but contained some burnt bone. Layer 10 (Spits 30- 32) was culturally sterile. The excavation ended without reaching bedrock, although the excavation units were not backfilled and there are plans to continue the excavation when funding is available.

Two squares were opened; SQ 1 (2x2 m) on the east side near the entrance and SQ 2 (1 x 2 m) on the west side and further inside the cave (Figure 4.5). The LDP is -33 cm for SQ1 and –47 cm for SQ2. Thick earthenware sherds, probably from burial jars, litter the current surface of the cave. SQ1, which is near the northern wall of the cave, also has limestone rubble at its surface (Plate A15).

Comparing the stratigraphy of the 2003 excavation with the 1979-80 excavation of Callao Cave, we can correlate the different contexts. Cuevas recognized two cultural horizons. In Cultural Horizon 1, the ceramic deposit ended at 50 cm, while flakes were still being recovered down to 80 cm. This implies a distinct preceramic below the pottery horizon. This separate preceramic horizon was not observed in the 2003 excavation as Layer 4 still has both sherds and stone flakes, then both suddenly ceased together in Layer 5. But Layer 8 is preceramic and occurs in both excavations.

Layer 1 in SQ1 extended from Spits 2 to 11 and contained earthenware sherds (mixed red-slipped, black and brown), human teeth and phalanges, human skull fragments and a few stone flakes. Some black sherds had incised designs on their rims and carinations (Plate A16). Land snail shells, two deer teeth and two pig teeth were also recovered. Layer 1 has a radiocarbon determination of 3530 ± 34 uncal. BP or 39003690 cal. BP (AMS date Wk-15648) on charcoal from Spit 10.

Cuevas also did not report a hiatus between his Cultural Horizons 1 and 2. His report indicates that from 80 cm to 230 cm there were no cultural materials and this correlates with layers 6 and 7 in my excavation. The differences in depth could mean a greater depth of sediment in the central area of the cave, compared to near the wall where my excavation was laid out. Cuevas’s Cultural Horizon 2, which contains burnt bones but no artefacts, could be correlated with layers 8 and 9 of my excavation. Layer 8 also contains burnt bones, associated with chert flake tools and hearths.

In layer 2a (Spits 12-20), no sherds were recovered but there was an increase in the numbers of chert and andesite flakes. A piece of charcoal attached to a chert flake from a depth of 70 cm yielded an AMS radiocarbon determination of 6124 ± 48 uncal. BP or 7260-6990 cal. BP (Wk-14879). Riverine gastropod shells dominated in this layer. Faunal remains include three deer teeth and a pig tooth. Small bones of bats and rat (Rattus sp.) were recovered throughout the whole stratigraphic sequence. Layer 3 was composed of an upper decomposing calcium carbonate, probably travertine, and a lower volcanic ash. No cultural remains were recovered from this layer.

Dalan Serkot Cave Dalan Serkot Cave (Plate A14) is located in the village of San Roque, which is south of the National Museum Field Station. The cave has coordinates of 17° 39.87’N, 121° 49.20 E, and an elevation of 165 meters. The mouth of the cave faces southeast. Dalan Serkot is about twenty minutes’ walk from Rabel Cave excavated by Wilfredo P Ronquillo in the late 1970s.

The stratigraphic sequence in SQ2 is similar to that of SQ1. Layer 1, which consists of spits 2-9, contained part of a human skull, teeth and phalanges, earthenware sherds of redslipped, black and brown pottery, and a few chert and andesite flakes. Chinese glass beads and a carnelian barrel bead were recovered in spit 3. These beads could have fallen from the surface through the cracks previously discussed. A probable chopper tool was found at 25 cm (Spit 5) below surface (Plate A17). Shells mostly of land snails and a few riverine shells, were recovered together with small animal bones (bats). Layer 2b contained cultural materials only in its upper part (Spit 10-13). These are mostly flakes, riverine and land snail shells, and small animal bones. No earthenware sherds were recovered from this layer. Spits 14-20 of this layer contained only small animal bones and shells. The sediment in this layer is mottled and contains bat droppings. This could be the reason why there were so few cultural remains recovered, since humans will avoid using any cave area with bats actively inhabiting it, particularly the deeper and darker sections. Layer 3 is the same calcium carbonate layer as in SQ1 but contained bat bones and one riverine shell. The excavation ceased without hitting bedrock.

Figure 4.5 Plan of the Dalan Serkot excavation 38

as bones of probable freshwater fish. Hunting of mediumsized animals such as wild boar, deer, monkey and reptiles is evident in the archaeological record. A simple flake technology using chert and later andesite and basalt was employed throughout (see Chapter 7 for discussion).

Table 4.3 Radiocarbon dates from the 2003 excavations Site Context, Date Oxcal 2 Lab depth sigma No. below calibration surface range Eme Cave Eme Cave Callao Cave

Callao Cave

Dalan Serkot

Dalan Serkot

Charcoal, Preceramic layer, 67 cm Charcoal, ceramic layer, 48 cm Charcoal, AMS, Preceramic layer, 112 cm Charcoal, AMS, ceramic layer, 40 cm Charcoal, AMS, Preceramic layer, 70 cm Charcoal, AMS, ceramic layer, 44 cm

3569± 52 BP

3990-3690 BP

WK14883

1908± 74 BP

2010-1690 BP

WK14882

25,968± 373 BP

These subsistence strategies as well as the lithic technology continued into the ceramic period, after c. 3500 BP. Pottery was brought into the foothills of the Sierra Madre, probably from the Cagayan Valley. The ceramic analysis will be discussed in Chapter 8. A number of new implements and ornament types were also introduced during this period, at around 3000 BP. Biconical and conical spindle whorls were recovered from Callao and Arku Caves. Fired-clay lingling o earrings were retrieved from Callao, Dalan Serkot, Arku and Musang Caves, while shell beads were recovered in Arku and Callao. Arku, Dalan Serkot and probably Callao Caves were used for interment of secondary burial jars about 3000 years ago. The polished stone adzes found in Arku cave were burial offerings rather than utility tools. They could have been exchanged materials later used as burial goods. No stone adzes were found in the other caves but they do occur in the red-slipped pottery layers in many of the Cagayan Valley sites.

WK14881

3335± 34 BP

3650-3470 BP

WK17010

6214 ±48 BP

7260-6990 BP

WK14879

3530 ±34 BP

3900-3690 BP

WK15648

The archaeological excavations in these caves show that postdepositional processes have occurred and that there has been some movement of artefacts through each stratigraphic profile. Cave sites, although generally having good stratigraphic sequences, can easily be affected by depositional and post-depositional disturbance (Spriggs 1999). The stratigraphic integrity of each site and their archaeological site formation processes will be discussed in the next chapter.

Discussion Except for Arku Cave, the Peñablanca cave sites all have similar archaeological sequences. Unfortunately, correlating the different cave sites based on radiocarbon dates is not straightforward. Thiel’s dates for the both ceramic and preceramic periods deposit in Musang Cave are suspect, since they are dated on riverine shells. Radiocarbon samples derived from marine shell can be contaminated by old radiocarbon, a process known as the ‘reservoir effect’ (Stuiver et al.:1986:980). This makes the dates too old (Stuiver et al.:1986:982), and it is necessary to correct them with a correction factor of 500 to 1500 years subtracted from the laboratory date (Spriggs 2003: 59). The correction is not straightforward since there is some regional variability in the marine reservoir age (Reimer and Reimer 2001:461). This problem could make the radiocarbon date about a thousand years younger than was reported by Thiel. Two dates from Rabel Cave that exceed 4000 BP for the ceramic period could also be problematic. I believe that the radiocarbon dates of c. 3600 BP for the oldest sherds in Callao and Dalan Serkot are more acceptable, since they conform to the radiocarbon dates for ceramics from the Cagayan Valley open sites (Snow et al. 1986; Tsang 2001). Spriggs (2003:60) also cautions that it is ‘unwise to push the Luzon Neolithic back beyond about 4000 BP at the most.’ Callao Cave has the oldest dated cultural remains from Luzon, dating to c. 25,000 BP. Unfortunately a cultural hiatus intervenes in the Peñablanca area between 25,000 BP and around 10,000 BP, if we consider the carbon date on riverine shell from Musang Cave. There are preceramic deposits in Dalan Serkot and Callao at c.6000 BP, but Eme still has preceramic deposits as late as 3500 BP. The preceramic deposits in the Peñablanca cave sites points to a hunter-gatherer subsistence strategy. Riverine shells were collected from the Pinacanauan de Tuguegarao River, as well 39

CHAPTER 5 Understanding site formation in Peñablanca Cave Sites

Geomorphology of caves

Cave sites are still the main source of archaeological information for the Pleistocene and Early Holocene periods in Southeast Asia. Hunters and gatherers used caves as intermittent campsites during the Late Pleistocene, and increasing usage occurred for burial during the Holocene period, especially in the Neolithic and later (Anderson 1997, 2005).

Caves, according to White (1988:60), are ‘natural openings in the earth.’ There are different varieties of caves but this book is concerned with karstic caves formed by dissolution of limestone through water action. ‘No two caves are exactly alike in their bedrock infrastructure, exposition, size, internal karstic relations, or influence of external geomorphological phenomena’ (Farrand 2001:538). Sediments inside caves may come from the weathering detritus of dissolving limestone. This may include quartz sand and chert nodules as well as clay minerals that may be authigenic that formed directly in the cave environment or were transported into the cave by fluvial and Aeolian processes (White 1988). Caves are semiclosed systems wherein sediments are deposited and built up and rarely move out. ‘Caves serve as traps for a variety of sedimentary types, and normally, whatever is deposited within a cave system remains there and is commonly modified and reworked in place’ (Courty et al 1989:194). In the tropical region sediments are also deposited from organic sources, especially bat and bird guano.

There are different approaches to understanding site formation in cave contexts. The approach used here is soil micromorphology, seldom used until recently in Southeast Asia. It was first applied in the region at the Tingkayu site complex in Sabah (Magee 1988) and at Gua Gunung Runtuh in Malaysia (Zauyah 1994). Recently, the approach has been applied in Niah Cave (Lewis 2003, Stephens et al 2005), and at Tabon and Ille Caves in Palawan (Lewis 2004). I have also applied the approach to a pre-shell midden deposit at Nagsabaran site in the Cagayan Valley (Mijares 2005). A soil micromorphology approach is used in this study to understand the site formation processes in the three excavated caves in the Peñablanca area, Cagayan Valley, northern Luzon.

Archaeological work in cave sites normally focuses on the mouth or entrance and any contiguous rock shelters. Rarely did prehistoric people inhabit the deep and dark interior of the caves. A number of processes act on archaeological deposits in caves that can change, preserve or destroy them. One of the most significant is chemical transformation, primarily caused by water that dissolves organic and inorganic materials (Karkanas et al. 2000:916). Bat guano may be oxidized by water, causing it to break down into phosphate, which reacts with the carbonates present in the cave dissolving them. Water not only causes dissolution but also cementation. Calcitic deposits can cement sediments, which might include archaeological materials, especially along drip lines. According to Anderson (1997:610) ‘among the more common difficulties encountered in cave archaeology in Southeast Asia is the carbonate cementing of the deposits associated with the formation of dripstones or percolation of calcareous ground water.’

The Soil Micromorphology Approach The soil micromorphology approach was developed initially in the soil science discipline and only in the late 1980s did it become fully utilised in archaeology. The practice of soil micromorphology has been standardised with the publication of the Handbook for Soil Thin-section Description (Bullock et al 1985), together with the key to the handbook by Stoops (1998, 2003), published to supplement and systematise the descriptions. The seminal work of Courty and her colleagues (Courty et al 1989) has made the approach relevant to interpretation of archaeological site formation. Understanding the intricate processes that form archaeological sites is an important component in interpreting past cultures. Understanding site formation includes the discernment of depositional and post-depositional processes, which included geogenic, biogenic and anthropogenic agencies. The use of soil micromorphology, or the analysis of undisturbed soil samples in thin-section, will provide the high resolution needed to understand formation processes. This approach has been used in the fine-grained analysis of site formation in the Peñablanca cave sites, Northern Luzon, and it complements other approaches towards understanding human activities during the early Holocene period.

Shahack-Gross and his colleagues (Shahack-Gross et al 2004) have researched bat guano and raised the possibility of identifying the type of bat (insectivorous or fruit-eating) based on the identification of authigenic minerals. Insectivorous bat guano has a low pH (acidic) and forms tarakanite phosphatic minerals. On the other hand, fruit bat guano would have an alkaline pH, producing dahllite phosphatic minerals (Shahack-Gross et al 2004:1270). Cave sites inhabited by insectivorous bats will have low preservation of bone due to the acidity, while those by inhabited by fruit bats should have better preservation.

Soil micromorphological analysis is carried out on undisturbed soil samples in thin-section, using a petrographic microscope. The technique involves taking a sample of undisturbed soil in a tin box driven into the section of the deposit (Fitzpatrick 1984, Goldberg and Macphail 2003), recording orientation, impregnating it with resin, and then cutting a thin-section to 25-30 µm thickness (Murphy 1986). The thin-section is then examined under a petrographic microscope, and the features described.

The pH level of cave sediment can either enhance the preservation of, or destroy, organic deposits. Charcoal tends to be resilient to biological decay and not affected by the pH of the sediments (Balme and Beck 2002:159). While charcoal tends to be concentrated in fireplace or hearth areas, starch in sediments, on the other hand, tends to be destroyed by heat (Balme and Beck 2002:164). Recovery of macrobotanical remains, primarily in a charred state, is often the best means for reconstructing the environment and diet of prehistoric cave dwellers.

The description of an oriented soil sample in thin-section can be divided into four parts: structure, mineral and organic components, groundmass, and pedofeatures. Structure refers 40

glue were removed, the sample was pressed and exposed to ultraviolet light for drying for fifteen minutes.

to the size, shape, and arrangement of particles and voids in aggregates and non-aggregates. Mineral and organic components are identified in relation to their sizes and volumes. Groundmass is described based on the coarse/fine (C/F) grain-size ratio, texture, birefringence, and type of fabric. Colour is identified in both plain-polarized light (PPL) and cross-polarized light (XPL). Pedofeatures are ‘discrete fabric units present in soil materials recognisable from an adjacent material by difference in concentration in one or more components … or by difference in internal fabric’ (Bullock et al 1985:19).

The mounted block was then cut to about 2-mm thickness, and then slowly ground down to 30-µm thickness in a Logitech LP 50 lapping and polishing machine. Finally, a cover glass was placed on the resulting thin-section.

The Archaeological Stratigraphy This section will describe the stratigraphic profile of each cave site, based on the field descriptions made after each excavation. The stratigraphic descriptions include colour, texture (clay-silt-sand), compactness (consistency-plasticitystickiness), continuity (continuous-discontinuous), boundary definition (sharp 0-.5 cm, abrupt .5-2.5, clear 2.5-6, gradual 613 cm, diffuse >13 cm), bedding (massive, laminated, crossbedded, graded, deformed) and structure (crumb, granular, blocky, prismatic, platy) (Courty et al 1989).

Based on the description of the soil thin-sections, an analyst can proceed to decipher different depositional and taphonomic processes. These will include the geogenic, biogenic and anthropogenic processes responsible for site formation. As stated earlier, geogenic sediments may originate inside the cave or may be transported in from outside through different mechanisms. Geogenic processes include diagenesis, weathering and erosion, and include cementation with calcium carbonate minerals. Movement of iron minerals through illuviation and eluviation leads to coating of voids and grains, and to a presence of iron-rich amorphous nodules.

Stratigraphic Description of Eme Cave Layer 1. 2.

Biological components are derived from the animals and plants that inhabit the cave (Farrand 2001:539), or which fall in or washed in from outside. Animal bones may be deposited when an animal dies in the cave, and notably land snails are normally abundant near the mouth of the cave. Biogenic processes involve soil faunal activity as well as root movements that shape and reshape the structure of soil and sediments. Worms create channels as they move through the soil and at the same time deposit pellety excrement that can be identified in thin-sections (French 2003:45).

3.

4.

Anthropogenic contributions are varied and differ from site to site. They can include living surfaces, hearths, artefacts, and structures (French 2003). ‘Anthropogenic contributions can result form either intentional or unintentional human acts’ (Farrand 2001: 542). Middens are formed primarily from food refuse, lithic debris from tool manufacture and discard, and ash and charcoal from hearths. Sometimes humans unintentionally bring sediments from outside inside the cave, attached to their feet or garments (Farrand 2001:542).

5.

Description (Figure 5.1) Crumb (fine) structure, massive bedding, 10YR 3/1 very dark grey loamy sand; loose, sticky and plastic. Crumb (fine) structure with small gravel inclusions, massive bedding, gradual smooth boundary, continuous, 10YR 4/2 dark grayish brown silty clay loam, loose-non sticky-plastic, pH 7. Sub angular blocky (fine) structure, massive bedding with charcoal and limestone pebbles, abrupt wavy boundary, continuous, 10YR 4/4 dark yellowish brown silty clay loam, firm-non sticky-plastic, pH 7. Sub angular blocky (fine) structure, massive bedding with charcoal and limestone pebbles, clear wavy anthropic boundary, continuous, 10YR ¾ dark yellowish brown silty clay loam, firm sticky-plastic, pH 7. Crumb (fine) structure, massive bedding, abrupt wavy anthropic-sedimentary boundary, continuous, 10YR 4/4 dark yellowish brown silty clay loam, compact, non sticky-plastic, pH 7.

Stratigraphic Description of Dalan Serkot Cave Layer

Thin-section Preparation

1.

This section describes the process used in preparing a thinsection, based on the materials used in my research. The thinsections were prepared in the Petrological Preparation Laboratory, Department of Earth and Marine Science, Australian National University. The samples were first impregnated using polyester resin, which is added slowly to the container in which the samples were taken. The resin enters the sample by capillary action (Murphy 1986), and is aided by placing the sample in a vacuum. After impregnation, the samples were placed in a well-ventilated fume cupboard for six weeks. When the resin hardened, the samples were placed in an oven for curing.

2.

3.

Description (Figure 5.2) Crumb (fine) structure and massive bedding, 7.5 YR 4/3 brown sandy silty clay, firm - plastic and sticky, pH6. Crumb (fine) structure and massive bedding, continuous wavy abrupt anthropic boundary; 10YR 4/4 dark yellowish brown clay, firm- plastic and sticky for SQ1; 5YR dark reddish brown clay loam, firmplastic-sticky for SQ2, pH 6. Crumb (fine) and graded bedding, discontinuous, wavy abrupt boundary, 10YR 7/1 light grey sandy loam, loose, non plastic- non sticky, pH 5.

Stratigraphic Description of Callao Cave Layer

The impregnated samples were cut using a diamond saw blade into blanks of 75 x 50 mm, and about 10 mm thick. The cut blank was then hand-ground to a flat surface using different grades of wet and dry carborundum abrasive paper (280, 320 and 400). The block was then mounted on a glass, using Loctite light-cured adhesive. After the bubbles from the

1. 2.

41

Description (Figure 5.3) Crumb (fine) structure and massive bedding, 10 YR 4/3 brown silt loam, loose plastic and non sticky. Crumb (fine) structure, massive bedding, abrupt smooth boundary, continuous, 10YR 5/4 yellowish brown sandy clay loam firm-plastic sticky.

which have produced the iron rich nodules. These oxidation features in thin-section are orange to reddish orange to red/black (French 2003). These alternating oxidising and reducing conditions could also signal the alternation of wet and dry environments. Another post-depositional process was precipitation of calcite (Plate C1.11), primarily micrite, which either fills voids, or coats grains and bones.

3.

Crumb (fine) structure and massive bedding, abrupt wavy boundary, continuous 2.5 Y 4/4 reddish brown clay, compact-plastic non-sticky, pH 6. 4. Crumb (medium) structure and massive bedding, sharp wavy boundary, continuous, 10 YR 4/2 dark grayish brown silty clay loam, loose very plastic sticky, pH 8 ½. 4b. Crumb (coarse) structure and massive bedding, clear wavy boundary, continuous, 10 YR 2/2 very dark brown silty clay loam, loose plastic sticky, pH 8 ½. 5. Crumb (coarse) structure and massive bedding, abrupt wavy anthropic boundary, discontinuous, 10 YR 3/2 very dark grayish brown sand, loose non plastic non sticky, pH 7 ½. 6. Granular (medium) and grading bedding, abrupt wavy boundary, discontinuous, 10YR 4/3 brown alternating calcified and loose sand, loose –firm- non plastic-non sticky, pH 6 ½. 7. Granular (medium) structure and graded bedding, sharp wavy discontinuous sedimentary boundary discontinuous, 10 YR 4/6 dark yellowish brown, alternating calcified and loose sand, loose to firm- non plastic non sticky, pH 7. 8. Crumb (fine) structure and massive bedding, gradual wavy boundary, continuous 10 YR ¾ dark yellowish brown clay, firm plastic sticky, pH 5 ½. 9. Crumb (fine) structure and massive bedding, abrupt wavy anthropic boundary, continuous, 10 YR 6/3 pale brown sand, firm-plastic- sticky, pH 5 ½. 10. Crumb (fine) structure and massive bedding, sharp wavy boundary, continuous,

Anthropogenic deposits include charcoal as well as charred plant residues (Plate C1.7), some of which still have intact structures and may derive from hearths or campfires. The pedofeatures such as the dark red to black- and-orange brown nodules (Plate C1.9), are probably fragments of burnt oxidised sediment. The Eme 2 sample, which represents the preceramic assemblage, is similar to Eme 1. Layer 4 is dated to 39903690 cal. BP (Wk 14883). This section is also heavily bioturbated, with identified roots and channels due to faunal activity. One small sherd fragment was identified, and perhaps moved downward by the bioturbation of the sediments. A rice husk dating to 107 uncal. BP (Wk 14884) was also recovered from a flotation sample in this layer. This rice husk must have fallen from the surface down one of the cracks mentioned above. Biological remains such as plant residues of probably woody origin and charcoal were widely spread through the section. Fragments of burnt sediment similar to the pedofeatures described in Eme 1 were also found (Plate C2.10). The bones from this section appeared visibly older than those in the upper section, since most are partially impregnated with calcite crystals (Plate C2.7). Precipitation of calcium carbonates was prevalent in impregnation of the groundmass as well as textural features.

Soil Micromorphology Analysis The descriptions of the different features of the sediment thinsections follow the terminology of Bullock et al (1985) and Stoops (1998, 2003). Interpretation of the features was derived using the approaches of Courty et al (1989), Fitzpatrick (1984) and French (2003). (see Table 5.1 for summary of descriptions, Appendix B for detailed thinsection descriptions, Appendix C for plates, and Appendix G for definition of soil micromorphology terms.)

The Eme Cave samples appear to reveal continuity in occupation from the latest Palaeolithic through to the Neolithic and later ceramic periods. However, the deposits appear to be seriously disturbed by faunal bioturbation and shrink-swell processes, with evidence for later cultural materials being mixed down into the earlier horizons. The extent of this disturbance is unclear, but it has definitely led to the movement of fine fragments of pottery and charred plant remains. Evidence of hearth deposits was noted in both thinsections, attesting to localised occupation activities, at least on a short-term basis, but the stratigraphic integrity of the deposits studied is questionable.

Eme Cave Site Formation Sample Eme 1 represents the beginning of the ceramic horizon in the cave, dating to 2010-1690 cal. BP (Wk-14882). It is highly bioturbated due to invertebrate faunal activity (Plate C1.4), which has reworked the sediments. Sediment mixing was already suspected from the observation of a number of cracks in the matrix. These cracks have vertical orientation and depths of up to 70 cm and might have been caused by the swelling and subsequent drying and shrinkage of the cave sediment.

Dalan Serkot Cave Site Formation The section where Dalan Serkot 1 was obtained is the transition from a lower lithic assemblage (Layer 2) to an upper ceramic assemblage (Layer 1). In the field, layers 1 and 2 were identified as sandy silt loams, but a textural description using a petrographic microscope identified them as silty clay loam. Layer 2 is radiocarbon dated to 7260-6990 cal. BP (AMS date Wk-14879). The massiveness of the deposit was observed in the field. The blocky structure as viewed under the microscope shows that the sediments have undergone wetting and drying events as well as compaction. Cracking of the sediment is visible throughout Layer 1, which can also be attributed to wetting and drying as in Eme. The reddish brown colour of the sediment may be due to the high amount of iron oxide, and good drainage. Amorphous nodules of iron rich minerals can be observed throughout the profile.

The bioturbation can also be observed macroscopically from the presence of circular poroids and channels with loose infilling, and the crumb structure of the sediment. According to French (2003:43), the soil thin-section takes on a pellety appearance called excremental fabric because the soil has been excreted from the gut of earthworms. Sections of roots were also identified in the slide. Clay illuviation, as seen in the coating of aggregates, grains and bones (Plate C1.5), points to hydromorphic processes in which moving water has redeposited clay-sized materials within the sediments. In addition, these sediments have been subjected to alternating oxidising and reducing conditions, 42

Figure 5.1 Eme Cave stratigraphy: 1 and 2 are oriented samples

Figure 5.2 Dalan Serkot stratigraphy: 1 and 2 are oriented samples

Figure 5.3 Callao Cave stratigraphy: 1 to 5 are oriented samples

43

44

Legend

Site/section Callao 1 Callao 2 Callao 3 Callao 4 Callao 5 Eme 1 Eme 2 Dalan Serkot 1 Dalan Serkot 2

cl=channel ck=crack po=poroids pv=packing void

Microstructure Porosity crumbs 5%, ck, po,ch crumbs 10%, ck, pv crumbs/granular 15%, sub-angular blocky 5%, ck,ch sub-angular blocky 5%, ck, po crumbs 15%, ch ,ck,po,pv granular 20%, ck,po angular blocky 5%, ch, ck sub-angular blocky 5%, ch, ck,

Table 5.1

Table 5.1 Discussion of the oriented sample results

q=quartz pl=plagioclase a=amphibole bt=biotite py=pyroxene b=bone s=shell lm=limestone cl=calcite o=olivine ch=radiolarian chert m=muscovite ad=andesite bl=basalt

Minerals lm,o,pl,a,cl,py,ch,q,ad,bl ad,lm,a,py,ch,cl,b q,a,py,ch q,pl,a,ch q,a,ch,pl,a b,q,pl,py,a,o lm, cl,bt,o,a,pl,q,b,ch lm,q,a,bt,py,cl, ch,b,s lm, q,py,a,pl,cl,m, b, s c:f ratio 15/85 12/83 2/98 2/98 2/98 10/90 10/90 15/75 7/93

ch=charcoal r=roots pr=plant residue

Organic ch,pr ch, pr pr,ch pr,ch,r pr,ch ch, r, pr pr, ch, r ch, r, pr

Summary of Soil Micromorphology Features Groundmass stipple speckle stipple speckle stipple speckle granostriated granostriated stipple speckle stipple speckle stipple speckle porostriation

mc=micrite coating ir=iron nodules cc=clay coating pse=pseudomorph ex=excrement dp=depletion

Pedofeatures cc,mc,pse,dp,ex,ir cc,mc,pse,ir,dp cc,mc,pse mc,ir, ex,cc mc,cc, ir. ex, mc, cc, ir, ex cc,mc, pse, ir, ex mc, cc, ir, dp, ex, pse cc,mc,ir

pt=pottery bs=burnt soil as=ash

bs as bs as, pt, bs bs,

Cultural pt bs,

There are areas in the sections where iron minerals leach out and take on a yellowish colour.

that area cemented with calcium carbonate and thus more difficult to excavate.

There is some evidence of bioturbation caused by faunal activity. Worm excrement was observed infilling channels (Plate C3.4). Precipitation or slow growth of calcite crystalscan be observed, with impregnation of the groundmass and coating of grains and voids with micrite crystals. Bones have also been subject to partial impregnation by calcite crystals.

The Callao 1 section was taken at the upper boundary of Layer 4, which has earthenware sherds, human skeletal remains and a few stone artefacts. This section has been heavily bioturbated, as can be seen from the channels and circular poroids made by faunal activity (Plate C5.4), and its crumb microstructure. Clay illuviation can be observed coating void walls and on grains and aggregates. Calcium carbonate precipitation was observed, with an impregnation of the groundmass as well as some bone (Plate C5.8). The fabric pedofeature described as a yellowish nodule in PPL and isotropic could be authigenic minerals (phosphatic minerals?) or decomposed bone. The zone of depletion in the mid section of the slide shows the loss of Fe- rich minerals, probably due to water activity. Anthropogenic deposits observed in the thin-section include small fragments of earthenware (Plate C5.6). There were also some burning activities in this level as seen in the abundant charcoal and charred plant residue.

Anthropogenic features primarily consist of charcoal and fragments of burnt sediment (pedofeatures 1 and 2), probably produced in hearths. There are some red dirty clay nodules, which may have been brought into the cave either by humans or animals. The lower section (Dalan Serkot 2) does not contain any cultural remains. The sediment overlaying the travertine layer (Plate C4.4) would probably be the floor used by the Palaeolithic occupants of the cave. The lower section also does not contain pedofeatures that can be related to hearths or cultural deposits. What is interesting is the volcanic ash deposit (Plate C4.8) below the travertine. The igneous minerals in this ash are sub-angular in shape, and so could have been deposited by a low-energy water flow. Impregnated sediment from this layer was subjected to Energy Dispersive X-ray Analysis (EDXA) using a Jeol 6400 Scanning Electron Microscope (SEM). A number of grains were analysed with the data generated using an SEM Quantitative Analysis in order to infer major elements and minerals. This analysis shows that the grains are rich in silicates. Identified minerals are plagioclase, quartz, ilmenite, clinopyroxene, iron oxide (magnetite) and hornblende.

Callao 2 section was taken at the lower boundary of Layer 4. An oxidising event occurred after a burning regime in this section of the stratigraphy. The burning of probably wooden material (Plate C6.4) produced abundant amorphous nodules and transformed the sediment into an iron rich disaggregated material. The burning regime may be associated with a hearth event. Fabric pedofeature 1 may be fragments of burnt clay sediment, while fabric pedofeatures 2 and 3 may be mixtures of burnt organics and sediments (Plates C6.8 AND C6.9). The burning regime might account for the dark brown colour of the section. There are also traces of illuviation due to the transport of clay materials by water. This can be observed in the coatings of grains, aggregates and bones. Precipitation of calcium carbonate was also active, with the pseudomorphising and coating of bones with micrite and sparite (Plate C6.5).

The Dalan Serkot thin-section reveals charcoal possibly related to human burning activity, along with ‘foreign’ clay nodules. The latter appear to derive from a source external to the cave, and may represent a sedimentary layer or lower soil horizon brought in from outside. The intact layer of volcanic ash was discovered through thin-section interpretation, and preliminary dates suggest it was deposited before 6,200 BP. This would represent an eruption, probably from one of the local volcanoes, during a time when people are certain to have been living in the area. A similar volcanic ash or tephra of andesitic composition deposit was observed in nearby Pedro Pagulayan Cave (Wasson and Cochrane 1979:21). Containing zoned plagioclase, brown hornblende, magnetite and ilmenite. They also concluded that the volcanic ash was washed into the cave.

The Callao 3 section was taken from Layers 5 and 6, both devoid of cultural remains. These layers have a distinct dark colour, as observed in the field. This is due to the high amount of charcoal in the section. There are light coloured zones (Plate C7.6), which may be of organic origin. Though layer 5 is devoid of cultural remains, the burning events could be anthropogenic in origin. A series of layers was identified in the field as alternating calcified and loose sand. It contained no cultural deposit. These layers (Layers 6 to 7), which are about 50 cm thick, correspond to an estimated 20,000 years of site abandonment. The reasons for the absence of cultural material or abandonment of visitation for such a long period are still unclear. The sediment was initially identified in the field as probable sand-sized calcium carbonate. An Energy Dispersive X-ray Analysis (EDXA) on impregnated sediment from this layer using a Jeol 6400 SEM invalidated this initial identification of the mineral. The EDXA result shows that the grains are high in silica (Si) and aluminium (Al), and are mostly probably volcanic ash. Minerals identified are plagioclase, quartz, ilmenite and possible garnet (almandine). Whether this volcanic deposit is related to the volcanic ash in Dalan Serkot and Pedro Pagulayan caves still needs to be demonstrated but is probable. Further analysis is needed to verify the relationship.

This volcanic ash layer could be used as a marker for other cave sites in Peñablanca dating to before 6000 BP. Directly dating this ash as well as conducting geochemical analysis on it will be important for future research. ‘Tephra layers may be dated using either documentary records or radiometric methods applied both to the tephra itself and to incorporated materials, and may be correlated between sites on the basis of chemical and optical properties’ (James et al 2000:330). Building up a better understanding of the relatively recent volcanic history of the region can produce local absolutely dated temporal sequences that would be useful to archaeologists, as well as providing valuable information on the prehistoric landscape.

Callao Cave Site Formation The 2003 excavation square of Callao was near the eastern wall of the cave. The western side of the excavation was along the drip line of an overhang, making the sediments in 45

The last two thin-sections, Callao 4 and 5, cut across Layer 8. These contained stone artefacts of chert and are dated to 25,968 ± 373 BP (AMS-Wk 14881). In the thin-sections this layer is the same as the lower sub-layer of Callao 4 and the upper sub layer of Callao 5. Layer 8 has a mixed blocky and crumb structure. There is evidence of bioturbation with large horizontal channels partially in filled with faunal excrement, probably vermiform (Plate C8.4). Remains of roots were also observed (Plate C8.9). The layer also has been subjected to wetting and drying events with its blocky structure, granostriated fabric and high iron-rich nodules. Precipitation of calcium carbonate in the form of micritic coatings and infilling of voids is evident.

Several different processes that occur in cave sites have been identified. Evidence for the occurrences of different biogenic processes was identified in the thin- sections, including: 1. Animal activity, such as burrowing holes (circular ‘poroids’) and excrement 2. Root residues Sedimentary and pedological processes include: 1. Water carrying clay-size particles (coating granules and infilling voids) 2. Illuviation and eluviation of clay and iron minerals, with depletions in some areas and concentrations in others 3. Precipitation of calcium carbonate and the fossilisation (?) of bones with micrite crystals. Anthropogenic deposits include: 1. Artefacts, such as lithics and small earthenware sherds. 2. Traces of burnt features such as hearths. 3. Occupation layers identified through increased charcoal volume. There is a comparatively large amount of charcoal (5-20 percent of the section) visible in those layers known to reflect active human occupation, compared to layers that are culturally sterile (about 2 percent).

During the excavation a probable hearth feature was observed in this layer. The feature has signs of burnt reddish sediment (Plate C8.11) and charcoal. The observation of the thinsection also reveals burning evidence of such as charcoal, burnt plant residues (Plate C9.7), ash nodules (Plate C9.5), and dark reddish brown nodules. The dark reddish fabric pedofeature observed in Callao 4.1 is related to a hearth event, since there are fragments of burnt sediment. This in situ hearth deposit at Callao cave was seen in the field, and dates well into the late Palaeolithic. Disturbed deposits apparently related to this burning was identified in thin-section. After a hiatus in cultural activity (at least in the location investigated), later disturbed hearth deposits were found micromorphologically at the base of the layer containing Neolithic or later funerary remains, associated with pottery and stone artefacts. It is uncertain whether the remains found in this later period of cave use relate to occupation or to burial, or both.

Through soil micromorphology it has been possible to characterise further the cultural deposits in the caves, as well as to add information regarding the preservation and integrity of the deposits. Eme Cave is highly bioturbated, so caution must be observed in interpreting the archaeological record. It was through inspection of the thin-section that small earthenware fragments were identified in a supposedly aceramic layer. Callao and Dalan Serkot also sustained bioturbation, but to a lesser extent.

Peñablanca Cave Site Formation Understanding the intricate details of the processes that form archaeological sites is an important component in interpreting past cultures. Site formation includes both depositional and post-depositional processes, including sedimentary, biogenic and anthropogenic agencies. The analysis of undisturbed soil samples in thin-section provides the high resolution needed to explore such processes. This approach was used in the analysis of sediments from the Peñablanca cave sites in Northern Luzon, and complements other approaches to understanding human activities there during the Late Pleistocene to mid-Holocene.

This discussion of cave formation gives us a context for discussion of the botanical and faunal remains in the cave, which are presented in the next chapter.

