CHRONOLOGY OF POST-GLACIAL SETTLEMENT IN THE GOBI DESERT AND

CHRONOLOGY OF POST-GLACIAL SETTLEMENT IN THE GOBI DESERT AND
CHRONOLOGY OF POST-GLACIAL SETTLEMENT IN THE GOBI DESERT AND
THE NEOLITHIZATION OF ARID MONGOLIA AND CHINA
by
Lisa Janz
_____________________
Copyright © Lisa Janz 2012
A Dissertation Submitted to the Faculty of the
School of Anthropology
In Partial Fulfillment of the Requirements
For the Degree of
DOCTOR OF PHILOSOPHY
In the Graduate College
THE UNIVERSITY OF ARIZONA
2012
2
THE UNIVERSITY OF ARIZONA
GRADUATE COLLEGE
As members of the Dissertation Committee, we certify that we have read the dissertation
prepared by Lisa Janz
entitled Chronology of Post-Glacial Settlement in the Gobi Desert and the Neolithization
of Arid Mongolia and China
and recommend that it be accepted as fulfilling the dissertation requirement for the
Degree of Doctor of Philosophy
_______________________________________________________________________
Date: 10 October 2011
John W. Olsen
_______________________________________________________________________
Date: 10 October 2011
Steven L. Kuhn
_______________________________________________________________________
Date: 10 October 2011
Michael B. Schiffer
_______________________________________________________________________
Date: 10 October 2011
Mary C. Stiner
_______________________________________________________________________
Date:
Final approval and acceptance of this dissertation is contingent upon the candidate’s
submission of the final copies of the dissertation to the Graduate College.
I hereby certify that I have read this dissertation prepared under my direction and
recommend that it be accepted as fulfilling the dissertation requirement.
________________________________________________ Date: 10 October 2011
Dissertation Director: John W. Olsen
3
STATEMENT BY AUTHOR
This dissertation has been submitted in partial fulfillment of requirements for an
advanced degree at the University of Arizona and is deposited in the University Library
to be made available to borrowers under rules of the Library.
Brief quotations from this dissertation are allowable without special permission, provided
that accurate acknowledgment of source is made. Requests for permission for extended
quotation from or reproduction of this manuscript in whole or in part may be granted by
the copyright holder.
SIGNED: Lisa Janz
4
ACKNOWLEDGEMENTS
This research was made possible by financial support through the Social Sciences and
Humanities Research Council of Canada (SSHRCC) Doctoral Fellowship and a
Fellowship for East and Southeast Asian Archaeology and Early History from the
American Council of Learned Societies (ACLS) with funding from the Henry Luce
Foundation. Additional funding was provided by the Asian American Faculty, Staff, and
Alumni Association Research Grant, University of Arizona, and the following University
of Arizona, School of Anthropology funds: Emil W. Haury Educational Fund; WilliamShirley Fulton Scholarship; Carol Kramer Memorial Scholarship; and the Traditions,
Transitions and Treasures Fund. Generous logistical support was provided at the
Museum of Far Eastern Antiquities, Stockholm by Dr. Eva Myrdal, Anna Böstrom, and
Dr. Håkan Wahlquist at the Ethnographic Museum, and at the American Museum of
Natural History, New York by Dr. David H. Thomas, Paul Beelitz, Alex Lando, and
Kristin Mable, as well as other staff members in the Divisions of Anthropology and
Palaeontology. While in Mongolia, logistical support was kindly provided by Dr. B.
Gunchinsuren, Dr. C. Amartushvin, D. Bukhchulun, and D. Khashbat. Drs. D. Tumen
and Ya. Tserendagva were generous in their provision of important details about
Mongolian archaeological research past and present. Drs. Sergei Gladyshev and Andrei
Tabarev were my gracious hosts at the Institute of Archaeology and Ethnology Siberian
Branch, Novosibirsk, Russian Federation. Special thanks to Dr. Magnus Fiskesjö, who
helped coordinate my original correspondence with the museums in Stockholm. My
gratitude is also extended to Drs. William Honeychurch and Joshua Wright, under whom
I was able to gain field experience and organize my own collaborations in Mongolia.
Drs. Robert G. Elston, Joshua Wright, and Tiina Manne have given me inspiration and
the opportunity to debate and discuss new ideas. Bonner Odell, Violet Yufeng Long, and
Minami Cohen generously opened their homes to my daughter when I most needed a few
uninterrupted days. The following people have been truly inspirational mentors
throughout my university career: Drs. John W. Olsen, Steven L. Kuhn, Michael B.
Schiffer, Mary C. Stiner, and Fumiko Ikawa-Smith. I would like to separately thank Dr.
Olsen for his support, encouragement, and habit of pushing me to put in just a little more
effort when I least wanted to. Also, to Dr. Kuhn, whose consistent intellectual support
and feedback has encouraged me to be a better scientist. Most importantly, the support of
my family has been essential in completing this milestone which has been so long in the
making.
5
DEDICATION
With love, for James H. Lee, who has brought me the joy of his companionship since I
started my graduate work at the University of Arizona and who I hope will still be my
partner in life at the end of my career. And to Sophia V. and Vivianne E. Lee, who have
spent their entire lives until now sharing me with my dissertation.
6
TABLE OF CONTENTS
LIST OF FIGURES .......................................................................................................... 10
LIST OF TABLES ............................................................................................................ 12
ABSTRACT...................................................................................................................... 15
CHAPTER 1 – INTRODUCTION ................................................................................... 16
CHAPTER 2 – CHRONOLOGY OF TECHNOLOGICAL DEVELOPMENTS AND
NEOLITHIZATION IN POST-GLACIAL NORTHEAST ASIA ................................... 30
2.1. Key technological developments in Northeast Asia ............................................. 31
2.1.1. Terminology ................................................................................................... 33
2.1.2. Microblade technology ................................................................................... 34
2.1.3. Pottery............................................................................................................. 38
2.1.4. Grinding stones ............................................................................................... 40
2.1.5. Polished stone tools ........................................................................................ 43
2.2. Chronology of food-production in Northeast Asia ............................................... 46
2.2.1. Brief chronology of plant domestication ........................................................ 47
2.2.2. Detailed chronology of animal domestication ................................................ 53
2.2.3. Animal domestication in the Gobi Desert ...................................................... 68
2.2.4. Chronology of nomadic pastoralism in Northeast Asia ................................. 70
2.3. Current knowledge of the Neolitthic transition in Mongolia ................................ 81
2.3.1. Post-LGM settlement and subsistence in Mongolia ....................................... 83
2.4. Summary ............................................................................................................. 105
3.1. Chronometric dating............................................................................................ 112
3.1.1. Sample selection ........................................................................................... 113
3.1.2. Dating methods ............................................................................................. 114
7
TABLE OF CONTENTS - Continued
3.1.3. Results .......................................................................................................... 119
3.2.1. Definition of “Neolithic” and issues in terminology .................................... 126
3.2.2. Chronological variation in technology, subsistence, and land-use............... 129
3.2.2.1. Early Epipalaeolithic (Upper Palaeolithic/Late Palaeolothic) – 19.0 to
13.5k cal yr BP .................................................................................................... 129
3.2.2.2. Late Epipalaeolithic/Oasis 1 (Mesolithic/Early Neolithic) – 13.5 to 8.0k
cal yr BP .............................................................................................................. 140
3.2.2.3. Neolithic/Oasis 2 (Early to Middle Neolithic) – 8.0 to 5.0k cal yr BP…
............................................................................................................................. 155
3.2.2.4. Eneolithic/Oasis 3 (Late Neolithic/Early Bronze Age) – 5.0 to 3.0k cal yr
BP ........................................................................................................................ 181
3.2.3. Estimates of ages for undated sites ............................................................... 201
3.3. Discussion ........................................................................................................... 215
CHAPTER 4 – PREHISTORIC HUMAN LAND-USE IN THE GOBI DESERT ....... 217
4.1. Analysis of land-use ............................................................................................ 219
4.1.1. Categorization of sites and environments ..................................................... 221
4.1.2. Assemblage composition .............................................................................. 242
4.1.2.1. Concepts and methods .......................................................................... 244
4.1.2.2. Results ................................................................................................... 254
4.1.3. Lithic reduction strategies ............................................................................ 262
4.1.3.1. Modeled expectations and methods ...................................................... 263
4.1.3.2. Results ................................................................................................... 279
4.2. Discussion ........................................................................................................... 288
CHAPTER 5 – PALAEOENVIRONMENTAL CONTEXT ......................................... 294
5.1. Palaeoenvironmental chronology of the Late Pleistocene and Early Holocene in
Northeast Asia ............................................................................................................. 296
5.1.1. Middle to late MIS 3 (43.0-25.0k cal yr BP) ................................................ 297
5.1.2. Early to middle MIS 2 (25.0-19.0k cal yr BP) ............................................. 298
8
TABLE OF CONTENTS - Continued
5.1.3. Middle to late MIS 2 (19.0-11.5k cal yr BP) ................................................ 299
5.1.4. Early to late middle MIS 1 (11.5-3.0k cal yr BP) ........................................ 299
5.2. Regional Variability ............................................................................................ 301
5.2.1. East Gobi ...................................................................................................... 301
5.2.2. Gobi-Altai ..................................................................................................... 310
5.2.3. Alashan Gobi ................................................................................................ 316
5.3. Desert forests ....................................................................................................... 331
5.4. Discussion ........................................................................................................... 337
CHAPTER 6 – DISCUSSION AND CONCLUSIONS ................................................. 342
6.1. Palaeoenvironment and local ecology ................................................................. 343
6.2. “Dune-dweller” foraging strategies..................................................................... 351
6.3. Oasis 3 and the rise of nomadic pastoralism ....................................................... 368
6.4. Conclusion........................................................................................................... 374
APPENDIX A – RESULTS AND CONTEXT OF DATED SITES .............................. 378
A.1. Dated Samples .................................................................................................... 378
A.2. Context of Dated Sites........................................................................................ 380
A.3. Results of Chronometric Dating ......................................................................... 383
APPENDIX B – DETAILED SUMMARY OF DATED SITES ................................... 385
APPENDIX C – ARTEFACT TYPOLOGIES ............................................................... 421
C. 1. Artefact summary for each site.......................................................................... 421
C.2. Codes for artefact types ...................................................................................... 434
APPENDIX D – ASSEMBLAGE CHARACTERISTICS ............................................. 435
9
TABLE OF CONTENTS - Continued
APPENDIX E – SUMMARY OF MEANS AND FREQUENCIES .............................. 448
REFERENCES…………………………………………………………………………453
10
LIST OF FIGURES
FIGURE 1.1, Map of Northeast and Inner Asia indicating study area ………………. 18
FIGURE 1.2, Map of Gobi Desert locales, target regions, and subregions ………….. 26
FIGURE 2.1, Map of key locales discussed in Chapter 2 ...………………………….. 32
FIGURE 2.2, Map of archaeological sites mentioned in Chapter 2 ..........................… 35
FIGURE 2.3, Relative frequency of key artefact types in Northeast Asian chronology
......................................................................................................................................... 46
FIGURE 3.1, Dates (cal yr BP/ka [2 σ]) plotted for Gobi Desert sites ……………..... 121
FIGURE 3.2, Map of geographic locales mentioned in Chapter 3................................ 130
FIGURE 3.3, Map of archaeological sites mentioned in Chapter 3 .............................. 132
FIGURE 3.4, Dated Oasis 2 pottery ………………………..………………………… 168
FIGURE 3.5, Unifacial and bifacial points from Oasis 2 sites .........…………………. 172
FIGURE 3.6, Small chipped adze from Baron Shabaka Well ....................................... 175
FIGURE 3.7, Painted pottery from the Gurnai Depression …………………………... 188
FIGURE 3.8, Examples of Oasis 3 pottery from dated sites …………………………. 190
FIGURE 3.9, Example of “spongy-textured” paste common in East Gobi sites …….. 192
FIGURE 3.10, Example of large curved bifacial blade from Shabarakh-usu 4 ……… 194
FIGURE 3.11, Examples of Oasis 3-type bifaces from Shabarakh-usu 2 …………… 194
FIGURE 3.12, Microblade core preforms from Yingen-khuduk and Ulan Nor Plain .. 196
FIGURE 4.1, Distribution of Gobi Desert sites according to each environmental
parameter …………….................................................................................................... 227
11
LIST OF FIGURES – Continued
FIGURE 4.2, Distribution of East Gobi sites according to each environmental parameter
…………………………………………………………………………………………. 231
FIGURE 4.3, Distribution of Gobi-Altai sites according to each environmental parameter
…………………………………………………………………………………………. 233
FIGURE 4.4, Distribution of Alashan Gobi sites according to each environmental
parameter ……………………………………………………………………………… 235
FIGURE 4.5, Distribution of Gobi Desert sites according to ecozones ……………… 239
FIGURE 4.6, Distribution of sites in each target region according to ecozones .......... 240
FIGURE 4.7, Distribution of site types across Gobi Desert for Oasis 2 and Oasis 3 … 257
FIGURE 5.1, Map of Gobi Desert showing palaeoenvironmental locales and study
regions …………………................................................................................................ 304
FIGURE 5.2, Map showing the study region at about 6.0 kya ............................…….. 311
FIGURE 5.3a, Approximation of hydrology and vegetation in three Gobi Desert regions
for 10.0k cal yr BP, 8.0k cal yr BP, 6.0k cal yr BP, and 3.0k cal yr BP ……….... 340-341
FIGURE 6.1, High-fired red-ware from the Bronze Age site, Dottore-namak ............. 373
FIGURE B.1, Macrotools from Baron Shabaka Well ………………..………………. 460
12
LIST OF TABLES
TABLE 2.1, Earliest evidence for use of major domesticated animal species in Northeast
Asia ……………............................................................................................................ 67
TABLE 3.1, Results of chronometric dating …………............................................... 120
TABLE 3.2, Comparison of ostrich eggshell and pottery dates from Gobi Desert sites
…………………………………………………………………………………………. 124
TABLE 3.3, Ostrich eggshell dates from Shabarakh-usu .............................................. 125
TABLE 3.4, Summary of sites containing small unifacial points in the Gobi Desert, East
Mongolia, and Northeast China ..................................................................................... 173
TABLE 3.5, Characteristics of pottery associated with each period ............................ 189
TABLE 3.6, Summary of artefact chronologies ……………………………………… 202
TABLE 3.7, Age estimates for studied archaeological sites …………………….. 204-207
TABLE 4.1, List of categories used for analysis of site context ………................….. 223
TABLE 4.2, Relationship between environmental parameters …….......................….. 225
TABLE 4.3, Actual and expected distribution of sites according to each environmental
parameter ….................................................................................................................... 228
TABLE 4.4, Actual and expected distribution of sites in the East Gobi according to each
environmental parameter .............................................................................................. 232
TABLE 4.5, Actual and expected distribution of sites in the Gobi-Altai according to each
environmental parameter .............................................................................................. 234
TABLE 4.6, Actual and expected distribution of sites in the Alashan Gobi according to
each environmental parameter ....................................................................................... 236
13
LIST OF TABLES - Continued
TABLE 4.7, Actual and expected distribution of Gobi Desert sites according to each
ecozone ......................................................................................................................... 239
TABLE 4.8, Actual and expected distribution of sites in each region according to
ecozone ......................................................................................................................... 241
TABLE 4.9, Actual and expected distribution across Gobi Desert ecozones for Oasis 2
and Oasis 3 sites ............................................................................................................. 258
TABLE 4.10, Regional distribution of Oasis 2 and Oasis 3 for each ecozone ……….. 259
TABLE 4.11, Actual and expected distribution of residential site types for Oasis 2 and
Oasis 3 across the entire Gobi Desert and for each region ........................................... 260
TABLE 4.12, T-test results for Residential A and B lithic data sets ............................ 282
TABLE 4.13, T-test results for Oasis 2 and Oasis 3 lithic data sets .............................. 287
TABLE 4.14, Coefficient of variation for core volumes and platform surface areas of
Oasis 2 and Oasis 3 assemblages from the Alashan Gobi ............................................. 287
TABLE 5.1, Summary of regional climate change from 12.0k cal yr BP to 2.0k cal yr BP
…………………………………………………………...........................................….. 330
TABLE 6.1, Possible seasonal distribution of raw material and edible resources ....... 347
TABLE 6.2, Mean percentage of core types according to site type for Oasis 2 and Oasis
3, all Gobi Desert sites ……………………….........................................................….. 362
TABLE 6.3, Number of Gobi Desert Oasis 2 and Oasis 3 sites with pottery and grinding
stones according to site type .......................................................................................... 364
14
LIST OF TABLES - Continued
TABLE 6.4, Number of Gobi Desert Oasis 2 and Oasis 3 sites with neither pottery nor
grinding stones according to site type ............................................................................ 365
TABLE 6.5, Number of Gobi Desert sites with pottery and grinding stones according to
ecozone and period ....................................................................................................... 365
15
ABSTRACT
Prior to this study, knowledge of Gobi Desert prehistory was mostly limited to
early and mid-20th century descriptions of undated stone tool assemblages from
unanalyzed museum collections. This research focuses on the use of extensive existing
museum collections to establish a baseline chronology of technology, economy, and landuse for prehistoric Gobi Desert groups. Radiocarbon and luminescence dating is used to
establish an artefact-based chronology and provide a relative age for 96 archaeological
site assemblages. Interpretations of land-use derived from lithic analysis are compared to
detailed regional and local palaeoenvironmental records in order to contextualize
residential mobility and subsistence. Results indicate that a dramatic shift in land-use
after about 8000 years ago was related to a combination of widespread forestation and the
increased productivity of lowland habitats during a period of high effective moisture.
Hunter-gatherers organized their movements around dune-field/wetland environments,
but utilized a range of both high- and low-ranked foods such as large ungulates from
adjoining plains and uplands, and seeds and/or tubers from dune-fields and wetlands.
New radiocarbon dates indicate that the use of dune-fields and wetlands persisted into the
early Bronze Age, overlapping with the rise of nomadic pastoralism across Northeast
Asia. These findings illuminate the period just prior to the rise of nomadic pastoralism in
Northeast Asia and add considerable depth to our understanding of hunter-gatherer
adaptations within arid environments following the Last Glacial Maximum.