The sediments from these caves generally derive from a number of sources. The entrances of Eme and Dalan Serkot have rock and soil mounds due to roof fall. Inside the caves there are also pebble- to boulder-sized limestone pieces that litter the floors and occur throughout the accumulated sediments. Soil and clay from the landscapes above and outside the cave can be eroded in so that they contribute to the cave fill. This can particularly be observed in Callao Cave, which has a wide opening in its ceiling through which sediments can fall into the cave. Microscopic analysis of the thin-sections reveals a number of volcanic minerals such as plagioclase, amphibole, biotite and quartz in the sediment matrices. These minerals could have been derived from ash falls, and washed into the caves. Organic components of the sediments are contributed by bat and bird guano. Anthropogenic sources include lithic and ceramic artefacts, burnt materials from hearths, and plants and animals brought in the cave. 46

CHAPTER 6 The excavated bone and plant remains other bones were recovered, both small and medium in size, they were generally too fragmented to allow identification.

This chapter discusses the floral and faunal materials recovered from the cave excavations. These materials were analysed in order to reconstruct environmental conditions, as well as possible human diet. The specialist reports are in appendices H2, H3 and H4. Angel Bautista and Carmencita Mariano of the Zooarchaeology Section of the National Museum of the Philippines identified the faunal remains. Dr. Victor Paz of the University of the Philippines conducted the macrobotanical analyses of charred floral remains, with the assistance of graduate student Jane Carlos. Dr. Jeff Parr of Southern Cross University conducted the microbotanical analysis.

Bat bones and mandibles (Plate D1) were the most abundant faunal remains in all three cave sites, and were present throughout the stratigraphic sequence. Whether they were part of the human diet cannot easily be ascertained. In Callao Cave, the faunal identifications for the ceramic period consist of four deer teeth (Cervus sp.), nine pig teeth (Sus sp.) and one wild-boar tusk (Plates D2 and D3). These could represent one individual animal for each species. There was a single charred animal bone in the lower preceramic layer in Callao Cave, identified as deer (Cervus sp.)

There are a number of problems in attempting to reconstruct past environments and human diets from the Penablanca cave sites investigated. Faunal remain and other organic materials tend not to preserve well in acidic sediments. Cave sediments tend to shift from acidic to neutral through time, as acidic bat or bird guano interacts with alkaline water and other minerals in the sediment (discussed in Chapter 5). Organic materials in such cave sites have to be in a charred state if they are to be preserved. The macrobotanical analysis therefore focused on the identification of charcoal and other charred remains.

In Eme cave, the ceramic period deposit contained two pig and two deer teeth, and from the preceramic deposit fourteen deer and thirteen pig teeth. Dalan Serkot contained rat (Rattus sp.) bones through the whole sequence. The ceramic period faunal sediment also contained two deer and two pig teeth (Plate D4), and the preceramic period sediment two deer and one pig tooth. The identifications of the pig teeth by Bautista and Mariano (2004) are only at the genus level. These pig teeth could belong to the species Sus philippinensis (Groves 1981), as there were no other pig species in Luzon prior to the arrival of Austronesians, who presumably brought in the domesticated pig (Sus scrofa) (Groves 1997:181). An earlier observation that the bearded pig (Sus barbatus) might have migrated to Luzon from Borneo via Palawan during the Middle Pleistocene (Mudar 1985:72) links to the notion that the Luzon species is also part of the barbatus group. But the Sus barbatus in the Philippines is restricted to Palawan, and a revision of native wild pig species taxonomy in the Philippines by Colin Groves (1997:188) shows that Luzon wild pig is an endemic species.

There are also the problems of depositional and postdepositional processes that can affect the context of biological remains. Understanding such taphonomic processes involves ‘study of the generation, transport, deposition, diagenesis and preservation of fossils’ (Hunt and Rushworth 2005:465). As discussed in the preceding chapter, the sediments in the cave sites have undergone bioturbation, as well as shrinkage and swelling events, which could have caused mixture or movement of biological and cultural materials. This is particularly observed in Eme Cave, which was heavily affected by cracking due to drying and by invertebrate bioturbation (worms).

Faunal Identification One concern with interpreting any faunal assemblage is the size of the sample, because quantification by number of species and individuals has always been an important approach: ‘Quantitative analyses of temporal and spatial trends in taxonomic relative abundance have become common in zooarchaeological treatments of issues relating to subsistence and palaeoecology’ (Cannon 2001:185).

Molluscs were an important item of diet in the Peñablanca cave sites. Two species were recovered. Riverine shells of Thiara sp. (Plate D5) were probably collected from the Pinacanauan de Tuguegarao River. There is also a land snail (Plate D6) that is present in the environment near the cave that may also have been eaten. The shells were weighed and counted by species and differentiated by cultural layer. To make comparisons consistent the data are presented as total weights by cultural layer from a 2 x 2-sq m square (Figure 6.2). Callao Cave has the lowest density of shells, there being only 70 and 50 grams per 4 sq m of land snail and riverine shell, respectively, for the ceramic period layer. No shell was recovered from the preceramic layer in Callao Cave.

One type of behaviour that might have affected sample size and taphonomy would have been selective transport of animal parts by hunters. Depending on carcass size and carrying distance, a decision must be made either to carry a whole hunted animal, or to disarticulate the carcass. The decision over which parts to carry is ‘influenced by the values of the various parts, the difficulty of disarticulating parts from one another, and the difficulty of carrying each part home’ (Rogers and Broughton 2001:763). Different parts of an animal can be given a value depending on how much meat is attached to the bone—high for the rear and front limbs, but low for feet and crania.

During the preceramic period in general there were more riverine shells (653 g) than land snail shells (241 g). This trend was reversed during the ceramic period when land snails (619 g) became more common than riverine shells (49 g).

In the case of the excavated caves, long bones did not survive well and faunal identification was primarily based on recognition of specific animal teeth (Figure 6.1). Although

Eme Cave has the most abundant shells from the three cave sites. Riverine shells dominate in both preceramic and ceramic layers. Almost 2000 g of riverine shell were collected 47

from the preceramic layer, with only 416 grams of land snail shell. During the ceramic period about 740 g of riverine shell were recovered and 237 g of riverine shell. This is surprising since Eme cave is the highest of the three caves and therefore

is farthest from the river, yet it contained relatively more riverine shells than the other sites.

Callao

Dalan Serkot

Eme

Figure 6.1 Identified fuanal remains

Preceramic Layer Ceramic Layer Cervus sp bone

Preceramic Layer

Sus. sp tooth Ceramic Layer

Cervus sp tooth

Preceramic Layer Ceramic Layer

counts

0

2

4

6

8

10

12

14

16

2

Callao

Dalan Serkot

Eme

Figure 6.2 Comparative volumes of shells by 4m square Preceramic Layer Ceramic Layer

Thiara sp.

Preceramic Layer land snail shell

Ceramic Layer Preceramic Layer Ceramic Layer 0

200

400

600

800

1000

1200

1400

1600

1800

2000

grams

this study, as discussed in Chapter 4, a flotation sample of about ten litres of soil was collected per stratigraphic unit. Charcoal was also collected from the dry sieve by hand.

Macrobotanical Remains Archaeobotany focuses on the study of plant macro and micro remains to understand the relationship between plant resources and their utilization by humans in the archaeological past (Paz 2001,Wright 2003, 2005). The identification of charred botanical remains recovered from archaeological sites is one aspect of this analytical method (Paz 2005). For plant remains to survive the depositional and post-depositional processes that are typical of in the wet tropics, they have to be in a charred state.

Burnt botanical remains sometime have traces of plant structure, such as parenchymatous tissue (Paz 2001). Using such a macrobotanical approach, Victor Paz (2001) has researched the nature of the Austronesian dispersal through northern Wallacea, and the cultural transformations that occurred after this dispersal. In his dissertation, Paz showed that there are as yet no correlations between the presences of Neolithic pottery and cereal agriculture in northern Wallacea (Paz 1999).

There are a number of factors such as moisture content, atmosphere, length of exposure, and temperature that appear to interact and variably influence the carbonisation process (Wright 2003:582). The physical and chemical attributes of a plant also contribute to whether or not it will be preserved. Carbonised plant remains are normally collected through flotation. If the charcoal particles are sufficiently large they can also be collected by hand. The sampling strategy for flotation differs in each site, depending upon the time and budget constraints or the research design (Wright 2005). For

Charcoal and burnt plant remains were observed in the impregnated thin-sections of undisturbed soil samples from Eme, Dalan Serkot and Callao caves. Most of the burnt plant remains in these sections were identified visually as hard wood rather than food plants. Paz and Carlos (see Appendix H2) have analysed the plant remains recovered through flotation and charcoal collection from the Peñablanca sites. The samples were sorted using a 48

layer 8 and practically seal it. Although roots can still penetrate the layer and cause some level of bioturbation, this would only be minimal.

low magnification stereoscopic microscope (20-40x), and separated into wood fragments, parenchymatous tissues, nut fragments and seeds. These materials were counted and a sample measured. A few charred parenchymatous tissue samples were examined using a LEICA S440 scanning electron microscope (SEM). Cell dimensions were measured using Scion Image software, plotted in a scattergraph, and compared with the dimensions of known species in Paz’s reference collection at the Archaeological Studies Program, University of the Philippines.

Paz and Carlos examined seven parenchyma samples under SEM. All were altered and have an amorphous and glassy structure, making identification impossible. Charred wood was also examined by SEM (Figure 6.8)

Eme Macrobotanical remains In layer 3 in Eme Cave, or the ceramic period horizon, a large number of probably mineralised Boehmeria cf. platanifolia seeds (n=203) were identified. The wild ramie Boehmeria cf. platanifolia (Plate D9) is a relative of Boehmeria nivea (Chinese ramie), which is cultivated for its fibre. Charred parenchymatous tissues (Plate D9), charred pieces of nut and wood fragments (Plate D10) were also identified. Sclerotia fungus bodies are also common, these being able to survive periods of adverse environment in order to germinate during favourable moist conditions. Layer 4 (Preceramic) also has sclerotia bodies and a few Boehmeria cf. platanifolia seeds (Plate D11). Again, charred parenchymatous tissues (Plate D11), charred pieces of nut and seed (Plate 12), and wood fragments (Plate D10) were also present.

Figure 6. 8 SEM of Callao charred wood (bar=30 µm) (Courtesy of Paz and Carlos)

Dalan Serkot Macrobotanical remains Dalan Serkot has the fewest botanical remains. The upper ceramic period horizon (Layer 1) contains fragments of parenchymatous tissue, sclerotia bodies, a single mineralized Boehmeria seed, and a charred Spilianthes jacq sp. (Plates D16 and D17). The layer also contains charred wood fragments and spheroid seeds. A single piece of parenchymatous tissue was examined by SEM and its cell structure measured (Figure 6.8). A piece of charred wood was also subjected to SEM analysis. Similar to the Eme parenchyma samples, the Dalan Serkot sample also has small cell dimensions and does not match any of the samples from the reference collection.

Paz and Carlos subjected three parenchymatous tissue samples from the Eme ceramic layer to SEM analysis (Figure 6.3). The cell dimensions were measured, and compared with reference collection Colocasia (Figure 6.4), Dioscorea alata (Figure 6.5), Ipomea batatas (Figure 6.6) and Manihot esculenta (Figure 6.7). However, there was no overlap with these species in terms of cell size, and the Eme cell dimensions were smaller than those of the known plant species that they were compared to (see Appendix F2 for other charts). Paz and Carlos then surmised that these parenchymatous tissues might belong to wild plants due to its relatively small cell sizes.

Figure 6. 9 SEM of Dalan Serkot [a] parenchymatous tissue (bar=3µm) and [b] charred wood (bar=10µm) (Courtesy of Paz and Carlos) Figure 6.3 SEM of Eme parenchyma tissues (bar=3 µm) (Courtesy of Paz and Carlos)

The lower preceramic horizon at Dalan Serkot has no recovered parenchymatous tissue. One unidentified seed was observed (Plate D18).

Callao Macrobotanical remains

Microbotanical Remains

The flotation sample from layer 4 (ceramic period) contained wild ramie seeds (Boehmeria cf. platanifolia, n=15) and fungus (sclerotia). Two charred seeds of Spilianthes jacq sp. were also identified. This genus probably originated in the tropical Americas and could be an intrusion from above. Parenchymatous tissues and charred wood were also recovered from this layer, mainly through dry sieving (Plate D13). Charred nuts and seeds were also identified. The late Pleistocene deposit of layer 8 contains parenchymatous tissues and a number of fungi sclerotia (Plate D15). Paz and Carlos interpreted the presence of sclerotia in this layer as a possible sign of disturbance. Thick interbedded layers of cemented and loose sand of volcanic origin overlie

Phytolith analysis was recently developed as a tool for reconstructing past environments (Albert et al 1999, Fearn 1998, Kealhofer 2003, Piperno 1988), reconstructing past human diets (Jiang 1995), identifying stone-tool usage (Kealhofer et al. 1999), and understanding crop processing (Harvey and Fuller 2005). Phytoliths or ‘plant opals’ are amorphous silica mineral particles (SiO2 nH2O) of varying shapes and sizes formed inside plant bodies. All plants absorb dissolved silica from the soil and this silica is deposited in between plant cells, forming characteristic shapes.

49

Figure 6.4 Comparison with Colocasia cell size (Paz and Carlos Report) 300

C. esculenta

LB S-23

250

D alata-6

Short

200

D.alata-Cav

Appari

150

Lakapen

dalan serkot

100

eme

eme2

50

eme3 0 0

20

40

60

80

100

120

140

160

Long

300

Figure 6.5 Comparison with Dioscorea alata cell size (Paz and Carlos Report)

250

Short

200 150 100 50 0 0

20

40

60

80

100

120

140

160

180

Lakon Nabisaya Balatok Cavite Appari Taropoyo Madai-1-L7 Madai-L9 dalan serkot eme eme2 eme3

Long

Figure 6.6 Comparison with I. batatas cell size (Paz and Carlos Report)

Short

90 80

Lal-lo

70 60 50

Urdaneta Lal-lo2 dalan serkot

40 30 20

eme eme2

10 0

eme3

0

20

40

60

80

100

120

Long

Short

Figure 6.7 Comparison with Manihot cell size (Paz and Carlos Report) 45 40 35

Lal-lo

30 25 20

dalan serkot

Series2

eme

15 10 5

eme2 eme3

0 0

10

20

30

40

50

Long

50

60

70

80

1f (35-39 µm) – not clearly defined

When a plant dies or is burnt, the phytoliths are deposited in the soil and can be preserved for long periods of geological time.

Lentfer et al (2002:695) caution that sizes might overlap, but express confidence in this preliminary analysis.

Silica-rich plants such as grasses produce a lot of phytoliths with characteristic shapes that can be identified. For example, Fearn (1988) has identified Panicoid grasses with dumbbelland cross-shaped phytoliths, festucoid grasses with rondels and sinous phytoliths, and chloridoid grasses with saddleshaped phytoliths. According to Jiang (1995:483), rice (Oryza) phytoliths can be identified from three morpho-types: bulliform, ‘scooped’ bilobate, and glume epidermis phytoliths. Phytoliths can also be recovered from cave sites. In Tabun Cave, Israel, Albert and his colleagues identified wood and bark phytoliths from Mousterian ash deposits (Albert et al. 1999:1258).

The sediment samples for this research were processed in the Archaeology and Natural History laboratory at ANU using the protocols proposed by Jeff Parr (2001, 2002). The samples were then sent to Parr for identification of phytoliths. Only seven sediment samples were submitted – four are from Callao and three from Dalan Serkot. The sediment samples from Eme were not analysed owing to disturbance identified from soil micromorphological analysis. Of the samples analysed by Parr (see Appendix F3) only those from Callao had phytoliths, in low to medium densities. Dalan Serkot was devoid of phytoliths (Appendix F3 Tables 1 and 2). In Callao Cave, preceramic layer 8 (Callao Sample 11) has a high incidence of arboreal type phytoliths with associated charcoal. During excavation, a probable hearth was observed in this layer and was verified in the soil micromorphological analysis (Chapter 5). Grass phytoliths (Poaceae) account for 32 percent of identified microfossils from this layer, while the other 55 percent are arboreal types. Parr classified the hairycell arboreal type as probably belonging to the Moraceae family (Figure 6.10). The Moraceae are flowering plants that include Ficus (fig, banyan), Artocarpus (breadfruit and jackfruit), and Morus (mulberry). Moraceae were also utilised for their fibres.

Extracting phytoliths from soil and sediments can be a tedious and expensive process. The literature discusses different processes that can be used to remove clay and organics and to separate microfossils from heavier minerals (Fearn 1998, Lentfer and Boyd 1998, 1999, Parr 2001, 2002, Zhao and Pearsall 1998). The extraction process used for the Peñablanca samples followed Parr’s (2001, 2002) approach, with the use of a microwave digestion technique. The phytolith extraction was conducted by Gillian Atkins in the Archaeology and Natural History laboratory (RSPAS-ANU). Microwave digestion can be two-pressurised or focused, and Parr’s approach uses focused microwave digestion. The method entails placing the samples in a digestion tube with the addition of liquid HNO3 and HCL. This tube is then placed in a microwave sample preparation oven and digested for about thirty minutes. The sample is then decanted and filtered. Parr (2002:334) observes that the extraction of phytoliths is faster using this approach, which is also less expensive and eliminates the need for heavy liquid flotation.

The grass phytoliths are robust and include block, long and prickle cells (Figure 6.11). The arboreal phytoliths are sphere types as well as long cells and hair cells.

Once the phytoliths have been extracted, the sample is mounted in glass for microscopic analysis. Shapes or morphotypes are identified which will be the base for identifying plant origins. Besides phytoliths, starch grains can also be recovered from soil samples. Starch has been used in reconstructing palaeoenvironments, and is known to survive in tropical soils (Barton 2005, Lentfer et al 2002). Barton (2005:68) has been able to identify some of the starch grains recovered from Niah Cave to species level; Alocasia spp, Caryota mitis (palm), and yams (Dioscorea sp) and possibly Dioscorea alata) were all recovered there.

Figure 6.10 Possible Moraceae phytolith (Courtesy of Jeff Parr) Parr also identified four starch grains in this layer. They have diameters of 14-22 µm and 23-30 µm, corresponding to types 1d and 1e of Lentfer, Therin and Torrence (2002).

Lentfer and her colleagues (Lentfer et al 2002) have attempted to discriminate starch grains recovered from a modern field in Papua New Guinea into cultivated and wild types of vegetation. Using a number of statistical tools, such as principal component analysis of the sizes and shapes of starch grains, they were able to formulate size classes and types. The following are the subtypes based on size ranges, with their corresponding habitat groups (Lentfer et al 2002:695): 1a (1-5 µm) and 1b (6-9 µm) – gardens and coconut plantations 1c (10-13 µm) – logged forest 1d (14-22 µm) – palm forest of small volcanic islands 1e (23-30 µm) – forest and garden sites

Callao Cave Layer 5 was identified in the field as a dark layer, probably due to the burning of plants, but with no cultural material. This layer is dominated by grass phytoliths (Figure 6.12), accounting for 53 percent of the microfossils identified. Palms (Arecaceae), bamboo and sedges (Cyperaceae) (Figure 6.13) were also identified.

51

Figure 6. 11 Poaceae phytoliths: [a] block type, [b] long type and [c] prickle type (Courtesy of Jeff Parr)

Figure 6.12 Poaceae phytoliths: [a] bilobate type, [b] blocky and articulated, [c] bulliform, and [d] saddle type (Courtesy of Jeff Parr) Pinnularia spp., and Rhopolodia spp. These diatoms as seen together indicate a local freshwater system (Parr 2005 report, Appendix F3). Starch grains were also observed with a 6µm to 9µm size range, which falls within type 1b of Lentfer et al (2002).

Figure 6.13 [a] Bambusoid and [b] Cyperaceae (Courtesy of Jeff Parr) Of the palms found in this layer, there are some that are identical to Metroxylon sagu (Figure 6.14). Although Luzon is outside the current known distribution of Metroxylon sagu (Ruddle et al. 1978) (Figure 6.15), there are other palms used for sago making in Luzon. These include Arenga, Corypha and Caryota. Caryota cumingi, a sago-like palm, is abundant in the Sierra Madre and is gathered by Agta foragers (Griffin 1984, Griffin and Estioko-Griffin 1978). The starch contained in the trunk of this palm is used as a rich source of carbohydrates. There are also a number of other palm species in the Sierra Madre, such as endemic rattan palms (Calamus spp.).

Figure 6.15 Distribution of starch-bearing palm genera for sago making (from Ruddle et al. 1978) Parr has suggested that the abundant grasses, sedges, and the presence of diatoms, indicate the presence of a nearby wetland environment. The floor of Callao cave is located about 50 m above the Pinacanaun de Tuguegarao River, so these resources could have been brought in to the cave from the floodplain below the limestone formation. The reason why these grasses and sedges were brought in the cave during the period of Layer 5 is not clear, but a few hypotheses can be stated. As observed in the field and in the soil micromorphology thin-section, this layer is composed of burnt or partially burnt plants. Grasses and sedges were probably collected as fuel. The presence of human skeletal remains above this layer might also indicate an attempt at cremation. On the other hand, grasses and sedges could also be used to produce baskets and mats.

Figure 6.14 Phytolith identical to Metroxylon sagu (18 to 20 µm across taken at 400x magnification (Ccourtesy of Jeff Parr) Parr also identified a number of diatoms that resemble Achnanthes spp., Auloacoseira spp., Cyclotella spp.,

Layers 3 and 4 are ceramic period assemblages. Both layers have high occurrences of arboreal-type phytoliths. The palms (Arecaceae) in these layers closely resemble those of coconut 52

(Cocos nucifera L) (Figure 6.16). Grass (Poaceae) phytoliths also occur (Figure 6.17). In layer 3, bamboo phytoliths that resemble those of the species Schizostachym brachycladum (Figure 6.18a) were observed, as well as the distinctive black spherical and keel-shaped Myrtaceae phytoliths (Figure 6.18c) that resemble those of Syzigium brevicymum. A starch grain (Figure 6.18d) that is within the size range of 6 µm to 9 µm falls under type 1b of Lentfer et al (2002).

Figure 6.17 [a] Cyperaceae and [b] Poaceae phytoliths (Courtesy of Jeff Parr)

Figure 6.16 Phytolith resembling those of Cocos nucifera (7 x 11 µm taken at 400x magnification, courtesy of Jeff Parr)

Figure 6.18 [a] S. brachycladum, [b] Bambusoid, [c] Myrtaceae [c] starch 6 µm – 9 µm [d] starch polarized light (from Jeff Parr) Level II than in the lower Cultural Level I in Musang cave. Thiel (1990b:78) hypothesised a shift in subsistence from shellfish gathering during the preceramic period to hunting during the ceramic period. This view cannot be verified from the new data presented here, though Eme Cave had both shellfish gathering and hunting of animals during the preceramic period.

Discussion The hunter-gatherers of the preceramic period in the caves investigated probably had a broad-spectrum subsistence strategy, exploiting all possible resources in this rainforest environment. As stated above, the identification of animal remains is biased towards the presence of animal teeth. Wild pig and deer were the two most preferred medium-sized mammals to be hunted. Unfortunately, the animal bones from the 25,000- BP layer in Callao cave were burnt beyond identification, and no mollusc shells were recovered from this layer.

The analysis of floral remains, based on both microfossils and macrobotanical remains, points to the exploitation of wild forest products, specifically arboreal plants. No remains of domesticated cereals, tubers or yams were recovered. Starch grains were recovered in Callao Cave in both the preceramic and ceramic period deposits, but cannot be identified as to species. Parenchymatous tissues were recovered in most cave layers (except in preceramic Dalan Serkot). These could come from wild root crops, but cannot be identified to genus or species. The SEM examination of these samples shows that they fall outside the size ranges of known domesticated species. The presence of charred nuts in the ceramic layers in Eme and Dalan Serkot and in the preceramic layer of Eme adds to this picture of forest exploitation. Again, the species has not been identified, but according to Paz and Carlos the same type of nut has also been identified in an archaeological context in Ille Cave, northern Palawan Island.

The Holocene preceramic faunal assemblage contains pig, deer and riverine shell. Although only bats, rats, pigs and deer were identified in the three caves excavated in 2003, and in nearby Musang cave, some other cave sites in the Penablanca area have a greater diversity. For instance, Minori Cave yielded deer (Cervus sp.), pig (Sus sp.), rat (Rattus sp.), macaque (Macaca sp.), Malay false vampire bat (Magaderma spasma), monitor lizard (Varanus salvator), other lizards and snakes, felids (civets and cats), other carnivores, birds, and bony fish (Mijares 2002:34). Collection of Thiara sp. from the river was a significant component of the diet during the preceramic period, but this waned during the ceramic period. Thiel (1990b) observed a similar trend in Musang cave, which lies closer to the river. According to Thiel (1990b:77), ‘Cultural Level II only has about 5 percent of the amount of shells in Cultural Level I.’ On the other hand, there were more animal bones in Cultural

The high incidences of Boehmeria cf. platanifolia in the ceramic period layer in both Eme and Callao caves show a possible exploitation of this wild ramie for its fibre. In Callao Cave, a spindle whorl was recovered from this layer that was used to spin thick fibres. The ethnographic use of ramie in 53

Taiwan is still documented (Chen Chi-Lu 1969). Extracting fibre from ramie plants entails stripping the bark off the stem and soaking the inner fibres in running water. Other plants were also utilised for fibre or as matting materials. Moraceae phytoliths have also been recovered in most layers in the Callao cave sites, but the highest incidence is in the Ceramic period layers. The bark fibres of Moraceae can be extracted and used to make barkcloth. Grasses and sedges were probably used to make mats, baskets and nets. The environmental correlations for starch typology reported by Lenfer and colleagues (2002) show that palm forests (type 1d), and forests and garden (type 1e) environments existed in the region about 25,000 years ago. During the ceramic period at c. 3000 BP there was a change towards cultivated locations, with gardens and coconut plantations (type 1b). Lentfer and colleagues (2002) did not state their confidence limits in formulating these types. However, the environment of northern Luzon 25,000 years ago seems to have been a rainforest that might have included palm forests. There is no evidence for any cultivation in the area at the time. The environment might have changed during the ceramic period, with the introduction of agriculture by 4000 BP, particularly in lowland valley areas. But this scenario is doubtful in the areas in and around the caves. As we have seen, people here were primarily exploiting forest products and there is no evidence of cultivated plants. The Holocene hunter-gatherers of the Peñablanca Cave sites had relied heavily on hunting small to medium-sized animals, gathering shellfish from the river and exploiting wild starch resources from the forest for subsistence. Although arboriculture cannot be strongly argued in northern Luzon during the Preceramic period, it is also possible during the ceramic/Neolithic period (Kealhofer 2003, Latinis 2000). We can observe a high incidence of arboreal plants both in the micro and macro remains during this period. It is also clear that people were consuming wild root crops rather than cereals during the Neolithic (Paz 2002). If rice or other cereals were consumed during the ceramic period (c. 3000 BP) in the area, their remains should be present in the macrobotanical remains and the microfossil analysis. Paz (2002:279) has argued that rice has robust grains and husks, and if charred would survive and be identified in the archaeological record. Yet, there has been no evidence for cereal remains in the cave sites, unlike the situation in lowland sites such as Andarayan and Dimolit.

54

CHAPTER 7 The Unchanging Flakes? The presence of cortex, or weathered surface, on a flake tool is a good indicator of its stage in a sequence of reduction activities (Mauldin and Amick 1989:69-70). Such reduction stages extend from initial flaking in order to remove cortex from a core, to the ultimate bifacial thinning of a formal tool. Removal of cortex will allow a knapper to carry more preforms from a quarry site since the weight of the cortex has already been removed. Some flakes will have a cortexedstriking platform, this indicating a very early stage in reduction (Andrefsky 1998:77).

Flake Analysis Studies of prehistoric human adaptation and culture often rely heavily on stone tools. Their durable characteristics make them one the few cultural materials to survive over thousands or even millions of years. Knowledge of how a stone tool was manufactured, used and discarded can be used to infer past human behaviour. Stone tools can generally be categorized as simple or formal, the latter having undergone a chain of reduction and modification stages in order to produce a predetermined or ideal form. To make a formal tool, a knapper has to expend a relatively large amount of time and effort in selecting a quality stone material. With the time and effort invested in a formal tool, the user will tend to curate or conserve it.

The types of raw materials used in making stone tools can influence the forms and shapes of the end-products. Siliceous stone (SiO2) is the easiest material to work due to its homogeneity and isotropic attributes. ‘In flaking stone to make tools, the lithic materials normally favoured were the more homogenous and isotropic varieties of siliceous stone; that is to say, those materials that have the least directiondependent properties’ (Cotterell and Kamminga 1987:677).

On the other hand, simple tools are made with minimal effort and modification. There is minimal concern among the makers over the forms of the flakes, which appear in a variety of sizes and shapes (Binford 1983, Nelson 1991). After use, simple tools are discarded. Binford (1983) and Nelson (1991) describe this as expedient technology. ‘Expediency refers to minimised technological effort under conditions where time and place of use are highly predictable’ (Nelson 1991:64). These kinds of tools are seldom modified or retouched. ‘The extent of shaping by retouch will be conditioned by the task at hand, not by planned maintenance or reuse. Unretouched flakes and marginal retouched flakes are expected’ (Nelson 1991:80).

The availability and quality of raw material for making stone implements will also have an effect on the production sequence and subsequent human behaviour (Andrefsky 1994, Jelinek 1976, Straus 1980). This will entail familiarity by a knapper with the quality of the different rocks available in a particular environment. Knowledge of how a particular rock will fracture is an important component in the selection of stone material suitable for a particular task. The quality of stone material also plays a role in structuring tool production. Here, stone material quality refers to the ease with which the stone can be chipped and controlled in the shaping process. Very fine-grained homogenous stone materials tend to be more easily shaped and reduced than coarser-grained and flawed raw materials, and thus represent better-quality stone. Coarser-grained lithic materials are much more difficult to chip and shape in a controlled manner (Andrefsky 1994:29).

This discussion focuses on flake tools, since the stone implements recovered from the Peñablanca cave sites are primarily flakes, rather than pebble or core tools.

Physical Attributes What then is a flake? Flaked or chipped stone tools ‘are defined as those tools that have remains of an objective piece with recognizable ventral and dorsal surface’ (Andrefsky 1998:77).

Different regions will have varying types of rock suitable for stone-tool manufacture. Regions rich in high-quality stone materials such as chert will tend to have a greater range of tool types, and conversely, regions with poor-quality rocks tend to have fewer types of lithic implements. ‘An important underlying factor in the selection of raw material is the ease and extent to which different kinds of stone can be worked by the techniques available within the cultural tradition’ (Jelinek 1976:24).

Analysis of flake tools entails measurement of dimensions, identification of physical attributes, and identification of possible function or use. Different lithic analysts have used their own sets of attributes, with differing meaning given to terms, such that comparison between assemblages can be difficult. Michael Shott (1994) recommends devising a minimum set of attributes to make analyses comparable from different archaeological and experimental data sets. He proposes the use of the following parameters as a minimal set: weight, cortex area, dorsal scar count, platform angle, platform class, degree of completeness, and stone material (Shott 1994:99).

High-quality lithic stone materials such flint or chert can easily be formed into formal tools such as bifaces, points and retouched flake tools. Low-quality medium to coarse-grained raw materials can best be used for simple or expedient tools, such as unretouched flake tools, or flakes with minimal retouch.

The degree of completeness of the tool (Sullivan and Rozen 1985) is an important attribute in categorising debitage. A complete flake will have a striking platform, intact margins, and a feather or hinge termination. One difficulty is that a step-terminated flake can easily be misidentified as a broken flake, and not the result of a flaking failure. This can cause errors in analysis, since measurements of flake dimensions should only be done on complete flakes so as not to skew the statistical analysis.

Stone Tool Manufacture Understanding of the mechanics of stone tool manufacture were greatly improved with the publication Hayden’s book on Lithic Use-Wear Analysis (1979) and Cotterell and Kamminga’s work on Mechanics of Pre-Industrial Technology (1990). The mechanics of flake production can be

55

form of opaline silica (SiO2nH2O), found in plant phytoliths, together with water and the contact surfaces of the stone tool.

types of use-wear on stone tools: micro-chipping, striation, rounding or smoothing, and micropolishing. For the latter, Vaughan identified three stages: Generic weak polish - a dull polish, which is, nonetheless, somewhat brighter than the natural reflective background of the flint, yet definitely less reflective than a well-developed polish on the same surface. Smooth-pitted polish - individual small polish components, with more or less smooth surfaces, formed on the higher points of the microtopography of the generic weak polish contact area. Since the components are incompletely linked, numerous darker interstitial spaces are left between the components (pits). Well-developed polish - a further linkage of the polish components and the formation of diagnostic surface features on the polish surface occurs in the area of greatest contact on tool edges. Welldeveloped polish can be further classified, based on worked material, such as bone polish, antler polish, wood polish, etc. (Vaughan 1985).

Cornish and Watt (1966), in their experiment in polishing glass, state that polish is formed by a gradual uniform removal of glass. Chemical and physical interaction between the glass and the polishing agent weakens the glass matrix, forming a hydrolysed surface. The presence of water, and especially the plant silica, also facilitates the development of polish on stone implements (Fullagar 1991). The polish development on the experimental glass material is similar to the polish development on stone. During the use of a tool in many activities, it contacts materials that can cause abrasion and when sufficiently concentrated can take the form of ‘abrasive smoothing’ (some referred to as ‘polish’). As Kamminga notes: ‘The agents that cause abrasion on a stone tool during its use can derive from two major sources: foreign dust and sand that adheres to the worked material, and small particles, fragments or flakes that are detached from the contact surfaces of the tool’ (Kamminga 1979: 151).

Fullagar (1991) also recognises three stages of polishing, and emphasises the role of moisture as a polishing agent that can enhance polish development. For instance, scraping dry bamboo will give a different degree of polish from scraping fresh and moist bamboo. Fullagar’s (1991:6) polish stages are as follows:

The use of a microscope to identify use-wear was pioneered by Semenov (1964), who established how the kinematics of working hands relates to the formation of striations on the working edges of stone tools. Such striations may run parallel, diagonally or at right angles to the axis of the instrument, or its blade. They can go in one or several directions, be straight or curved, continuous or interrupted. Moreover, they have varied frequencies and lengths, as well as morphological characteristics (Semenov 1964). A transverse motion like scraping will produce striations perpendicular to the working edge, while a longitudinal motion such as sawing or cutting will produce parallel striations. Kamminga (1982:12) differentiates two types of striations: sleek and furrow. Sleek striations are extremely fine and have smooth and regular margins, while furrows are larger and have irregular margins.

Stage 1. From fresh fracture to abraded surface. This involves edge stabilisation and very slight rounding. Stage 2. From abraded surface to polish on the higher peaks. Predominant mechanisms in this stage are shearing and levelling of peaks, smearing, deepening of subsurface cracks and impaction of granular material in depressions. Stage 3. From polish on higher peaks to an extensive and stable polished surface. Predominant mechanisms during this stage are extension of subsurface cracks and flaking out of the surface, and gradual removal of surface defects. Stage 4. From an extensively polished surface to a completely polished and featureless surface. Water and an effective polishing agent are essential for this stage, and chemical interactions between them and the silica in the glass surface are the predominant polishing factors.

Lawrence Keeley (1980) used a high magnification microscope (200x) to identify general categories of worked material, using reflectivity, surface texture, topographic features, and distribution of polishes that a particular contact material will produce. However, there has been some doubt about his methodology and scientific rigor. Keeley’s former collaborator Mark Newcomer raised issues concerning whether a specific contact material can be identified using Keeley’s approach, and the validity of blind tests. Newcomer et al (1986:216) stated that we do not believe that microwear analysis should be abandoned, but that the investigation of stone-tool function should be based on a wide range of sources comprising evidence of tool morphology, edge damage and of course the usual suite of microwear features, including polishes.

Polish on stone tools can also be produced not only through use but also by post depositional factors (Levi-Sala 1986, Shea and Klenck 1993). Post-depositional surface modification (PDSM) can occur on a flake surface through soil movement, trampling and bioturbation. Levi-Sala (1986) forewarns that PDSM polish can be very similar to use-wear. Shea and Klenck’s (1993) experiments also show that trampling produces comparable kinds of wear to stone tool use, although they can be differentiated: ‘Use-wear is generally distinct from wear imposed by trampling and other sedimentary processes because use-wear is usually more concentrated on a specific part of the edge’ (Shea and Klenck 1993:178).