16
CHAPTER 1 – INTRODUCTION
Spanning southern Mongolia (Mongol Uls/Outer Mongolia) and a vast area of
northern China (including the western Inner Mongolia Autonomous Region or Nei
Menggu Zizhiqu, northern Gansu, and the northeastern Xinjiang Uyghur Autonomous
Region), the Gobi Desert is bounded by the Altai Mountains in the west and north, the
Mongolian steppe to the northeast, the Hexi Corridor and Qinghai-Tibet Plateau in the
southwest and the North China Plain far to the east (Figure 1.1). Today, it is covered by
extensive mountain ranges, plateaux, erosional basins, former lake beds, gravel plains,
desert-steppe, and dune-fields. Annual extreme temperatures range from +40oC in the
summer to -40oC in the winter and the average rainfall is less than 200 mm/year. The
continental desert environment separates two distinct ecological and cultural zones of
East Asia – the fertile Central Plains of China and the vast northern steppes of Central
and Northeast Asia. By the third century BCE, highly complex stratified agricultural and
nomadic pastoral societies had emerged to the south and north of the Gobi Desert
respectively, forever altering a base of hunter-gatherer subsistence economies typical of
Palaeolithic populations since the appearance of anatomically modern Homo sapiens in
Eurasia.
The transition from hunter-gatherer to either sedentary agriculturist or nomadic
pastoralist represents a key trajectory in the prehistory of East Asia. Although the two
modes of food production developed in close proximity, sharing very similar
technological traditions since the Upper Palaeolithic, intense political opposition
following the early Bronze Age polarized these two economic strategies within the
17
ideology of local inhabitants (see Janz, 2007). As such, the Gobi Desert serves as a
geographical barrier between two environmental zones that were the setting for
significant but divergent economic and cultural developments. Agriculture was central in
the establishment of early Chinese states, which is why the study of domestication its
contribution to the establishment of sedentary village communities and social complexity
is a focal point of Chinese archaeology. Likewise, the adoption of domesticated herd
animals and the spread of nomadic pastoralist economies allowed for the rise of powerful
pastoralist states, including the Mongol Empire, whose conquests and migrations shaped
the history of cultures across Eurasia – from Mongolia, to Eastern Europe, the Middle
East, and South Asia.
Despite our current knowledge of this long-standing mutual influence and
interaction between “the steppe and the sown”, knowledge of prehistoric hunter-gatherers
and the transition from hunting and gathering to herding is severely limited in East Asia.
It is surprising that the prehistory of the Gobi Desert has been recently so little studied,
since the strategic geographic nature of the region implies a central crossroads in the
divergent technological and cultural developments of Neolithic Northeast Asia.
Understanding the trajectory and origin of technological development and subsistence
economies in the Gobi Desert is essential to furthering our understanding of technological
and cultural transmission between the forerunners of powerful nomadic pastoralist
communities and emerging sedentary agriculturalist civilizations in China and eastern
Central Asia. Across this region, new technological achievements spread from the West,
such as bronze metallurgy and certain types of animal husbandry, reaching
18
agriculturalists settled along the Yellow River (Huang He) and contributing to the
florescence of the Chinese state.
Figure 1.1 Map of Northeast and Inner Asia indicating study area. Base map copyright
of maps.com, used by permission.
The study of Gobi Desert prehistory began with the discovery of hominid fossils
and stone tools by Western scientists in the early 1900s (Andersson, 1943). These
discoveries led to the inclusion of archaeologists in two important scientific expeditions
in Mongolia and China during periods of tumultuous local political upheaval in the 1920s
and 1930s. The Central Asiatic Expeditions were led by Roy Chapman Andrews under
the auspices of the American Museum of Natural History in New York City. Sven Hedin
led the Sino-Swedish Expeditions, which were funded by the Swedish State, the Chinese
government, Deutsche Lufthansa and several private donors (Hedin, 1943).
19
Recovered archaeological materials from these expeditions spanned all
occupation periods, from the Palaeolithic to modern times, but the vast majority of finds
were Stone Age and considered to represent the Mesolithic and Neolithic periods
(Nelson, 1926a, 1926b; Maringer, 1950, 1963). Artefacts from almost 500 Stone Age
sites were shipped to the United States and Sweden for analysis along with other
materials recovered by expedition scientists. Half of the materials collected during the
Sino-Swedish expeditions was left in China and eventually studied by Chen Xingcan in
the 1980s (see Fiskesjö and Chen, 2004). Most were derived from surface assemblages
and a few from excavated contexts. Heightened political tensions and civil war, followed
by the rise of isolationism among the communist governments, effectively terminated
work by Western scientists.
Collections made by Central Asiatic Expedition archaeologists Nels C. Nelson
(1925) and Alonzo Pond (1928) are currently housed at the American Museum of Natural
History. These collections have been studied briefly by a number of archaeologists,
including Richard Morlan (1976). A list of sites, artefact illustrations, and portions of
original journals and unpublished manuscripts were published by Walter Fairservis
(1993). However, no extensive analyses had been undertaken until several sites from the
Shabarakh-usu locality were studied as part of my Master’s thesis research (Janz, 2006)
and ostrich eggshell specimens from Shabarakh-usu and the Baron Shabaka Well locality
(Site 19) were radiocarbon dated in 2007 (Janz et al., 2009).
Collections made by the Sino-Swedish Expedition archaeologist, Folke Bergman,
were supposed to have been returned to China following study. For many years the Stone
20
Age remains were believed to have been repatriated to China, as had been a portion of the
Chinese Neolithic material excavated by Andersson, and most of the historic remains
uncovered at Etsin Gol (Edsengol/Ejina He). During his tenure as Director of the
Museum of Far Eastern Antiquities in Stockholm in the early 2000s, Magnus Fiskesjö
discovered Bergman’s collections which had been stored at the museum while being
studied and published by Johannes Maringer. Indeed, the collections had not been reexamined since Maringer’s work in the 1940s (Maringer, 1950, 1963). My initial
assessment of Nelson’s and Pond’s finds suggested that valuable data could be salvaged
from the assemblages, and the renewed availability of the Sino-Swedish archaeological
collections made the project even more appealing. The sheer geographic extent
represented offered a sample that could not be reproduced under modern logistical and
political constraints. With recent advances in chronometric dating on ostrich eggshell
and pottery, the collections could be used to establish connections between Gobi Desert
groups and contemporaneous environmental and cultural influences.
Between the collections from the Central Asiatic and Sino-Swedish expeditions, I
studied over 100 sites from three Gobi Desert regions – the East Gobi, the Gobi-Altai,
and the Alashan Gobi, producing a sample size of approximately 6,000 artefacts. In
February 2010, I was able to examine additional collections at the Institute of
Archaeology, Mongolian Academy of Sciences in Ulaanbaatar, and the Institute of
Archaeology and Ethnology, Russian Academy of Sciences, Siberian Branch in
Novosibirsk. Analysis of these collections reinforced my initial impressions of
chronology, but these assemblages were not included in this analysis.
21
Although my dissertation research focused on the analysis of existing museum
collections, fieldwork experiences have contributed to my understanding of Northeast
Asian prehistory. My fieldwork in Mongolia has included the excavation of Bronze Age
ceremonial structures in Arkhangai aimag, full-coverage survey and mapping of
prehistoric and historic sites in Dundgov’ aimag (Middle Gobi province), excavation at
the open-air Palaeolithic Tolbor locality in Bulgan aimag, and an exploratory survey of
the Shabarakh-usu locality in Ömnögov’ aimag (South Gobi province). I have also
participated in the excavation of Kurma XI, a Serovo/Glazkovo period cemetery in the
Lake Baikal region of Siberia. These experiences have been invaluable in building my
knowledge of archaeological landscapes and interpretation of land-use, and have helped
me to situate the Gobi Desert within a larger body of Northeast Asian archaeology.
Prior to this study, only two post-Last Glacial Maximum (LGM) archaeological
sites in the Gobi Desert proper had been dated (Derevianko et al., 2003; Janz et al., 2009)
and so our knowledge of the region consisted mostly of broad generalizations (Maringer,
1963; Fairservis, 1993; Derevianko et al., 2003). While recent studies have focused on
the relationship between the effects of climate change on desert hunter-gatherers and the
development of agriculture in northwestern China (Bettinger et al. 2007, Bettinger et al.,
2010a, 2010b), very little is known about the subsistence and settlement strategies of
desert inhabitants living on the margins of agricultural China. In order to understand how
domestic herd animals and nomadic pastoralism were first introduced to Mongolia and
northern China we must consider the organizational strategies of post-LGM hunter-
22
gatherers, who would have been the first to adopt domesticated herd animals and
incorporate them into existing subsistence strategies.
Based on the geographic range of collections and the distinct variability of their
geological and environmental contexts, three target regions were selected for further
research as part of the present study. Localized variability in Gobi Desert palaeoclimatic
records illustrates that major regional variation in the influence of monsoon regimes and
subtle deviations in vegetation, hydrology, and geography could have fostered the
development of divergent archaeological records. Each target region was collected under
the direction of an individual archaeologist, and regional samples were expected to be
relatively cohesive due to consistency in collection strategies. The three target regions
are defined below and illustrated in Figure 1.2.