The problems with Keeley’s approach were his failure to recognise the importance of edge scarring or fractures resulting from use, and his heavy reliance on polish (Davenport 2003, Kamminga 1980). Yet, despite these problems with the validity of particular use-wear analysis, this approach is still regarded as a valuable contribution to lithic analysis. For instance, Vaughan (1985) identified four

Odell and Odell-Vereecken (1980) have been proponents of using low-magnification microscopes to determine use-wear on stone tools. At magnifications of 20-40x, they identified polish, edge-rounding, striations and different microflaking 57

terminations, and related them to the hardness of the contact material. In their experiments, they recognised different microflake scars for different tool actions. For instance, longitudinal motions such as cutting and sawing usually produce scars on both surfaces of the edge, with striations if visible near the edge and parallel to it. Transverse motions produce unifacial scarring. Odell and Odell-Vereecken (1980:101) concluded that small feather scar termination occurs on edges used on soft and medium-soft materials. Striations are usually faint and polish could be seen if the tool is used long enough. Medium to large hinge-scarring terminations with striations and polish could indicate medium-hard materials. Medium to large step-scar terminations indicate hard materials. Polish and striations are identifiable though they might be removed due to extensive scarring.

(1980), and Richard (1988) used basalt, a medium-grain volcanic rock. Vitreous and non-vitreous basalt flakes react differently to polishing, and non-vitreous specimens are not a good material for micropolish analysis (Richards 1988). Fullagar (1984) used quartz, which is the most common tool stone in Australia, for his experiments. Rhyolite, a coarse-grained volcanic stone, was used by Foix and Bradley (1985) for their studies. They cited the difficulties in flaking rhyolite, which requires a large force for flake detachment. Rhyolite has use-wear patterns that vary from small feather and step terminations to amorphous fractures with no observable patterning. Rounding and smoothing of edges was also observed. In their conclusion, they stated that: There appears to be greater resistance to change with coarser materials in comparison to the finer grained materials; however, change will tend to be more apparent on the finer grained rhyolites. This and the tendency for granular detachment which have amorphous fractures in the coarser rhyolites make use-wear analysis in higher magnification a necessity (Foix and Bradley 1985:119).

A fresh unutilised flake will normally have sharp edges, or as Kamminga (1982:17) describes it, will exhibit an angular cross section. When this flake is used the edge becomes blunted, and with continued use becomes rounded. The spot where edge rounding occurs can aid in identifying the kinematic motion of tool-use (Vaughan 1985, Richards 1988). In transverse (scraping) motion, edge-rounding is common on the leading surface. In longitudinal (cutting) motion both surfaces of the edge usually have equal rounding. Vaughan (1985:26) found the following general trends in his experiments: 1. Greater length of use produces more rounding. 2. Coarse-grained flint takes longer to reach a given degree of rounding than finegrained flint used for the same task. 3. Harder worked materials induced a given degree of rounding much more quickly than softer contact materials.

I (2001, 2002) used chert and andesite flakes on bamboo, rattan and meat. The experiment utilised andesite materials from the Cagayan Valley and chert from New Mexico. The actual experiment was conducted in Albuquerque, New Mexico, in 2000. Microscopic analysis was conducted at both low and high magnification. On working with andesite flakes, I (2002:73) stated that: At 200x or higher, one can see polish on ferromagnesian minerals, such as hornblende and pyroxene. Hornblende and Ca2(Mg,Fe2+)4(Al,Fe3+)(Si7Al)22(OH)2) pyroxene (X2Si2O6) minerals are silicates. Being silicates, they seem to behave similarly to chert (SiO2) in forming polish and striations.

To develop knowledge of use-wear attributes, experimental programs were employed. The resulting observations were then applied to archaeological flakes in order to infer usage and probable contact material. Different stone materials were fashioned into tools and used on different contact materials. The preferred stone materials tended to be those that exhibited isotropic properties and were fine-grained: The majority of recent experimental microwear studies are concerned with the development of wear patterns on flint, cherts and obsidian, while lesser efforts have been directed towards ‘coarser’ lithic materials, including quartzite, quartz, basalt and rhyolite (Richard 1988:7).

I compared the use-wear attribute of experimental flakes used to work bamboo, rattan and meat. He observed that bamboo tended to produce more use-wear than rattan, which, as the softer of the two materials, produced both generic weak and smooth-pitted polish on chert flakes. The flakes used to cut meat (deer and pig) had feather fracture scars. I (2002:54) concluded that: Bamboo and rattan as contact materials produced microwear traces like those of other hard materials. However, specific microwear signatures of both contact materials remain to be identified.

Obsidian (Corruccini 1985, Hay 1977, Kamminga 1982), and chert (Davenport and Kamminga 2002, Keeley 1980, Mijares 2002, Vaughan 1985) are the two preferred stone materials used in experimental research. Shale was the material used by Akoshima (1979) in his experiments, for which he used both low magnification to document edge scarring, and high magnification to identify polish and striations. The development of polish in shale is similar to that in chert, but may vary with the texture of the shale and the hardness of the contact material. ‘Very coarse shale seldom develops clear polish, while fine-grained shale easily does’ (Akoshima 1979:11).

Davenport (2003) recently differentiated the use-wear produced on flakes used on bamboo and rattan, based on polish distribution. ‘Although both rattan and bamboo are very silicious and therefore produce a high gloss, rattan is flexible and bamboo rigid, hence the distribution of the polish is different’ (Davenport 2003:63). He found polish on the flakes used to thin rattan running as deep as 5 mm from the edge. Flakes used on bamboo, on the other hand, have a narrow band of intense gloss less than 0.5 mm deep.

Flake Analysis in the Philippines

Medium- and coarse-grain stone is rarely used in such experiments. Kamminga (1982), Odell and Odell-Veerecken

Flaked stone assemblages in Island Southeast Asia, particularly the Philippines, are characterised by flakes and 58

flake tools, rather than by the pebble tools characteristic of the mainland Palaeolithic and Hoabinhian industries (Bellwood 1997). The first attempt to conduct systematic analysis of stone artefacts in the Philippines was made by Robert Fox (1970) for the Tabon Cave lithic assemblage. Fox (1970) published his preliminary analysis, but passed away before he could publish his intended final report. The Tabon Cave lithic industry has a deep time depth, beginning around 50,000 BP and continuing until at least 9000 BP. The flakes are of chert and Fox classified them according to a simple typology of primary flakes, flake tools with evidence of utilisation, flake tools with retouch, waste flakes and unused lumps of chert. He noted that ‘there was no basic change in the method-of-manufacture of the tools excavated from the first appearance of man in Tabon Cave until the final period of occupation; covering, as indicated, about 40,000 years’ (Fox 1978:64). Fox further stated that ‘there were no conscious attempts to shape the tools, no evidence of core preparation.’

flake tools used edges through a microscope indicates that some could have been used on bamboo and/or rattan vines. The used edges also exhibit angles, which could have been potentially used for whittling, cutting, and scraping wood, hide, and meat (Henson 1978:130). Wilfredo Ronquillo (1981) also worked on the lithic assemblage from Rabel Cave. To address the technological component of the analysis he measured a number of flake physical attributes. To infer probable flake function he used low magnifications ranging from 10x to 40x. According to Ronquillo: Functional analysis is interested in the usebehaviour of the artisan while technological approach focuses on the manufacturingbehaviour of the stone-tool maker (Ronquillo 1981:5).

Most of the original collection of artefacts that Fox studied has been missing since his death (Pawlik and Ronquillo 2003). A re-excavation of Tabon Cave was conducted in 2000-2001 and a new collection of excavated stone artefacts is now available for study. I undertook use-wear analysis as well as metrical measurement of these artefacts in 2001, and concluded that most of the flake tools had been used to work ‘hard’ material, probably wood or bamboo (Mijares 2004:1819).

He added that ‘the production of flake tools was the main objective of the knapper and the utilized flake tools are maintenance tools’ (Ronquillo 1981:11). Continuing the tradition of analysing lithic assemblages from the Peñablanca Caves, I (Mijares 2001, 2002) more recently conducted physical and use-wear analyses of flaked lithics from Minori Cave. Since no one had yet published any studies relating to use-wear on andesite flakes, I conducted experimental work, using chert and andesite flakes, on bamboo, rattan and meat. The results of these working experiments were then applied to the use-wear analysis of the archaeological lithics from Minori Cave. In 1986-1987, Johan Kamminga (Davenport and Kamminga 2002) conducted a large series of stone tool-use experiments, mostly with chert, siliceous tuff, and obsidian, among the Agta of Peñablanca. The Agta subsist primarily by hunting and gathering and some swidden farming along the Sierra Madre in northern Luzon (Griffin 1984). Kamminga employed these Agtas to use the experimental flakes on different materials and different activities. Subsequently, Daniel Davenport (2003) analysed these experimental tools to characterise the different types of use-wear, identifying wear patterns.

In the early 1970s, Warren Peterson worked on the eastern coast of Northern Luzon at the site of Dimolit (Peterson 1974). Dimolit contained ceramics associated with flake tools. Peterson conducted low-magnification microscopy at 8x-80x. He noted that some of the flakes displayed edge gloss, with two distinct types of distribution along their cutting edges. The first type is composed of silica polish that forms patches, lines and streaks along the edge and flake scar ridge. The second is more intense polish found only along the edge. Peterson conducted replicative tool-use experiments to identify the reasons for the two different distributions. He used jasper flakes on different contact materials, particularly grass, bamboo and rattan, and concluded that the experimentation indicated that the polish patterns differ on flakes used on bamboo and ones used for reaping grain” (Peterson 1974:151).

Peter Coutts and Jane Wesson (1980) also analysed flake tool assemblages from archaeological excavations on Panay Island in the central Philippines. They classified the flakes into the basic categories of tools and flaking debitage. Metrical measurements of each flake were taken as well as the angle of the working edge. The flake tools were also subjected to lowpower microscope analysis to identify edge damage. Based on this study, Coutts and Wesson (1980) wrote a critical appraisal of the flaked-stone industries in the Philippines. They identified two basic stone industries in the Philippines: a so-called smash-and-grab industry, in which cobbles were smashed with a hammer stone and the flakes selected for use; and a ‘blade industry.’ A true blade industry remains unsubstantiated in the Philippines since no prepared blade cores have been identified. Thus the so-called blades could simply be ‘blade-like’ flakes.

According to Barbara Thiel (1980, 1990b), Musang Cave had a flaked lithic industry that dated from c. 12,000 to 4000 BP. Thiel (1980) conducted edge-wear examination at 2x to 20x magnification to infer function, concluding that the flaked tools she identified had functioned as scrapers, spokeshaves, knives, grass-cutting blades, gravers, drills, and saws. She did not elaborate on how she identified these functions in relation to particular edge-wear attributes. Florante Henson analysed the flaked lithic industry of Laurente Cave (Henson 1978). He measured physical attributes of the artefacts and subjected the data to statistical procedures, such as analysis of variance and statistical intercorrelation, to test his hypothesis that there were no changes in lithic technology through time. Henson also used a microscope to infer possible tool-use. He stated: The analysis of data establishes the fact that the flake tools at Laurente Cave had been used as tools. These flake tools could have been used on animals. The examination of

Methodology The stone implements recovered from the three Peñablanca cave excavations described in this thesis were individually wrapped in bubble plastic after excavation to avoid postrecovery damage. They were then cleaned by soaking them in 59

microscope being feather, hinge, step, crescent break. 2. Degree of Edge Rounding is estimated as slight, intensive or absent. 3. Presence of Polish, identified according to the degree of light reflected from the surface under the low and high magnification stereo-microscopes as generic weak polish, smooth-pitted polish, and well-developed polish (Vaughan 1985): 4. Direction of Striations on the Working Edge, identified as parallel, perpendicular, oblique, and irregular. 5. Hand Kinematics, which can be either longitudinal (cutting, sawing) or transverse (scraping, whittling) (Semenov 1964) as inferred from the direction of the striations and the presence or absence of polishing on one or both faces.

a 10 percent HCL solution for thirty minutes (Levi-Sala 1996, Mijares 2002, 2004) and soaked in soapy water, lightly brushed to remove sediment, and finally rinsed with water and air-dried. The lithics were then sorted for further analysis. Flakes that had a good working edge, and were sufficiently large to be held by the thumb and the index finger (Mijares 2004), were selected for analysis along lines similar to those used previously for the Minori Cave lithics (Mijares 2001, 2002). The selected flakes were measured using sliding and spreading calipers, and a goniometer for measuring the edge angle. Length, width, thickness, weight, angle of working edge and length of working edge were all recorded, as were nominal or non-metric attributes such as the raw material, type of flake termination, presence of cortex and shape of the working edge (Mijares 2002). The use-wear analysis combined the techniques of both low magnification microscopy (cf. Odell and Odell-Vereecken 1980) and high magnification microscopy (cf. Vaughan 1985). This general approach is an aspect of ‘integrated function analysis’ first promoted by Kamminga and Richard Fullagar (1984) and later elaborated by Fullagar (1991).

The contact material is approximated according to the degrees of edge rounding, polish intensity, quantity of striations, and the type of edge scarring. Contact materials can be generally identified into soft and hard categories. If no use-wear is found, the flake is considered unutilised.

An Olympus SZX-9 reflected-light stereoscope with 6-60x magnifications was used for the low-magnification microscopy, and an Olympus BXFM reflected-light microscope with 100x, 200x, and 500x magnifications for the high-magnification microscopy (Figure 7.1). The following attributes for each item were then recorded, when present (Mijares 2002). 1.Type of Use Scar Termination, similar to flake terminations but smaller in scale, the main types visible under a low-power stereo-

Technological Analysis The analysis of the lithics commenced with the physical measurements and identification of nominal attributes of the flakes. The selected materials were grouped separately according to whether they belonged to the lower preceramic horizon or the upper ceramic horizon in terms of cultural context (Table 7.1). Eme Cave has the largest number of flakes analysed.

Figure 7.2 Olympus BXFM (left) and Olympus SZX-9 (right) microscopes (University of the Philippines Lithic Laboratory) To make the comparisons across cultural horizons and between cave sites, mean values were used in the data presentation. The mean values for length and width for the three sites, divided into preceramic and ceramic groups, and range from 2.96 to 3.27 cm and 2.66 to 3.04 cm, respectively. For thickness, the mean values range from 0.81 to 1.05 cm, while the edge angle mean values range from 46° to 50°. Mean values for weight are from 6.26 to 9.31 grams. Working-edge lengths (LWE in Figure 7.3)

Table 7.1 Number of flakes measured Ceramic Preceramic Horizon Horizon Eme 59 177 Callao 22 25 Dalan 16 21 Serkot

60

The lithics were all made either from volcanic rocks such as andesite and basalt, or from sedimentary rocks, primarily chert. Andesite was the most common raw material in Eme Cave, during both the preceramic and ceramic periods, with chert being marginally less common. In Callao Cave, chert was the only raw material used during the Upper Paleolithic period (c 25,000 BP), and remained dominant during the Ceramic period when it accounted for 63.4 percent, while andesite comprised the other 36.6 percent. At Dalan Serkot, andesite was the predominant material during the preceramic period, but became secondary to chert during the later ceramic period.

Serkot has about 60 percent of the flakes exhibiting no cortical surface, which could mean they had undergone further reduction, or were made from preform cores. Preform cores or prepared cores had undergone reduction to remove cortical surface. Flake termination and the shape of the working edge are two related attributes. Different flake terminations are produced according to the amount of energy applied to the striking platform and the follow-through motion of the knapper. A skilled flake knapper can produce a sharp feather-terminated flake. A hinge, step or plunging termination is produced when too little or too much energy is applied to the striking platform (Cotterell and Kamminga 1979, 1987, 1990, Kamminga 1982). Most of the flakes analysed were predominantly feather-terminated, pointing to the relatively high skill of the knappers who made them (Figure 7.12). Most of the assemblages are dominated by convex working edges, except for the Callao ceramic horizon and the Eme preceramic horizon, both of which have more irregular shaped edges (Figure 7.13).

The remaining coverage of cortex was measured to estimate the stage of reduction of each implement. More than 60 percent of the flakes in all six lithic assemblages exhibit surviving cortex over 50 percent or less of their dorsal surfaces (Figure 7.11). This shows that these flakes were used during relatively early stages in the reduction sequence (Andrefsky 1998:109). Differentiating the chert materials from the Dalan Serkot preceramic horizon with other materials showed a different result. Chert material from Dalan

Figure 7.10 Com parative m aterials used in m aking flakes

andesite

120

100

80

basalt

60

40

20

0

Ceramic Layer

Preceramic Layer Callao

Ceramic Layer

Preceramic Layer

Ceramic Layer

Dalan Serkot

Preceramic Layer

chert

Eme

Figure 7.11 Percentage of surface cortex observed

None

80 70

percent

60 50

50%

10 0 Ceramic Layer

Preceramic Layer Callao

Ceramic Layer

Preceramic Layer

Dalan Serkot

63

Ceramic Layer

Preceramic Layer Eme

Figure 7.12 Types of flake terminations

feather

90 80 70 60

hinge

50 40 30 20

plunging

10

percent

0

Ceramic Layer

Preceramic Layer

Callao

Ceramic Layer

Preceramic Layer

Ceramic Layer

Dalan Serkot

Preceramic Layer

step

Eme

Figure 7.13 Types of working edge

percent

70 60 50 40 30 20 10 0

concave irregular straight

Ceramic Layer

Preceramic Layer

Callao

Ceramic Layer

Preceramic Layer

Dalan Serkot

Ceramic Layer

Preceramic Layer

convex pointed

Eme

Vaughan 1985). Use-wear analysis on flake tools of volcanic rock such as andesite was first conducted by Mijares (2002, 2002). This study confirmed that use-wear analysis can be performed on andesite materials, and polishes and striations can be identified on crystal inclusions, such as ferromagnesian minerals (Plates E8, E12, E20, E24, E51).

Use-wear Analysis Most edge scars caused by damage on the flakes have feather terminations (Figure 7.14, Plates E1, E5, E58, E59), followed by hinge terminations (Plates E27, E31, E48, E60, E83), then crescent breaks (Plates E3, E13, E21, E33, E54 E117) and, lastly, step terminations. The majority of the flakes exhibit some level (slight to intensive) of rounding on the working edge (Figure 7.15, Plates E19, E52, E70).

Polish was observed mostly on the internal surfaces, but at least 50 percent of the flakes in each assemblage also have polish on their external surfaces (Figure 7.16). Generic weak polish (Plates E2, E10, E36, E45, E59, DE5, E88) is the most commonly observed type on the internal surfaces, with the exception of flakes in the Callao and Eme ceramic horizons, that have greater occurrences of smooth pitted polish (Plates E4, E49, E61, E69, E102 E110). A few flakes exhibit welldeveloped polish (Plates E40, E41, E71, E98, E116) on the internal surfaces. Again, generic weak polish is the most common type of polish on the external surfaces, with the exception of flakes in the Callao and Dalan Serkot ceramic horizons (Figure 7.17). The latter have greater occurrences of smooth pitted polish. Striation within the polished area of a flake is more difficult to see, and only few flakes have been observed to have them (Figure 7.18, Plates E16, E40, E100). The most common such striations are parallel and perpendicular to the working edge. Oblique striations were also observed. These data, augmented by the presence or absence of polish on one or both surfaces, help in the reconstruction of tool use. Except for the Callao preceramic horizon flakes, which have more scraping/whittling hand movement, the other assemblages have more cutting/sawing hand movement (Figure 7.19).

High-magnification microscopy identifies the extents of polishing and striation that develop when a flake is utilised. Longitudinal hand actions such as sawing and cutting produce striations parallel to the edge. On the other hand, transverse hand actions such as whittling and scraping produce striations perpendicular or at a high angle to the edge. The presence or absence of polish on one or both faces can also assist, in that longitudinal action will lead to polish on both faces, whereas transverse action will lead to polish on either just one surface, or on both but in unequal intensities (such as a combination of smooth-pitted and generic weak). In this study, the categories of use-polish are those of Vaughan (1985:29-30): ·generic weak polish ·smooth pitted polish ·well-developed polish Polishing can be easily seen on chert, and most experimental use-wear analyses have been conducted on chert and other siliceous stone materials (Kamminga 1982, Keeley 1980, 64

crescent break

Figure 7.14 Types of edge damage scar terminations 60

feather

50

percent

40

30

hinge

20

10

step

0

Ceramic Layer

Preceramic Layer

Ceramic Layer

Callao

Preceramic Layer

Ceramic Layer

Dalan Serkot

Preceramic Layer

none

Eme

Figure 7.15 Comparative flake edge rounding 100 90 80 70 60 50 40 30 20 10 0

absent

slightly

percent

Ceramic Layer

Preceramic Layer

Callao

Ceramic Layer

Preceramic Layer

Ceramic Layer

Dalan Serkot

Preceramic Layer

intensive

Eme

Figure 7.16 Types of polish on internal flake surfaces

generic weak

60

50

smoothpitted

40

30

20

welldevelped

10

percent

0

Ceramic Layer

Preceramic Layer Callao

Ceramic Layer

Preceramic Layer

Ceramic Layer

Dalan Serkot

Preceramic Layer Eme

absent

Figure 7.17 Types of polish on external flake surfaces 60

50

generic weak 40

30

smooth-pitted

20

10

well-develped

percent

0

Ceramic Layer

Preceramic Layer

Callao

Ceramic Layer

Preceramic Layer

Dalan Serkot

Ceramic Layer

Preceramic Layer

Eme

65

absent

The combination of the four use-wear attributes can help in inferring the possible contact materials. Initially, they were assessed as either hard or soft materials. Hard contact materials could be wood, bamboo or rattan. Since no bone tools, or bones with obvious cut marks, were recovered from the three sites, the possibility that the flakes were used to shape bone is minimal. Soft contact materials include meat, particularly from the locally hunted pigs and deer.

from a ramie plant, the bark of the stem must be stripped using a sharp knife (Chen Chi-Lu 1969:101). A simple chert flake tool could do this job. Since wild ramie is a fibrous plant, use-wear attributes would be similar to those used on rattan with a typical smooth-pitted polish and hinge or step scar terminations. Davenport (2003:53-54) discussed the use-wear attributes of stone implements used on palm wood and in different activity tasks. Scraping palm wood with chert caused step-terminated fractures, edge rounding and polish. The basalt flake used in scraping palm wood has more substantial edge rounding and exhibits more polishing.

A typical flake used on a hard contact material will exhibit hinge, step or crescent break-scar terminations, with slight rounding and smooth pitted to well-developed polish. It may or may not have visible striations. Flakes used on soft contact materials will have typical feathered scar terminations, with slight rounding and generic weak polish. Each flake was assessed based on the four attribute classes: for instance, a flake with a hinge scar termination and smooth pitted polish is assessed as having been used on a hard contact material.

Around half of each flake assemblage, except the Dalan Serkot ceramic horizon flakes, was used on a hard contact material (Figure 7.20). Some flakes did not show any usewear at all, despite having good potential working edges (Plate E93, E94).

Additional probable contact materials that were not part of the original experiment but were identified in the archaeological record are palm and wild ramie. In order to extract the fibre

Preceramic Layer

no visible striation

Dalan Serkot

Ceramic Layer Preceramic Layer

Callao

Eme

Figure 7.18 Directions of striations

Preceramic Layer

oblique

Ceramic Layer

perpendicular parallel

Ceramic Layer 0

10

20

30

40

50

60

70

80

90

percent

Figure 7.19 Comparative hand kinematics

Callao

Dalan Serkot

Eme

Preceramic Layer Ceramic Layer

cutting/sawing Preceramic Layer Ceramic Layer

scraping/whittling

Preceramic Layer Ceramic Layer 0

5

10

15

66

20

25

Number of Cases

Figure 7.20 Types of contact m aterials 60

hard

50

40

30

soft

20

percent

10

0

Ceramic Layer

Preceramic Layer

Callao

Ceramic Layer

Preceramic Layer

Dalan Serkot

Ceramic Layer

Preceramic Layer

unutilized

Eme

a few blade-like flakes from this period, they are very few and show no further modification. The more ‘formal’ stone implements from the previous Late Pleistocene seem to have ceased, and a simpler, more expedient lithic technology persisted (Mijares 2002).

Discussion The flake implements recovered from Callao, Eme and Dalan Serkot caves in general show little change in form or use from the lower preceramic horizons through the upper ceramic horizons. Simple flakes were used from the last glacial maximum c. 25,000 years ago until at least 2000 years ago.

The same raw materials and the same simple hard-hammer percussion technique persisted, even after the introduction of pottery from the Cagayan Valley about 3500 years ago. At Eme Cave, flake tools were still associated with earthenware pottery at around 1900 BP. The previous research by Thiel, Henson and Ronquillo in other Peñablanca caves confirms the long-continuing existence of this unchanging flake industry.

The earliest lithic assemblage, at Callao Cave, was mainly manufactured on chert. The flakes were manufactured with simple percussion techniques. However, the recovery of more blade-like flakes in the preceramic period in Callao Cave could signify some variation in the lithic tradition through time. The possible evidence for a use of spear or arrow points from two blade-like flakes hints at a more formal lithic technology.

Use-wear analysis of flakes from the preceramic horizons shows that about half were used on hard contact materials, possibly bamboo, palm and rattan, which are ubiquitous in the region. These activities might have included the manufacture of spears, bamboo knives, traps, or the making of mats. Some flakes were used in meat processing, as they exhibit soft contact use wear attributes. Bones of pig (Sus) and deer (Cervus) were associated with the assemblages. This range of use-wear was also observed on the flakes recovered from the ceramic horizons.

Unfortunately, we have yet no evidence from the Philippines for stone points made with the prepared-platform techniques reported by Glover from late Pleistocene Leang Burung 2 in South Sulawesi, or the bifacial techniques reported by Bellwood for the Tingkayu industry from Sabah, or the backing and serrating techniques used in the Holocene Toalian industry of South Sulawesi (Bellwood 1988, Glover 1977, 1981). Callao Cave is also the only Cagayan Valley assemblage dating to c. 25,000 BP found so far. The assemblage is also small, thus limiting our analysis and interpretation. We need to verify this lithic technology in other cave sites of the same time period.

Reaping knives with phytolith gloss, of the type used for harvesting grains, were not observed. Such a gloss was described by Peterson (1974: 151) at Dimolit in Palanan Bay, Isabela, as a: …distinctive sheen distribution pattern, striations perpendicular to the edge and comet-shaped pits, with the tail leading away from the edge.

At around 6000 BP, there was a change in Cagayan to using both chert and volcanic rocks, particularly andesite. The Pinacanauan de Tuguegarao River, which bisects the Callao Limestone Formation, carries many cobblestone-sized volcanic rocks from outcrops in the Sierra Madre. Most of the flakes, especially those of andesite, carry varying amounts of cortex. The cortical surface of each pebble was probably used as striking platform in producing these flakes. This can be seen because most flakes have cortexed striking platforms. The addition of volcanic rocks might signal a diminishing access to chert raw material in the area.

Daniel Davenport (2003:76) described the use-wear on experimental chert flakes used in harvesting rice as having high but diffused gloss distributed along the working edge, commonly 3 to 5mm deep. The flake tools from Peñablanca were manufactured using a simple percussion technique. The aim was to produce a good working edge that could be used for a number of tasks. There was no need to produce specialised tools, and the simplicity and expediency of the technology made such flake tools very adaptable in the region’s tropical karst environment.

All of the flakes from the three caves studied show no intentional retouch during this period (6000 –3500 BP). This signifies that the prehistoric Holocene people were not concerned with curating flakes, and they had sufficient raw material simply to knock off a new flake rather than retouch one that had become blunt or dull from use. Though there are 67

CHAPTER 8 Earthenware Ceramics and non-lithic Materials: Their Relevance to Interaction terms of shape, Thiel said that there were pots (restricted vessels) with everted straight rims, bowls (unrestricted vessel) with everted rims and large pots (restricted vessels) with everted and thick rim (1990b:70-71).

The Peñablanca Pottery Tradition As discussed in Chapter 3, in the late 1970s to early 1980s, nine cave sites near Peñablanca were excavated. The earthenware sherds from Callao Cave, Rabel Cave, Alejandro Malanos Cave, Pedro Pagulayan Cave, Laurente Cave and Minori Cave were counted and sorted based on surface colour, but were not subjected to more formal analysis. Rabel Cave was reported to have mostly black pottery, but also contained some red-slipped pottery (Ronquillo 1978). Minori and Pedro Pagulayan Caves also produced red-slipped pottery (Mijares 2002, Ronquillo 1977).

The next section discusses the methodological approach used in this study.

Methodology The pottery from the three newly excavated Peñablanca caves is here compared with that from the Conciso and Irigayen open sites in Lal-lo, Cagayan Valley (Figure 8.1), particularly in terms of form and surface finish. Ogawa (2002a, 2002b) has formulated a typology, which will be used in the following comparison, for the non-decorated red-slipped pottery from Irigayen and the black pottery from Conciso.

The Lattu-Lattuc Cave earthenware was subjected to a formal analysis by Dalupan, who reported both red-slipped and black pottery, the latter with incised decoration (Dalupan 1981). There were also yellowish red and brown coloured sherds. She also found three types of surface finish: plain, slipped and polished, and slipped and unpolished. The incised designs were found along lips, angled shoulders (carinations), and on body sherds. Dalupan did not conduct petrographic analysis, although she mentioned quartz, biotite and feldspar inclusions.

Petrographic ceramic analysis has seldom been used in archaeological research in the Philippines. Initial attempts to investigate ceramic petrography were made by Susan Naranjo (1984), for a case study during a SPAFA training course. Snow and Shutler (1985) conducted thin-section analysis on ceramic materials from Fuga Moro Island, near Palaui Island off the north coast of Luzon, and on some ethnographic pots from Luzon. The thin-section petrographic analysis of rim sherds from Nagsabaran, Magapit and Andarayan was conducted at ANU, and the results compared with those for sherds excavated from the three Peñablanca cave sites. Nagsabaran (Tsang, Santiago, and Hung 2001) and Magapit Hill sites (Aoyagi et al 1993) are also open sites in Lal-lo, Cagayan Valley. Nagsabaran contains both black and red-slipped pottery while Magapit has only red-slipped pottery, some of which have designs. The Andarayan site is also along the Cagayan River and contains only red-slipped pottery (Snow et al. 1986). A sherd from the modern pottery-making village of Atulu, 25 km north of Tuguegarao City (Figure 8.1), was also subjected to thin-section analysis, as a sample from known clay and sand sources.

Vessel Surface Treatment and Form Surface Treatment The excavated sherds could be separated into three categories based on surface colour: red-slipped, unslipped black, and unslipped brown. The red slip is a surface coating of diluted red clay or hematite, applied to the pot before firing. Firing using organic materials such as rice husks could have produced the unslipped black colour by producing a reducing atmosphere. Surface decoration was either incised or stamped.

Figure 8.1 Location map of pottery sites in Cagayan Valley Arku Cave functioned as a secondary burial location. In her analysis of materials excavated from here, Thiel (1990a) identified four surface colours: red-slipped, which is the most common; orange-brown; black; and dark red-brown. There were pots and bowls in different sizes, and cylindrical jars with straight or slightly curved sides and flat bases (burial jars?). A few sherds have incised designs on rims, incised (stamped?) circles, and mat impressions.

Form Rim sherds were separated into open unrestricted vessel and restricted and necked vessel, similar to the types recognised by Summerhayes’ (2000) typology for Lapita pottery in western Melanesia, and by White and Henderson (2003) for pottery in northeast Thailand. The basic differentiation between unrestricted and restricted vessels is taken from Shepard 1976 (Shepard 1976, Snow and Shutler 1985). Rim

At Musang Cave, Thiel (1990b) reported brown, black and red-slipped sherds, which according to her were made using a paddle and anvil and with sand temper (Thiel 1990b:70). In 68

J9- Everted rim with a convex inner surface and an outer thickening to the lip (vessel shape uncertain). J10- Everted rim of a restricted vessel with a thickened neck and a rounded lip. J12- Everted rim of a restricted vessel with an inflected rather than angular neck contour. J-13- Inverted rim of a restricted vessel with a thickened lip. B1- Everted rim of an unrestricted vessel, with a slightly thickened and widened lip.

forms can be further categorised based on direction, being either everted, direct (straight), or inverted (Summerhayes 2000, White and Henderson 2003). Rim profiles can be parallel, convergent (thinned) or divergent (thickened). Lip profiles can be flat, rounded or pointed. Rim thickness was measured using a pair of spreading calipers. The pot orifice diameter was also estimated using a series of concentric circles drawn at one-cm intervals. The following list gives Ogawa’s typology for those redslipped rims from the Irigayen open site in the Cagayan Valley that have equivalent forms in the excavated Peñablanca Cave sites (Ogawa 2002c). Each type has variations, as presented in Ogawa’s original article (Figure 8.2).

Petrographic Methodology Ceramic petrography has been a useful approach in archaeological research, particularly for characterising ceramic fabrics, inclusions, and manufacturing techniques. Archaeologists have used this approach to identify possible sources of raw materials. Sourcing can be used to group ceramics by site or region of manufacture (Fitzpatrick, Dickinson, and Clark 2003, Freestone 1991). In addition, sherds can be grouped according to temper characteristics, as stated by Williams (1983:304): The method is based on the premise that pottery made in a restricted geological area, and hence probably from the same formation, should contain sand grains of approximately the same size, all subject to the same degree of sorting and weathering. Petrographic analysis can also give some light on the manufacturing techniques used in the production of pottery (Peacock 1970:379). The petrographic analysis of ceramic thin-sections includes quantitative and qualitative descriptions of grain inclusions and fabrics. For the coarser sized inclusions, analysis includes the identification of the mineral/rock fragment, size distribution, sorting and shape. Sizes are measured along the longest axes of 100 or 200 grains (Darvill and Timby 1982, Kilmurry 1982). A comparator chart can be used to estimate the degrees of sorting by shape of the grains (Freestone 1991, Streeten 1982).

Figure 8.2 Red-slipped pottery rim forms from Irigayen (Ogawa 2002c) J1- Everted rim of restricted vessel, not thickened, and slightly concave internally. J8-Everted rim of restricted vessel, thickened at the neck, with a convergent lip. J11- Slightly everted rim of restricted and neckless vessel. The lip has a very short eversion and the lip is rounded. B7 Everted rim of unrestricted vessel, with a very short and slightly inverted or vertical lip The following gives Ogawa’s typology for those black ware rims from Conciso that have equivalent forms in the excavated Peñablanca Cave sites (Figure 8.3) (Ogawa 2002b):

Minerals are initially divided into isotropic and anisotropic groups. Isotropic minerals remain dark in cross-polarised light (XPL) and when the stage is rotated while anisotropic minerals remain generally light (Courty et al 1989, Nesse 2004). Minerals in plane-polarized light (PPL) exhibit distinct characteristics in relief, shape, cleavage, colour and pleochroism. Optical identification of anisotropic minerals in XPL is done in terms of properties such as birefringence, extinction angles, and twinning or zoning. Very fine-grained materials, such as the clay minerals, are beyond the capability of a petrographic microscope (Peacock 1970:379 ). The fabric of any ceramic can therefore be described in terms of its clay mineral colour, birefringence, and the amount and type of porosity. Petrographic analysis of the thin-sectioned sherds was conducted using a petrographic microscope with 35x, 100x and 500x magnification. The size/frequency distribution for each mineral inclusion was calculated by measuring 100 grains along their longest axes (Darvill and Timby 1982, Williams 1983). The shapes of the grains were recorded from a visual chart. A visual chart was also used to estimate the degree of sorting (Mackenzie and Adams 1994). Colours in plane polarised (PPL) and cross-polarised light (XPL) and birefringence were described for the clay matrix on both interior and exterior surfaces (Freestone 1991). Porosity was

Figure 8.3 Black pottery forms from Conciso (Ogawa 2002b) J1- Similar to J1 for the red-slipped pottery, but lacking the concave inner surface to the everted rim. J5- Everted rim with a thickened lip 69

The brown sherds from Dalan Serkot contain small notches around some carinations (Figure 8.5.6), and possible impressions of grass matting (Figure 8.5.4).

estimated using a frequency chart for voids (Bullock et al 1985). These different attributes were then used to characterise the fabric. Comparison by site between the different ceramic fabrics was then conducted.

The black pottery rims (Figures 8.6) from Dalan Serkot have notches on their lips, or sometimes around the carination. Similarly notched black pottery sherds are reported from the shell middens of Nagsabaran, Bangag and Conciso (Tanaka 2002, Tsang, Santiago, and Hung 2001 Ogawa, 2002c).

The Ceramic Analysis Vessel Surface Treatment and Form Surface Treatment The Peñablanca earthenware body sherds were counted and weighed according to surface colours (Figure 8.7). Black pottery dominated the Eme cave assemblage. Callao and Dalan Serkot Caves were dominated by brown pottery, and there were more black pottery sherds than red-slipped. The large number of brown body sherds in Dalan Serkot Cave perhaps reflects their origin from broken large vessels placed in the cave originally as burial jars. The small vessels are redslipped, black and brown; the large vessels are all brown; while all the unrestricted vessels are black or brown. Turning to rim sherds (Figure 8.8), black surfaces dominate the unrestricted vessel forms of both Callao and Dalan Serkot, and the small jars of Dalan Serkot and Eme. However, there were more red-slipped than brown or black from smallrestricted vessels in Callao Cave, and quite a lot of redslipped unrestricted vessels. Eme lacked red-slipped pottery entirely. The large restricted vessels in both Callao and Dalan Serkot consisted mostly of brown pottery, but Eme had no large restricted vessels.

Surface Decoration

Figure 8.5 Decorated sherds from Dalan Serkot: 1 and 6 small incised design along the carination, 2 and 4 grassimpressed design, 3 notch design, and 5 incised design below the neck

While no decorated earthenware sherds were recovered from Eme, both Callao and Dalan Serkot yielded a few. From Callao, a single carinated and red-slipped sherd carried, below the carination, a zone of two parallel incised lines forming a triangular frieze with punctate filling (Figure 8.4.1). Similar designs occur on the red-slipped pottery from Magapit and Nagsabaran. Callao also yielded a red-slipped and carinated lid (Figure 8.4.3), part of a red-slipped ring foot (Figure 8.4.4), and a sherd with three apparently decorative perforations that might belong to a ring foot (Figure 8.4.2).

Figure 8.6 Dalan Serkot black pottery rims with notch design

Statistics relating to vessel forms The diameter of each vessel at the lip was measured from the rim sherds, as shown in Figure 8.9. The range of small jar diameters measured from Eme Cave rims is 14-28 cm, with a mean of 22.5 cm. For Dalan Serkot and Callao the size range and mean for small restricted vessels are 6-32 cm and 17.5 cm, and 12-30 cm and 21.93 cm, respectively.

Figure 8. 4Callao Cave red-slipped pottery: 1) incised and punctuate design, 2) part of ring foot with perforations, 3) carinated lid, 4) part of a ring foot 70

Figure 8.7 Body sherd colours by site and w eight (gram s) 18000 16000

B lack

14000 12000

B ro wn

10000 8000 6000

Redslipped

4000 2000 0

Callao

Dalan Serkot

Eme

Figure 8.8 Comparative surface finishing of rim sherds per vessel type 120 100 80

Red

60

Black

40

Brown

20 0 percent

unrestricted

restricted

large restricted

unrestricted

Callao

restricted

large restricted

unrestricted

Dalan Serkot

restricted

large restricted

Eme

has a bigger rim diameter for red-slipped pottery than the black. The Eme Cave black pottery has the greatest rim diameter among the three sites.

For unrestricted vessel lip diameters, Callao has a range of 16-26 cm and a mean of 20.9 cm, and Dalan Serkot a range of 14-28 cm and a mean of 21.14 cm. The large restricted vessels in both Callao and Dalan Serkot have diameters of 30 and 31 cm, respectively. The large restricted vessels were probably used as burial jars since they were found with human skeletal remains in the same stratigraphic context

Sherd thicknesses were measured, as shown in Figure 8.11. Means for Callao and Dalan Serkot bowls are 0.88 and 0.91 cm, respectively—the difference here is insignificant. For small jars, mean thickness for Callao is 1.11 cm, Dalan Serkot 0.87 cm and only 0.65 for Eme. The large restricted vessel sherds in both Callao and Dalan Serkot have the same mean thickness of 1.4 cm, considerably thicker than for the smaller vessels.

If we group the rim diameters by surface finish (Figure 8.10), there is no difference in orifice size of the brown pottery in the three cave sites. Rim diameters of the black and redslipped pottery are almost equal in Dalan Serkot, but are significantly smaller than those from Callao Cave. The latter

Figure 8.9 Mean rim diameter by site and vessel type 35

30

centimeter

25

20

Diameter 15

10

5

0

unrestricted

restricted Callao

large restricted

unrestricted

restricted Dalan Serkot

71

large restricted

unrestricted

restricted Eme

large restricted

Figure 8.10 Comparative mean diameter of rims per surface color 25

diameter

20

15

diameter

10

5

0

red-slipped

Black

brown

red-slipped

Callao

Black

brown

Black

Dalan Serkot

brown Eme

Figure 8.11 Mean rim sherd thicknesses, by site per vessel type 1.6 1.4

centimeter

1.2 1 0.8

Thickness 0.6 0.4 0.2 0

unrestricted

restricted

large restricted

unrestricted

Callao

restricted

large restricted

Dalan Serkot

unrestricted

restricted

large restricted

Eme

Rim forms

5. Another everted rim like 4 above, except that the lip is only thickened on the exterior side (n=1) (Ogawa’s J-5 in Figure 8.3).

The earthenware sherds recovered from Eme Cave were all black and brown pottery. The black pottery has six major rim shapes (Figure 8.12):

The brown or orange brown pottery has three rim forms (Figure 8.13):

Figure 8.12 Eme black pottery 1. An indirect and long rim with a divergent profile and a flat lip (n=3) (Ogawa’s J-12 in Figure 8.3). 2. An everted rim with a convergent profile and lightly rounded lip. The interior of the rim is convex (n=1) (Ogawa’s J-1 in Figure 8.3). 3. An everted rim with a divergent profile. The upper part of the rim is convex towards the interior. The lip is thickened and rounded on the exterior wall (n=2) (Ogawa’s J-9- in Figure 8.3). 4. An everted rim with a divergent profile. The lip is thickened towards both interior and exterior, and the lip is flat (n=2) (Ogawa’s J-5 in Figure 8.3).

Figure 8.13 Eme brown pottery 1. A direct and converging rim. The lip is rounded. 2. An everted and convergent rim. The lip is rounded (similar to Ogawa’s J1 for black pottery). 3. An everted rim with an abrupt divergent profile. It has a flat lip (similar to Ogawa’s J5 for black pottery). The Callao red-slipped rim sherds included the following (Figure 8.14):

72

3. A restricted vessel with an everted rim and convergent profile. The external wall is slightly thickened (similar to Ogawa’s J8 for red-slipped pottery). 4. A restricted vessel with an everted rim, parallel in profile and with a flat rounded lip. 5. An everted rim with a convergent profile slightly thickened externally (similar to Ogawa’s J9 for black pottery). 6. A converging unrestricted vessel rim with a pointed lip (similar to Ogawa’s B7 for red-slipped pottery). 7. An unrestricted vessel with a thickened and pointed lip.

Figure 8.14 Callao red-slipped Pottery 1. An everted and tapered lip of uncertain vessel profile. The lip is rounded. 2. An everted rim with a divergent profile. The lip is rounded and thickened on the exterior (n=1) (Ogawa’s J-8 in Figure 8.2). 3. An everted rim with a rounded lip (n=1) (Ogawa’s J-1 in Figure 8.2). 4. An everted rim with a diverging lip, thickened on the exterior. The lip profile is flat (n=2) (Ogawa’s J-1 in Figure 8.2). 5. A converging unrestricted vessel rim with a pointed lip (n=3) (Ogawa’s B-7 in Figure 8.2). 6. A diverging unrestricted vessel rim with a rounded lip (n=1) (Ogawa’s B-7 in Figure 8.2).

Figure 8.16 Callao brown pottery The Dalan Serkot sherds belong to unrestricted vessels, small and large restricted vessels, the latter probably used as burial vessels The Dalan Serkot red-slipped restricted vessels have two rim shapes, as follows (Figure 8.17):

The Callao black pottery rims include the following (Figure 8.15):

Figure 8.17 Dalan Serkot red-slipped pottery 1. A thick everted rim with a tapered and rounded lip (n=1). 2. An everted rim with a relatively short eversion, with a rounded lip (n=7) (Ogawa’s J-11 in Figure 8.2).

Figure 8.15 Callao black pottery 1. An everted rim with a divergent profile. The lip is thickened on both the interior and exterior sides and is rounded (n=1) (Ogawa’s J-9 in Figure 8.3). 2. Another everted rim with a divergent profile. The lip is flat and only thickened on the exterior side (n=3) (Ogawa’s J-9 in Figure 8.3). 3. An everted rim with a convex interior surface and rounded lip (n=1) (Ogawa’s J-1 in Figure 8.3). 4. A dish or a lid with a flat base and an everted rim with a rounded lip (n=5).

The Dalan Serkot black pottery rims are as follows (Figure 8.18): 1. An everted and thickened short rim, thickened on both the interior and exterior sides and with a rounded lip (n=5) (Ogawa’s J-12 in Figure 8.3). 2. An everted rim, similar to that above but thickened only on the exterior wall (n=3) (Ogawa’s J- 10 in Figure 8.3). 3. An everted rim with a divergent profile and rounded lip (N=). 4. An everted rim, similar to 1 above but not thickened around the inside of the neck (n=2) (Ogawa’s J13 in Figure 8.3). 5. An unrestricted vessel direct rim that is divergent and has a flat lip (Ogawa’s B-1 in Figure 8.3).

The Callao brown pottery has the following rim forms (Figure 8.16): 1. A restricted vessel with an everted rim with parallel profile and rounded lip. 2. A restricted vessel that has an everted rim with high curvature and an abrupt convergent profile. The lip is rounded. 73

1. A divergent rim with a rounded lip 2. A parallel rim profile with a rounded lip. 3.A parallel rim profile with a heavy thickening of the external wall at the lip. 4. An abrupt convergent rim with a rounded lip. 5. A parallel rim with a rounded lip. There is a thickening at the mid internal wall. 6. A divergent rim with a rounded lip.

Petrographic Analysis Thin-sections were made on sherd samples from the three Peñablanca Cave sites and sherds from Cagayan Valley alluvial and shell middens sites. One sherd per surface colour and per site was thin-sectioned. Individual thin section description is presented in Appendix F and G, and the summary on Table 8.1. Major inclusions (temper) were recorded by frequency of grain size per type of inclusion. Individual thin-sections were compared within the surface colour groups (in this case, red-slipped and black only) in an attempt to identify correlations between the cave and valley sites.

Figure 8.18 Dalan Serkot black pottery rims The following are the Dalan Serkot brown pottery rim forms (Figure 8.19):

Comparison of Petrographic Analyses This section discusses the result of the comparison among the different sherds subjected to petrographic analysis to infer correlation among these pottery sites. Some trends were observed when the mineral and rock fragment inclusions for the red-slipped sherds were compared with those in the other groups (Figure 8.21). The thin-sections of the red-slipped sherds show a predominance of quartz and plagioclase grains. Presences and counts for minerals associated with igneous rocks—clinopyroxene, amphibole, biotite, and muscovite— were varied, but amphibole, which is typical of extrusive and plutonic rocks, is the most dominant.

Figure 8.19 Dalan Serkot brown pottery 1. A restricted vessel with an everted rim and divergent profile. The lip is tapered, with thickening in the external wall (similar to J8 of Ogawa’s red-slipped pottery). 2. An everted rim of a restricted vessel with high curvature. The rim has a divergent profile and the lip is pointed (similar to J10 of Ogawa’s black pottery). 3. A thick everted rim with a pointed and rounded lip. 4. A large unrestricted vessel that has a parallel profile and flat rounded lip. There is a slight thickening along the top of the interior wall. 5. Another large unrestricted vessel, but with a divergent profile and a flat rounded lip. The internal wall is also slightly thickend.

The size frequencies for the inclusions in the red-slipped sherds show a bimodal distribution, with medium-sized sand particles dominating. There is quite a lot of variability in the fine sand-grade sizes, and the Nagsabaran red-slipped sherd stood out since it has more medium-sized sand particles and few very fine sand particles. Comparing the overall matrices of the red-slipped sherds, there are two subgroups, based on the colour and birefringence of the interior core fabric. The first has a dark brown to black colour (PPL) with weak birefringence in the interior core. The second has a yellow brown colour (PPL) with moderate birefringence in the interior core. The first subgroup contains the Dalan Serkot, Nagsabaran and Andarayan red-slipped sherds. The second includes the Callao and Magapit red-slipped sherds.

There are a number of large restricted vessel in Dalan Serkot that might have been used as burial jars (Figure 8.20). These are brown in colour.

The black pottery group (Figure 8.22) also contains quartz and plagioclase as the major inclusions. It has the same igneous rock minerals as the red-slipped pottery, such as clinopyroxene, amphibole, biotite and muscovite. The Peñablanca black pots have more pyroxene mineral inclusions, whereas amphibole minerals are more common in the Nagsabaran black pottery. The grain size counts for the black pottery are also bimodal in distribution, with coarser grain sizes predominating. The trends for all sites are close, with Eme black pottery exhibiting a finer average grain size.

Figure 8.20 Dalan Serkot large brown rims

74

75

Poorly sorted

Poorly sorted

Moderately sorted

Poorly sorted

Poorly sorted

Moderately sorted

Moderately sorted

Moderately sorted

Moderately sorted

Poorly sorted

Callao Red

Dalan Serkot Red

Nagsabaran Red

Andarayan Red

Magapit Red

Callao Black

Dalan Serkot Black

Eme Black

Nagsabaran Black

Atulu

Sub-angular to angular and rounded Sub-rounded to rounded and sub-angular

Sub-rounded to rounded and sub-angular Sub-rounded to sub-angular

Sub-angular to angular

Rounded and sub-angular

Rounded and sub-angular

Rounded and sub-angular

Sub-rounded to sub-angular

Rounded to subrounded and sub-angular

Table 8.1 Descriptions of ceramic matrix fabric Ceramic Degree of Grain shape sorting

Oriented channels and vugh, 5%

Oriented channel voids, 5%

Oriented channel voids, 2 %

Oriented channel voids, 3%

Oriented channel voids, 5 %

Oriented channel voids and vughs, 2% Oriented channel void, 2%

Oriented channel voids and vughs, 5% Oriented channel voids, 5%

Oriented channel voids and vugh, 2%

Matrix porosity

Moderate for interior and exterior

Moderate for interior and weak for exterior Moderate for interior and weak for exterior Weak for both interior and exterior

Weak for both interior and exterior

Moderate for interior and exterior

Very weak for interior and weak for exterior Weak for interior and exterior

Weak for interior and exterior

Moderate for interior and exterior

Birefringence

Orange brown

Yellowish brown

Reddish brown

Orange brown

Reddish brown

Yellow brown

Dark brown

Black

Dark orange brown

Interior PPL Yellow brown

Colour

Dark yellowish brown

Reddish brown

Red orange

Red orange

Black

Dark brown

Black

Black with streaks of red brown Black

Interior XPL Dark yellowish brown

Red orange

Reddish brown

Yellow brown

Reddish black

Reddish brown

Red orange

Red orange

Red orange

Reddish brown

Exterior PPL Reddish orange

Dark red orange

Black

Dark yellowish brown

Black

Black

Dark red orange

Dark red orange

Dark red orange

Red orange

Exterior XPL Red orange

CHAPTER 9 The Prehistory of Northeastern Luzon Pleistocene Cabalwanian industry is a concept fraught with problems. Most of the pebble and flake tools were surface finds, and associations with extinct megafauna are problematic. I agree with Wasson and Cochrane (1979) that the flake-based Cabalwanian industry is contemporaneous with that of the Peñablanca caves. The only way to resolve the chronological issue of the Cabalwanian Industry is to undertake excavations to recover stone implements and megafauna in situ, and to date them independently using luminescence or uranium series techniques (Morgenstein et al 2003:513).

Unravelling the prehistory of northern Luzon has warranted a review of past archaeological endeavours as well as the presentation of new information from the Late Pleistocene through the Holocene period. Although numerous sites have been excavated, previous attempts to synthesise the archaeology of the area have been lacking. To add to the known archaeological record, three cave sites were excavated in the Callao Limestone Formation in Peñablanca. These caves generally have good stratigraphic sequences and extend back into deep time around 25,000 years ago. Most open sites in the Cagayan Valley itself do not have continuous stratigraphic sequences from preceramic into ceramic. The shell midden sites along the lower reaches of the Cagayan Valley only contain early to late ceramic period (c.4000-1000 BP) remains. Most of the sites bearing alleged mid-Pleistocene assemblages such as the Cabalwanian comprise only surface finds and lack ceramic phase components.

Around 25,000 BP Based on the current archaeological record from Callao Cave, we now have evidence for human occupation in northern Luzon c. 25,000 years ago. The possibility of finding an older cultural horizon (c. 40,000 BP) in Luzon is promising, since modern humans have inhabited Palawan Island since 47,000 years ago (Dizon 2003). These Late Pleistocene hunters and gatherers could have crossed from Palawan, which is part of Sundaland, into the rest of the Philippines by raft. Elsewhere, presumably modern humans also reached some of the remoter islands in Southeast Asia around 35,000 years ago, or earlier, such as Talaud, the northern Moluccas, northern Sulawesi and East Timor (Bellwood, 1979, 1997; Bellwood et al 1998; Tanudirjo 2001; Glover 1977).

In the controlled excavations within the cave sites, a number of cultural materials were recovered (Tables 9.1 and 9.2). Flake tools were recovered from the lowest cultural horizons (preceramic) and onwards through the ceramic phase. There seems to have been some change in lithic technology between 25,000-year-old deposit and 12,000 years and later deposits. Pottery and related Neolithic materials were introduced in the ceramic phase, with occasional examples of shell beads, clay ling ling o and spindle whorls. Floral and faunal remains were also recovered, which illuminate past environments and human diet. Animal species available in the region did not change between the preceramic and ceramic periods.

The lower layers of Callao Cave have evidence of human occupation dating to 25,968 ± 373 uncal. BP (Wk-14881). This layer contains chert flake tools, mostly blade-like flakes. There is also limited evidence for a more formal lithic technology that could have involved the use of points. Most of the chert flakes were use on hard contact materials, possibly wood (Moraceae family), bamboo or rattan.

The cultural and biological materials recovered from the excavated sites were subjected to a number of physical and laboratory analyses. Field observations and soil micromorphology analyses were essential since recognising disturbance is crucial for cultural analysis and interpretation. Radiocarbon dating has been employed to establish the chronological sequence of the sites and assemblages, and comparison with other sites of the same time period.

The Tabon Cave lithic assemblage from Palawan also has only flakes made on chert using a simple percussion technique (Fox 1970), plus a few retouched flake scrapers. Use-wear analysis of the Tabonian chert flakes shows that most were used on hard contact material (Mijares 2004). The chert assemblage from the lower layer of Callao cave could easily fit within the Tabonian tradition in terms of technology and use-wear. We can infer that the oldest discovered hunters and gatherers of Luzon probably migrated via Palawan at least 30,000 years ago. The archaeological survey of Batanes shows no preceramic occupation, so arrival from Taiwan is not likely (AFAP 2002, Bellwood 2005, Bellwood and Dizon 2005).

Flake tools were analysed to infer changes, if any, in manufacturing techniques and functions through time. Ceramic analysis entailed reconstruction of vessel forms, especially rims and vessel profiles, and petrographic analysis was conducted in order to infer possible sources. Specialists in each relevant discipline analysed biological remains. Reconstruction of past human diet may be inferred from faunal remains, charred botanical remains, and phytoliths recovered from sediments. This suite of analyses also provides evidence to interpret past environments. The analysis of the cultural and biological remains has given us a clearer picture of the archaeological record in the Peñablanca Cave sites. Situating this archaeological record within the larger regional context facilitates the reconstruction of prehistory in northeastern Luzon.

This lower layer in Callao Cave also has traces of hearths, identified in the field and in soil thin-section, the latter containing reddish orange nodules of burnt sediment as well as a high incidence of charcoal. Faunal remains were also recovered, but unfortunately they were burnt beyond recognition for species identification. No human remains were recovered from this time period to inform about skeletal morphology or diet.

The Preceramic hunters and gatherers

The exploitation of forest products like wild tubers can be glimpsed in the presence of charred parenchymatous tissues and starch grains. Unfortunately, these plant remains could not be identified to the species level. These late Pleistocene

The archaeological record of Pleistocene northern Luzon is poorly understood. This is because very few sites belong to this period, and very few are dated. The alleged mid78

foragers exploited forests and open grasslands. Grasses and fibres from Moraceae tree bark were probably utilized as raw materials for mats, baskets and nets.

Cabalwanian. The Cabalwanian has about seven percent pebble-cobble tools in its assemblage; the remainder are flake tools (Fox and Peralta 1974).

There is large time gap on our archaeological record for the Peñablanca area. After the 26,000 BP level in Callao Cave, the next dated assemblage is only 12,000 BP, from Musang Cave (Thiel 1990b). The reason for this large time gap could simply be archaeological visibility, since most cave sites were not excavated to bedrock. Even the previous 1980 excavation of Callao Cave, which reached 7.5 meters below surface, failed to reach bedrock. There is a C14 gap across the LGM for many sites all over Island Southeast Asia—maybe because sea levels were low and coastlines far away from surviving archaeological sites. An alternative explanation is that foragers during this hiatus were hunting and gathering in the alluvial plains of the Cagayan Valley. The thin lower archaeological deposit at Callao Cave might therefore represent a brief incursion into the foothills of the Sierra Madre.

The model that Peterson (1974) and Thiel (1980) proposed, in which internally generated population increases are said to have caused changes in subsistence economy among these late Pleistocene to mid Holocene forager communities, cannot be substantiated by the data at hand. Peterson favoured a narrow- to broad-spectrum trajectory, and Thiel a broad- spectrum to specialized subsistence strategy. But neither hypothesis stands up in the face of the current archaeological record. We do not have any reliable evidence of population growth during this period from which to posit any population pressure. Furthermore, there were no major changes in lithic technology, and the available floral and faunal evidence shows a continuing broad-spectrum subsistence strategy. Contrary to Headland and Bailey (1991), hunter-gatherers I believe did subsist in Holocene rainforests, as in Mainland Southeast Asia (Endicott and Bellwood 1991). Their success in exploiting and living in the forest would have depended on their foraging skills and knowledge of different ecosystems. Resources in rainforest can be patchy and seasonal, but foragers can strategise and optimise their movements to correspond to the shifting food resources (Winterhalder 2001).

There is still need to search for sites within this time period to enhance our knowledge of this era in Luzon. Callao Cave may still reveal more secrets since we have not yet reached the basal layer. The 26,000 BP layer is only 130 cm below the current surface, and we know from the earlier excavations that the deposits are at least seven meters deep.

12000 to 3500 BP

The Sierra Madre Agta of the recent ethnographic past would plan their group movements based on the seasons and available food sources (Allen 1985; Griffin and EstiokoGriffin 1978; Mudar 1985). They hunted in the montane forest during the rainy season and in the lowland forest during the summer, deriving most of their food supply from fishing, shellfish gathering, wild yams, nuts and Caryota palms. Mid-Holocene foragers could easily have done the same.

Between 12,000 and 3500 BP the Peñablanca caves seem to have been used by foraging groups as ephemeral hunting and collecting campsites. Perhaps people moved in and out seasonally from the rich estuarine/riverine environment of the Cagayan Valley. Like many foragers at the time in Southeast Asia they practised a broad-spectrum subsistence strategy (Barker 2005, Barker et al. 2002, Gorman 1971), hunting pig (Sus philippinensis), deer, monkey, reptiles, felids and birds, as well as engaging in shellfish gathering. Wild tubers (as seen in the charred parenchymatous tissue in the Eme Cave Preceramic horizon), seeds and nuts are likely to have been exploited.

The coming of the Austronesians The hunting and gathering groups continued to exploit the forest and alluvial plain resources into the mid-Holocene. At around 4000 to 3500 years ago, a different group of people with a different subsistence economy and technology arrived. These were the first speakers of Malayo-Polynesian languages within the Austronesian family to reach the Philippines (Ross 2005, Pawley 2002, Blust 1985). Both the linguistic and the archaeological evidence suggest that they migrated from the Islands of Formosa (including Ludao and Lanyu), possibly via the Batanes and Babuyan Islands. As a result of contacts with these incoming Malayo-Polynesians, the populations of indigenous hunters and gatherers soon adopted Austronesian languages (Headland and Reid 1989, Reid 1994) to the extent that only a few lexical traces of their former languages remain, according to Reid.

Stone-tool manufacture remained simple and expedient throughout this period. Chert and newly added volcanic rocks (andesite and basalt) were subjected to simple percussion techniques to produce amorphous flakes. The volcanic rocks were probably derived from the riverbed, and there are a few chert sources along the Callao Limestone Formation such as those near Dalan Serkot Cave. Retouching or further modification of flakes was no longer practised in this period, perhaps due to the ready availability of volcanic pebbles. Retouching of coarse-grained rocks would also produce a duller and thicker edge compared to a newly produced flake, so the practise was apparently discontinued. The Peñablanca cave sites have produced very few pebble tools. Use-wear analysis of the flake tools shows that most were used on hard materials such as bamboo, palm and rattan. A number of flakes also show use-wear characteristic of soft contact materials, such as meat and possibly grasses. If early Holocene foragers were also practising an arboreal subsistence economy, as Latinis (2000) has proposed for their contemporaries in eastern Indonesia and Near Oceania, the evidence is not visible in the Peñablanca cave sites. Arboriculture might have been practiced on the alluvial plains and terraces of the Cagayan Valley, where industries with more chopper and chopping tools existed, such as the

The Austronesians brought with them a suite of cultural materials that included pottery and polished adzes (Bellwood 2005, 1997), cultural markers now used widely to demarcate this new period, for which there had been no previous occurrences in the Philippines. The introduction of ground and polished-adze technology was particularly significant. Roger Duff (1970) recognised the clear relationships between Taiwan and Philippine adzes, with the latter being regarded as an elaboration of the former. The quadrangular-sectioned adzes of Duff’s Type 2A are the most common form in Southeast Asia, and occur both in Luzon and Taiwan. Duff’s 79

types 1A and 1B, with quadrangular cross-sections and stepped and shouldered butts, respectively, are also found in both Taiwan and Luzon. In Cagayan Valley, stone adzes were found associated with both red-slipped and black pottery. In the Peñablanca Cave sites, only Arku Cave (Thiel 1990a) yielded stone adzes and these were associated with burial contexts, and so could have been grave goods.

clay lingling o earrings. But economic exchange of cereals was either absent or so far not archaeologically visible. The analysis of both micro and macro botanical remains has revealed a subsistence strategy based on foraging wild root crops (unidentified parenchymatous tissues), possibly Caryota palms, nuts and other arboreal forest products. These plant carbohydrates supplemented the hunting of pig and deer and the collection of shellfish from the river. If the ceramicphase Holocene hunter-gatherers in the Peñablanca Caves were exchanging forest products for cereal foods such as rice, then they were not consuming them inside the caves. Possibly they were not acquiring cereals at all, especially if rice production was not abundant enough for exchange or if the lowlanders were cultivating other crops. Paz (2004) has suggested that rice might have been replaced widely in the Philippines by cultigens such as yams (Dioscorea alata) and taro, and Latinis (2000) has pointed to the possible importance of an arboreal-based subsistence strategy. In the foothills of the Sierra Madre, the middle Holocene (ceramic phase) hunter-gatherers might have been gathering plants for fibre, such as wild ramie (Boehmeria cf. platanifolia), rattan, the bark of mulberry trees, and grasses and sedges. These fibres could have been initially processed in the caves, e.g., by stripping the bark of a wild ramie stem with a simple unretouched flake. They could have traded these forest products with lowlanders in exchange for pottery and other items, as extant hunter-gatherers like some Luzon Agta still do today (Griffin 1984, 1985, Griffin and Estioko-Griffin 1978, Headland 1986, Peterson 1978).

Bellwood (1997, 2005) and Hung (2004, 2005), in correlating the pottery at c.4000 to 3000 BP from Taiwan and Luzon, point to great affinity in form, surface finish and decoration. Of particular concern is the red-slipped surface finish that they trace to southeastern Taiwan, the region that may have been the immediate source-area for the ancestral MalayoPolynesian movements. The Malayo-Polynesian subgroup itself was formed by innovations that occurred after departure from Taiwan (Ross 2005). Red-slipped pottery has been found in many sites in the Cagayan alluvial floodplains, dating possibly from 4850-3650 cal. BP at Pamittan (Tanaka and Orogo 2000), although this is a single isolated determination and all other Cagayan dates for this phase are closer to 3600 cal. BP. The Cagayan redslipped pottery corresponds well with that from the newly discovered red-slipped pottery site at Chaoliaqiao in southeastern Taiwan (Hung pers. comm. 2005). The migration from Taiwan into Luzon by early MalayoPolynesians was not necessarily a massive single-event phenomenon. And it probably did not emanate from just from one area in Taiwan. The presence of black pottery together with the red-slipped pottery in sites such as Nagsabaran, Irigayen, Bangag, and Catugan in the Cagayan Valley, and Callao and Dalan Serkot Caves in Peñablanca, dating to c. 3500 BP, points to a possible additional contribution from sites such as Fengpitou in southwestern (rather than southeastern) Taiwan. Fengpitou has produced black pottery dating to as early as c. 3800 BP (Chang 1969). The close affinity of the spindle whorl found in Callao Cave with one from Fengpitou (see Appendix H1) adds to this observation.

By 2500 BP, the sea had lowered to its present level, falling from about three meters above present level at c. 4500 BP (Berdin et. al 2003). This marine retreat probably contributed to the increase in estuarine and freshwater shellfish (Batissa childreni and Corbicula fluminea) in the lower reaches of the Cagayan River. These were then collected and consumed by the inhabitants of the Cagayan Valley in very large numbers beginning around 2000 BP. The shell-midden formation phase in the Cagayan valley also signalled the decline of the red-slipped and the predominance of the black pottery. This can also be seen in Eme Cave, where the ceramic period is contemporaneous with the shellmidden formation phase and black pottery tradition in the valley. Eme Cave has only black and plain pottery.

Besides pottery, other cultural materials apparently brought in by Malayo-Polynesian populations, not necessarily all at the same time, were clay and stone lingling o earrings, baked clay spindle whorls, and shell and stone beads. In terms of economic subsistence, Bellwood (2005, 1997) believes that the Austronesians brought with them domesticated animals such as pigs and cultigens such as rice. The Cagayan Valley has broad alluvial plains protected by mountain ranges, regularly enriched by inundation by the Cagayan River, which makes it ideal for rice cultivation today.

We now need to examine the question of which people – indigenous hunter-gatherers or incoming Malayo-Polynesians – were importing the earthenware vessels into the Penablanca cave sites. Was there a replacement of the preceramic hunter and gatherer population by incoming pottery-using agriculturists? To answer these questions we have to relate the pottery with the associated cultural materials. The caves in the Peñablanca area mostly have two cultural horizons: a lower preceramic one with flake tools, and an upper ceramic one with similar flake tools (Ronquillo 1981; Thiel 1989; Mijares 2002). Analysis of the flake tools from both of these cultural horizons (Chapter 7) shows that there was no technological change from the late preceramic into the ceramic phase, and both successive populations used their stone tools to work the same contact materials. On the other hand, the red-slipped earthenware sherds in the Cagayan Valley open sites occur mostly with polished adzes and axes, rarely with flake tools. The latter occur mostly in the younger shell-midden layers with the black pottery.

There is so far only limited evidence of rice cultivation in Luzon during this period, mainly the AMS-dated rice husk inclusions in the red-slipped pottery from Andarayan (Snow et al 1986). To date we have not found any Cagayan sites similar to the deeply buried alluvial plain sites within the Tainan Science-based Industrial Park in southwest Taiwan, from where abundant charred rice and foxtail millet samples have been collected (Tsang 2005). In the Peñablanca cave sites, we also do not have any evidence of cereal-based subsistence, either locally cultivated or acquired through exchange. From the ceramic analysis we know that earthenware vessels were coming into the caves from the Cagayan valley, possibly through exchange, which would also have included shell beads, spindle whorls, and 80

81

25,968±373 uncal. BP

7260-6990 BP

3990-3690 BP

Callao

Dalan Serkot

Eme

lithics

Andesite, basalt, and chert flakes with amorphous shapes, most used on hard contact material

Andesite and chert flakes, amorphous shapes, most used on hard contact material

Chert flakes, with prevalent blade-like forms, most used on hard contact material

fauna animal shells

Deer, pig, abundant riverine shells

Deer, pig, abundant riverine shells

Charred bones, no recovered

14

C date

2010-1690 BP

3650-3470 BP

3900-3690 BP

Site

Eme

Callao

Dalan Serkot

Brown, black and redslipped pottery, some with incised decoration

Brown, black and redslipped pottery, some with incised decoration

Black and plain pottery, no decoration

Ceramic Andesite, basalt, chert flakes with amorphous shapes, most used on hard contact material Andesite and chert flakes, amorphous shapes, most used on soft contact material Andesite, basalt and chert flakes amorphous shapes, most used on hard contact material

Lithics

macrobotanical

Pig, deer, few riverine shells

Deer, pig, few riverine shells

Boehmeria cf. platanifolia, parenchymatous tissue, charred nuts, wood and seeds Boehmeria cf. platanifolia, parenchymatous tissue

Possibly sago-like palms, bamboo, grasses and sedges, Moraceae, starch grains None found

Not analysed

Phytoliths

Not analysed

None found

Moraceae, grasses, starch grains

phytoliths

Boehmeria cf. platanifolia seeds, parenchymatous tissue, nuts and wood

macrobotanical

Boehmeria cf. platanifolia, parenchymatous tissues, seeds, nuts, wood

Unidentified seeds

Charred parenchymatous tissues

Pig, deer, riverine shells

fauna

Table 9.2 The major features of each excavated site during the Ceramic Period

C date

14

Site

Table 9.1 The major features of each excavated site during the preceramic period soil micromorphology

Massive blocky structure, iron oxides, charcoal and burnt sediment feature (hearth), precipitation of calcium carbonate

Crumb structure, burnt plant materials and sediments, charcoal, precipitation of calcium carbonate and illuviation of silt and clay

Highly bioturbated with faunal activity, clay illuviation, charred plant remains.

Soil micromorphology

Crumb structure, heavily bioturbated with faunal activity, charcoal, impregnated with calcite crystals

Blocky structure, travertine, volcanic ash.

Blocky and crumb structure, precipitation of calcium carbonates, burnt sediments and charcoal (hearths)

from that time and thereafter with the indigenous huntergatherer groups of the upland Sierra Madre.

The continuity of flake tool technology from the preceramic into the ceramic period in the cave sites could signify that the same hunter and gatherer groups were using the caves. It is thus possible that they were the ancestors of the modern Agta Negrito population of the region.

Cave sites, because of their good stratigraphic contexts and good preservation, have been major sources of information for prehistoric reconstruction. Developments in the Cagayan Valley open sites are at least partially reflected in the archaeological records of the caves that played such an important role in the lives of the hunter-gatherers. The Peñablanca Cave sites have thus contributed information relevant for interpreting the entire sequence and chronology of the Cagayan Valley archaeological landscape.

Based on my excavated and analytical data, I suggest that there was no replacement of people during the Neolithic in the Peñablanca Caves, but rather that continuing hunter-gatherer population obtained their earthenware vessels, clay earrings, spindle whorls and beads from other populations resident in the Cagayan Valley itself. The Austronesian-speaking people, who migrated into Luzon c.4000 BP, were exchanging goods

Table 9.3 A summary of Northern Luzon Culture History Period Material Culture Late Pleistocene (25,000 Chert flake tools, blade-like flakes, and BP) points Terminal Pleistocene to Chert, andesite and basalt flakes. Mid Holocene (12,000 to 3500 BP) mid-Holocene (3500 – 1500 BP)

Flake tools in the cave sites, adzes in the valley sites. Introduction of pottery (redslipped, black and plain), spindle whorls, baked clay lingling o earrings.