The East Gobi (collected by Alonzo W. Pond)
The East Gobi is a desert-steppe environment of basins, small lakes (nuur/nur/nor),
steppe, and mesas, being dissected by numerous drainage channels, riverbeds, and dry
gullies. The easternmost bend of the Yellow River borders the western edge of this
region, while the northeastern edge is bordered by the Hushandake Sandy Land. Isolated
patches of dune-fields are distributed across the region. The archaeological collections
that I studied came from three major subregions: Southwest, is located just above the
north-easternmost bend of the Yellow River, and is typified by plains and mountains, the
latter of which seem to have been an area of lithic raw material procurement; Shara
23
Murun River, covers the area along the Shara Murun river system, extending over
badlands, valleys, dune-fields, small lakes/lake basins, and mesas across the neighbouring
plains, and bordered by southern uplands; and farthest east, the Great Lake Basin is
flanked on the eastern edge by mountains over 1500 m a.s.l. (Hill, n.d.: 21, 22), and is
characterized by badlands, plains, valleys, many mesas, numerous small lakes, and dunefields along the southern and eastern margins (see Figure 1.2 for map of subregions).
The southern edge of the East Gobi falls under the influence of the East Asian
summer monsoonal system and is located along what is today a temperate desert to
temperate steppe transitional zone. During its period of northward migration, the East
Asian Summer Monsoon system would have more heavily influenced the East Gobi than
the Gobi-Altai and the Alashan Gobi. Increasing humidity, warmer temperatures, and a
decrease in aeolian activity are evident in different parts of this region between about
13.0-10.0k cal BP, and woodlands invaded steppe-dominated environments by the midHolocene (Wang et al., 2001; Li et al., 2002; Yang et al., 2004).
The Gobi-Altai (collected by Nels C. Nelson)
The Gobi-Altai is a desert to desert-steppe environment surrounding the easternmost
foothills of the Altai Mountains. The region is characterized mostly by sparsely
vegetated gravel pavements (gov’ is the Mongolian root word for the Gobi), but is
interspersed with dune-field accumulations. Long-lived, internally-drained lake basins,
shallow brackish former or seasonal lakes, and wetlands are scattered throughout lowland
24
habitats. Large alluvial fans and scattered west-to-east-trending ranges divide the GobiAltai and Alashan Gobi regions. The collections are derived from three subregions:
Shabarakh-usu constitutes the area around the Neolithic type-site of the same name, and
is characterized by uplands and basins where many dune-field accumulations are situated
around dry or seasonally filled shallow lakes and riverbeds, some with Haloxylon
ammodendron “forests” (zakh/dzag or saxaul); the Arts Bogd-Ulan Nuur (Nor) Plain is an
area particularly rich in high quality lithic raw materials such as jaspers and chalcedonies,
and is characterized by a few scattered dune accumulations around numerous shallow
seasonal or dry lakes and former wetlands; and the Valley of the Gobi Lakes, a large
valley scattered with dune-field accumulations and east-west trending sub-valleys
situated between the southern slope of the Khangai (Khangay) Mountains and the
southeastern-most extent of the Altai massif, home to four major semi-saline lakes (Böön
Tsagaan Nuur, Adagin Tsagaan Nuur, Orok Nuur, and Tsagaan Nuur) fed by rivers
draining from the Khangai Mountains. Due to its northerly position, the valley is largely
outside the influence of the Asian summer monsoon system.
An arid desert reached its maximum extent in the Gobi-Altai during the LGM and
then again after 13.0k cal BP (Feng et al., 2007). Climatic amelioration, recognized
primarily in palaeoclimatic records by a high in effective moisture, reached its height
around 8.5k cal BP and continued until about 5.0k cal BP (Owen et al., 1998; Herzschuh,
2006; Feng et al., 2007). The region would likely have remained an arid grass and shrubland environment, interspersed with more stable and better vegetated river valleys, stream
channels, and playas (Owen et al., 1997).
25
The Alashan Gobi (collected by Folke Bergman)
The Alashan Gobi is a semi-desert region characterized by dune-fields, dissected
badlands, and gobi pavements. Dune-fields are more extensive here than in the East Gobi
or Gobi-Altai regions. Archaeological collections are derived from several subregions of
the northern Alashan Desert: the Eastern Alashan, just west of the Yellow River and
foothills of the Langshan Mountains; the Galbain Gobi, an extensive basin of dunes,
badlands, and dry lake basins which stretches along the border between Mongolia and
Inner Mongolia; the Goitso Valley, situated along the southern margin of the Galbain
Gobi depression, where near surface-level groundwater hosts many small oases and
supports rich pasture land; the Ukh-tokhoi/Khara Dzag Plateaux, an upland zone along
the northern margin of the Badain Jaran Desert, rich with high quality tool stone and
characterized by drift sand and relict higher elevation marshlands; the Juyanze region,
located in the far west, represents lakes and oases of the terminal palaeolake system fed
by major river drainages from the Qilian Mountains; and, finally, the Gurnai Depression,
a major erosional basin situated between the Ruoshui/Heihe drainage system and the
Badain Jaran Desert.
Climate change, as indicated by local palaeoenvironmental records, is more
similar to the Gobi-Altai region than the East Gobi. Terminal Pleistocene and early
Holocene desert environments were characterized by increased humidity before 7.0k cal
BP, with a drought event occurring between 7.0-5.0k cal BP (Chen, 2003). The middle
Holocene was much less arid with a warm and wet climate between ca. 5.6-3.9k cal yr
26
BP, when arid to desert-steppe vegetation dominated (Herzschuh et al., 2004; Mischke et
al., 2005).
Figure 1.2 Map of Gobi Desert locales, target regions, and subregions.
I collected data from Gobi Desert assemblages with two primary objectives:
understanding local chronologies; and reconstructing Gobi Desert settlement systems for
the terminal Pleistocene to late middle Holocene. The role of external stimuli, such as
climate change and the development of nearby emergent farming and herding economies
were explored in order to contextualize contemporary patterns of Gobi Desert
technological and economic change.
27
A study of plant and animal domestication processes and economic change is
especially important in order to contextualize Gobi Desert land-use and subsistence
within broader patterns of post-LGM technological and economic developments.
Chapter 2 details the chronology of post-LGM technology and economy in Northeast
Asia and presents the history of and existing literature on Gobi Desert archaeology. An
examination of the origin and intensification of herding practices in Northeast Asia and
the rise of nomadic pastoralism in Mongolia are especially relevant to the issue of
transitional hunter-gatherer strategies at the end of the Neolithic and the beginning of the
Bronze Age.
Accelerator mass spectrometry (AMS) radiocarbon dating of ostrich eggshell
artefacts, and pottery (O’Malley et al., 1998; Kuzmin and Shewkomud, 2003; Janz et al.,
2009), as well as luminescence dating of pottery (Aiken, 1985; Feathers, 2003) provide
chronological control for key sites. The results of chronometric dating are presented in
Chapter 3, along with a study of defining technological developments in neighbouring
regions of Northeast Asia. The use of contextual data in building local chronologies is
imperative for recognizing and contextualizing inter-regional trends, and helps to
construct a comparative framework for both directly and indirectly dated assemblages.
Recognition of diagnostic stone tool and ceramic types associated with dated
assemblages, enabled cross-dating at many additional sites.
Based on contextual data and chronometric dates, I initially formulated testable
hypotheses in Chapter 4. Lithic analysis was used to reconstruct patterns of raw material
use and draw inferences about residential mobility and relative length of occupation in
28
various ecological zones. The relative length of site occupation was determined based on
patterns of lithic reduction and use/retouch. Less residentially mobile groups can more
easily stockpile raw materials, while more highly mobile hunter-gatherers tend to
organize technological systems around formal, diachronically curated, and intensively
used tools (Shott 1986; Parry and Kelly 1987; Torrence 1989; Kuhn 1994; Wallace and
Shea, 2006). A general picture of duration, site function and temporal shifts in residential
mobility for the entire Gobi Desert, and for each target region, were based on relative site
density, variability in artefact assemblages, and the results of lithic analysis.
Detailed analyses of existing local and regional trends in climate and
palaeoenvironment are presented in Chapter 5 and related to local trends in land-use for
each target region. Although woodlands are restricted in modern times, recent research
on the ecology of relict desert forests suggests that riparian and upland forests were much
more widespread in prehistoric times. Published palaeoenvironmental data allowed for a
synthesis of early to late middle Holocene environments in each of the target regions. In
Chapter 6 the resulting palaeoenvironmental synthesis was used to interpret land-use,
mobility, and subsistence as they relate to ecological expressions of climate-mediated
environmental change, including the seasonal availability and density of local resources.
Lastly, the relationship between changes in residential mobility and technology among
hunter-gatherer groups, and the emergence of pastoralist economies is addressed.