82

Subsistence Not clear, but included hunting of animals and collecting of wild tubers. Broad-spectrum subsistence that included hunting of small- and medium-sized mammals, gathering shellfish, wild tubers, nuts, caryota palms. Broad-spectrum subsistence in the cave sites, which included hunting of small and medium size animals, gathering shellfish, wild tubers, nuts, caryota palms. Limited cultivation of rice in the valley, possible cultivation of yams and taro. Collection of shellfish and shell midden formation along the Cagayan River by 2000 BP.

—. 2005. The Use of Caves in Peninsular Thailand in the Late Pleistocene and Early and Middle Holocene. Asian Perspectives 44:137-153.

Reference: Adams, J., and H. Faure. 1997. Preliminary Vegetation Maps of the World since the Last Glacial Maximum: An Aid to Archaeological Understanding. Journal of Archaeological Science 24:623-647.

Andrefsky, W. J. 1994. Raw-Material Availability and the Organization of Technology. American Antiquity 59:21-34. —. 1998. Lithics: Macroscopic approaches to analysis. Cambridge: Cambridge University Press.

Adi Taha. 1983. Recent Archaeological Discoveries in Peninsular Malaysia 1976-1982. Journal of the Malaysian Branch of the Royal Asiatic Society 53:47-63.

Aoyagi, Y., M. Aguilera, H. Ogawa, and K. Tanaka. 1993. Excavation of Hill Top Site, Magapit Shell Midden in Lal-lo Shell Middens, Northern Luzon, Philippines. Man and Culture in Oceania 9:127155.

—. 1993. Recent Archaeological Discoveries in Peninsular Malaysia (1991-1993). Journal of the Malaysian Branch of the Royal Asiatic Society 66:67-83. —. 2000. Archaeological investigation in Ulu Kelantan, Peninsular Malaysia. PhD Thesis, Australian National University, Canberra.

Aoyagi, Y., H. Ogawa, and K. Tanaka. 1997. Excavation, and Ornaments Discovered at the Magapit Shellmidden Site, in Northern Luzon. The Journal of Sophia Asian Studies 15:167-180.

AFAP. 2002. "Report on the Archaeological Research in Batanes and Ilocos Norte Provinces, northern Philippines," in unpublished report prepared by Asian Fore-Arc Project. Canberra.

Bailey, R., G. Head, M. Jenike, B. Owen, R. Rechtman, and E. Zechenter. 1989. Hunting and Gathering in Tropical Rain Forest: Is It Possible? American Anthropologist 91:59-82.

Akoshima, K. 1979. An Experimental Study of Microwear Traces on Shale Artifacts. MA Thesis (English Summary), Tohoku University.

Bailey, R., and T. N. Headland. 1991. The Tropical Rain Forest: Is It a Productive Environment for Human Foragers? Human Ecology 19:261-285.

Albert, R., O. Lavi, L. Estroff, S. Weiner, A. Tsatskin, A. Ronen, and S. Lev-Yadun. 1999. Mode of Occupation of Tabun Cave, Mt Carmel, Israel During the Mousterian Period: A Study of Sediments and Phytoliths. Journal of Archaeological Science 26:1249-1260.

Balasse, M. 2002. Reconstructing Dietary and Environmental History from Enamel Isotopic Analysis: Time Resolution of Intra-tooth Sequential Sampling. International Journal of Osteoarchaeology 12:155-165. Balme, J., and W. Beck. 2002. Starch and Charcoal: Useful Measures of Activity Areas in Archaeological Rockshelters. Journal of Archaeological Science 29:157-166.

Allen, M. 1985. "The Rainforest of Northeastern Luzon and Agta Foragers," in The Agta of Northeastern Luzon: Recent Studies. Edited by P. B. Griffin and A. Estioko-Griffin, pp. 45-68. Cebu: San Carlos Publication.

Barbosa, A. 1979. "A Preliminary Archaeological Report on the Lattuc-Latuc Cave," in Manuscript of National Museum. Manila.

Almeida, A and G. Zybzweski. 1967. “A contribution to the study of the prehistory of Portuguese Timor-lithic industries,” in Archaeology at the eleventh Pacific Science Congress: Papers presented at the XI Pacific Science Congress, Tokyo, AugustSeptember 1966. Edited by W. G Solheim II, pp. 55-67. Social science Research Institute, University of Hawaii, Asian Pacific Archaeology series 1.

Barker, G. 2005. The Archaeology of Foraging and Farming at Niah Cave, Sarawak. Asian Perspectives 44:90106. Barker, G., H. Barton, P. Beavit, M. Bird, P. Daly, C. Doherty, D. Gilbertson, C. Hunt, J. Krigbaum, H. Lewis, J. Manser, S. McClaren, V. Paz, P. Pipper, B. Pyatt, R. Rabett, T. E. G. Reynolds, J. Rose, G. Rushworth, and M. Stephens. 2002. Prehistoric Foragers and Farmers in South-east Asia: Renewed Investigation at Niah Cave, Sarawak. Proceedings of the Prehistory Society 68:147-164.

Al-Shorman, A. 2004. Stable carbon isotope analysis of human tooth enamel from the Bronze Age cemetery of Ya'amoun in Northern Jordan. Journal of Archaeological Science 31:1693-1694.

Barton, H. 2005. The Case for Rainforest Foragers: The Starch Record at Niah cave, Sarawak. Asian Perspectives 44:56-72.

Anderson, A. 2000. Slow boats from China: Issues in the prehistory of Indo-Pacific seafaring. Modern Quaternary Research in Southeast Asia 16:13-50.

Beavit, Paul, Edmund Kuriu, and Gill Thompson. 1996. Confirmation of an Early Date for the Presence of Rice in Borneo; Preliminary Evidence for Possible Bidayuh/Aslian Links. Borneo Research Bulletin. 27:29-28.

—. 2005. Crossing the Luzon Strait: archaeological chronology in the Batanes islands, Philippines and the regional sequence of Neolithic dispersal. Journal of Austronesian Studies.1(2) 25-46. Anderson, D. 1990. Lang Rongien Rockshelter. Philadelphia: University of Pennsylvania Museum.

Bellwood, P. 1979. Man's Conquest of the Pacific: The Prehistory of Southeast Asia and Oceania. New York: Oxford University Press.

—. 1997. Cave Archaeology in Southeast Asia. Geoarchaeology 12:607-638. 83

Approaches. Edited by G. Bartstra, pp. 233-274. Rotterdam: Balkema.

—. 1988. Archaeological Research in South-Eastern Sabah. Kota Kinabalu: Sabah Museum Monograph 2. —. 1992. "Southeast Asia before Prehistory," in The Cambridge History of Southeast Asia, vol. 1. Edited by N. Tarling, pp. 55-136. Cambridge: Cambridge University Press.

Bellwood, P., and M. Omar. 1980. Trade patterns and political developments in Brunei and adjacent areas, AD 700-1500. The Brunei Museum Journal 4:155-179.

—. 1993. Cultural and Biological Differentiation in Peninsular Malaysia: The Last 10,000 Years. Asian Perspectives 32(1)37-60.

Bellwood, P., and A. Sanches-Mazas. 2005. Human Migrations in Continental East Asia and Taiwan: Genetic, Linguistic, and Archaeological Evidence. Current Anthropology 46:480-484.

—. 1995. "Austronesian Prehistory in Southeast Asia: Homeland, Expansion and Transformation," in The Austronesians: Historical and Comparative Perspectives. Edited by P. Bellwood, J. Fox, and D. Tryon, pp. 96-106. Canberra: Department of Anthropology, Research School of Pacific and Asian Studies, ANU.

Bellwood, P., J. Stevenson, A. Anderson, and E. Dizon. 2003. Archaeological and Palaeoenvironmental Research in Batanes and Ilocos Norte Provinces, Northern Philippines. Bulletin of the Indo-Pacific Prehistory Association 23:141-161.

—. 1997. Prehistory of the Indo-Malaysian Archipelago, second edition. Honolulu: University of Hawaii Press.

Berdin, R., F. Siringan, and Y. Maeda. 2003. Holocene Relative Sea-level Changes and Mangrove Response in Southwest Bohol, Philippines. Journal of Coastal Research 19:304-313.

—. 2001. Early Agriculturist Population Diasporas? Farming, Languages, and Genes. Annual Review of Anthropology 30:181-207.

Beyer, H. O. 1947. Outline Review of Philippine Archaeology by Islands and Province. Philippine Journal of Science 77:205-374.

—. 2002. "Farmers, Foragers, Language, Genes: the Genesis of Agricultural Societies," in Examining the farming/language dispersal hypothesis. Edited by P. Bellwood and C. Renfrew, pp. 17-30. Cambridge: McDonald Institute of Archaeological Research, University of Cambridge.

—. 1948. Philippine and East Asian archaeology and its relation to the origin of the Pacific Islands population. Bulletin 29: National Research Council of the Philippines. Binford, L. 1983. "Organization and Formation Process: Looking at Curated Technologies," in Working in Archaeology. Edited by L. Binford pp. 269-286. New York: Academic Press.

—. 2004a. "Colin Renfrew's Emerging Synthesis: Farming, Languages and Genes as Viewed from the Antipodes," in Traces of ancestry: studies in honour of Colin Renfrew. Edited by M. Jones, pp. 31-39. Cambridge: McDonald Institute for Archaeological Science, University of Cambridge.

Blust, R. 1976. Austronesian culture history: some linguistic inferences and their relations to archaeological record. World Archaeology 8:19-43. —. 1985. The Austronesian Homeland: A linguistic perspective. Asian Perspectives 26:45-67.

—. 2004b. "The origins and dispersals of agricultural communities in Southeast Asia," in Southeast Asia. Edited by I. Glover and P. Bellwood, pp. 21-40. London: RoutledgeCurzon.

Brosius, J. P. 1991. Foraging in Tropical Rain Forests: The Case of Penan of Sarawak, East Malaysia (Borneo). Human Ecology 19:123-150.

—. 2005. The First Farmers: The Origins of Agricultural Societies. Oxford: Blackwell Publishing.

Brown, P., T. Sutikna, M. Morwood, R. Soejono, Jatmiko, E. W. Saptomo, and R. A. Due. 2004. A new small-bodied homonin from the Late Pleistocene of Flores, Indonesia. Nature 431:1055-1061.

Bellwood, P., and J. Diamond. 2005. On explicit ’replacement’ models in Island Southeast Asia: a reply to Stephen Oppenheimer. World Archaeology 37(4): 503-506.

Brothwell, D. R. 1960. Upper Pleistocene human skull from Niah Cave. Sarawak Museum Journal 9:325-349.

Bellwood, P., and E. Dizon. 2005. The Batanes Archaeological Project and the "Out of Taiwan" Hypothesis for the Austronesian Dispersal. Journal of Austronesian Studies 1(1):1-32.

Bullock, P., N. Fedoroff, A. Jongerius, G. Stoops, and T. Tursina. 1985. Handbook for Soil Thin-section Description. England: Waine Research Publication.

Bellwood, P., R. Gillespie, G. B. Thompson, J. S. Vogel, I. Ardika, and I. Datan. 1992. New dates for prehistoric Asian rice. Asian Perspectives 31:161170.

Cabanilla, Israel. 1972. Neolithic Shellmound of Cagayan: The Lal-lo Excavation. In Field Report National Museum. Manila.

Bellwood, P., and I. Glover. 2004. "Southeast Asia: Foundations for an archaeological history," in Southeast Asia. Edited by P. Bellwood and I. Glover, pp. 4-20. London: RoutledgeCurzon.

Cannon, M. 2001. Archaeofaunal Relative Abundance, Sample Size, and Statistical Methods. Journal of Archaeological Science 28:185-195. Chang, K. 1964. Prehistoric and Early Historic Culture Horizons and Traditions in South China. Current Anthropology 5:359+ 368-375.

Bellwood, P., G. Nitihaminoto, G. Irwin, Gunadi, A. Walayu, and D. Tanudirjo. 1998. "35,000 years of prehistory of northern Moluccas," in Bird's Head 84

Congress of the Indo Pacific Prehistory Association, Taipei, Taiwan, 2002.

—. 1969. Fengpitou, Tapenkeng, and The Prehistory of Taiwan. New Haven: Yale University Publications.

de la Torre, A. 2000. "Preliminary Report of the Lal lo, Cagayan Archaeological Project: Clemente Irigayen Property Site (II-1995-O), Sta Maria, Lallo Cagayan," in Excavation of Lal-lo Shell Middens. Edited by H. Ogawa, pp. 93-128. Tokyo: Tokyo University of Foreign Studies.

—. 1995. "Taiwan Strait Archaeology and ProtoAustronesian," in Austronesian Studies Relating to Taiwan. Edited by P. Li, C.-h. Tsang, Y. Huang, D. Ho, and C. Tseng, pp. 161-184. Taipei. Chen Chi-Lu. 1969. Material Culture of the Formosan Aborigines. Taipei: The Taiwan Museum.

de Vera, Elenita. 1983. Minori Cave Vertebrate Skeletal Remains: A Preliminary Study. In Manuscript of the National Museum of the Philippines. Manila.

Chesner, C., W. Rose, A. Deino, R. Drake, and J. Westgate. 1991. Eruptive history of Earth's largest Quaternary caldera (Toba, Indonesia) clarified. Geology 19:200-203.

de Vos, J., and A. Bautista. 2003. Preliminary notes on the vertebrate fossils from the Philippines. Proceedings of the Society of Philippine Archaeologists 1:42-62.

Corby, G. 1951, Geological and Oil Possibilities of the Philippines, Manila, Bureau of Printing. Cornish, D., and I. Watt. 1966. "The Mechanism of Glass Polishing." Symposium on the Surface Chemistry of Glass, American Ceramic Society, Washington DC, 1966.

Denham, T. 2004. The roots of agriculture and arboriculture in New Guinea: looking beyond Austronesian expansion, Neolithic packages and indigenous origins. World Archaeology 36:610-620.

Corruccini, J. A. 1985. Moisture and the Formation of Obsidian Striations. Lithic Technology 14:33-35. Cotterell, B., and J. Kamminga. 1979. "The Mechanics of Flaking," in Lithic Use-wear Analysis. Edited by B. Hayden, pp. 99-112. New York: Academic Press.

Detroit, F., E. Dizon, C. Falgueres, S. Hameau, W. Ronquillo, and F. Semah. 2004. Upper Pleistocene Homo sapiens from Tabon cave (Palawan, The Philippines): description of new discoveries. Human Palaeotology and Prehistory 3:705-712.

—. 1987. The Formation of Flakes. American Antiquity 52:657-708.

Diamond, J., and P. Bellwood. 2003. Farmers and Their Languages. Science 300:597-603.

—. 1990. Mechanics of pre-industrial technology. Cambridge: Cambridge University of Press.

Dizon, E. 2000. Archaeology of Batanes Province, Northern Philippines: the 1996-1997 status report. Bulletin of the Indo-Pacific Prehistory Association 19:11524.

Courty, M. A., P. Goldberg, and R. Macphail. 1989. Soils and Micromorphology in Archaeology. Cambridge: Cambridge University Press.

—. 2003. New Direct dating of Human Fossils from Tabon Cave, Palawan, Philippines. Proceedings of the Society of Philippine Archaeologists 1:63-67.

Coutts, P. J. F., and J. P. Wesson. 1980. Model of Philippine Prehistory: A Review of Flaked Stone Industries. Philippine Quarterly of Culture and Society 8:203259.

Dizon, E., F. Detroit, F. Semah, S. Hameau, W. Ronquillo, and E. Cabanis. 2002. Notes on the Morphology and Age of the Tabon Cave Fossil Homo sapiens. Current Anthropology 43:660-666.

Cox, M. 2005. Indonesian Mitochondrial DNA and Its Opposition to a Pleistocene Era Origin of ProtoPolynesians in Island Southeast Asia. Human Biology 77:179-188.

Drucker, D., and H. Bocherens. 2004. Carbon and Nitrogen Stable Isotope as Tracers of Change in Diet Breadth during Middle and Upper Palaeolithic in Europe. International Journal of Osteoarchaeology 14:162-177.

Cuevas, M. 1980. "Preliminary Report on the Archaeological Excavation conducted at Callao Caves," in Manuscript of the National Museum. Manila.

Duff, R. 1970. Stone Adzes of Southeast Asia. Christchurch: Canterbury Museum.

Dalupan, M. L. 1981. The Prehistoric Occupation of Lattu Lattuc Cave: The Evidence of Ceramics. MA Thesis, Ateneo de Manila University, Quezon City.

Durkee, E., and S. Pederson. 1961. Geology of Northern Luzon. American Association Petroleum Geologists Bulletin 45:137-168.

Darvill, T., and J. Timby. 1982. "Textural Analysis: A Review of Potentials and Limitations," in Current research in ceramics: Thin-section studies, vol. 32. Edited by I. Freestone, C. Johns, and T. Potter, pp. 73-86. London: British Museum.

Dyen, I. 1965. A lexicostatistical classification of the Austronesian languages. International Journal of American Linguistics, Memoir 19. Endicott, K., and P. Bellwood. 1991. The possibility of independent foraging in the rain forest of Peninsular Malaysia. Human Ecology 19:151-185.

Davenport, D. R. 2003. A Functional Analysis of Southeast Asian Pacific Island Flaked Stone Tools. B.A. Honours Thesis, The Australian National University, Canberra.

Farrand, W. 2001. Sediments and Stratigraphy in the Rockshelters and Caves: A Personal Perspective on Principles and Pragmatism. Geoarchaeology 16:537-557.

Davenport, D. R., and J. Kamminga. 2002. "Functional analysis of island Southeast Asia-Pacific lithic assemblages Paper presented to the." 17th 85

Goldberg, P., and R. Macphail. 2003. Short Contributions: Strategies and Techniques in Collecting Micromorphology Samples. Geoarchaeology 18:571-578.

Fearn, M. L. 1998. Phytoliths in sediments as indicator of grass pollen source. Review of Palaeobotany and Palynology 103:75-81. Fitzpatrick, E. A. 1984. Micromorphology of Soils. London: Chapman and Hall Ltd.

Gorman, C. 1971. The Hoabinhian and after: subsistence patterns in Southeast Asia during the late Pleistocene and early Recent periods. World Archaeology 2:300-320.

Fitzpatrick, S., W. Dickinson, and G. Clark. 2003. Ceramic petrography and cultural interaction in Palau, Micronesia. Journal of Archaeological Science 30:1175-1184.

—. 1972. Excavation at Spirit Cave, North Thailand: Some Interim Interpretation. Asian Perspectives 13:79108.

Fix, A. 2002. "Foragers, farmers, and traders in the Malaysian Peninsula: origins of cultural and biological diversity," in Foragers-Traders in South and Southeast Asia. Edited by K. Morrison and L. Junker, pp. 185-202. Cambridge: Cambridge Press.

Griffin, P. B. 1984. "Forager Resources and Land Use in the Humid Tropics: The Agta of Northeastern Luzon, the Philippines," in Past and Present in Hunter Gatherer Studies. Edited by C. Schrire. New York: Academic Press, Inc.

Foix, L., and R. Bradley. 1985. "Rhyolite: Studies in Usewear Analysis," in Views of the Jordana Mogollon, vol. 12. Edited by C. Beck, pp. 112-120: Eastern New Mexico University Contributions to Anthropology.

—. 1985. "A Contemporary View of the Shift from Hunting to Horticulture: The Agta Case," in Recent Advances in Indo-Pacific Prehistory. Edited by V. N. Misra and P. Bellwood, pp. 349-352. New Delhi: Oxford and IBH.

Fox, R. 1970. The Tabon Cave. Manila: Monograph of the National Museum No. 1.

Griffin, P. B., and A. Estioko-Griffin. 1978. Ethnoarchaeology in the Philippines. Archaeology 31:34-43.

—. 1971. "Ancient Man and Pleistocene Fauna in Cagayan Valley, Northern Luzon Philippines," in Manuscript of the National Museum. Manila.

Groves, C. 1981. Ancestors for the Pigs: Taxonomy and Phylogeny of Genus Sus. Canberra: Department of Prehistory, Research School of Pacific Studies, Australian National University.

—. 1978. "The Philippine Paleolithic.," in Early Paleolithic in South and East Asia. Edited by F. Ikawa-Smith, pp. 50-85. Paris: Mouton Publisher. —. 1979. "The Philippines in Prehistoric Times," in Readings in Philippine Prehistory, vol. 1. Edited by M. Garcia. Manila: Filipiniana Book Guild, Inc.

— 1985 Plio-Pleistocene mammals in island Southeast Asia. Modern Quaternary Research in Southeast Asia 9:143-54.

Fox, R., and J. Peralta. 1974. "Preliminary Report on the Paleolithic Archaeology of Cagayan Valley, Philippines, and the Cabalwanian Industry." First Regional Seminar on Southeast Asian Prehistory and Archaeology, Manila, 1974, pp. 100-147.

—. 1997. Taxonomy of wild pigs (Sus) of the Philippines. Zoological Journal of Linnean Society 120:163191. Harrisson, T. 1975. Tampan: Malaysia's Palaeolithic Reconsidered. Modern Quaternary Research in Southeast Asia 1:53-69.

Freeman, D. 1955. Iban Agriculture. London: Her Majesty's Stationery Office.

Harvey, E., and D. Fuller. 2005. Investigating crop processing using phytolith analysis: the example of rice and millets. Journal of Archaeological Science 32:739-752.

Freestone, I. 1991. "Extending Ceramic Petrology," in Recent Developments in Ceramic Petrology, vol. 81. Edited by A. Middleton and I. Freestone, pp. 399-410. London: British Museum.

Harris, Edward. 1989. Principles of Archaeological Stratigraphy. London. Academic Press.

French, C. 2003. Geoarchaeology in Action: Studies in soil micromorphology and Landscape evolution. London: Routledge.

Ha Van Tan 1997. The Hoabinhian and Before. Bulletin of the Indo-Pacific Prehistory Association 3:35-41.

Fullagar, R. 1991. The Role of Silica in Polish Formation. Journal of Archaeological Science 18:1-24.

Hay, B. 1977. Use-scratch Morphology: A Functional Significance Aspect of Edge Damage on Obsidian tools. Journal of Field Archaeology 4:491-492.

Garong, A. 2002. "Culture in Trash. An archaeological excavation of Conciso Property shell midden, Catayauan, Lal-lo, Cagayan Valley, Northern Philippines," in Excavation of Lal-lo Shell Middens. Edited by H. Ogawa, pp. 11-32: Tokyo University of Foreign Studies.

Hayden, B. Editor. 1979. Lithic Use-Wear Analysis. New York: Academic Press. Heaney, L. 1985. Zoogeographic evidence for middle and late Pleistocene land bridges to the Philippine islands. Modern Quaternary Research in Southeast Asia 9:127-144.

Glover, I. 1977. The Late Stone Age in Eastern Indonesia. World Archaeology 9:42-46.

Headland, T. N. 1986. Why Foragers do not Become Farmers: A Historical Study of Changing Ecosystem and its Effect on the Negrito HunterGatherer in the Philippines, University of Hawaii.

—. 1981. Leang Burung 2: An Upper Palaeolithic Rock Shelter in South Sulawesi, Indonesia. Modern Quaternary Research in Southeast Asia 6:1-38.

86

Isorena, E. 2004. The Visayan Raiders of the China Coast, 1174-1190 AD. Philippine Quarterly of Culture and Society 32:73-95.

Headland, T. N., and R. Bailey. 1991. Introduction: Have hunter-gatherers ever lived in tropical rain forest independently of agriculture? Human Ecology 19:115-122.

James, P., D. Chester, and A. Duncan. 2000. "Volcanic Soils: their nature and significance for archaeology," in The Archaeology of Geological Catastrophes. Edited by W. McGuire, D. Griffiths, P. Hancock, and I. Stewart, pp. 317-338. London: The Geological Society.

Headland, T. N., and L. Reid. 1989. Hunter and Gatherers and their Neighbors from Prehistory to the Present. Current Anthropology 30:43-46. Heine-Geldern. 1945. "Prehistoric research in the Netherlands-Indies," in Science and Scientists in the Netherlands-Indies. Edited by P. Honig and F. Verdoorn, pp. 129-167. New York.

Jelinek, A. 1976. "Form, Function, and Style in Lithic Analysis," in Cultural Change and Continuity: Essays in Honor of James Bennett Griffin. Edited by C. Cleland, pp. 10-33. New York: Academic Press.

Henson, F. 1977. "The Laurente Cave Archaeological Project: A Preliminary Report," in Manuscript of the National Museum. Manila.

Jiang, Q. 1995. Searching for evidence of early rice agriculture at prehistoric sites in China through phytolith analysis: an example from central China. Review of Palaeobotany and Palynology 89:481485.

—. 1978. The Flake Tool Industry. MA Thesis, University of the Philippines, Quezon City. —. no date. "An Analysis of the Laurente Cave Stone Tool Technology," in Manuscript of the National Museum. Manila.

Junker, L. 2002a. "Economic specialization and inter-ethnic trade between foragers and farmers in prehispanic Philippines," in Forager-Traders in South and Southeast Asia: Long Term Histories. Edited by K. Morrison and L. Junker, pp. 201-241. Cambridge: Cambridge University Press.

Higham, C. 1989. The Archaeology of Mainland Southeast Asia. Cambridge: Cambridge University Press. —. 2004. "Mainland Southeast Asia from the Neolithic to the Iron Age," in Southeast Asia: From prehistory to history. Edited by I. Glover and P. Bellwood, pp. 41-67. London: RoutledgeCurzon.

—. 2002b. "Southeast Asia: introduction," in ForagerTraders in South and Southeast Asia: Long Term Histories. Edited by K. Morrison and L. Junker, pp. 131-166. Cambridge: Cambridge University Press.

Higham, C., and R. Thosarat. 1994. Khok Phanom Di: Prehistoric Adaptation to the World's Richest Habitat. Forth Worth, Dallas: Harcourt Brace College Publishers.

Kamminga, J. 1980. Stone Tools Analyzed. Science 210:5859.

Hung, H. c. 2004. A sourcing study of Taiwan stone adzes. Bulletin of the Indo-Pacific Prehistory Association 24:57-70.

—. 1982. Over the Edge: Functional analysis of Australian stone tools: University of Queensland Museum Occasional Papers in Anthropology.

—. 2005. Neolithic Interaction between Taiwan and Northern Luzon: The Pottery and Jade Evidence from Cagayan Valley. Journal of Austronesian Studies. 1(1):106-130.

Kamminga, J., and R. Fullagar. 1984. "Use-wear analysis in Australia," in Archaeology of ANZAAS Canberra. Edited by G. K. Ward, pp. 258-59. Canberra: Canberra Archaeological Society.

Hunt, C., and G. Rushworth. 2005. Pollen taphonomy and airfall sedimentation in a tropical cave: the West Mouth of The Great Cave of Niah in Sarawak, Malaysia Borneo. Journal of Archaeological Science 32:465-473.

Karkanas, P., O. Bar-Yosef, P. Goldberg, and S. Weiner. 2000. Diagenesis in Prehistoric Caves: the Use of Minerals that Form in situ to Assess the Completeness of the Archaeological Record. Journal of Archaeological Science. 27: 915-929.

Hutterer, K. 1977. "Reintreperting the Southeast Asian Paleolithic," in Sunda and Sahul: Prehistoric Studies in Southeast Asia, Melanesia and Australia. Edited by J. Allen, J. Golson, and R. Jones, pp. 31-71. London: Academic Press.

Kealhofer, L. 2003. Looking into the Gap: Land Use and the Tropical Forests of Southern Thailand. Asian Perspectives 42:72-95.

Ipoi, Datan, and P. Bellwood. 1991. Recent Research at Gua Sireh (Serian) and Lubang Angin (Gunung Mulu National Park), Sarawak. Bulletin of the IndoPacific Prehistory Association 1:386-405.

Kealhofer, L., R. Torrence, and R. Fullagar. 1999. Integrating Phytoliths within Use-Wear/Residue Studies of Stone Tools. Journal of Archaeological Science. 26:527-546.

Iizuka, Y, Peter Bellwood and H.c Hung. 2005 A nondestructive mineralogical study of nephritic artefacts from Itbayat Island, Batanes, northern Philippines Journal of Austronesian Studies. 1(1) 80-105.

Keeley, L. 1980. Experimental Determination of Stone Tool Uses: A Microwear Analysis. Chicago: University of Chicago Press. Kilmurry, K. 1982. "The Manufacture of Stamford Ware: An Application of Thin-sectioning and Neutron Activation Analysis," in Current research in ceramics: Thin-section studies, vol. 32. Edited by I. Freestone, C. Johns, and T. Potter, pp. 105-111. London: British Museum.

Iizuka, Y., and H. c. Hung. 2005. Archaeomineralogy of Taiwan Neprite: Sourcing Study of Nephritic Artifacts from the Philippines. Journal of Austronesian Studies 1(1)33-79. 87

Caves: unpublished report for the Evans Fund, the Society of Southeast Asian Studies, and the Niah cave Project (AHRB).

Koenigswald, G. H. R. v. 1956. Fossil Mammals from the Philippines. Quezon City: The National Research Council of the Philippines. Koenigswald, H. h. R. v. 1958. Preliminary Report of a Newly-Discovered Stone Age Culture from Northern Luzon, Philippine Islands. Asian Perspectives 2:69-70.

Lien, C.-m. 1995. "Some Aspects of Peinan Settlement Patterns," in Austronesian Studies Relating to Taiwan. Edited by P. Li, C.-h. Tsang, Y. Huang, D. Ho, and C. Tseng, pp. 353-377. Taipei.

Koomoto, M. 1983. "General Survey in Batan Island," in Batan Island and Northern Luzon, pp. 17-68. Japan: University of Kumamoto, Faculty of Letters.

Lin, M., C.-C. Chu, R. Broadberry, L.-C. Yu, -. J.-H. Loo, and J. Trejaut. 2005. "Genetic Diversity of Taiwan's Indigenous Peoples," in The Peopling of East Asia. Edited by L. Sagart, R. Blench, and A. Sanches-Mazas, pp. 230-247. London: RoutledgeCurzon.

Kress, J. 2004. "The Necrology of Sa'gung Rockshelter and Its Place in Its Place in Philippine Prehistory," in Southeast Asia Archaeology: Wilhelm G. Solheim II Festchrift. Edited by V. Paz, pp. 239-275. Quezon City: University of the Philippines Press.

Mackenzie, W. S., and A. E. Adams. 1994. A Colour Atlas of Rocks and Minerals in Thin-section. London: Manson Publishing.

Krigbaum, J. 2003. Neolithic subsistence patterns in northern Borneo reconstructed with stable isotopes of enamel. Journal of Anthropological Archaeological 22:292-304.

Maeda, Y., F. Siringan, A. Omura, R. Berdin, Y. Hosono, A. Atsumi, and T. Nakamura. 2004. Higher-thanpresent Holocene mean sea levels in Ilocos, Palawan and Samar, Philippines. Quaternary International 115-116:15-26.

Kuhn, S., and M. Stiner. 2001. "The antiquity of huntergatherers," in Hunter-gatherers: an interdisciplinary perspective. Edited by C. PanterBrick, R. Layton, and P. Rowley-Conwy, pp. 99142. Cambridge: Cambridge University Press.

Magee, J. 1988. "Chapter 2. Oriented soil samples from Tingkayu Basin," in Archaeological Research in South-eastern Sabah. Sabah Museum Monograph 2. Edited by P. Bellwood, pp. 31-37. Kota Kinabalu: Sabah Museum and State Archives.

Lampert, C., I. Glover, R. Hedges, C. Heron, T. Higham, B. Stern, R. Shoocongdej, and G. B. Thompson. 2003. Dating resin coating on pottery: the Spirit Cave early ceramic dates revised. Antiquity 77:126-133.

Majid, Z. 1982. The West Mouth, Niah, in the prehistory of Southeast Asia. Sarawak Museum Journal 31:1200.

Latinis, K. 2000. The Development of Subsistence System Models for Island Southeast Asia and nearOceania: The Nature and Role of Arboriculture. World Archaeology 32:41-67.

—. 1990. The Tampanian problem resolved: Archaeological evidence of a late Pleistocene lithic workshop. Modern Quaternary Research in Southeast Asia 11:71-96.

Lentfer, C. J., and W. E. Boyd. 1998. A Comparison of Three Methods for the Extraction of Phytoliths from Sediments. Journal of Archaeological Science 25:1159-1183.

—. Editor. 1994. The Excavation of Gua Gunung Runtuh and the Discovery of the Perak Man in Malaysia. Kuala Lampur: Department of Museums and Antiquity Malaysia.

—. 1999. An Assessment of Techniques for the Deflocculation and Removal of Clays from Sediments Used in Phytolith Analysis. Journal of Archaeological Science 26:31-44.

Majid, Z., M. Saidin, S. C. Soon, and Z. Hassan. 1994. "Artifacts from the Gua Gunung Runtuh Excavation," in The Excavation of Gua Gunung Runtuh and the Discovery of the Perak Man in Malaysia. Edited by Z. Majid, pp. 149-168. Kuala Lampur: Department of Museums and Antiquity Malaysia.

Lentfer, C. J., M. Therin, and R. Torrence. 2002. Starch Grains and Environmental Reconstruction: a Modern Test Case from West New Britain, Papua New Guinea. Journal of Archaeological Science 29:687-698.

Mathisen, M. E. 1980. Plio-Pleistocene geology of Central Cagayan Valley, Northern Luzon Philippines. PhD Dissertation, Iowa State University.

Levi-Sala, I. 1986. Use Wear and Post-depositional Surface Modification: A Word of Caution. Journal of Archaeological Science 13:229-244.

Mathisen, M. E., and C. F. Vondra. 1983. The fluvial and pyroclastic deposits of Cagayan Basin, Northern Luzon, Philippines-an example of non-marine volcaniclastic sedimentation in an interarc basin. Sedimentology 30:369-392.

—. 1996. A Study of Microscopic Polish on Flint Implements. BAR International Series 629. Lewis, H. 2003. The potential of soil micromorphology in Southeast Asian archaeology: preliminary work at Niah Cave Sarawak, Malaysia, and Ille Cave, Palawan, Philippines. Hukay. Journal of the University of the Philippines Archaeological Studies Program 5:60-72.

Mauldin, R., and D. Amick. 1989. "Investigating Patterning in Debitage from Experimental Bifacial Core Reduction," in Experiments in Lithic Technology, vol. 528. Edited by D. Amick and R. Mauldin, pp. 66-88. Oxford: BAR International Series. Meacham, W. 1988. On the Improbability of Austronesian Origins in South China. Asian Perspectives 16:89106.

—. 2004. The soil micromorphological potential of cultural sediments from cave sites in Island Southeast Asia: progress report on work at Niah, Ille and Tabon 88

Theory, vol. 3. Edited by M. Schiffer, pp. 57-100. Tucson: University of Arizona.

Mihara, S., K. Miyamoto, T. Kurosaka, T. Nakamura, and H. Koike. 2004. AMS 14C Dating using Black Pottery and Fiber Pottery. Radiocarbon 46:407412.

Nesse, W. 2004. Introduction to Optical Mineralogy, 3rd edition. New York: Oxford University Press.

Mijares, A. S. 2001. An Expedient Lithic Technology in Northern Luzon, Philippines. Lithic Technology 26:138-152.

Nguyen Khac Su, Pham Minh Huyen, and Tong Trung Tin. 2004. "Northern Vietnam from the Neolithic to the Han period," in Southeast Asia. Edited by I. Glover and P. Bellwood, pp. 177-208. London: RoutledgeCurzon.

—. 2002. The Minori Cave Expedient Lithic Technology. Quezon City: University of the Philippines Press. —. 2004. Lithic Analysis of recently Excavated Tabon Cave Flakes. Proceedings of the Society of Philippine Archaeologists 2:15-19.

Nguyen Viet. 2005. The Da But Culture: Evidence for Cultural Development in Vietnam during the Middle Holocene. Bulletin of the Indo-Pacific Prehistory Association. 25: 89-93

—.2005. The Nagsabaran Shell Midden Site: A soil micromorphology approach. Hukay. Journal of the University of the Philippines Archaeological Studies Program. Hukay 8:1-12

Nguyen, X. H. 1998. "Rice remains from various archaeological sites in North and South Vietnam," in Southeast Asian Archaeology 1996. Edited by M. Klokke and T. de Brujin. Leiden: Centre for Asian Studies, University of Hull.

Mijares, A. S., and S. C. Jagoon. 2001. Archaeological Survey of Itbayat Island, Batanes Province, northern Philippines. Philippine Quarterly of Culture and Society 29:296-308.

O’Connor, Sue, and Peter Veth. 2005. Early Holocene shell fish hooks from Lene Hara Cave, East Timor establish complex fishing technology was in uses in Island Southeast Asia five thousand years before Austronesian settlement. Antiquity 79:249-256.

Morgenstein, m. E., S. Luo, T.-L. Ku, and J. Feathers. 2003. Uranium-series and Luminescence dating of Volcanic Lithic Artefacts. Archaeometry 45:503518.

O’Connor, Sue, Matthew Spriggs and Peter Veth. 2002. Excavation at Lene Hara Cave establishes occupation in East Timor at least 30,000-35,000 years ago. Antiquity 76:45-50.

Morrison, K. 2002. "Historicing adaptation to history: foragers-traders in South and Southeast Asia," in Foragers-Traders in South and Southeast Asia. Edited by K. Morrison and L. Junker, pp. 1-20. Cambridge: Cambridge University Press.

Odell, G., and F. Odell-Vereecken. 1980. Verifying the Reliability of Lithic Use-Wear Assessments by 'Blind Tests'. Journal of Field Archaeology 7:87120.

Morwood, M., R. Soejono, R. Roberts, T. Sutikna, C. Turney, E. Westaway, W. Rink, J. Zhao, G. van den Bergh, R. A. Due, D. Hobbs, M. Moore, M. Bird, and L. Fifield. 2004. Archaeology and age of new Homonin from Flores in eastern Indonesia. Nature 431:1087-1091.

Ogawa, H. 2000a. "Excavation of the Mabangog Cave, San Mariano, Lal-lo, Cagayan, Philippines," in Excavation of the Lal-lo Shell Middens. Edited by H. Ogawa, pp. 194-212. Tokyo: Tokyo University of Foreign Studies.

Movius, H. 1944. Early Man and Pleistocene Stratigraphy in Southern and Eastern Asia. Papers of the Peabody Museum 19.

—. 2000b. "Introduction: The excavation of Lal-lo shell midden," in Excavation of Lal-lo Shell Middens. Edited by H. Ogawa, pp. 1-15. Tokyo: Tokyo University of Foreign Studies.

—. 1948. The lower Paleolithic cultures southern and eastern Asia. Transactions of American Philosophical Society 38:329-420.

—. 2002a. Archaeological Research on the Lower Cagayan River. Tokyo: Tokyo University of Foreign Studies.

Mudar, K. 1985. "Bearded Pigs and Beardless Men: Predator-Prey Relationships Between Pigs and Agta in Northeastern Luzon, Philippines," in The Agta of Northeastern Luzon: Recent Studies. Edited by P. B. Griffin and A. Estioko-Griffin, pp. 69-84. Cebu.

—. 2002b. "Chronological Context of Non-Decorated Black Pottery Phase from Lal-lo Shell Middens-Analysis of the Black Pottery from Conciso Site in the Lallo Shell Middens, Northern Luzon, Philippines," in Archaeological Research on the Lower Cagayan River. Edited by H. Ogawa, pp. 103-115. Tokyo: Tokyo University of Foreign Studies.

Mulvaney, D. J., and R. Soejono. 1972. The AustralianIndonesian Archaeological Expedition to Sulawesi. Asian Perspectives 13:163-178.

—. 2002c. "Chronological Study on the Red-slipped Pottery of Lal-lo Shell Midden Special Reference on the non decorated Red-slipped Pottery under the Shell Middens," in Archaeological Research on the Lower Cagayan River. Edited by H. Ogawa, pp. 79-101. Tokyo: Tokyo University of Foreign Studies.

Murphy, C. P. 1986. Thin-section of Soils and Sediments. Great Britain: A B Academic Publishers. Naranjo, S. 1984. "A Technical Report on the Training on Material Testings and Evaluation with Case Study on the Analysis of Four Potsherds from Selected Sites," in Manuscript of the National Museum, Manila.

Omoto, K. 1981. The Genetic Origin of the Philippine Negritos. Current Anthropology 22:421-422.

Nelson, M. 1991. "The Study of Technological Organization," in Archaeological Method and 89

Peacock, D. 1970. The Scientific Analysis of Ancient Ceramics: A Review. World Archaeology 1:375389.

—. 1985. "The Negritos: genetic origins and microevolution," in Out of Asia. Edited by R. Kirk and E. Szathmary, pp. 123-131. Canberra: Journal of Pacific History.

Peterson, J. 1978. The Ecology of Social Boundaries: Agta Foragers of the Philippines. Urbana: University of Illinois Press.

Oppenheimer, S. 2004. The 'Express Train from Taiwan to Polynesia': on the congruence of proxy lines of evidence. World Archaeology 36:591-600.

Peterson, J., and W. Peterson. 1977. "Implication of Contemporary and Prehistoric Exchange Systems," in Sunda and Sahul: Prehistoric Studies in Southeast Asia, Melanesia and Australia. Edited by J. Allen, J. Golson, and R. Jones, pp. 533-564. London: Academic Press.

Oppenheimer, S., and M. Richards. 2002. "Polynesians: Devolved Taiwanese Rice Farmers of Wallacean Maritime Traders with Fishing, Foraging and Horticultural Skills?" in Examining the farming/language dispersal hypothesis. Edited by P. Bellwood and C. Renfrew, pp. 287-298. Cambridge: McDonald Institute of Archaeological Research, University of Cambridge, Cambridge.

Peterson, W. 1974a. Anomalous Archaeological Sites of Northern Luzon and Models of Southeast Asian Prehistory. Ph D Dissertation, University of Hawaii, Manoa.

Orogo, A. 1982. "Minori cave Chamber A Archaeological Excavation," in Manuscript of the National Museum. Manila.

—.1974b. Summary report of two archaeological sites from North-eastern Luzon. Archaeology and Physical Anthropology in Oceania 9(1):26-35.

Papathanasiou, A. 2003. Stable Isotope Analysis in Neolithic Greece and Possible Implications on Human Health. International Journal of Osteoarchaeology 13:314-324.

Piperno, D. 1988. Phytolith Analysis. New York: Academic Press. Poloni, E., A. Sanches-Mazas, G. Jacques, and L. Sagart. 2005. "Comparing linguistic and genetic relationship among East Asian populations: a study of the RH and GM polymorphism," in The Peopling of East Asia. Edited by L. Sagart, R. Blench, and A. Sanches-Mazas, pp. 250-272. London: RoutledgeCurzon.

Parr, J. 2002. A comparison of heavy liquid floatation and microwave digestion techniques for the extraction of fossil phytoliths from sediments. Review of Paleobotany and Palynology 120:315-336. Parr, J., V. Dolic, G. Lancaster, and W. E. Boyd. 2001. A microwave digestion method for the extraction of phytoliths from herbarium specimens. Review of Paleobotany and Palynology 113:203-212.

Reid, L. 1994. "Unravelling the linguistic histories of Philippine Negritos," in Language Contact and Change in Austronesian World. Edited by T. Dutton and D. Tryon, pp. 443-476. Berlin: Mouton de Gruyter.

Pawley, A. 2002. "The Austronesian Dispersal: Languages, Technologies and People," in Examining the farming/language dispersal hypothesis. Edited by P. Bellwood and C. Renfrew, pp. 251-274. Cambridge: McDonald Institute of Archaeological Research, University of Cambridge.

Reimer, P.J. and R Reimer. 2001. A marine reservoir correction database and on-line interface. Radiocarbon 43(2a):461-463.

Pawlik, A. F. 2004. The Paleolithic Site of Arubo 1 in Central Luzon, Philippines. Bulletin of the IndoPacific Prehistory Association 2:3-12.

Reynolds, T. E. G. 1993. Problems in Stone Age of SouthEast Asia. Proceedings of the Prehistory Society 59:1-15.

Pawlik, A. F., and W. Ronquillo. 2003. The Paleolithic in the Philippines. Lithic Technology 28:79-93.

Richards, T. 1988. Microwear Patterns on Experimental Basalt Tools. BAR International Series 460.

Paz, V. 1999. Neolithic human movement to Island Southeast Asia: the search for archaeobotanical evidence. Bulletin of the Indo-Pacific Prehistory Association. 2:-151-158.

Rogers, A., and J. Broughton. 2001. Selective Transport of Animal Parts by Ancient Hunters: A New Statistical Method and an Application to the Emeryville Shellmound Fauna. Journal of Archaeological Science 28:763-773.

—2001. Archaeobotany and Cultural Transformation: patterns of early plant utilisation in northern Wallacea. PhD Thesis, University of Cambridge, Cambridge.

Ronquillo, W. 1977a. "Report on the Preliminary Archaeological Activities at Pedro Pagulayan Cave, San Roque, Penablanca, Cagayan Province," in Manuscript of the National Museum. Manila.

—. 2002. "Island Southeast Asia: Spread or Friction Zone?," in Examining the farming/language dispersal hypothesis. Edited by P. Bellwood and C. Renfrew, pp. 275-286. Cambridge: McDonald Institute of Archaeological Research, University of Cambridge.

—. 1977b. "Report on the Preliminary Archaeological Excavation of Alejandro Malanos Cave, San Roque, Penablanca, Cagayan province," in Manuscript of the National Museum. Manila. —. 1978. "Progress Report on the Archaeological Excavations of Rabel cave, San Roque, Penablanca," in Manuscript of the National Museum. Manila.

—. 2005. Rock Shelters, Caves, and Archaeobotany in Island Southeast Asia. Asian Perspectives 44:107118.

90

Shoocongdej, R. 2000. Forager Mobility Organization in Seasonal Tropical Environments of Western Thailand. World Archaeology 32:14-40.

—. 1981. The Technological and Functional Analysis of Lithic Flake Tools from Rabel Cave, Northern Luzon Philippines. Vol. 13. Manila: Anthropological Paper.

Shott, M. 1994. Size and Form in the Analysis of Flake Debris: Review and Recent Approaches. Journal of Archaeological Method and Theory 1:69-110.

—. n.d. "Preliminary Results of the Technological Analysis of Rabel Cave Lithic Materials," in Manuscript of the National Museum. Manila.

Shutler, R., and J. C. Marck. 1975. On the dispersal of the Austronesian horticulturists. Archaeology and Physical Anthropology in Oceania 10:81-113.

Ronquillo, W., and R. Santiago. 1977. "Archaeological Caves and Open Sites Exploration at Penablanca, Cagayan Province," in Manuscript of the National Museum. Manila.

Shutler, R., and M. E. Mathisen. 1979. Pleistocene studies in the Cagayan Valley of northern Luzon, Philippines. Journal of the Hongkong Archaeological Society 8:105-114.

Ronquillo, W., R. Santiago, S. Asato, and K. Tanaka. 1993. The 1992 Archaeological Reexcavation of the Balobok Rockshelter, Sanga Sanga, Tawi Tawi Province Philippines: A Preliminary Report. Journal of Historiographical Institute:1-40.

Simanjuntak, T., and I. N. Asikin. 2004. Early Holocene Human Settlement in Eastern Java. Bulletin of the Indo-Pacific Prehistory Association 24:13-19.

Ross, Malcom. 2005. The Batanic languages in relation to the early history of the Malayo-Polynesian family. Journal of Austronesian Studies. 1(2):1-24.

Simanjuntak, T., and H. Forestier. 2004. "Research Progress on the Neolithic in Indonesia: Special Reference to the Pondok Silabe Cave, South Sumatra," in Southeast Asian Archaeology: Wilhelm G. Solheim II Festschrift. Edited by V. Paz, pp. 104-118. Quezon City: University of the Philippines Press.

Rots, V., and B. Williamson. 2004. Microwear and residue analyses in perspective: the contribution of ethoarchaeological evidence. Journal of Archaeological Science 31:1287-1299.

Snow, B. E., and R. Shutler. 1985. The archaeology of Fuga Moro Island: new approaches for the isolation and explanation of diagnostic ceramic assemblages in Northern Luzon, Philippines. Cebu: San Carlos Publications, University of San Carlos.

Ruddle, K, Johnson D, Townshed P, Rees J. 1978 Palm Sago: A Tropical Starch from Marginal Lands. Canberra, Australian National University Press Sagart, L. 2005. "Tai-Kadai as a sub-group of Austronesian," in The Peopling of East Asia. Edited by L. Sagart, R. Blench, and A. SanchesMazas, pp. 161-176. London: RouthledgeCurzon.

Snow, B. E., R. Shutler, D. E. Nelson, J. S. Vogel, and J. R. Southon. 1986. Evidence of rice cultivation in the Philippines. Philippine Quarterly of Culture and Society 14:3-11.

Sagart, L., R. Blench, and A. Sanches-Mazas. Editors. 2005. The Peopling of East Asia. London: RoutledgeCurzon.

Solheim, W. I. 1972. Northern Thailand, Southeast Asia, and World Prehistory. Asian Perspectives 13:145159.

Saidin, M., and H. D. Tija. 1994. "Gua Gunung Runtah: The Cave," in The Excavation of Gua Gunung Runtuh and the Discovery of the Perak Man in Malaysia. Edited by Z. Majid, pp. 9-21. Kuala Lampur: Department of Museums and Antiquity Malaysia.

—. 1981. "The Philippine Prehistory," in The People and Art of the Philippines. Edited by G. Casal, R. Jose, E. Casino, and W. I. Solheim, pp. 17-83. Los Angeles: Museum of Cultural History, University of California.

Sanches-Mazas, A., E. Poloni, G. Jacques, and L. Sagart. 2005. "HLA Genetic Diversity and Linguistic Variation in East Asia," in The Peopling of East Asia. Edited by L. Sagart, R. Blench, and A. Sanches-Mazas, pp. 273-296. London: RoutledgeCurzon.

Sørensen, P. 1972. "The Neolithic Cultures of Thailand (and North Malaysia) and their Lungshanoid Relationship," in Early Chinese Art and Its Possible Influence in the Pacific Basin, vol. 2. Edited by N. Barnard, pp. 459-506. New York: Intercultural Arts Press.

Semenov, S. 1964. Prehistoric Technology. London: Adams and Mackay.

Spoer, A. 1973. "Zamboanga and Sulu: An Archaeological Approach to Ethnic Diversity," in Ethnology Monograph Department of Anthropology, University of Pittsburgh. Pittsburgh.

Shahack-Gross, R., F. Berna, K. Panagiotis, and S. Weiner. 2004. Bat guano and preservation of archaeological remains in cave sites. Journal of Archaeological Science 31:1259-1272.

Spriggs, M. 1999. Archaeological Dates and Linguistic Subgroups in the Settlement of the Island Southeast Asian-Pacific Region. Bulletin of the Indo-Pacific Prehistory Association 18:17-24.

Shea, J., and J. Klenck. 1993. An Experimental Investigation of the Effects of Trampling on the Results of Lithic Microwear Analysis. Journal of Archaeological Science 20:175-194.

— 2000. Spread of Southeast Asian Pleistocene and Neolithic maritime cultures. Modern Quaternary Research in Southeast Asia 16:51-75.

Shepard, A. 1976. Ceramics for the Archaeologists. Washington D.C.: Carnegie Institution of Washington.

—2003. Chronology of the Neolithic Transition in Island Southeast Asia and the Western Pacific: A View from 2003. The Review of Archaeology 24:57-80. 91

—. 2004. "The Continuity and the Gap of the Occupation of Shell Midden Sites in the Lower Reaches of the Cagayan River, Northern Luzon," in Southeast Asian Archaeology: Wilhelm G. Solheim II Festschrift. Edited by V. Paz, pp. 158-183. Quezon City: University of the Philippines Press.

Stephens, M., J. Rose, D. Gilbertson, and M. Canti. 2005. Micromorphology of Cave Sediments in the Humid Tropics: Niah Cave Sarawak. Asian Perspectives 44:42-55. Stoops, G. 1998. Key to the ISSS ' Handbook for Thinsection Description. Natuurwet Tijdschr 78:193203.

Tanaka, K., and A. Orogo. 2000. The Archaeological Excavation at Pamittan Site, Barangay Lanna, Solana, Cagayan Province, Philippines. Journal of Environmental Studies 8:113-141.

. —2003. Guidelines for the Analysis and Description of Soil and Regolith Thin-section. Wisconsin: Soil Science Society of America, Inc. Straus, L. G. 1980. The role of raw materials in lithic assemblage variability. Lithic Technology 9:68-72.

Tanudirjo, D A. 2001. Islands in Between: Prehistory of the Northeastern Indonesian Archipelago. PhD Thesis, Australian National University, Canberra.

Streeten, A. 1982. "Textural Analysis: An Approach to the Characterization of Sand Tempered Ceramics," in Current research in ceramics: Thin-section studies, vol. 32. Edited by I. Freestone, C. Johns, and T. Potter, pp. 123-134. London: British Museum.

—. 2004. "The Structure of Austronesian Migration into Island Southeast Asia and Oceania," in Southeast Asian Archaeology: Wilhelm Solheim II Festchrift. Edited by V. Paz. Quezon City: University of the Philippines Press.

Stiuver, M. GW. Perason and T Braziunas. 1986. Radiocarbon age calibration of marine samples back to 9000 cal yr BP. Radiocarbon 28(2b):9801021.

—. 2005. Long and continuos or short term and occasional occupation? The human use of Leang Sarru rock shelter in the Talaud Islands, northeastern Indonesia. Bulletin of the Indo-Pacific Prehistory Association. 25:15-20.

Sullivan, A. I., and K. Rozen. 1985. Debitage Analysis and Archaeological Interpretation. American Antiquity 50:755-779.

Thiel, B. 1980. Subsistence Change and Continuity in Southeast Asian Prehistory. PhD Thesis, University of Illinois at Urbana-Champaign.

Summerhayes, G. 2000. Lapita Interaction. Vol. 15. Terra Australis. Canberra: Department of Archaeology and Natural History and Centre for Archaeological Research, The Australian National University.

—. 1988. Austronesian Origin and Expansion: The Philippine Archaeological Data. Asian Perspectives 16:119-129. —. 1989. Excavation at the Lal-lo Shell middens, Northeast Luzon, Philippines. Asian Perspectives 17:71-94.

Sung, W. 1965. The Archaeological dates in western Taiwan. Taiwan Historica (in Chinese) 16(4):144155.

—. 1990a. Excavation at Arku Cave, Northeastern Luzon Philippines. Asian Perspectives 21:229-264.

Szabó, K., and S. O'Connor. 2004. Migration and complexity in Holocene Island Southeast Asia. World Archaeology 36:621-328.

—. 1990b. Excavation at Musang Cave, Northeastern Luzon, Philippines. Asian Perspectives 28:61-81.

Szabó, K., M. Swete Kelly, and A. Penalosa. 2004. "Preliminary Results from Excavation in the Eastern Mouth of Ille cave, Northern Palawan," in Southeast Asia Archaeology: Wilhelm G. Solheim II Festchrift. Edited by V. Paz, pp. 209-224. Quezon City: University of the Philippines Press.

Toth, N., D. Clark, and G. Liganue. 1992. The Last Stone Axe Makers. Scientific America 267:88-93. Tsang, C.-h. 1986. The Colonization of the P'eng-hu Islands. PhD Thesis, Harvard University. —.1992. Archaeology of P’eng hu Islands, Taipei. Academia Sinica, Institute of History and Philology, Taipei.

Tanaka, K. 1998. Preliminary Report of the Archaeological Excavation of Catugan Shell-midden (Dumbrigue Site), Lal-lo, Cagayan, Philippines. Bulletin of Chiba Keiai Junior College 20:149-177.

—. 1995. "New Archaeological Data from Both Sides of the Taiwan Straits and Their Implication fro the Controversy about Austronesian Origins and Expansion," in Austronesian Studies Relating to Taiwan. Edited by P. Li, C.-h. Tsang, Y. Huang, D. Ho, and C. Tseng, pp. 185-226. Taipei.

—. 2000. "Problems and Excavation of San Lorenzon III Shell Midden (Siriban Site) Lal-lo Cagayan, Philippines," in Excavation of Lal-lo Shell Middens. Edited by H. Ogawa: Tokyo University of Foreign Studies.

—. 2005. "Recent discoveries at a Tapenkeng culture site in Taiwan: implications for the problem of Austronesian origins," in The Peopling of East Asia. Edited by L. Sagart, R. Blench, and A. Sanches-Mazas, pp. 63-74. London: RoutledgeCurzon.

—. 2002. Ceramic Chronology in Northern Luzon: Typological Analysis of Pottery from the Lal-lo Shell-midden. PhD Thesis, University of the Philippines, Quezon City. —. 2003. The Excavated Pottery of the San Lorenzon III Shell- midden Site, Lal-lo, Northern Luzon Philippines. Journal of Southeast Asian Archaeology 23:93-111.

Tsang, C.-h., and R. Santiago. 2001. "Recent Archaeological Studies in Northern Luzon and Their Implication for the Prehistoric Relationship between Taiwan and the Philippines." 92

Main in Malaysia. Edited by Z. Majid, pp. 123140. Kuala Lampur: Department of Museums and Antiquity Malaysia.

International Conference on Austronesian Cultures: Issues relating to Taiwan, Taipei, 2001. Tsang, C.-h., R. Santiago, and H. c. Hung. 2001. "Preliminary Report on the 2nd Excavation of Nagsabaran Site," in Annual report of the AsianPacific Project of Academia Sinica. Taipei.

Zhao, Z., and D. Pearsall. 1998. Experiments for Improving Phytolith Extraction from Soils. Journal of Archaeological Science 25:587-598.

Tsirk, A. 1979. "Regarding Fracture Initiations," in Lithic use-wear analysis. Edited by B. Hayden, pp. 8396. New York: Academic Press. Underhill, P. 2005. "A Synopsis of Extant Y Chromosome Diversity in East Asia and Oceania," in The Peopling of East Asia. Edited by L. Sagart, R. Blench, and A. Sanches-Mazas, pp. 297-313. London: RouthledgeCurzon. Vaughan, P. 1985. Use-wear Analysis of Flake Stone Tools. Tucson: University of Arizona Press. Vondra, C. F., M. E. Mathisen, D. Burggraft, and E. Kvale. 1981. "Plio-Pleistocene Geology of Northern Luzon, Philippines," in Hominid Sites: Their Geologic Settings. Edited by G. J. Rapp and C. F. Vondra, pp. 255-310. Colorado: American Association for the Advancement of Science 63. Walker, P. 2005. Greater Sciatic Notch Morphology: Sex, Age, and Population Differences. American Journal of Physical Anthropology 127:1-7. Wasson, R., and R. Cochrane. 1979. Geological and Geomorphological Perspectives on Archaeological Sites in the Cagayan Valley, Northern Luzon, Philippines. Modern Quaternary Research in Southeast Asia 5:1-26. Wernstedt, F and Spencer, J. 1967. The Philippine Island World: A Physical, Cultural, and Regional Geography, Los Angeles. University of California Press. White, J., and W. Henderson. 2003. Pottery anatomy; review and selection of the basic nomenclature as a step towards a searchable rim form database for the Sakoh Nakkon Basin. Bulletin of the Indo-Pacific Prehistory Association 23:35-50. White, W. B. 1988. Geomorphology and Hydrology of Karst Terrain. Oxford: Oxford University Press. Williams, D. F. 1983. "Petrology of ceramics," in The Petrology of archaeological artefacts. Edited by D. Kempe and A. Harvey, pp. 301-329. Oxford: Clarendon Press. Winterhalder, B. 2001. "The behavioural ecology of huntergatherer," in Hunter-gatherer: an interdisciplinary perspective. Edited by C. Panter-Brick, R. Layton, and P. Rowley-Conwy, pp. 12-38. Cambridge: Cambridge University Press. Wright, P. 2003. Preservation or destruction of plant remains by carbonization? Journal of Archaeological Science 30:577-583. —. 2005. Flotation samples and some paleoethonobotanical implications. Journal of Archaeological Science 32:19-26. Zauyah, S. 1994. "Characteristic of Soil Layers in Gua Gunung Runtuh," in The Excavation of Gua Gunung Runtuh and the Discovery of the Perak 93

94

APPENDICES

95

Appendix A

The Excavation Plates

Plate A1 Location of the Eme Cave Complex

Plate A2 Excavation of Eme Cave

Plate A5 Excavation of Callao Cave

Plate A3 Eme Cave stratigraphic profile

Plate A6 Stratigraphic profile of Callao Cave

Plate A4 Eme Cave brown and black pottery sherds

96

Plate A11 Fengpitou (left, Photo by C. Hung) and Callao (right, Photo by J. Cameron) spindle whorls

Plate A7 Callao shell beads

Plate A8 Callao clay lingling o earring Plate A12 Callao Cave flake tools ( 25,000 BP)

Plate A9 Callao red-slipped, brown and black pottery rims

Plate A13 Excavation of Dalan Serkot Cave

Plate A10 Callao human pelvic bones

Figure A14 Dalan Serkot stratigraphic profile

97

Plate A15 Dalan Serko:t human phalanges

Plate A16 Dalan Serkot sherds with incise design on the rims

Plate A17 Dalan Serkot : A probable chopper tool of andesite

98

pedal cracks. The circular poroids have loose continuous infilling of crumbs, granules and rock fragments. Ground Mass and C/F ratio The matrix has a stipple speckle fabric with moderate birefringence. Under PPL the colour is yellowish red and at XPL it is reddish brownish. The C/F ratio is 10/90 with a texture of silty clay loam. Minerals Sub rounded to sub angular quartz grains 200-800 µm in size dominate the mineral content. Limestone fragments are sub rounded and 30-7000 µm in size range. Spherical radiolarian chert, about 1000 µm, is also identified. Other minerals are calcite grains (2000 µm), biotite (200-300 µm), olivine 100-350 µm, amphibole 300-500 µm, and plagioclase 100-600 µm. There are also 1200-2000 µm bone fragments. A round ash nodule (9 mm), which is pinkish gray in PPL and grey in XPL, was observed in the upper area of the section (Plates C2.7). An earthenware fragment was identified in the section It is black in PPL and dark red-black in XPL (Plate C2.9). It contains minerals such as quartz (200750 µm), pyroxene (400 µm), biotite (400 µm) and muscovite (300 µm). It has vughy and channel voids. Organic Materials Plant residues, (300-400 µm) with their structure still recognizable were observed. Charcoal 200-500 µm in size can be seen across the section covering 2 % of the area. Remains of partially decomposed roots (Plate C2.5) were identified in the mid portion of the section. Fabric Pedofeature There is one pedofeature in this section, red orange in PPL and red brown in XPL, of a very fine clay matrix (Plate C2.9). It has 50µm quartz inclusions. Textural Pedofeature Calcite micropan 300-500 µm thick can be observed in one of the peds. A coating of clay and micrite can be seen on the bone fragments (Plate C2.6). Crystalline Pedofeature Pseudomorphic calcite crystals partially impregnate the bones (Plate C2.7). Cryptocrystalline/amorphous Iron rich sub rounded nodules, about 30-40 µm (Plate C2.4). Excrement Pedofeature Rounded yellow orange excrement 100 µm in size can be observed infilling poroid channels.

Appendix B Descriptions of the Soil Micromorphology Thin-sections Eme 1 General Description Sample Eme 1(Plate C1.1) was taken between layers 3 and 4 (pH = 7). Layer 3 is a sandy silt loam with sherds, lithic and faunal remains. Layer 4 is sandy clay with lithics and faunal remains but no pottery. This thin-section (6 x 4 cm) was divided into two sub sections: the upper 4 cm consisting of a very porous deposit, and the lower 2 cm of denser deposit. The upper sub section is also looser in the right lateral area. The general colour is yellowish brown. The whole slide is about 15 % porous, with moderately sorted rounded to sub rounded peds and rock fragments. Porosity is estimated using a visual chart. Structure and Porosity Moderately developed crumb (100-500 µm) structure (Plate C1.2 and C1.3) with rounded micro aggregates. There are two zones of porosity, a compound packing void in the right lateral margin, and channels in the matrix in the left lateral and lower sections. The channels are not accommodating. Intra-pedal cracks are also observed in most peds and are accommodating. There are circular poroids (Plate C1.4) about 2-4 mm wide, possibly due to faunal activity. These are continuously filled by loose crumbs and sub-rounded micro aggregates, and rock fragments. Ground Mass and C/F ratio The ground mass has a yellowish red colour in plane polarized light (PPL) and dark yellowish red in cross polar light (XPL). It is moderately sorted with a stipple speckle fabric and high birefringence. The texture is silty clay loam with a C/F ratio of 10/90. Minerals There are bone fragments (Plate C1.5) from 500 µm to a maximum of 8mm in size. Minerals identified include quartz of rounded to sub rounded shape and 30-500 µm in size. Other minerals are plagioclase 200-350 µm, olivine 100200 µm, pyroxene 50-350 µm and amphibole 150-300 µm. Organic Materials There are an estimated 5 % organic materials in the sub layer. These include plant residues (Plate C1.7) from 1000-2000 µm in size and plant roots (Plates C1.8) with hair, and pollen grains about 2000 µm in diameter. A plant residue with its structure still intact, probably of wood was observed in the lower section. Charcoal fragments (Plate C1.6) about 500 to 1000 µm in size were observed covering 5 % of the area of the slide. Fabric Pedofeatures There are three fabric pedofeatures dispersed across the section. The first are dark nodules (Plate C1.9) about 1000-5000 µm in size. These are dark red to black in PPL and XPL with low birefringence. They cover about 10 % of the area of the section. The second fabric pedofeature is an orange brown in PPL and pale orange brown in XPL colour with high birefringence. The fabric is dusty reddish clay with 50-500 µm mineral inclusions of quartz and plagioclase. The third pedofeature is sub rounded iron rich nodules with a fabric similar to the matrix. The sizes range from 1000-1500 µm. Textural Pedofeature A typic coating of clay and micrite around most aggregates and rock fragments was observed. Crystalline Pedofeature Typic hypidiotopic microsparite crystals fill voids and connect peds and grains, about 100 µm (Plate C1.11). The walls have digitate sparite crystals. Calcite nodules, which consist of typic sparitic calcite with equigranular idiotopic structures, were identified. Cryptocrystalline/amorphous Rounded amorphous iron rich nodules about 150-500 µm in size made of fine clay material (Plate C1.10). They are red-black in PPL and black in XPL. Excrement Pedofeature Grains of rounded gray orange colour in PPL and filling circular poroid voids were identified.

Dalan Serkot 1 General Description The Dalan Serkot 1 sample (Plate C3.1) was taken between layers 1 and 2 (pH=6). Layer 1 is a sandy silt loam and associated with sherds, human teeth and phalanges, faunal remains, and a few flake tools. Layer 2, on the other hand, has primarily a flake tool assemblage with few faunal remains, in a silty clay loam sediment. The section viewed macroscopically has a massive structure with a large void in the upper left margin. The sediment colour is yellowish red. Structure and Porosity A well developed angular blocky structure, with intra pedal cracks and channels (Plates C3.2 and C3.3). The intra pedal cracks are accommodating. Channels are partially filled with silty clay and micrite. Porosity is estimated at about 5 %. Ground Mass and C/F ratio The ground mass is a stipple speckle fabric with moderate birefringence. Colour is reddish brown in PPL and dark reddish grey in XPL. The texture is silty clay loam, poorly sorted with a C/F ratio of 15:75. Minerals Limestone (600 µm-4mm) and quartz (50-1500 µm) that are sub rounded to round and sub angular in shape are abundant in the section. Other minerals identified are amphibole (60-600µm), biotite (100 µm), pyroxene (200-350 µm), calcite (1000µm), and radiolarian chert (350-500 µm). Bones (350 µm 5mm) and shells (8mm) were observed across the section. Organic Materials Charcoal fragments 700-1000 µm with sub angular and sub rounded shapes are abundant and estimated at about 7 % of the sample. Punctuations 50-85 µm are evenly distributed. Fabric Pedofeatures There are four fabric pedofeatures identified in the section. The first are reddish black in PPL and opaque in XPL that have accommodating cracks (Plate C3.7). They have quartz and amphibole inclusions in a fine grain matrix. The second has a dark red colour in PPL and red in XPL consisting of dirty red clay. Inclusions include quartz and amphibole, and have channel voids. The third is a sub rounded dirty clay nodule (255-850 µm). The last is subrounded nodule, with dark yellow colour (PPL) with punctuations, yellowish grey in XPL with low birefringence (200-3000 µm). Textural Pedofeature Micrite coatings of channel walls were visible. There are also typic coatings of clay and micrite on grains and nodules, and calcitic hypo coatings of voids. Some of the channels are partially in filled with silty clay (Plate C3.6). Crystalline Pedofeature Some of the bones are partially or completely pseudo morphosized with sparite crystals (Plate C3.5).

Eme 2 General Description The sample Eme 2 (Plate C2.1) was taken between layers 4 and 5. Layer 4 has been described above, while layer 5 is a sandy silt loam with no cultural materials. The thin-section, about 6 x 4 cm, does not have distinct sub layers. It is loose on the mid-right lateral side. The sediment colour is brown to strong brown. The peds are sub angular to sub rounded (2-7mm) and have faunal channels 4 mm wide. Limestone fragments as big as 7 mm can be observed. Structure and Porosity The section has a moderately developed granular structure (Plates C2.2 and C2.3). There are sub rounded to rounded aggregates 40-60 µm and crumbs 20-40 µm. The mid section has a porosity of 20 % and most peds have intra

99

Cryptocrystalline/amorphous Rounded iron rich nodules that are reddish black to black in PPL and opaque in XPL (100-350 µm) were observed. Depletion Pedofeature Some of the ground mass has lost its iron rich content and shows a yellowish colour in PPL. It has a diffuse boundary with the reddish brown adjacent material. Excrement Reddish yellow faunal excrement was seen in a channel (Plate C3.4), consisting of cylinders with rounded ends (1200 µm), circular (255- 340 µm) in cross section.

Infilling of micritic crystals of voids and nodules composed of micrite are seen. Cryptocrystalline/amorphous Fe rich nodules of red-black to black 200-800 µm in size were observed. Dalan Serkot 2.3 General Description This sub section is a massive volcanic ash deposit that unconformably overlaid clay sediment (Plates C4.8). Structure and Porosity Microstructure is single space porphyric with vughy voids 150-250 µm in diameter. Ground Mass and C/F ratio The groundmass, which is the volcanic ash, is gray in PPL and very dark gray in XPL. It has a low birefringence and undifferentiated fabric. Texture is sandy loam with a C/F ratio of 10/90. Minerals Mineral identified are mostly sub-angular in shape. They are quartz (20-250 µm), plagioclase (10-200µm), pyroxene (90-150 µm), and amphibole (40-150 µm). No bone or shell fragments were observed. Textural/Crystalline Pedofeature The voids have sparite typic coatings of 100 µm thicknesses. The crystals are idiotopic in shape.

Dalan Serkot 2.1 General Description The Dalan Serkot 2 sample (Plate C4.1) was taken between layer 2 and layer 3. Layer 3 is devoid of cultural remains and was identified as travertine in the field. In order to study both the top and bottom part of the section, it was cut and divided into 2 parts. This section is the upper part of Dalan Serkot 2 and is 4x4 cm in size. The section is massive matrix supported clast deposit with a reddish brown colour. Laminae 2 mm thick of calcium carbonate is visible in the upper part of the section. Below is a mixture of sediment and calcium carbonate 7mm thick. Structure and Porosity The mid section is composed of moderately developed sub angular to angular blocky structure (Plates C4.2 and C4.3). The peds are 300-1000 µm in size. Channels and intra pedal cracks are partially to fully accommodating. The lower section consists of weakly developed granular microstructure. The granules are about 50- 350 µm. The porosity is estimated at 5 %. Ground Mass and C/F ratio The fabric shows porostriation with moderate birefringence. The matrix has dark red (PPL) and dark reddish brown (XPL) colour. The C/F ratio is 7/93 with a clay texture. Minerals Limestone rock fragment are macroscopically visible (2-5 mm). Quartz grains, which are rounded to sub rounded, dominate the coarse fraction (50900 µm). Other minerals identified are pyroxene (50-450 µm), amphibole (200 –450 µm), muscovite (500 µm), plagioclase (50-450 µm) and calcite (170-2000 µm). Only one bone (500 µm) was observed in the section. Organic Materials Punctuations (40-200 µm) of probable organic remains are visible in the whole section. Only about 2 % of the area contains charcoal. Fabric Pedofeature Rare dirty red clay nodules were observed, about 200 µm in size. Textural Pedofeatures Typic clay coatings of grains and aggregates (Plate C4.5) were observed. Xenotipic cap-linking aggregates and infilling of channels 85-170 µm thick by sparite are visible in the upper and bottom sections of the slide. Calcitic hypocoating of voids was also observed (Plates B4.6 and B4.7). Crystalline Pedofeature The calcium carbonate laminae have a massive typic sparitic deposit (Plate C4.4). The thick layer is probably a precipitate or a travertine. Below it is a partial impregnation of the soil matrix with calcite crystals. Cryptocrystalline/amorphous Iron rich nodules about 300-450 µm in size were observed in the section.