This study used existing museum collections to establish a baseline chronology of
technology and land-use for post-LGM hunter-gatherers in the Gobi Desert. A holistic
and in-depth approach to collections analysis allows us to move beyond seriation, and
29
speculations about climate-mediated culture change. A wide range of data was employed
in the evaluation of Gobi Desert archaeology, including contextual knowledge of
technological and economic developments in neighbouring regions, chronometric dating,
archival data, qualitative artefact analysis, quantitative lithic analysis, and a synthesis of
detailed regional and local palaeoenvironmental data. The resulting observations
represent the first substantive step in addressing issues of residential mobility, responses
to climate-mediated environmental change, economic strategies, and subsistence. This
study establishes a foundation upon which new researches in museum collections
analysis and excavation can be used to refine and enrich our knowledge of Gobi Desert
prehistory, as well as our understanding of cross-cultural interactions within the
transitional economies of Neolithic Mongolia, China and beyond.
30
CHAPTER 2 – CHRONOLOGY OF TECHNOLOGICAL
DEVELOPMENTS AND NEOLITHIZATION IN POST-GLACIAL
NORTHEAST ASIA
Following the Last Glacial Maximum (LGM) a number of technological
trajectories were shared among hunter-gatherers in Northeast Asia. Even as late as the
Neolithic, geographically delineated decorative and technological styles suggest close
interaction or cultural affinities across economic boundaries (Maringer, 1950; Larichev,
1962; Debaine-Francfort, 1995; Cybiktarov, 2002; see also illustrations in Chard, 1974).
Post-LGM material culture in the Gobi Desert is consistent with this observation and
bears evidence of affinities with areas to both the north and south (Maringer, 1950;
Okladnikov, 1962).
These widespread stylistic similarities are not remarkable since Gobi Desert
hunter-gatherers lived in close proximity to and probably had some knowledge of or
contact with their agriculturalist and pastoralist neighbours. At the same time, it is
unfortunate that reconstructions of Mongolian and Gobi Desert archaeology have been
based largely on analogies with neighbouring regions. Pottery styles reminiscent of those
from better known cultures in Siberia or agricultural China has resulted in an emphasis on
migration theories and influenced interpretations of economic activity (Maringer, 1963;
Okladnikov, 1962; Chard, 1974:82). This circumstance has led to a state in which the
actual chronology and nature of local economic and technological developments are
relatively unknown. In order to assess the relationship between technological and
31
economic developments amongst Gobi Desert groups and other contemporaneous
populations within China and Mongolia, it is essential to establish a clear chronological
understanding of developmental processes, including the adoption of local and foreign
plant and animal domesticates throughout Northeast Asia.
This chapter highlights both the broad similarities and the notable divergences
between late Pleistocene and early Holocene groups in Northeast Asia. The purpose is to
outline cross-regional generalizations, while promoting increased understanding of
distinct regional chronologies and unique variations of such broadly characterized
concepts as the “Neolithic package,” especially for the Gobi Desert. In short, this chapter
summarizes those developments in order to contextualize regional chronologies of
technology, land-use, and subsistence throughout the terminal Pleistocene until the late
middle Holocene.
2.1. Key technological developments in Northeast Asia
Northeast Asia extends across eastern Siberia and the Russian Far East, Mongolia, Korea
and northern China, encompassing a wide range of cultures across varied ecological
zones that include sub-boreal and temperate mixed forests, forest-steppe, steppe, desertsteppe, alpine, basin-range, maritime, lacustrine, and riverine environments. Despite the
vast geographic range covered, there is also continuity amongst the Stone Age
technological traditions of this region. The most notable of these are the widespread use
of microblade technology beginning sometime before 24.0 ka (Kuzmin et al. [Eds.],
2007), and the early use of ceramic vessels which predates 16.0ka (Keally et al., 2003).
32
Highlights of developmental trajectories throughout key regions of Northeast Asia are
illustrated in Figure 2.3. This study focuses primarily on developmental trajectories
across the Upper Yensei River and Lake Baikal regions of Siberia (including Cis- and
Trans-Baikal; see Figure 2.1), northern China, and Mongolia, between which the
strongest cultural affinities have been noted (Maringer, 1950, 1963; Okladnikov, 1962;
Larichev, 1962).
Figure 2.1 Map of key locales discussed in Chapter 2. Base map copyright of maps.com,
used by permission.
33
2.1.1. Terminology
Assemblages characteristic of the period following the Last Glacial Maximum (LGM; ca.
21.3-19.0 kya1 [after Herzschuh, 2006]) are variously referred to as Terminal Upper
Palaeolithic, Late Palaeolithic, Palaeolithic-to-Neolithic transition, Mesolithic or
Epipalaeolithic. After Zhang (2000), the term Epipalaeolithic is used to designate the
terminal Old Stone Age period, which was followed by the New Stone Age/Neolithic.
Microblade core reduction sequences are typical of the Epipalaeolithic and although in
some regions they pre-date the LGM, they are more regularly used during following the
LGM period. This florescence of microblade technology may be related to post-LGM
repopulation by microblade-using groups, as an increase in the number of radiocarbon
dates following the LGM suggests elevated population density across northern China and
other regions (Barton et al., 2007). Although a variety of new tool types were introduced
during the Epipalaeolithic, lithic technology across Northeast Asia was overwhelmingly
and increasingly microblade-based.
In regions where agriculture did not develop until historic times, the
Epipalaeolithic was followed by the widespread use of pottery and characterized by a
more complex and varied lithic tool kit. The spread of and increased reliance on
processing technologies like ceramic vessels and grinding stones are typical of the New
Stone Age in Northeast Asia (Derevianko and Dorj, 1992; Lu, 1998; Cybiktarov, 2002;
Cohen, 2003; Kuzmin and Shewkomud, 2003). While emerging sedentary agricultural
1
All uncalibrated radiocarbon dates were converted to calibrated years BP using the online version of
CalPal-2007 online (Danzeglocke et al., 2011) and are written as “k cal yr BP.” The original published
dates or ages are given in brackets. Calibrated radiocarbon dates are reported as published and written as
“k cal yr BP.” Luminescence dates are given as “ka.” “Kya” is used when discussing the timing of
specific events. “Kya”, “ka”, and “k cal yr BP” are all considered to be roughly equivalent.
34
communities in North China began to abandon microblade technology and focus on the
production of polished stone and bone tools (Lu, 1998; Guo, 1995a; Jia, 2007),
formalized microblade technology continued in much of Northeast Asia and was
complimented by well-executed bifacially pressure-flaked tools like projectile points
(usually referred to as “arrowheads”) and knives (Derevianko and Dorj, 1992; Nelson
[Ed.], 1995; Keally et al., 2003; Janz, 2006).
According to Russian terminology, the adoption of pottery signals the beginning
of the Neolithic, but extremely early post-LGM dates for pottery in Japan and the Russian
Far East make this appellation less than ideal. In order to maintain coherency in this
chapter, the dichotomous terms Epipalaeolithic and Neolithic will be used when
necessary, as outlined above, but the reader should remember that the terms refer to the
presence/absence of ceramics rather than differences in economy. The existing
terminology is discussed in more detail in Chapter 3.
2.1.2. Microblade technology
Microblade technology is the hallmark of the Northeast Asian Epipalaeolithic. The
production of microblades (parallel-sided flakes less than 10 mm in width; after Seong,
1998) from prepared cores was employed as the primary method of lithic reduction in
Northeast Asia just prior to or following the LGM. According to the existing data, some
form of microblade technology may have been used prior to 25.0 kya in different regions
of Northeast Asia, especially Siberia, but was distributed more widely between about
25.0-20.0 kya (Keates, 2007; Kuzmin, 2007; Gladyshev et al., 2010).
35
Figure 2.2 Map of archaeological sites mentioned in Chapter 2. Base map copyright of
maps.com, used by permission. 1. Xiachuan; 2. Nanzhuangtou; 3. Zhoukoudian; 4.
Hutouliang; 5. Yujiagou; 6. Cishan; 7. Banpo; 8. Lingkoucun; 9. Kexingzhuang; 10.
Dadiwan; 11. Xishanping; 12. Xindian; 13. Dahezhoung; 14. Changshan; 15. Pigeon
Mountain Basin; 16. Zhongri; 17. Donghuishan; 18. Ukh-tokhoi/Khara-dzag Plateaux
region; 19. Shabarakh-usu; 20. Chikhen Agui; 21. Altan Bulag; 22. Ulan Khada; 23.
Tamsagbulag; 24. Xinglongwa; 25. Xiaoshan; 26. Dabasu; 27. Yinggeling.
Goebel (2002) suggested that precursors to microblade technology developed
early in the northerly reaches of Northeast Asia and dispersed southward as populations
were pushed back by extreme glacial climates. Standardized microlithic technology
would then have developed in the south and been re-introduced following the LGM. This
model is supported by slightly earlier dates for transitional forms of microlithic
technology in the Altai Mountains of Siberia, and a lack of evidence for the use of
microlithic technology in the same region immediately following the LGM. Still, the
many scattered finds of developed and standardized microblade technology in Northeast
Asia are not sufficiently dated to test Goebel’s hypothesis (see Kuzmin et al. [eds.],
36
2007). Possible transitional forms were also found in northern China at Shuidonggou
Locality 2, and dated to 28.7-33.9k cal yr BP (29.0-24.0k yr BP) (Madsen et al., 2001).