Callao 1 General Description The Callao 1 sample (Plate C5.1) was taken between layers 3 and 4. Layer 3 is associated with sherds, shell beads, clay earring and human remains. Layer 4 also has human remains with ceramics and few flake tools and faunal remain. Layer 3 was identified in the field as a silt loam, while layer 4 is a silty clay loam. Layer 3 has a neutral pH of 6 and layer 4 is moderately alkaline (pH=8.5). The matrix of Callao 1 is strong brown in colour with sub angular peds, and is moderately sorted and has about 5 % general porosity. Structure and Porosity The microstructure is a moderately developed crumb microstructure (Plates C5.2 and C5.3). The crumbs were about 50-150 µm in size. There are nonaccommodating inter-pedal channels about 600 µm wide. Most sub angular peds have intra pedal cracks that are accommodating. Faunal activity can be observed, with a number of circular poroids (Plate C5.4) 5 mm wide and with continuous infillings of crumbs and rock fragments. Ground Mass and C/F ratio The ground mass has stipple speckle fabric with high birefringence. The colour is brownish yellow in PPL and yellowish brown in XPL. Texture is silty clay loam with a C/F ratio of 15/85. Minerals Rock fragments of andesite (1000-2500 µm), basalt (1800 µm) and limestone (800-1000 µm) can be observed. Minerals identified are olivine (150 µm), plagioclase (200-1400 µm), amphibole (100-250 µm), calcite (500 µm), pyroxene (300 µm), radiolarian chert (500-700 µm) and quartz (300-800 µm). There is a pottery fragment (12 mm) at the top right edge of the slide (Plate C5.6). It has a dark reddish brown colour (PPL) and is black in XPL, with oriented voids. It contains quartz, plagioclase, amphibole and biotite inclusions. The section contains bones 300- 8000 µm in size and shell fragments (1000µm). Organic Material Charcoal occupies about 5 % of the area and burnt plant residues (probably of wooden origin) about 200 to 2000 µm in size (Plate C5.7). Fabric Pedofeatures There are two fabric pedofeatures identified (Plates C5.10). The first are red orange sub rounded nodules (600 µm) that contain dusty clay. The others are yellow orange nodules that are yellow gray to isotropic in XPL (200 µm). Textural Pedofeature The circular poroids have typic hypocoating of Fe rich clay. There are typic coating of impure clay (Plate C5.9) and micrite on grains and aggregates. There are also laminated calcite coatings around bones. Crystalline Pedofeature Calcite nodules are abundant in the sample. Some of the bones have partial pseudomorphisis (Plate C5.8) of calcite crystals (micrite and sparite). Cryptocrystalline/amorphous Red to red-black Fe rich nodules are abundant, that are sub angular to sub rounded (700-2000 µm) in shape. Geodic Fe nodules are also visible. Depletion Pedofeature In the mid section of the slide, there is a lighter brown colour of the sediment, that might have lost Fe minerals as compared the peds besides it. Excrement Pedofeature Lenticular (Plate C5.5) and sub rounded faunal excrement can be seen, particularly inside circular poroids.

Dalan Serkot 2.2 General Description The lower portion of the sample is composed of two sub layers. The upper is massive yellowish red with blocky aggregates. A light gray deposit unconformably underlies the matrix (Plate C4.7). The upper matrix is a continuation of the main matrix of the Dalan Serkot 2.1 sample. Structure and Porosity A mixed structure is evident in the matrix, with a moderately developed subangular blocky structure and moderately developed granular microstructure. Channels, which are partially accommodating, separate the aggregates. Intra pedal cracks are also present. Porosity is estimated at 5 %. Ground Mass and C/F ratio Porostriation was observed in the groundmass, which has moderate birefringence and is moderately sorted. The colour is dark red in PPL and dark reddish brown in XPL. Texture is also clay with 7:93 C/F ratio. Minerals Chert fragments 300-650 µm in size are visible as well as quartz grains (50500 µm). Other minerals are plagioclase (100-300 µm, amphibole (200-300 µm), pyroxene (80-250 µm), and calcite (650 µm). A single shell fragment (350 µm) can be seen. Organic Material Punctuations of probable organic origin or Fe/Mg oxides are spread through out the matrix. Their sizes are about 40-100 µm. No charcoal was identified. Fabric Pedofeature Rare dusty red clay nodules were observed. Textural Pedofeature Typic iron rich clay coatings on grains and aggregates were observed. Crystalline Pedofeature

Callao 2 General Description The Callao 2 sample (Plate C6.1) was taken between layers 4 and 5. Layer 4 was described above. Layer 5 only contains bat bones and loose sand as

100

observed in the field. The thin-section (6x4 cm) looks massive macroscopically. It has matrix-supported clast and is moderately sorted, and has about 2 % general porosity. There is no sub layer, but two zones were identified. These are a yellowish red matrix and a brown matrix in the mid right lateral area. Structure and Porosity Crumbs and granules (Plates C6.2 and C6.3) make up the matrix and have complex packing voids. Inter-pedal cracks are present in most peds and nodules. Porosity is estimated at 10 %. Ground Mass and C/F ratio The ground mass has stipple speckle fabric and is moderately sorted. Texture is a sandy silt loam with a C/F of 12/83. The main zone has yellowish red colour (PPL) and dark reddish brown in XPL with low birefringence. The second zone has brownish yellow colour in PPL and is dark yellowish brown in XPL, with medium birefringence. Minerals There are 4 mm andesite rock fragments as well as 1500 µm limestone fragments. Other minerals identified are amphibole (300-500 µm), pyroxene (400-800 µm), radiolarian chert (600-2000 µm), and calcite (400-800 µm). There is an estimated 5 % abundance of bones, which are 40 µm to 13mm in size. Organic Materials The slide has frequent (20 %) burnt organic material. This includes identifiable wood tissue residue (500 µm) (Plate C6.4), disaggregated plant residue, charcoal and punctuations (50 µm). Fabric Pedofeatures There are four fabric pedofeatures identified. The first (1200-1600 µm) are red-orange colour (PPL) with fine-grained materials (Plate C6.7). The second are a mixture of red orange and dark red fabric (Plate C8). The third are layers of yellowish red to dark red colour fabric that may have been burnt organic materials (Plates C6.9). And the fourth are rounded nodules that are yellow in PPL and isotropic (300-1200 µm), which could be bones (Plate C6.10). Textural Pedofeatures There are a number of coatings occurring through out the thin-section. They include hypocoating of calcite aggregates, typic clay coating of grains and bones (Plate C6.6), and compound coating of aggregates with silty clay, Fe compounds and micrite. Crystalline Pedofeature A large crystalline pedofeature at the upper end of the thin-section about 13 mm in size was observed. The feature was partially burnt. It could be pseudomorphising of bone with micrite and sparite (Plate C6.5). Cryptocrystalline/amorphous Rounded to sub rounded monomorphic Fe compounds are abundant in the slide. They are reddish brown in PPL and 900-1200 µm in size. Depletion Pedofeature The second zone at the mid right lateral could be a depletion zone that has lost the iron rich content that its surrounding area contains.

Typic clay coating on grains and typic micrite coating on aggregates were observed. Crystalline Pedofeature There is complete pseudomorph of bone by micrite. The zones have partial impregnation with micrite crystals. Cryptocrystalline/amorphous Reddish brown with sub angular to sub rounded nodules are distributed in the section. Some have geodic forms.

Callao 3

Callao 4.2

Callao 4.1 General Description Callao sample 4 (Plate C8.1) was taken between layers 7 and 8; the former has no cultural deposit while the latter has flake tools. Layer 7 is identified in the field as loose sand with a pH of 7, while layer 8 is silt loam and moderately acidic (pH of 5.5). There are two sub-layers and laminae of dark reddish brown aggregates (2001400 µm) that separate the sub layers. The upper sub layer, which is yellowish brown in colour with the laminae, will be designated 4.1. The sublayer is massive with preferred horizontal channels. Structure and Porosity The sub layer has a moderately developed sub angular blocky structure (Plate C8.2 and C8.3). Most peds have accommodating cracks. Peds are separated by partially accommodating channels. Porosity is estimated at 5 %. Ground Mass and C/F ratio The fabric shows reticulate striations with moderate birefringence. The matrix is yellowish red in PPL and dark reddish gray in XPL, with moderate sorting. The texture is clay with 4:96 C/F ratio. Minerals Quartz (50-600 µm) and plagioclase (350-450 µm) predominate in the grain inclusions. There are also amphibole minerals (90-350 µm). Rare chalcedony (1500 µm) and volcanic ash granules (450 µm) were observed. No bone or shell was found. Organic Materials Punctuations and a few charcoals (2 %) were observed. Plant residues are also present (Plate C8.5) Fabric Pedofeatures Three fabric pedofeatures are identified. The first are the laminae cited above, which are dark red to red orange in PPL and dark reddish brown in XPL (100800 µm). They are made of fine materials (Plate C8.11). The second are dirty red clay nodules that are abundant in the sub layer. There are also pedofeatures. These are yellowish (PPL) and isotropic (1200 µm). Textural Pedofeatures Typic coating of voids and some aggregates of microsparite and clay are visible. Acicular micritic coating of a yellowish nodule (PPL) is notable (Plate C8.6). Cryptocrystalline/amorphous Iron rich nodules, black and reddish black 150-200 µm, were observed.

General Description The lower sub layer is reddish brown massive sediment. Horizontal channels are visible. Structure and Porosity It has a mixed structure (Plates C8.7 and C8.8), moderately developed sub angular and blocky for the upper part and weakly to moderately developed crumb microstructure in the lower part. Intra pedal cracks are visible in most peds, while channels as wide as 400 µm separates the aggregates. There is one Spherical poroid which has loose continuous infilling of crumbs and granules at the upper left of the section. Ground Mass and C/F ratio The matrix fabric is granostriated with moderate birefringence. It has a reddish yellow colour (PPL) and is brown in XPL. Texture is clay with C/F ratio of 2/98. Minerals Quartz (400-600 µm) and plagioclase (300-600µm) again are the most common minerals. Chert or Spherical radiolarian tests (300-1000 µm) were identified as well as amphibole minerals (800µm). Organic Material Charcoal (100-400 µm) occupies about 5 % of the sub layer area, and punctuations are widespread. Plant residues with structure still discernable can be seen. A long piece of plant residue (1600 µm), probably of a root, was observed (Plates C8.9). Fabric Pedofeature Rare dirty red clay nodules are seen. Textural Pedofeature There are dense incomplete infillings of clay, organic materials and mineral grains which are circular and lenticular in shape (250-5000 µm. A few typic micrite coatings of aggregates and voids are observed. Cryptocrystalline/amorphous Black ferruginous nodules about 300 µm are spread out of the sub layer. Excrement Pedofeature The circular to lenticular infilling of channels is probably faunal excrement (Plate C8.4).

General Description Callao sample 3 (Plate C7.1) was taken between layers 5 and 6, both only containing bat bones. Layer 5 has a pH of 7.5. There are no distinct sub layers for this sample, but it has distinct zones of light coloured deposit. The matrix has a dark brown colour while the zones are light brown. Spherical to oval chambers are visible throughout the section, with 10 % estimated porosity. These chambers are 1-5 mm in diameter and unfilled. Structure and Porosity The matrix has mixed structure consisting of well-developed crumb/granular microstructure, and a moderately developed lenticular microstructure (Plates C7.2 and C7.3). There are a few peds with cracks. The zones have a welldeveloped crumb microstructure (Plates C7.6) with about 15 % porosity. Ground Mass and C/F ratio The matrix has a yellowish red colour in PPL and is dark reddish grey in XPL, with low birefringence. The ground mass exhibits a stipple speckle fabric. The C/F ratio is 2/98 with clay texture. The zones have a yellowish brown colour in PPL and greyish brown in XPL, with moderate birefringence. Minerals Quartz grains (10-500 µm) are distributed across the section. Minerals identified are amphibole (200-300 µm) and pyroxene (200-400 µm). There are a few chert nodules (200-800 µm). Organic Materials There are abundant organic remains in the section, estimated at 10 % of the slide. Plant residues (100-2200 µm) with structures still visible and charcoal (100-400 µm) are giving the dark colour of the section. Charcoal is estimated to constitute 20 % of the section. Punctuations are also wide spread. An isotropic and grayish nodule was observed which could be an ash deposit (Plate C7.4). Fabric Pedofeature Dusty red clay nodules 300-600 µm in size were observed distributed in the section (Plate C7.5). There are also nodules that are yellow orange and dark reddish grey in XPL, which could be bones. Textural Pedofeature

101

Callao 5.1 General Description The Callao sample 5 (Plate C9.1) was taken between Layer 8, the flake tool assemblage, and layer 9, which is devoid of flakes. There are two sub- layers in this section: the upper 2.4 cm, which is reddish brown, and the lower brownish yellow. The upper sub-layer is a continuation of the lower section of Callao 4. It also has a massive structure with oriented channels. Structure and Porosity The upper sub layer has a mixed structure, moderately developed sub-angular blocky and moderately developed crumb (Plates C9.2 and C9.3). There are partially to fully accommodating channels that are up to 200 µm wide. Intra pedal cracks are also visible. Circular poroids of faunal origin are seen (400600 µm). Ground Mass and C/F ratio The matrix fabric is granostriated with moderate birefringence. It is a reddish yellow in PPL and brown in XPL. Texture is clay with C/F ratio of 2/98. Minerals Quartz (100-300 µm) grains are observed in the sub layer. Amphibole minerals (100-800 µm) are also identified. Chert/radiolarian tests (200-300 µm) are seen. Organic Material Plant residues with visible structures (200-1500 µm) are notable (Plate C9.7). Charcoal is estimated at 2 % of the sub layer and punctuations (50-200 µm) are also identified. Ashy deposits (1000µm) are observed (Plates B9.5). Fabric Pedofeature Rare dirty red clay nodules (200 µm) are seen. Textural Pedofeature/ Crystalline Pedofeature Typic coating and capping of aggregates with micrite crystals and clay (Plate C9.6) are observed. Cryptocrystalline/amorphous There are rounded to sub rounded reddish black nodules of ferruginous origin observed (250 µm). Excrement Pedofeature Rounded yellow excrements were observed filling circular poroids. Channels are also filled by lenticular excrement up to 3.5 mm long (Plate C9.4). Callao 5.2 General Description The lower sub layer generally has a brownish yellow colour. It has massive structure separated by horizontal channels. Structure and Porosity It has a mixed structure, weakly developed granular (100-300 µm) and moderately developed sub angular blocky (Plates C9. and C9.9). Channels, which are mostly horizontal in direction have a maximum width of 1200 µm and are partially accommodating. There are also vughs 600 µm in diameter, and intra pedal cracks. Ground Mass and C/F ratio The fabric is granostriated with moderate birefringence. The main matrix has a yellowish red colour in PPL and is reddish brown in XPL, with moderate sorting. Texture is clay with 2/98 C/F ratio. Minerals Quartz is still the major grain inclusion (50-700 µm) with plagioclase (250400 µm). There are still traces of amphibole (400-550 µm) and chert (200400 µm). Organic Materials Punctuations were observed in the sub layer. Strands of possible root residue were observed. It is colourless in PPL but yellowish and has a high birefringence in XPL (Plate C9.13). Fabric Pedofeatures There are three fabric pedofeatures identified. The first are red clay aggregates composed of 50 µm granules (Plate C12). The second are light yellowish nodules (Plate C9.11), which are almost isotropic and have a size range of 800-1300 µm. These could be traces of bones. The third are yellow orange nodules, which can be clean clay (Plate C9.12). Textural Pedofeature/Crystalline Pedofeature Typic coating and link capping of aggregates by micrite crystals was observed. There is more coatings and capping in this sub layer compared to the upper sub layer. Cryptocrystalline/amorphous There are iron rich sub angular red-black nodules spread throughout the sub layer (Plate C9.14). Excrement Pedofeature Lenticular in filling of channels, which are probably excrement around 2400 µm long.

102

Appendix C

Soil Micromorphology Plates (Bar is equal to 200 microns)

Plate C2

Plate C1

C1.6. Charcoal fragment

C1.1 Eme 1 Section (Flat bed Scanned)

C1.2. Moderately developed Crumb structure (PPL)

C1.3. Same as above (XPL) with stipple speckle fabric

C1.4. Part of a circular poroid due to faunal activity

C1.5. Bone with clay coatings

C1.7. Plant residue with structure still discernable

C2.1. Eme 2 Section

C1.8. Root at XPL

C2.2. Moderately developed granular structure (PPL)

C1.9. Fabric pedofeature 1, dark red to black with quartz inclusion (XPL)

C2.3. Same as above in XPL, stipple speckle fabric

C1.10. Iron rich nodule (PPL)

C2.4. Amorphous nodule (PPL)

C1.11. Micrite and micro sparite texture pedofeature infilling voids. (XPL)

C2.5 Roots (PPL)

103

Plate C3

C2.6. Coating of clay and micrite of aggregates (XPL)

C2.7.Partial impregnation of bone with micrite (XPL)

C3.6. Infilling of channels with silty clay

C3.1 Dalan Serkot 1 Section

C3.7. Fabric pedofeature 1, reddish black with mineral inclusions (XPL) Plate C4

C2.8. Ash nodule in PPL

C2.9. Earthenware sherd, black with mineral inclusions (PPL)

C3.2. Well developed angular blocky structure (PPL)

C3.3. Same as above in XPL. Stipple speckle fabric

C4.1. Dalan Serkot 2 Section

C2.10. Fabric pedofeature 1, red orange with quartz inclusion (PPL)

C3.4. Faunal excrements (PPL)

C4.2. Moderately developed sub angular to angular blocky structure (PPL) C3.5. Partially impregnated bone with micrite crystals

C4.3. Same as above (XPL) porostriated fabric 104

Plate C5

C5.6. Pottery fragment (PPL) C4.4. Travertine laminae (XPL)

C4.5. Clay coating on grain (PPL)

C4.6. Micrite crystal coating of walls. (XPL)

C4.7. Transition to volcanic ash layer (XPL)

C5.1. Callao 1 Section

C5.7. Plant residue and bone (PPL)

C5.2. Moderately developed crumb structure (PPL)

C5.8. Partial pseudomorphism of bone by calcite crystals

C5.3. Same as above (XPL) stipple speckle fabric

C5.9. Typic clay coating of grain.

C5.4. Part of circular poroid with infilling C4.8. Volcanic ash in XPL

C5.5. Excrement (orange) and charcoal (PPL)

105

C5.10. Fabric pedofeatures 1 (red) and 2 (yellow-orange) (PPL)

Plate C6

Plate C7

C6.6. Typic clay coating of bone

C6.1. Callao 2 Section

C6.2. Complex packing void of crumbs and granules (PPL)

C6.3. Same as above in XPL, stipple speckle fabric

C6.7. Fabric pedofeature 1, red orange nodule (PPL)

C6.8. Fabric pedofeature 2, mixture of orange and dark red fabric

C6.9. Fabric pedofeature 3 (PPL), layered yellowish red to dark red fabric

C6.4. Plant residue (PPL)

C7.1. Callao 3 Section

C7.2. Well developed crumb structure (PPL)

C7.3. Same as above in XPL, stipple speckle fabric

C7.4. Ash nodule in PPL C6.10. Fabric pedofeature 4, yellow nodule (PPL)

C6.5. Pseudomorphism of bone with calcite crystal (XPL)

C7.5. Fabric pedofeature 1 (red clay nodule) and 2 (yellow orange nodule) and charcoal in PPL

106

C7.6. Zone area in XPL, high birefringence

C8.5. Plant residue (PPL)

Plate C8

C8.11. Pedofeature 1, reddish orange nodule (PPL) Plate C9

C8.6. Acicular micrite coating of a grain (PPL)

C8.1.Callao 4 section

C8.7. Moderately developed sub angular blocky structure and crumb structure (Lower sub layer) in PPL

C9.1.Callao 5 section

C8.2. Moderately developed sub angular blocky structure (upper sub layer) in PPL

C8-8. Same as above in XPL, granostriated fabric

C9.2. Moderately developed sub angular blocky and crumb structures (upper sub layer) in PPL

C8.3. Same as above in XPL, reticulate striation fabric

C8.9. Plant root residue in PPL

C9.3. Same as above in XPL, granostriated fabric

C8.4. Lenticular excrements in channels (PPL)

C8.10. Same as above in XPL

C9.4. Sperical excrements in filling circular poroids (PPL)

107

C9.5. Ash nodule in PPL

C9.11. Fabric pedofeature 2, yellowish nodule (PPL)

C9.6. Typic clay coating of grain (PPL)

C9.12. Fabric pedofeature 3, yellow orange nodule (PPL)

C9.7. Plant residue (PPL)

C9.13. Plant residue (roots?) XPL

C9.8. Moderately developed sub angular blocky structure (lower sub layer) in PPL

C9.14. Amorphous nodules (PPL)

C9.9. Same as above in XPL, granostriated fabric

C9.10. Fabric pedofeature 1, red clay nodule (PPL)

108

Appendix D Flora and Fauna Plates

Faunal Remains

Plate D1 Peñablanca caves bat mandibles Plate D4 Dalan Serkot pig incisor and molar

Plate D2 Callao Cave deer teeth

Plate D5 Riverine shells

Plate D3 Callao Cave wild pig tooth and tusk

Plate D6 Land snail shells

109

Eme Cave Macrobotanicals (Photos by Paz and Carlos) Ceramic Period Plate D7 [a]Parenchymatous tissue and [b] Boehmeria cf.

platanifolia

Plate D8 [a] Wood fragments and [b] nut

Preceramic Period

Plate D9 [a] Fungi sclerotia, [b] Boehmeria cf. platanifolia, and [c] parenchymatous tissue

Plate D 10 [a] Seed and [b] nut fragments

Callao Cave Macrobotanicals Ceramic Period

Plate D11 [a] Parenchymatous tissues, [b] wood fragment, and [c] Boehmeria cf. platanifolia

110

Plate D12 [a] Seed, and [b] Spilanthes Jacq. sp

Preceramic Period

Plate D13 [a] Seeds, [b] parenchymatous tissues, and [c] fungi sclerotia

Dalan Serkot Cave Macrobotanicals Ceramic Period

Plate D14 [a] Fungi sclerotia, [b] Boehmeria cf. platanifolia, and [c] Spilanthes

Plate D15 [a] Parenchymatous tissues, [b] wood fragment, and [c] spheroid seeds

111

Appendix E

Flake Analysis Plates

Eme Flakes

Ceramic Period

Plate E2. G-228 at 100x, generic weak polish Plate E1. G-228 at 20x, feather termination flake scars

Plate E4. G-535 at 500x, smooth pitted polish Plate E3. G-535 at 6x, crescent break

Plate E5. G-536 at 20x, feather termination flake scars

Plate E6. G-536 at 100x, smooth-pitted polish

Plate E7. G-543 at 6x, crescent breaks

Plate E8 G-543 at 500x, smooth-pitted polish

Plate E9. G-556 at 10x, small feather termination flake scars

Plate E10. G-556 at 200x, generic weak polish

Plate E11. G-204 at 16x, crescent breaks Plate E12. G-204 at 500x, smooth pitted polish

112

Plate E13. G-505 at 16x, crescent break Plate E14. G-213 at 200x smooth-pitted polish

Plate E15. G-510 at 8x, hinge termination flake scars Plate E16. G-510 at 500x, smooth pitted polish with parallel striations

Plate E17 G-511 at 8x, crescent breaks Plate E18 G-511 at 500x, smooth pitted polish

Plate E19. G-518 at 10x, crescent breaks

Plate E20. G-518 at 500x, smooth pitted polish

Preceramic Period

Plate E22 G-495 at 500x, smooth pitted polish Plate E21. G-495 at 20x, crescent break

Plate E23. G-565 at 8x, crescent breaks

Plate E24. G-565 at 500x, smooth pitted polish

113

Plate E25. G-370 at 16x, hinge termination flake scars

Plate E26. G-370 at 500x, smooth-pitted polish

Plate E28. G-378 at 100x, smooth pitted polish Plate E27. G-378 at 20x, hinge termination flake scars

Plate E30. G-427 at 100x, smooth pitted polish Plate E29 G-427 at 20x, feather termination flake scars

Plate E33. G-441 at 16x, crescent breaks

Plate E34. G-441 at 200x, smooth pitted polish

Plate E35. G-401 at 10x, feather termination flake scars

Plate E36. G-401at 200x, generic weak polish

Plate E37. G-475 at 16x, hinge termination flake scars Plate E38 G-475 at 500x, smooth pitted

114

Plate E31. G-418 at 10x, hinge termination flake scars Plate E32. G-418 at 100x, smooth pitted polish

Plate E39. G-246 at 20x, hinge termination flake scars

Plate E40. G-246 at 100x external surface, well developed polish with parallel striations. Possible silica gloss

Plate E41. G-246 at 200x internal surface, well developed polish

115

Callao Cave Flakes Ceramic Layer

Plate E42. J-7295 at 10x, hinge termination flake scars Plate E43. J-7295 at 500x, smooth pitted polish

Plate E44 J3-7393 at 20x, hinge termination flake scars Plate E45 J3-7393 at 100x, generic weak polish

Plate E47. J3-7290 at 200x, generic weak polish

Plate E46. J3-7290 at 20x, feather termination flake scars

Plate E49. J3-7401 at 200x, smooth pitted polish

Plate E48. J3-7401 at 20x, hinge termination flake scars

Plate E50. J3-7402 at 8x, hinge termination flake scars

Plate E51. J3-7402 at 500x, smooth pitted polish

Plate E52. J3-7227 at 20x, crescent break with pronounce rounding

Plate E53. J3-7227 at 200x, smooth pitted polish

116

Plate E54. J3-7291 at 12x, crescent breaks Plate E55. J3-7291 at 200x smooth pitted polish

Plate E56. J3-7296 at 10x, hinge termination flake scars Plate E57. J3-7296 at 500x, smooth pitted polish

Plate E58 J3-7369 at 20x, feather termination flake scars Plate E59. J3-7369 at 200x, generic weak polish

Plate E60. J3-7395 at 16x, hinge termination flake scars Plate E61. J3-7395 at 200x, smooth pitted polish

Preceramic Layer

Plate E63. J3-7540 at 200x, generic weak polish Plate E62. J3-7540 at 20x, feather termination flake scars

Plate E64. J3-7543-20x, hinge termination flake scars

Plate E65. J3-7543-200x, generic weak polish

117

Plate E67. J3-7556 at 100x, smooth pitted polish

Plate E66. J3-7556 at 25x, hinge termination flake scars

Plate E69. J3-7446 at 500x, smooth pitted polish

Plate E68. J3-7446 at 16x, hinge termination flake scars

Plate E70. J3-7488 at 20x, hinge termination flake scars, intensive rounding

Plate E71. J3-7488 at 200x, well developed polish

Plate E72. J3-7553 at 20x, hinge termination flake scars

Plate E73. J3-7553 at 500x, well developed polish

Plate E74. J3-7552 at 20x, feather termination flake scars

Plate E75. J3-7552 at 100x, generic weak polish

Plate E76. J3-7541 at 25x, feather and hinge termination flake scars

Plate E77. J3-7541 at 200x internal surface, generic weak polish

Plate E78. J3-7541 at 200x external surface, smooth pitted polish 118

Dalan Serkot Flakes Ceramic Layer

Plate E79. I8-213 at 20x, feather termination flake scars

Plate E80. I8-213 at 200x, generic weak polish

Plate E81. I8-216 at 12x, hinge termination flake scars

Plate E82. I8-216 at 500x, smooth pitted polish

Plate E83. I8-361 at 16x, hinge termination flake scars Plate E84. I8-361 at 200x, smooth pitted polish

Plate E85. I8-362 at 10x, feather termination flake scars

Plate E86. I8-362 at 200x, generic weak polish

Plate E88. I8-204 at 200x, generic weak polish

Plate E87. I8-204 at 20x, feather termination flake scars

Plate E90. I8-211 at 200x, generic weak polish

Plate E89. I8-211 at 10x, feather termination flake scars

119

Plate E91. I8-212 at 12x, feather termination flake scars Plate E92. I8-212 at 200x, smooth pitted polish

Plate E93. I8-280 at 10x, crescent break Plate E94. I8-280 at 100x generic weak polish

Plate E95. I8-342 at 12x, crescent breaks Plate E96. I8-342 at 100x, generic weak polish

Plate E97. I8-368 at 16x, hinge termination flake scars Plate E98. I8-368 at 200x, well developed polish, possible silica gloss Preceramic Layer

Plate E99. I8-309 at 10x, hinge termination flake scars Plate E100. I8-309 at 500x, smooth pitted polish with parallel striation

120

Plate E101. I8-408 at 12x, hinge termination flake scars

Plate E102. I8-408 at 100x, smooth pitted polish

Plate E104. I8-420 at 200x, smooth pitted polish

Plate E103. I8-420 at 20x, crescent breaks

Plate E105. I8-374 at 20x, crescent breaks

Plate E106. I8-374 at 200x, smooth pitted polish

Plate E107. I8-449 at 16x, crescent breaks

Plate E108. I8-449 at 100x, smooth pitted polish

Plate E109. I8-431 at 10x, hinge termination flake scars and crescent break

Plate E110. I8-431 at 200x, smooth pitted polish

Plate E111. I8-454 at 10x, hinge termination flake scars Plate E112. I8-454 at 100x, smooth pitted polish

121

Plate E113. I8-333 at 20x, feather termination flake scars Plate E114. I8-333 at 200x, generic weak polish

Plate E115. I8-319 at 6x, hinge termination flake scars Plate E116. I8-369 at 200x, well developed polish

Plate E118. I8-416 at 100x, smooth pitted polish

Plate E117. I8-416 at 12x, crescent breaks

122

Appendix F

The red-slipped pottery from Nagsabaran (Plate G3) contains quartz, plagioclase and rock fragments as major inclusions. Minor mineral inclusions are pyroxene, amphibole and muscovite. The matrix description is as follows: Degree of temper sorting: moderately sorted Temper grain shapes: rounded and sub-angular Temper porosity: oriented channels voids (5 %) Birefringence: very weak for the interior and weak for the exterior Colour: black in both PPL and XPL for the interior core; redorange in PPL and dark red-orange in XPL for the exterior surface.

Descriptions of Petrographic Thin Sections Red-slipped Pottery The Callao Cave red-slipped pottery (Plate G1) primarily has quartz and plagioclase grain inclusions, with a few pyroxenes, amphibole, biotite and garnet minerals. There are also a few rock fragments, possibly of basalt or andesite. The following is a description of the matrix: Degree of temper sorting: poorly sorted Temper grain shapes: rounded to sub-rounded and sub-angular grains Matrix porosity: oriented (parallel to the surface) channel voids and vughs (2 % of the surface area of the slide) Birefringence: moderate for interior and exterior Colour: yellow brown in PPL and dark yellowish brown with black stripes in XPL for the interior core of the sherd; reddish orange in PPL and red orange in XPL for the exterior surface.

Nagsabaran red-slipped sherd temper grain size frequency 18

Callao red-slipped temper grain size frequency

16

14

counts

18

quartz 16

pyroxene

14

amphibole

12

counts

quartz

12

20

10

biotite

8

pyroxene

10

amphibole

8

muscovite

6

plagioclase

4

basalt/ande site

2

plagioclase 6

0

basalt/andesite

coarse

4

medium

fine

very fine

silt

size 2

0

coarse

medium

fine

very fine

silt

More than 50 % of the grain counts for Andarayan red-slipped pottery (Plate G4, Figure 8.24) are quartz. Plagioclase and rock fragments of basalt and andesite were also identified. Amphibole, pyroxene and biotite were also present. The matrix description is as follows: Degree of temper sorting: poorly sorted Temper grain shapes: rounded and sub angular Matrix porosity: oriented channels and vughs (2 %) Birefringence: weak for interior and exterior Colour: dark brown in PPL and black in XPL for the interior core; red orange in PPL and dark red orange in XPL for the exterior surface.

size

The Dalan Serkot red-slipped pottery (Plate G2,) also contained predominately quartz and plagioclase inclusions with rock fragments. Other mineral inclusions consisted of amphibole, clinopyroxene, biotite, and muscovite. The matrix description is as follows: Degree of temper sorting: poorly sorted Temper grain shapes: sub-rounded to sub-angular grains Matrix porosity: oriented channels and vughs voids (5 %) Birefringence: weak for interior and exterior Colour; dark orange brown in PPL and black with streaks of red brown in XPL for the interior core of the sherd, reddish brown in PPL and red orange in XPL for the exterior surface.

Andarayan red-slipped sherd temper grain size frequency 25

Dalan Serkot red-slipped sherd temper grain size frequency

20 16

quartz

14

pyroxene

12

quartz pyroxene

10

counts

counts

15

amphibole biotite 10

plagioclase

amphibole

basalt/andesite

biotite

8

muscovite 6

5

plagioclase basalt/andesite

4

0

coarse

2

medium

fine size

0

coarse

medium

fine

very fine

silt

size

123

very fine

silt

The Magapit red-slipped sherds (Plate G5) have mostly quartz grain inclusions with plagioclase. A few rock fragments were observed and pyroxene and amphibole were also identified. The matrix is as follows: Degree of temper sorting: poorly sorted Temper grain shapes: rounded and sub-angular Matrix porosity: oriented channels (2 %) Birefringence: moderate for both interior and exterior Colour: yellow brown in PPL and dark brown in XPL for the interior core; red orange in PPL and dark red orange in XPL for the exterior surface.

Birefringence: moderate for interior and weak for exterior Colour: orange brown in PPL and red orange in XPL for the interior; reddish black in PPL and black in XPL for the exterior

Dalan Serkot black pottery sherd temper grain size frequency 16

14

Magapit red-slipped sherd temper grain size frequency

12

quartz

10

pyroxene counts

30

amphibole

8

biotite plagioclase

25 6

basalt/andesite

4

20

counts

quartz 2

pyroxene amphibole

15

plagioclase

0

coarse

basalt/andesite

medium

fine

very fine

silt

size

10

The Eme black pottery (Plate G8) inclusions are primarily quartz, which comprises almost 75 % of the grain-size counts. Plagioclase grains are fewer, and some rock fragments were observed. Other minerals identified were pyroxene, amphibole and muscovite. The matrix description is as follows: Degree of temper sorting: moderately sorted Temper grain shapes: sub-rounded to sub-angular Matrix porosity: oriented channel voids (2 %) Birefringence: moderate in the interior and weak in the exterior Colour: reddish brown in PPL and red orange in XPL for the interior; yellow brown with streaks of black in PPL and dark yellow brown with streaks of black in XPL for the exterior.