Human populations apparently contracted, abandoning large parts of Northeast
Asia, during the LGM (Vasil’ev et al., 2002; Barton et al., 2007). Such a population
decline or geographic shift would have affected the overall distribution of archaeological
sites, as well as the geographic range of microblade use. A gap between about 28.518.2k cal yr BP in the archaeological record of northern Mongolia (Gladyshev et al.,
2010) supports a late Pleistocene occupational hiatus, with repopulation commencing
early in the post-LGM period. It has been suggested that microblade technology may
have been adopted as a means of creating standardized flakes for well-made composite
hunting tools in the face of increasing long, harsh winters and declining large herbivore
populations (Elston and Brantingham, 2002). In contrast, human populations in northern
China, presumably under stress from harsh winter climates and less predictable prey,
actually appear to have favoured a highly expedient technological strategy using poor
quality and easily available local raw materials like quartz (Barton et al., 2007). Despite
differences in reduction strategies, which may be related to differential access to lithic
raw materials, the manufacture of inset flakes for use in composite tools indicates an
underlying shift in hunting technology that emerged prior to the LGM, but became
increasingly important during the terminal Pleistocene and early Holocene.
Microblade technology is representative of terminal Pleistocene and early
Holocene sites over a vast area. Developed microblade technology, with a wider array of
core forms, had been incorporated into tool kits across Northeast Asia and Alaska by
37
15.4-11.5k cal yr BP (13.0-10.0k yr BP) (Aikens and Akazawa, 1996; Ackerman, 2007).
Post-LGM occupations at the Tolbor locality in northern Mongolia are typified by an
emphasis on microblade core technology with retouched microblades and other small
tools (Gladyshev et al., 2010).
The decline of microblade technology in Northeast Asia was highly variable, but
appears to have disappeared earliest in regions of declining mobility and/or agricultural
production. By about 12.9k cal yr BP (11.0k yr BP) microblade technology was replaced
by flake technologies in much of Japan and Korea, though it lasted somewhat longer in
northern Japan (Aikens and Akazawa, 1996; Seong, 1998; Sato and Tsutsumi, 2007). In
central China, microblade core reduction strategies were gradually supplanted by
polished stone and bone tools after about 8.9k cal yr BP (8.0k yr BP) as plant cultivating
groups became increasingly sedentary (Lu, 1998). By 8.0k cal yr BP developed
microblade assemblages in North China and Northeast China typically included pottery,
milling/grinding stones, as well as both polished and bifacially retouched stone tools.
Although microblade use declined or nearly disappeared in some regions, especially after
5.8k cal yr BP (5.0k yr BP), the technology was still widely employed in other parts of
Northeast Asia (Lu, 1998). The decline of microblade technology in northern China
seems to be partially associated with an increase in sedentism (see Lu, 1998). Huntergatherers in arid northern China and Mongolia continued to rely heavily on microblades
until at least the early Bronze Age (Aikens and Akazawa, 1996; An, 1992a, 1992b; Lu,
1998; Xia et al., 2001; see also Chapter 2).
38
2.1.3. Pottery
Pottery is often associated with sedentary agricultural societies, but ceramics were first
used by hunter-gatherers in the Russian Far East, Japan, and southern China. The initial
adoption of pottery in Northeast Asia, beginning by 16.5k cal yr BP (Keally et al., 2003)
is extraordinarily early in comparison to elsewhere in the world – after 11.0k cal yr BP in
Saharan Africa and after 9.5k cal yr BP in western Asia (Close, 1995; Moore, 1995; Rice,
1999). Environmental contexts for the first pottery use in Northeast Asia tend to be
characterized by mixed coniferous-broadleaved forests and forest-steppe (Keally et al.,
2003; Kuzmin and Shewkomud, 2003). Various types of tool kits were associated with
early Northeast Asian ceramic sites, including: assemblages predominated by partially
ground and chipped stone adze/axes, blades, and large bifacial points in northern, eastern
and central Japan; microblade complexes in western Japan; and a combination of
microblade core reduction, bifacial shaping and some polishing in the Russian Far East
(Keally et al., 2003; Kuzmin and Shewkomud, 2003).
The variety of artefact assemblages found with the first pottery makes it difficult
to associate the early use of ceramic vessels with a particular pattern of technological
change or subsistence strategy. Vessel forms are difficult to reconstruct from the small
shards that have been recovered, but analyses of carbon and organic residues on the
interior and exterior surfaces indicate that they were often used for cooking (Keally et al.,
2003). The most broadly applicable explanation for the invention (as distinct from
adoption) of pottery at different times across the globe is that new processing
technologies were required by distinct changes in subsistence during the terminal
39
Pleistocene or early Holocene (for summaries see Pavlů, 1997; Rice 1999). Such
explanations have been applied to the adoption of pottery in Northeast Asia (IkawaSmith, 1976, 1986; Underhill, 1997; Keally et al., 2003; Kuzmin and Shewkomud, 2003).
The earliest pottery in southern China may be associated with evidence for intensive bone
marrow and grease extraction. It has been argued that pottery was first introduced to
facilitate bone grease rendering (Elston et al., 2011), although earlier evidence for grease
rendering in Europe is not associated with pottery (Manne and Bicho, 2009; Manne et al.,
2012). There is currently insufficient data to associate early Northeast Asian pottery with
any one subsistence strategy.
Outside of the Russian Far East, Japan, and southern China, pottery-use may have
been highly localized until after 10.0k cal yr BP. In the Trans-Baikal region of Siberia,
numerous dates on both charcoal and potshards indicate that ceramics were probably first
used around 13.0k cal yr BP (O’Malley et al., 1998; Kuzmin and Orlova, 2000), but the
weight of associated radiocarbon dates for Siberian assemblages indicate that pottery was
probably not widely used until after 9.0k cal yr BP (Kuzmin and Orlova, 2000). The
earliest ceramic assemblages in North China date to 11.1 ka for Yujiagou (Xia et al.,
2001) and about 11.9k cal yr BP for Nanzhuangtou (Wu and Zhao, 2003; Kuzmin, 2006);
but, as in Siberia such early sites are rare. Dates from the earliest ceramic site in Korea,
on the western end of Jeju Island, suggest that pottery use in Korea does not pre-date 7.9k
cal yr BP (Nelson, 1993; Bae and Kim, 2003; Kuzmin, 2006). New direct dates on
pottery from northern China and Mongolia indicate that the earliest Gobi Desert pottery
40
dates to at least 9.6k cal yr BP (see Chapter 3). After about 7.0k cal yr BP, pottery was
an important technology across most of Northeast Asia.
2.1.4. Grinding stones
Grinding or milling stones have been found across the globe in much earlier contexts than
pottery. Recent research has identified starch grains from various wild plants, including
cattail and fern, on grinding tools from a variety of European Palaeolithic contexts dated
to between about 32.0 to 28.1k cal yr BP (Revedin et al., 2010). Grinding stones appear
to have been used even earlier for tasks like grinding ochre in southern Africa
(McBrearty and Brooks, 2000). Such implements are usually small and appear to have
been chosen based on the natural form of suitable cobbles. Grinding tools occur
sporadically in Palaeolithic contexts, making it difficult to as clearly link the origins of
such technology with post-LGM developments as pottery (but see Soffer, 2000). It is
clear that the use of grinding stones in much of Northeast Asia intensified during the
early Holocene, and assemblages came to include more formalized forms with reduced
portability (Nelson [Ed.], 1995; Shelach, 2000; Cohen, 2003; Keally et al., 2003; Janz,
2006). Such formalized technology is usually thought to be associated with increased
plant processing and sedentism, though the record of such developments in the Gobi
Desert is more complex (see also Hoopes, 1995; Bright and Ugan, 1999; Eerkens et al.,
2002; Eerkens, 2003, 2004).
41
The earliest evidence of increasing reliance on grinding stones in Northeast Asia
comes from China and is probably related to food processing as well as grinding of the
mineral pigment ochre. Xiachuan, a site in Shanxi, features formal microblade
technology, some evidence of polishing, and at least 27 stone grinding slabs varying in
size from 8 to 35 long and 8.5 to 21 cm wide (Cohen, 2003). Judging by the youngest of
six radiocarbon dates for the stratum within which the microlithic assemblage was found,
an age of about 20.0k cal yr BP can be proposed (Tang, 2000; but see Elston et al., 2011).
Comparably early sites in North China are not yet known.
Dates on excavated remains from around the foothills of the Helan Mountains,
west of the Yellow River, indicate that thin, lightweight grinding stones were used prior
to 13.5k cal yr BP (11.6k yr BP) (Elston et al., 1997). Similar grinding stones are
typically associated with microlithic assemblages in the western Gobi Desert and differ
from formal rollers and saddle querns, found both in agricultural regions of northern
China and in the East Gobi. East of the Yellow River grinding slabs are associated with a
developed microlithic assemblage at Hutouliang (ca. 12.9k cal yr BP or earlier), while at
Nanzhuangtou (ca. 11.9k cal yr BP) slabs and rollers were found along with pottery,
possible domesticated animals and a tool kit that did not include microlithic technology
(Cohen, 2003; Liu, 2004). Chipped adze/axes are also often associated with such
assemblages (Lu, 1998). Similar dates are reported in Japan, where large querns and
grinding stones are associated with rounded stone axes and chipped stone axes between
15.0 and 13.3k cal yr BP in the southwest (Keally et al., 2003; Habu, 2004). After 11.0k
42
cal yr BP, grinding stones and mortars were used more frequently across Japan (Underhill
and Habu, 2006).
Less information is available for sites in Northeast China and Korea, where there
appears to be less evidence of transitional Palaeolithic-to-Neolithic economies. In these
regions, grinding stones are usually part of the Neolithic package, which includes
regional variants of microblade technology, pottery, and polished stone tools. The
Dabasu site in Jilian Province might be an exception, as the assemblage is thought to date
to about 10.0k yr BP (11.5k cal yr BP) and contains one grinding slab in association with
many small flake tools, microblades and microcores, along with chipped adzes and
knives (Lu, 1998). Most early assemblages in Northeast Asia that contain grinding
stones are dated to between 8.0 and 7.0k cal yr BP (Nelson, 1993; Guo, 1995a; Jia, 2007).
In the Gobi Desert, grinding stones are only associated with post-LGM
assemblages although the timing of their introduction has not been established. East
Gobi sites are distinctive in that they contain much higher numbers of grinding stones,
including more formal styles like saddle querns, pestles, and rollers (Fairservis, 1993).
Such forms are seldom found in the more western Gobi Desert and Alashan Gobi sites
rarely contain grinding stones of any type (Maringer, 1950; Fairservis, 1993; Janz 2006).
This variation in distribution is likely related to differences in subsistence and may
indicate some association in the East Gobi with emerging agricultural communities.
43
2.1.5. Polished stone tools
The earliest evidence of polishing as a means of shaping and finishing stone tools is in
the form of chipped tools or pebbles with polishing marks. Chipped adze/axes with
localized polishing may be transitional to fully polished forms that occur in the later
Neolithic, but chipped and polished specimens persisted in the Gobi Desert, where fully
polished axes are rare. Formal grinding stones generally exhibit shaping through directed
polishing and are contemporary with other polished stone tools. Polished bone artefacts
become important during the Neolithic, and in some cases people seem to have relied
most heavily on bone as a raw material; however, the technology is representative of
earlier Upper Palaeolithic technological developments. For that reason, only polished
stone tools are discussed in this section.
Polishing is commonly associated with the manufacture of Neolithic macrotools
in Northeast Asia, particularly adzes and axes, but the earliest evidence for the use of
edge-ground adze/axes comes from the Upper Palaeolithic in Japan and may be older
than 34.0k cal yr BP (>27.0k/30.0k yr BP) (Oda and Keally, 1992). Such tools were
widespread in Japan and were used widely prior to 27.5k cal yr BP (23.0k yr BP), when
they were largely abandoned until after about 18.0k cal yr BP (15.0k yr BP) (Oda and
Keally, 1992). Small gravel fragments with polished scratches, found at the
Shandingdong (Upper Cave) site in Zhoukoudian, Beijing are considered to be the first
evidence of polishing technology at about 30.0k cal yr BP, but the evidence is
contentious (Wu and Zhao, 2003). In the Russian Far East, the earliest ceramic
44
assemblages (~16.5k cal yr BP) included polished points (arrowheads), knives and
bifaces (Kuzmin and Shewkomud, 2003).
In many regions of Northeast Asia, polished stone tools are reported in
combination with ceramics, but in many cases the technology predates the use of ceramic
vessels. This point is supported in Siberia by finds of polished slate shanks for composite
fishhooks, an adze, and a perforated stone object from pre-pottery layers at Ulan-Khada
in the Lake Baikal region – a scenario that is repeated in other habitation sites around
Lake Baikal (Khlobystin, 1969; Goriunova and Khlobystin, 1991; Weber, 1995).
Similarly, at the pre-ceramic Hutouliang site in Hebei Province, a point associated with
dates of about 12.5k cal yr BP (10,690 + 210 BP) provides the earliest clear evidence of
polished stone in northern China (Wu and Zhao, 2003).
After 8.0k cal yr BP, polished stone tools were increasingly important across
much of Northeast Asia. Polished stone tools are part of the typical Neolithic package
which appears in Northeast China and Korea between 8.0 and 7.0k cal yr BP (Nelson,
1993; Guo, 1995a; Lu, 1998; Jia, 2007), but there is little evidence of polished stone tools
prior to this period. In parts of Northeast China where flake tools and microblades were
more prominent, polished stone tools may comprise up to 30% of the total assemblage.
Many sites with evidence of long term settlement are characterized by tool kits with even
higher frequencies of polished tools such as adzes, knives, and rollers (Lu, 1998). By this
time, sedentary farming villages were typical of habitation sites in the Central Plains of
China and subsistence was focused on millet as a cultivated staple along with exploitation
45
of domesticated pigs, dogs, and possibly chickens (Liu, 2004). Grinding and polishing
stone and bone were the primary methods of tool manufacture (Lu, 1998).
Considering the frequent negative correlation between polished stone and
microblade technology, it is probable that polished tools met a new need amongst some
early to mid-Holocene groups for which microblade technology was either less proficient
or unsuitable. The simultaneous decline of microblade technology and florescence of
polished stone tools have been associated with the rise of agriculture (Lu, 1988), but a
similar correlation does not exist amongst hunter-gatherer groups in Japan and Siberia.
One possibility that has not been explored is that shifts in land-use and mobility would
have dramatically increased or decreased the availability of various raw materials.
Despite being more time consuming, polishing allows stone adze/axes or knives to be
resharpened with little loss of surface area. The technology is thereby more conservative
of raw material than microblade core reduction sequences. This is especially true in
circumstances when wear on tool edges is highly intensive, such as woodworking or in
conditions where enormous amounts of food had to be processed in a short time (Hayden,
1989). Additionally, raw material requirements for ground and polished tools are more
flexible, making the use of local resources more viable. Polished bone tools may also
have been more available with agriculture and increased sedentism as bone would have
remained from animals slaughtered and/or processed at a long-term habitation site. As
such, it can be hypothesized that a reliance on polished stone is probably related to both
the functionality of ground edges and a shift in the accessibility of various raw material
types.
46
5.0
10.0
15.0
20.0
25.0
kya
Microblades
Pottery
Grinding stones
Polished
Stone
Figure 2.3 Relative frequency of key artefact types in Northeast Asian chronology.
2.2. Chronology of food-production in Northeast Asia
As in many other parts of the globe, one of the defining developments in early to midHolocene Northeast Asia is the emergence of production economies dependant on
domesticated plants and animals. As in other Old World centres of agriculture and
urbanization (e.g., North Africa and the Middle East), politically separate pastoralist and
agricultural communities had emerged by the mid-Holocene. Domesticated plants and/or
animals that formed the bases of production economies in Northeast Asia were either
47
domesticated locally or domesticated in other parts of the world and later introduced.
Many species integral to the economic base of pastoralists (e.g., horses, cattle, sheep,
wheat) were first domesticated in the Middle East or Central Asia, while most of those
exploited by sedentary agriculturalists (e.g., pigs, millet, rice) were domesticated from
species indigenous to East Asia.
Some researchers have proposed that early to mid-Holocene Gobi Desert huntergatherer groups practiced cultivation of grass seeds, based on the increased use of
grinding tools and pottery in what are defined as Neolithic sites (Derevianko and Dorj,
1992; Cybiktarov, 2002; but see Janz, 2007). Nomadic pastoralism developed in the
region sometime between 3.5-3.0k cal yr BP, and although the economy was based on
domestic herd animals, local diets were still complemented by a range of wild resources
(this chapter). In order to assess possible relationships between Gobi Desert huntergatherers and contemporary developments in domestication and production economies
elsewhere in Northeast Asia, a clear understanding of regional domestication processes is
necessary.
2.2.1. Brief chronology of plant domestication
It is becoming clear that prior to domestication people around the world exploited a wide
range of local plant species (Revedin et al., 2010), but there is a widespread pattern of
evidence for intensive processing and storing of plant foods in the early Holocene that is
not characteristic of the Palaeolithic record and culminates in the genetic alteration of
certain species. Plant cultivation is the key marker of food production in modern
48
agricultural zones of Northeast Asia, providing the first evidence of intensive exploitation
and unequivocal domestication of a food resource. Both millet and rice agriculture
characterize early agricultural communities in the Central Plains of China, but due
partially to ecological constraints on early varieties of domesticated rice, millet was
initially the primary cereal crop in Northeast Asia. The process and timing of
domestication, particularly as understood in the sense of genetic alteration, is poorly
understood in East Asia and the relationship between the reliance on or extensive use of
certain plants and their actual domestication is still in the early stages of research
(Crawford, 2006; Jiang and Liu, 2006; Fuller et al., 2007).