5

0

coarse

medium

fine

very fine

silt

size

Black Pottery The Callao Cave black pottery (Plate G6) contains quartz and plagioclase inclusions with traces of pyroxene, amphibole and biotite minerals. A few rock fragments were observed. The matrix description is as follows: Degree of temper sorting: moderately sorted Temper grain shapes: sub-angular to angular grains Matrix porosity: oriented channel voids (5 %) Birefringence: weak for both interior and exterior Colour: reddish brown in PPL and black in XPL

Eme black pottery sherd temper grain size frequency 30

Callao black pottery sherd temper grain size frequency 25

quartz

20

pyroxene counts

20

18

amphibole

15

muscovite plagioclase

16

counts

10

14

quartz

12

pyroxene

basalt/andesite

5

amphibole

10

biotite

8

0

coarse

plagioclase

6

medium

fine

very fine

silt

size

basalt/andesite

4

Nagsabaran black pottery is composed of quartz and plagioclase with rounded basalt/andesite fragments (Plate G9). Other mineral inclusions are pyroxene, amphibole and biotite. The matrix has the following attributes: Degree of temper sorting: moderately sorted Temper grain shapes: sub-angular to angular and rounded Matrix porosity: oriented channels (5 %) Birefringence: weak for both interior and exterior Colour: yellowish brown in PPL and reddish brown in XPL in the interior; reddish brown in PPL and black in XPL in the exterior.

2

0

coarse

medium

fine

very fine

silt

size

Dalan Serkot black pottery (Plate G7) inclusions are predominantly quartz and plagioclase grains with pyroxene, amphibole and biotite minerals. Rock fragments were also observed. The matrix description is as follows: Degree of temper sorting: moderately sorted Temper grain shapes: sub rounded to rounded and sub angular grains Porosity: oriented channels with vughs (3 %)

124

Nagsabaran black pottery sherd temper grain size frequency 20

counts

18

16

quartz

14

pyroxene

12

amphibole

10

biotite

8

plagioclase

6

basalt/andesite

4

2

0

coarse

medium

fine

very fine

silt

size

Ethnographic pottery from Atulu The ethnographic Atulu earthenware sherd (Plate G10) subjected to thin sectioning has mostly quartz inclusions with plagioclase, amphibole, pyroxene and rock fragments. The matrix has the following attributes: Degree of temper sorting: poorly sorted Temper grain shapes: sub-rounded to rounded and sub-angular Matrix porosity: oriented channels and vugh voids (5 %) Birefringence: moderate for interior and exterior Colour: orange brown in PPL and dark yellowish brown in XPL for interior; red orange in PPL and dark red orange in XPL for exterior.

Atulu sherd temper grain size frequency 25

20

quartz 15

counts

pyroxene amphibole plagioclase

10

basalt/andesite

5

0

coarse

medium

fine

very fine

size

125

silt

126

Appendix G

Ceramic Petrographic Plates

External= the outer surface of a sherd cross-section Internal= the middle area of the sherd cross-section Bar=200 microns

Internal in XPL

External in XPL

Plate G1 Callao red-Slipped Quartz and plagioclase grain inclusions, with a few pyroxenes, amphibole, biotite, and garnet minerals

Internal in XPL

External in XPL

Plate G3. Nagsabaran red-slipped Quartz, plagioclase, pyroxene, amphibole and muscovite mineral inclusion

Plate G5. Magapit red-slipped Quartz, plagioclase, pyroxene, and amphibole mineral inclusions

Internal in XPL

External in XPL

External in XPL Internal in XPL Plate G2. Dalan Serkot red-slipped Quartz and plagioclase inclusions with rock fragments, amphibole clinopyroxene, biotite, and muscovite

External in XPL

Internal in XPL

Plate G 4. Andarayan red-slipped Quartz, plagioclase, amphibole, pyroxene and biotite mineral inclusions

127

Internal in XPL

Internal in XPL

Plate G6. Callao black pottery Quartz and plagioclase inclusions with traces of pyroxene, amphibole and biotite minerals

Plate G8. Eme black pottery Quartz, plagioclase, pyroxene, amphibole, and muscovite mineral inclusions

Plate G 10. Atulu modern pot Quartz inclusions with plagioclase, amphibole, pyroxene minerals, and rock fragments

External in XPL External in XPL

Internal in XPL External in XPL Internal in XPL

Internal in XPL

Plate G7. Dalan Serkot black pottery Inclusions are quartz and plagioclase grains with pyroxene, amphibole, and biotite minerals

Plate G9. Nagsabaran black pottery Quartz, plagioclase, pyroxene, amphibole, and biotite mineral inclusions

External in XPL

External in XPL

128

Appendix H1Report on an analysis of spindle whorl J-2514 Callao Cave, Peñablanca, northern Luzon, Philippines.

The spindle whorl was made from clay with a high mineral content. Traces of quartz were noted. It has a JIS (Munsell Notation) of 7 (Hue), 7/1 (Value). The whorl is hand built, and undecorated except for some incised lines, some of which appear to form a cross on the base.

Judith Cameron Archaeology and Natural History, Research School of Pacific and Asian Studies, The Australian National University, ACT 0200

The whorl weighs approximately 48.46 g, which falls within the range of prehistoric spindle whorls. It measures 4.00cm in diameter and 3.50 cm in height. The central perforation of the whorl is perfectly straight and measures 0.8 cm in diameter although the top of the whorl is broken.

The functional attributes of spindle whorl (J-2514) from Callao Cave were measured and are summarised in Table 1.

Shape is the principal diagnostic feature of the Callao Cave whorl as it enables more specific correlations to be drawn. Elsewhere I have demonstrated the cultural specificity of certain spindle whorls (Cameron 2002). Conical whorls of this type first appear in the Early Neolithic at archaeological sites in the Yangzi River with the earliest parallels in the basal layers (layer 4) of Hemudu.

The three significant functional attributes of the spindle whorl are material composition, shape and weight.

Material composition

Munsell colour

Weight (g)

Shape

Height (cm)

Width (cm)

Central perforation (cm)

pottery

Hue 7 Value 7/1

48.46

conical

3.50

4.00

0.80

Table 1. Functional attributes of spindle whorl (J-2574) Callao Cave Cameron, J. 2002. "Textile Technology and Austronesian Dispersals," in The Archaeology of Lapita dispersal in Oceania. Edited by G. Clark, A. Anderson, and T. Vunido, pp. 177-183. Canberra: Pandanus Book. Chang, K. 1969. Fengpitou, Tapenkeng, and The Prehistory of Taiwan. New Haven: Yale University Publications. Cummins, J. S. 1962. "The travels and controversies of Friar Domingo Navarrete, 1618-1686," in Hakluyt Society. Cambridge.

Parallels also occur at sites in Taiwan from the Fengpitou culture (c.2500 –500 BC) through to the Beinan culture (ca. 1500-800 BC). At Fengpitou, 13 whorls of this type were found where they occurred in all layers. Two whorls [P (1), V (1] were excavated from layer 1, 2 [K (3), P (2)] were found in layer II. Two conical whorls of this type were also found in layer III [N (1), V (3)]. Four were found in layer IV (B (1), K (1), S (1), V (4) and three (P (1), U (1) and V (3) were recovered from the surface. The whorl shown in Figure 65 (6) of Chang’s site report (Chang 1969) on Fengpitou is precisely the same as that from Callao Cave. The weight of the whorl is also diagnostic as it indicates the tensile strength of the fibres that were spun. As the whorl is relatively heavy (98 g), we can conclude that it was used to spin very strong fibres. When weight is combined with size, we can say that the whorl was not used to spin light fibres for clothing, but cords. Stiff elongated strands from leaves and hard leaf stems are generally used to manufacture cords and the most most likely fibres to have been spun by the whorl is Musa textilis (abaca, Manila hemp) which has long been used in the Philippine islands to make black “hemp” rigging and cables. Navarrette (16181686) refers to the usage of these fibres called cabo Negro for this purpose on the island of Mindoro in the 17th Century (Cummins 1962).

129

quantity of charred wood and parenchymatous tissues (except layer 5), untransformed fungi sclerotia (Cenococcum type), probably mineralized Boehmeria cf. platanifolia seeds (except Sq.1 spit 17) and small gastropod shells. Layer 3 (8 li) – This is the only floated sample in Eme with fragments of earthworm cocoon, indicating a high probability that this layer was a subsurface at one time. More than 500 pieces of sclerotia bodies were recovered and also a significant number (203) of Boehmeria cf. platanifolia seeds. An unidentified material #1 (light brown, elongated and curled, about 2mm long) also came in a substantial number (220). Untransformed roots came in a maximum thickness of 2mm.

Appendix H2 Eme, Callao and Dalan Serkot Caves: analysis of plant macro-remains Victor Paz and Jane Carlos Archaeological Studies Program University of the Philippines Diliman, Quezon City

Introduction

Layer 4 (7.5 li) – A single rice hull was found in this layer, dated to 100 BP by AMS. There were 82 sclerotia bodies, 5 Boehmeria cf. platanifolia and 7 bone fragments probably mineralized due to discoloration. The unidentified material #1 found in layer 3 also appears (123 pcs). Untransformed roots at a maximum of 1.5 mm in thickness.

This report presents the analysis of plant macro remains from samples coming from Eme Rockshelter, Callao and Dalan Serkot Caves. The sites mentioned above are located in Peñablanca, Cagayan Province. These sites are part of the fifty-four (54) caves in the area, explored in 1977 by the National Museum headed by Wilfredo Ronquillo and Rey Santiago. Including the said three sites, only eleven have been excavated to date. A team led by Armand Mijares in August to November 2003 excavated the three sites. The excavation is part of the Mijares’s dissertation that aims to study the area’s transition from the Paleolithic to the Neolithic period.

Layer 5 (7 li) – 72 sclerotia bodies, 5 Boehmeria, 1 bone fragment and a single unidentified seed (white, spheroid with a pointed end, textured surface, 7mm in length) were recovered. Maximum untransformed root thickness of 3 mm. Square 1 spit 17 (2 li) – 9 sclerotia bodies, 3 bone fragments and a single insect part were found in this spit which is part of Layer 4. Untransformed roots with a maximum thickness of 2.5 mm.

Eme Rockshelter is a complex comprising three caves. Elevated at 220 masl, it’s coordinates are 17º 41.25’N and 121º 49.71’E. It is located at about 2 km southwest of the National Museum Field Station at Callao Resort. Callao Cave lies lower at 84 masl. Maharlika Cuevas and his team first excavated this cave in 1979-1980. Dalan Serkot Cave is in Bgy. San Roque, south of the field station. Elevated at 165 masl, its coordinates are 17º 39.87’N and 121º 49.20’E (Mijares 2005).

Spit 9 hearth – This lone feature (part of Layer 3) also appears to have intrusions with the presence of 98 sclerotia, 37 Boehmeria seeds (mineralized) and 5 insect parts. Maximum thickness of untransformed roots is 1.5 mm.

The Samples

Dry Sieved Samples Twenty-nine (29) dry sieved samples from two squares were analyzed. One spit (#16) in Square 2, which weighed only 1g, did not yield any plant remain. The sampled spits were part of Layers 3 and 4. The significant plant remains here aside from the wood and parenchymatous fragments are the charred nut fragments present in most of the samples. The still to be determined pieces of nuts have three locules. Lastly, charred and uncharred bones were found only in Square 1.

Samples from the three sites were received in batches, as these were available, straight from the excavation of each site. There were a total of 10 flotation samples, and these came in small plastic bags or in the cloth handkerchiefs where these were originally left to dry. The flot samples varied in volume, ranging from 2 – 21 liters. The other samples (mostly charcoal) from the dry sieving comprised the bulk, totalling 69 for the three sites. These were also in small plastics with the proper labels of site, square and spit numbers and weight. All in all, there were 79 samples from 6 squares, 7 layers, 65 spits and 2 features.

Square 1 – All spit samples had charred wood fragments but only spits 7, 9-14 had charred parenchymatous tissues. Charred pieces of nuts were recovered in spits 5-8, 10, 11, 13-15. Both charred and probably fossilized (due to discoloration) bones also found in spits 3, 9, 10, 12-15.

Methodology All samples were analysed at the University of the Philippines Archaeological Studies Program (UP-ASP). First, these were sorted using a low power microscope (20-40x), separating the wood fragments, parenchymatous tissues, nut fragments, seeds and some faunal remains like gastropod shells, bone fragments, insect parts and earthworm cocoon fragments. The number of wood, parenchymatous tissues, seeds, etc. was counted and some were measured (see tables 1-3 for details). These were then placed in small plastic containers with proper labels of their contexts. Some of the sorted materials were then photographed using the low-power microscope-mounted Olympus CAMEDIA C5050 digital camera in the ASP Lithics Laboratory.

Three parenchyma samples from spits 7, 10 and 11 were mounted for SEM analysis. All three samples were fit for further determination.Square 2 - All spit samples had charred wood fragments but the charred parenchymatous tissues were only found in four spits (9, 12, 14 & 15). Charred pieces of nuts were recovered in spits 8 -12. Callao Cave The three floated samples from 2 layers and a hearth feature all contained charred parenchymatous tissues and untransformed fungi sclerotia bodies (Cenoccocum type). Only Layer 8 had no charred wood fragments.

Selected charred parenchymatous tissue samples were examined using a LEICA S440 scanning electron microscope (SEM) at the UP College of Engineering. The fragments of charred tissue were first broken to expose fresh samples measuring, at the least, about 2 cubic millimeter. These were mounted on Cambridge style stubs. Mounting was done using double-sided tape, putty and aluminum foil. Aluminum foil was used as a conductive nest needed to generate a passable image. No sputter coating was performed on the samples. This was an intentional practice that sacrifices quality of image for the opportunity of possibly dating any samples deemed important.

Layer 4 (9 li) – sclerotia (83), Boehmeria cf. platanifolia (15), 2 charred elim. Spilianthes Jacq.sp., 1 unidentified seed #1 (white, spheroid with a pointed end, textured surface, 7mm in length); earthworm cocoon fragments, small gastropod shells, bone fragments and a high number of insect parts. Maximum (untransformed) root thickness of 2.5 mm. Layer 8 (5 li) – The presence of 400+ untransformed sclerotia bodies is notable. Maximum untransformed root thickness of 1mm. Hearth in Sq.1 spit 22 (21 li) –sclerotia bodies (147) and insect parts (3). Maximum stem thickness of 2mm and untransformed roots at 1mm thick.

The Scion Image software was used to measure and analyse the charred tissues. The cell dimensions (in micrometers) -- long, short, perimeter and thickness of cell wall were measured. The measurements were then plotted in a scatter graph to check against the values from the reference collection developed by Paz and now housed in the ASP.

Dry Sieved Samples All the twenty dry sieved charcoal samples analyzed from two squares yielded plant remains. Both squares also had animal bone fragments. Square 1 – All spits except one (#13) with charred wood pieces; all spits except one (#4) with parenchymatous tissues. Spit 2 with 1 probably fossilized (discolored) bonefragment.

The system of identifying the seeds, parenchyma and other plant macroremains followed the system of graduated confidence used by Paz (2001). The seed and parenchyma reference collection developed by Paz and now housed at the ASP was referred to in the process of determination.Results

Spit 4 had wood fragments resembling bamboo because of the long cells in the transverse section. However, cross section analysis of the samples under high magnification of 200-500x revealed cells which are

Eme Cave All five floated samples from three different layers and two spits produced a high

130

circular and thick-walled, as compared to the cells of the experimentally charred bamboo specimen which are rounded squares and thin-walled. The tissues most likely belong to a fruit exocarp.

starting to be indistinct. The cell shape is generally rounded with some elongated and there are few intercellular spaces.

Discussion

Square 2 – All spits with charred wood fragments and half of the spits sampled with parenchymatous tissues. Spit 7 with a piece of nut and a single bone fragment and spit 15 had 3 charred bone fragments. The single piece of charred nut is the only one recovered in this site. Seven parenchyma samples from the lower levels of both squares (spits 22 and 23) were mounted for SEM analysis. The samples however, were amorphous and glassy, making these charred tissues undiagnostic and therefore determination could not be advanced further.

Eme Cave Three laboratory dates were obtained for this site. Charcoal in Layer 3 (spits 3-10) is radiocarbon dated to 1908 +74 BP (Wk-14882) and Layer 4 (spits 11-18) has a radiocarbon date of 3569 +52 BP (Wk-14883) (Mijares 2005). The intrusion of materials in Layer 3 from the upper parts is confirmed by the deep vertical cracks observed by the excavating team. Mijares (2005) attributes this to perhaps the periodic drying and shrinkage. These vertical cracks extended to Layer 4, so the same explanation could be applied here. The presence in Layer 4 of a single rice hull dated 100 BP by AMS confirms the intrusion from the upper layers. The generally diminishing number of sclerotia bodies as the layer(s) go deeper is likewise observed, further evidence of the ‘trickling down’ of materials from the upper layers.

Dalan Serkot Cave Two floated samples from this site had charred wood fragments with Layer 1 having 11 pieces and Layer 2 only a single piece. Both layers also had seed coat fragments, small gastropod shells, insect parts and one each of an unidentified seed (dark brown to black, textured surface, less than 1mm in length).

The intrusion could be extended to Layer 5 since untransformed roots were recovered here, having a maximum thickness of 2mm. With this may have come the less than 2 mm Boehmeria seeds. However, the single unidentified white seed may originally be from this context, considering its 7 mm length.

Layer 1 – This layer has fragments of parenchymatous tissues (2 pcs). Sclerotia bodies are present and a single mineralized Boehmeria seed and a charred elim. Spilanthes Jacq. sp. Four bone fragments were also recovered. What looks like untransformed leaves or stem fragments, came in a maximum thickness of 1mm.

The presence of a substantial number of charred nut fragments in Layers 3 and 4 dated 1908 +74 BP to 3569 +52 BP suggests the exploitation of forest resources. It is interesting to note that similar nut fragments were also found in Ille Cave, Northern Palawan where both authors had recently excavated. At Ille, the context was below the shell midden or below the 10,500 BP date.

Layer/level 2 – No parenchymatous tissues were found in this layer. Less than 1mm thick untransformed rootlets were also present as well as around 23mm thick materials that look like stems.

While determination of these nut remains await reference match, it is interesting to note that the same nut type was utilized in two different locations in the archipelago. This may be telling us that this nut type was a common staple of early communities in the Philippines.

Dry Sieved Samples A total of twenty dry sieved charcoal samples from the squares were analyzed from this site. Like Callao Cave, both squares yielded plant remains and charred bone fragments. Square 1 – Only 3 spits (13,16,18) without wood fragments and only spit 5 has no parenchymatous tissues. Charred bone fragments found in spits 4, 6 and 13. A parenchyma sample from spit 3 was mounted for SEM analysis. Square 2 – All sampled spits had charred wood fragments except spit 9. Only spits 6, 8 and 9 had parenchymatous tissues. One (1) charred bone fragment was recovered in Spit 6. Two samples of parenchyma from spit 8 were mounted for SEM analysis but were found unidentifiable due to the glassy nature and the fused state of the cells.

Callao No vertical cracks were observed in the squares excavated, however, the high number of untransformed sclerotia bodies (esp. in Layer 8) may suggest a disturbance in the layers. This is very likely since untransformed roots (maximum thickness of 1mm) were found in the flotation sample of Layer 8. Elim. Spilanthes Jacq. sp.’s probable origin (as a genus) is in tropical America; its presence in the form of two charred pieces in Layer 4, which is of the Neolitihic period, hence is problematic. However, the determination is of a very low level of confidence so this presence cannot really be used as a basis for a post-16th century deposition (Paz 2001). Perhaps this is still a case of intrusion from the upper levels. Nevertheless, the results of this analysis confirm the need for caution when dealing with collecting samples for dating and associating small artefactual remains to a specific stratigraphic layer.

SEM Analysis

Dalan Serkot

Four parenchyma samples (one from Dalan Serkot and three from Eme) had enough features that could benefit from SEM analysis. Since no vascular tissues were identified, only the cell dimensions (long, short, perimeter and thickness of cell wall) were plotted in a scatter graph. This is to find out if these coincide with the values from the reference collection that include Dioscorea alata, Colocasia, Ipomea batatas and Manihot esculenta. The measurements from Dalan Serkot (square 1 spit 3) and Eme (square 1 spits 7, 10 and 11) did not overlap with these species in terms of cell size – long, short and perimeter. Only a few values for cell wall thickness overlapped with some measurements from the different species. The clustering of the measurements was noted within the same sample and also generally between all four samples that grouped in the lower values of the graph. Eme Rockshelter The sample from square 1 spit 7 has rounded cells with no intercellular spaces while that from spit 10 has round shaped cells and very few intercellular spaces. Only a few cells were measured in the spit 10 sample because most have fused already. Rounded cells were also observed in the sample from spit 11 though this was limited to the smaller sized cells; the bigger cells were angular.

Compared to the two sites, Dalan Serkot has the fewest plant macroremains recovered. Its being more of a secondary burial site rather than a habitation (Mijares 2004) is probably a good explanation for this result. A charred elim. Spilanthes Jacq. Sp was recovered; it is definitely transformed but the fact that only a single material was found poses a problem. SEM Analysis As mentioned in the results, the dimensions of the cells were smaller than those from the reference collection (see charts below). The samples from Dalan Serkot, Eme and Eme 2 are probably from the same species considering their overlapping cell dimensions. The Eme 3 sample differs, it has the smallest cell size. While the confidence level for the similarity of the cross-site samples is low, there is certainty that we have in the assemblage at least two types of forest product that were utilized for their parenchymatous organs. These are wild plants based on their relatively small cell sizes. It is not certain though is these remains are fragments of fruits or vegetative organs.

Dalan Serkot The sample has cells that appear warped, most likely due to charring while still fresh and with some moisture content. Though visible, the cells are

131

Summary

* Non-prefixed - the highest level of confidence wherein a binomial determination is made whenever there are clear photographic reference(s) of the seed, and/or an illustration reference(s) of the seed, or both. With photographic references, and illustrations at hand, the existence of a reference collection is still important, but not essential to non-prefix identification. The exact fit of the taxonomic features, the geographic distribution, and the species citation in the local flora are firm prerequisites for a non-prefix determination.

In summary, the data shows that there is movement of materials from the upper to the lower layers due to bioturbation (roots and insects among others) and the vertical cracks. The 100 BP dating of the single rice hull in the context dated 3569 +52 BP confirms the intrusion and at the same time shows that upland rice utilization is present in the area, however at a later age. It can also be stated that there is a glaring absence of domesticated plants and known wild root crops.

*prob.: what are required for the determination ‘prob.’, are the prerequisites of flora citation, geographic area compatibility and an agreement with the taxonomic details. The existence of either an image or illustration is necessary, as well as a specimen in a reference collection. It differs however, from the ‘non-prefixed’ determination in that only one out of three - image, illustration and reference collection exist, and the references sometimes do not give an exact or good fit.

Acknowledgement Acknowledgement is due to the following: Armand Mijares for providing the samples for analysis; Dr. Amorsolo and Mark Romano of the UP College of Engineering Department of Materials, Metallurgical and Mining for the use of the scanning electron microscope and for the assistance in the SEM photography.

cf.: In this category, all six categories may or may not exist. The archaeological material resembles an image/illustration/reference sample or a previous identification of the archaeobotanist, but there is no exact morphological fit. Three out of the five other categories fit the archaeological material but the researcher has doubts about the exact fit of these categories with the material.

References Burkill, Isaac Henry. 1966. A Dictionary of the Economic Products of the Malay Peninsula. Malaysia: Art Printing Works. Hather, Jon G. 1993. An Archaeobotanical Guide to Root and Tuber Identification. Oxford: Oxbow Books. __________ 2000. Archaeological Parenchyma. London: Archetype Publications Ltd. Mijares, Armand S. 2005. Preliminary Report on the Archaeology of Peñablanca Cave Sites, northern Luzon, Philippines. Journal of Austronesian Studies. 1(2) _______________ July 2004. The Archaeological Excavation of Eme Cave, Callao Cave, and Dalan Serkot Cave. Test Pit 4:10-11. Moody, K., C.E. Munroe, R.T. Lubigan and E.C. Paller, Jr. 1984. Major Weeds of the Philippines. UPLB:Weed Science Society of the Philippines. Paz, Victor J. 2001. Archaeobotany and Cultural Transformation: Patterns of Early Plant Utilisation in Northern Wallacea. PhD dissertation. University of Cambridge. ___________ 2004. Of nuts, seeds and tubers: The archaeological evidence from Leang Burung 1. Modern Quaternary Research in Southeast Asia 18:191-220

elim. :this is the lowest level of confidence when trying to determine a binomial identification of the archaeological material. It indicates that the material has a chance of being the species proposed but the determination was derived without any images, illustrations or reference collection sample. It is based on the taxonomic description of a plant and its fit with its geographic range. It is listed with other species of the same genus in a flora or reference botanical work on the region studied, elimination of the other species was done and the archaeobotanist decides on a likely candidate from the remaining species based on the fit it has with other existing information. Suffix ‘type’: This is applied when the level of confidence is very low. It means that the shape of the specimen fits the geographic distribution, some of the morphological characters of a plant, and that it may also be in the relevant flora. It is used only at the family, and genus level of determination. Form shape description: None of the six listed types of information exist, but the archaeological specimen is distinctly a seed, a nut fragment or any other plant parts. The material may then be described by its general shape i.e. Spheroid, Angular/Triangular, Long etc. A number is attached based on its position chronologically with other seeds looked at and determined. Sometimes, under this categorization, a very tentative identification is added, mostly at the family level. This is to facilitate future researchers, with a better stock of references

Levels of confidence Based on different levels of confidence, the determination system used the following prefixes and suffixes: non prefixed, prob., cf, elim, suffix ‘type’ and form shape description.

COMPARATIVE MEASUREMENT CHARTS

(Peñablanca Sites in oval) Colocasia

C. esculenta Radial 60

Bahi-bahi

50

Cavite

Short

40

dalan serkot

30

eme

20

eme2 10

eme3 0 0

20

40

60

Long

132

80

100

Ce ll Size 300

C. esculenta LB S-23

250

D alata-6 D.alata-Cav

200

Short

Appari 150

Lakapen dalan serkot

100

eme eme2

50

eme3 0 0

20

40

60

80

100

120

140

160

Long

Dioscorea alata

Dioscorea alata (Transverse) 300 250

Lakon Nabisaya Balatok Cavite Appari Taropoyo dalan serkot eme eme2 eme3

Short

200 150 100 50 0 0

50

100

150

Long

D.alata radial 120 100

Balatok Nabisaya Lakon dalan serkot eme eme2 eme3

Short

80 60 40 20 0 0

50

100

Long

133

150

Ipomea batatas

I. batatas (transverse) 90 80 70

Lal-lo Urdaneta Lal-lo2 dalan serkot eme eme2 eme3

Short

60 50 40 30 20 10 0 0

20

40

60

80

100

120

Long

Manihot

Manihot 45

Lal-lo

40 35

Series2

Short

30

dalan serkot

25 20

eme

15 10

eme2

5 0 0

20

40

60

80

eme3

Long

M.Utilissima without matirix 60

Tangential

Short

50

Transverse

40 30

dalan serkot

20

eme

10

eme2

0 0

20

40

60

Long

134

80

100

eme3

reference material from PNG and Australia. Actual counting and plant identification procedures replicated those methods used in previous studies for West New Britain PNG, the Torres Strait Islands, Byron Bay and the Bora-Coddrington areas of eastern Australia (Parr, 2002; Parr, 2004; Parr and Carter, 2003; Parr et al., in press).

Appendix H3 A Report on the Analysis of Microbotanical remains from Callao Cave Dr Jeff Parr Plantstone Carbon Sequestration Centre for Geoarchaeology and Palaeoenvironmental Research Southern Cross University, Sydney

Site Dalan Serkot – The samples provided contained no identifiable microfossils at all. There were however microscopic charred particles as indicated in Table 1 and 2. Site Callao – Microfossil preservation in the samples provided ranged from low to medium densities.

General observations: Preservation of microfossils in the samples provided for the sites Callao and Dalan Serkot ranged from very poor to reasonable for analysis. Because some samples had low microfossil counts both actual counts and percentages are provided to help your assessment of these samples (Tables 1 and 2). In addition, I have provided a breakdown of the microfossil types observed (Table 3) as well as some images that may assist yourself and/or others should you seek a second opinion.

Callao 1 Layer 3 has the highest count of arboreal phytolith types of all the samples studied accounting for around 70% of the total phytolith assemblage for this layer. This high percentage of arboreal type phytoliths coincides with a high charred material component thus may indicate a possible association between woody plants and burning. Other phytolith types present include Palms (Arecaceae, closely resemble those of Cocas nucifera L) 7%, bamboo (Bambusoid) that closely resemble those that are found in the species Schizostachym brachycladum (Blanco) 7% (figures 1a,b), distinctive black spheres (e.g. figure 1c) and keel shaped Myrtaceae phytoliths that resemble those of Syzigium brevicymum (Diels) Merr. 3% and grasses (Poaceae) 14% of the total phytolith assemblage. Other microfossils are found in insignificant numbers and as are discarded at this stage. The one starch grain observed was in the 6µm to 9µm range described by (Lentfer et al., 2002) as type 1b associated with cultivated locations, gardens and coconuts plantations.

Microscopic charred material remained visible throughout all samples even those with low microfossil counts. This phenomenon occurs because charcoal is made of a complex molecular structure (Schmidt et al., 1999) that is not subject to the same dissolution as silica microfossils under certain conditions (Bartoli and Wilding, 1980).

Methods of identification: Identification of microfossils was carried out using a range of databases relevant to the Asia Pacific region, or from

Figure 1. [a] S.brachycladum , [b] Bambusoid, [c] Myrtaceae [c] starch 6µm – 9µm [d] starch polarized light.

Callao 3 Layer 4 is a layer dominated by arboreal phytolith types that represent around 55% of the total phytolith assemblage. There are more Palm (Arecaceae) 15% and less grass 11% phytoliths in this layer than occurs in any other layer for the Callao site. Charred material and other microfossil counts in this layer are very low. Callao 5 Layer 5 is dominated by grass phytoliths representing around 53% of the overall phytolith count (Tables 1 and 2). Palm (spheres are identical to Metroxylon sagu), bamboo (Bambusoid) and sedge type (Cyperaceae) phytoliths also occur at this level (Tables 1 and 2). Some of the grass and sedge phytolith types resemble those associated with coastal wetland species found in West New Britain, PNG. Moreover a strong presence of diatoms supports the interpretation that many of the grass phytoliths and the Cyperaceae phytoliths were from a localised wetland system (Table 3). Diatoms include those that resemble Achnanthes 135

spp., Auloacoseira spp., Cyclotella spp., Pinnularia spp., and Rhopolodia spp. that together indicate a local freshwater system (Lane, 1999; Taffs, 2001). Unless the this layer itself represents an open area freshwater lake or spring fed pond system the abundance of grasses and sedges together with diatoms indicate that they may have been introduced to the site. That is to say, it is possible that these herbaceous plants were harvested from a local wetland system and taken back to the site for some type of use. The diatoms have probably been attached to these plants and as such also introduced to the site as an unintentional process. Alternatively, the diatoms may simply indicate the transport of water to the site. The former is most likely but why these plants would have been introduced to the site is not clear. Also of interest is the presence of a few starch grains and unidentified plant cells coated in resinous like material (Tables 1 and 2). Starch grains are in the 6µm to 9µm range described by (Lentfer et al., 2002) as type 1b associated with cultivated locations, gardens and coconuts plantations. The resin-coated cells are quite common in peatland

layers where wet conditions have prevented oxidation of plant material and have also been observed in situations where a woody component of resin duct has survived (Parr, 2002). Arboreal type phytoliths are much lower in this layer than in all other layers as are charred particle counts (Tables 1 and 2). Callao 11 Layer 8 dated at c. 25,000 years. The microfossil counts for this layer were low (Tables 1 and 2). Nevertheless, there are some indications of site processes. For example, as with was noted with the sample Callao 1 Layer 3 there are a high percentage of arboreal type phytoliths present, which coincides with a high charred material component thus indicating a possible association between woody plants and burning. Robust grass phytoliths such as block, long and prickle cells make up 32% of the phytolith counts for this layer, however, arboreal sphere types as well as a couple of long cells and a hair cell that resembles those found in Moraceae species represent 55% the total phytolith count. As with Callao 5 Layer 5 there is a few starch grains and the unidentified resin coated plant material. Two starch grains were in the 14-22µm range described by (Lentfer et al., 2002) as type 1d associated with palm forests of small volcanic islands. Another two starch grains were in the 23-30µm range or type 1e described as being associated with both forest and garden sites (Lentfer et al., 2002). References Cited: Bartoli, F. and Wilding, L.P., 1980. Dissolution of biogenic opal as a function of its physical and chemical properties. Soil Science Society of America, 44: 873878. Lane, C.M., 1999. The use of diatoms as biological indicators of water quality in two coastal dune lakes, Northern New South Wales. Unpublished Honours Thesis, Southern Cross University, Lismore. Lentfer, C.J., Therin, M. and Torrence, R., 2002. Starch Grains and Environmental Reconstruction: a Modern Test Case from West New Britain, Papua New Guinea. Journal of Archaeological Science, 29: 687-698. Parr, J.F., 2002. West Byron Wetland Vegetation History Report, Southern Cross University, Unpublished report, Lismore. Parr, J.F., 2004. Morphometric and visual fossil phytolith identification using a regionally specific digital database. Phytolitharian, 16(2): 2-10. Parr, J.F. and Carter, M., 2003. Phytolith and starch analysis of sediment samples from two archaeological sites on Dauar Island, Torres Strait. Vegetation History and Archaeobotany, 12(2): 131-141. Parr, J.F., Kerr, G., Arthur, J. and Taffs, K.H., in press. Fires and their implications for the acidic peatlands of northeastern NSW. In: A. Korstanje and P. Babot (Editors). Altamira, and BAR (British Archaeological Series). Schmidt, M.W.I., Skjemstad, J.O., Gehrt, E. and KögelKnabner, I., 1999. Charred organic carbon in German chernozemic soils. European Journal of Soil Science, 50: 351 - 365. Taffs, K.H., 2001. Diatoms as indicators of wetland salinity in the Upper South East of South Australia. The Holocene, 11(3): 281-290. 136

137

Table 3. Actual counts and types of phytoliths, charred material, starch grains, diatoms and resinous plant material observed in the Callao layers.

Table 2. Percentage counts of phytoliths, charred material, starch grains, diatoms and resinous plant material observed in the Callao and Serkot sites.

Table 1. Actual counts of phytoliths, charred material, starch grains, diatoms and resinous plant material observed in the Callao and Serkot sites.

Appendix I Glossary of Soil Micromorphology Terms (Based on Bullock et al 1985, Stoops 2003 and Courty et al 1989) Accommodation. Degree to which opposite faces exhibit complimentary shape Authigenic. Crystals that belong to the time of formation of the rock, or to some subsequent change within the rock Anthropic. Soil produced by human activity Birefringence. A quantitative measure of double refraction which is the difference between the maximum and the minimum refractive indices of the mineral Chambers. More or less equidimentional, smooth-walled pores interconnected by channels Channels. Tubular smooth voids with a cylindrical cross section Compound packing voids. Voids resulting from loose packing between non.accommodating aggregates Crumb. porous, more or less equant peds that are bounded by rounded faces Diagenesis. The chemical alteration that takes place in sediments, turning them into rocks Dusty clay. Composed of clay containing micro particles Eluviation. Lost of clay particles though translocation Idiotopic. Consisting of euhedral crystals Illuviation. Accumulation of clay particles Isotropic. Transparent or coloured in plane polar light but remains dark in cross-polar light Geodic Nodule. Nodules with hollow interior Granostriated. Clay domains are oriented parallel to the walls of resistant fabric units producing a halo of interference colour around the grains in cross polar light. Hypidiotopic. Consisting of subhedral crystals Hypocoating. Matrix pedofeature referred to a natural surface in the soil and immediately adjoining it Limpid clay. Uniform clay without inclusion of micro particles Micrite. Fine calcite (5µm) Porostriation. Clay domains in the micromass are oriented parallel to the surface of the pore Porphyric. Larger fabric units occurring in a dense groundmass of smaller units Pseudomorphs. An alteration of mineral grains whose original outline is still recognisable and the created pore space is either preserved or filled with neoformed material Punctuations. Small dark or opaque grains Renticulate striated. Two sets of birefringent streaks intersect at right angle Stipple speckle. Characterized by randomly arrange oriented clay Vughs. More or less equidimensional, irregular voids, smooth or rough normally inter-connected to void of comparable size Xenotipic. Consisting of anhedral crystals

138

About the author Armand Salvador B Mijares is currently an Assistant Professor at the Archaeological Studies Program, University of the Philippines. He received his PhD in Archaeology and Palaeoanthropology from the Department of Archaeology and Natural History, Australian National University in 2006.

139