Considering the relative abundance of archaeological data, millet (including the
species Echinochloa crus-galli, Panicum miliaceum, and Setaria italica) appears to have
been the first and most important domesticated cereal crop in North China. Evidence for
the large scale storage of broomcorn or common millet (Panicum miliaceum) is present at
the Chinese Neolithic site of Cishan between 10.3-8.7k cal yr BP, but morphological
evidence of domestication was not identified (Lu et al., 2009). Further west, at the
Dadiwan site, increased human reliance on millet is indicated by the apparent
provisioning of domesticated dogs with millet and meat by 7.9-7.2k cal yr BP, and pigs
by 6.5-4.9k cal yr BP (Barton et al., 2009). Charcoal concentrations in conjunction with
an increase in Poaceae pollen at Xindian in the middle Yellow River valley indicate
“slash-and-burn” methods of cultivation beginning around 7.7 ka (Li et al., 2009). While
the early dates from Cishan suggest that broomcorn millet was domesticated first in
North China, barnyard millet (Echinochloa crus-galli) was domesticated locally in
49
northern Japan (Crawford, 1997, 2006). Lithic assemblages and site structure at
Xinglongwa in Northeastern China also hint at intensive exploitation of cultivated plant
foods by 8.0k cal yr BP (7240 + 95 yr BP), but the only botanical remains reported are
wild walnut (Juglans mandshurica Maxim) and/or chestnuts (Guo, 1995a; Shelach,
2000). Even though the earliest substantial evidence for agriculture in North China dates
to sometime between 10.0 and 8.0k cal yr BP (8000-6000 BC) the practice of cultivation
probably began much earlier (Crawford, 2006; Lee et al., 2007).
Millet remains are found in mid-Holocene Northeast Asian archaeological sites
across North China, Northeast China, Korea, and Japan. The earliest known evidence for
the use of domesticated millet in Korea consists of foxtail millet (Setaria italica) grains
dated to 5.3k cal yr BP (3360 BC) (Crawford and Lee, 2003). Foxtail millet was used
much later in Neolithic sites in the Central Plains and was relatively rare in early contexts
(0.4-2.8% of grain crops by 8.0k cal yr BP [7235 + 105 yr BP]) (Lu et al., 2009).
Although wild barnyard millet is common in the late Early Jomon site of Hamanasuno in
Hokkaido, it is not until about 4.5k cal yr BP (2500 BC) that seeds are consistent with the
larger domesticated barnyard millet in the same region (Crawford, 2006). Considering
that the earliest evidence of domesticated millet in Korea and Japan are of varieties
uncommon in the Central Plains of China, it is probable that each species was
domesticated locally. It is not clear what influence the increasing frequency of food
producing economies may have had on local decision making processes associated with
domestication, but the influence of foreign developments in cultivation may have
contributed to local developments.
50
During the Longshan period, beginning about 4.5k cal yr BP (2500 BC), exotic
crops were being added to the existing agricultural suite of cultigens. The intensification
of agriculture that occurred at this time may be associated with extensive provisioning of
domestic animals (Crawford, 2006). By the middle Holocene (7.0-4.6k cal yr BP),
various agricultural communities had been established along the Yellow River and its
tributaries in North China, growing crops like soybean (Gylcine soja), adzuki bean
(Vigna sp.), hemp (Cannabis sativa), Chinese cabbage (Brassica chinensis), buckwheat
(Fagopyrum sp.), along with weedy plants such as chenopods (Chenopodium sp.),
canola/rapeseed (Brassica rapa), and other mustards (Brassicaceae) (Crawford et al.,
2005; Lee et al., 2007; Li et al., 2009). Rice (Oryza sativa) was domesticated south of
the Central Plains and was also found in agricultural contexts during the Longshan
period, as was wheat (Triticum aestivum), which would have been introduced from the
west (Crawford et al. 2005). The earliest evidences of wheat in China come from Gansu
Province, at the Donghuishan site in Minle County (5.0-4.5k cal yr BP [3000-2500 BC]),
and Xishanping (4.6k cal yr BP) (Li, 2002; Li et al., 2007). Such finds may illustrate a
path of introduction from Central Asian pastoralists.
Similar dates are observed in other regions of Northeast Asia. Rice was
introduced to Korea between 5.0-3.5k cal yr BP (3000 to 1500 cal BC), where
agricultural groups cultivated a similar suite of crops to those in the Central Plains of
China. By 3.0k cal yr BP agricultural production intensified in Korea and included wheat
(Triticum aestivum) and barley (Hordeum vulgare) (Crawford and Lee, 2003). A similar
situation occurred in neighbouring Northeast China, where sedentary communities
51
cultivated millet, rice, barley and sorghum (Sorghum sp.) by 3.2k cal yr BP (3000 BP)
(Jia, 2007).
Agriculture seems to have spread more slowly north of the fertile river valley of
the Chinese Central Plains. Shelach (2000) contends that soybeans and millet may have
been domesticated locally in Northeast China independently of influences from the south,
but it was not until about 5.0k cal yr BP (3000 BC) that groups in the eastern part of what
is now the Inner Mongolia Autonomous Region, People’s Republic of China (PRC) show
evidence of partial reliance on agriculture (Guo, 1995a; CICARP, 2003; Jia, 2007). It is
possible that the use of domesticated plants in Northeast China predates this evidence, as
plant processing technology associated with semi-subterranean houses, organized in what
appears to be a sort of village community is found at Xinglongwa, dated to 8.0-7.2k cal
yr BP (CICARP, 2003).
Evidence for use of domestic plants among sedentary or semi-sedentary
communities in neighbouring mainland territories is broadly contemporaneous with
Northeast China. In Korea, plant taxa from archaeological sites associated with
cultivation indicate an early reliance primarily on Chenopodium sp. (36.8%), foxtail
millet (35.5%), some use of broomcorn millet (9.1%), as well as wild millet at about 5.3k
cal yr BP (Crawford and Lee, 2003). Increasing use of foxtail millet culminated in a
focus on millet and soybean agriculture after about 3.3k cal yr BP (1300 BC) (Lee, 2001;
Crawford and Lee, 2003; Underhill and Habu, 2006). Foxtail millet cultivation is dated
to at least 4.7k cal yr BP (4150 + 60 BP) in the temperate Primorye region of the Russian
Pacific Coast, where wild foods continued to be exploited as evidenced by the presence
52
of wild millet, acorn and walnut (Kuzmin, 1998; Vostretsov et al., 2003; Jia, 2007). A
similar pattern is suggested in eastern Mongolia at the Tamsagbulag site, where cattle and
millet were exploited along with fish and other wild species (Derevianko and Dorj,
1992).
Wild resource use in concert with low-level food production is perhaps best
understood in Japan, where plant remains are abundant and well studied. Mainland
domesticates, including rice, were not adopted in Japan until after about 2.5 kya (RowleyConwy, 1984). Although a fully formed agricultural production economy was introduced
late, barnyard millet was domesticated locally and it is becoming clear that native
Japanese plant species – chestnut, sumac, adzuki bean, perilla (shiso/beefsteak plant) –
supplemented the diet and provided raw materials (Crawford, 2006; Matsui and
Kanehara, 2006). Across Japan a general pattern was established of fishing and marine
mammal hunting in the summer, collecting nuts and other plants in the autumn, hunting
terrestrial mammals in the winter, and collecting shellfish in the spring; however, this
general pattern of subsistence progressively included more intensified use of some
resources at the regional scale (Habu, 2004).
Low-level cultivation, complimented by the hunting and gathering of wild
resources, is more typical of the early Holocene in Northeast Asia than is the
development of sedentary agricultural communities. Use of domesticated plants became
more widespread by 5.0k cal yr BP and intensified after 3.0 kya. It is not clear to what
extent nomadic pastoralist groups in Northeast Asia exploited domesticated plant
resources prior to the Iron Age, when sedentary communities emerged within the larger
53
pastoralist system (Janz, 2007). The wealth of cultivated plant remains in sedentary
communities and the apparent reliance on wild species amongst more mobile foragers
indicates a correlation between decreased mobility and plant domestication.
2.2.2. Detailed chronology of animal domestication
The chronology of animal domestication in Northeast Asia is less well understood at the
regional level than plant domestication; however, new genetic data on major
domesticated species is making important contributions. Several key species in Northeast
Asian food-producing economies, including pigs, chickens, yaks, and water buffalo, were
domesticated in East Asia. Camels, dogs, and cattle may have been domesticated
independently in East Asia, but the record for these species is either more difficult to
interpret or, in the case of camels, very sparse. Several western species also became
important to subsistence in Northeast Asia and the timing of their introduction is an
important consideration. An overview of current knowledge about animal domestication
and adoption is offered in order to better contextualize evidence for various species in the
faunal record and the possibility of domestication in regions with few faunal remains.
Dogs
Dogs ae the earliest known domesticated species (faunal or floral) in the world.
Archaeological finds of domesticated dogs occur earliest in Europe, where cranial
fragments from Russia and Germany date to the terminal Pleistocene, and more recently
54
in Belgium at 35.5k cal yr BP (31,680 + 250 yr BP) (Nobis, 1979; Olsen, 1985; Sablin
and Khlopachev, 2002; Germonpré et al., 2009). Genetic data indicate that domestication
began around either 40.0 kya or 15.0 kya (Savolainen et al., 2002), and although
domestication in multiple geographic regions is suggested there is strong evidence of
dominance in the role of Middle Eastern wolves (von Holdt et al., 2010). Evidence of
genetic contribution from East Asian wolves indicates that local species also played an
important contributing role to modern genetic profiles (Olsen and Olsen, 1977; Vilà et al.,
1997; Savolainen et al., 2002; Pang et al., 2009).
Dogs were associated with early Neolithic sites in China (after 13.0k cal yr BP)
and there is direct evidence of provisioning domestic dogs with millet by 7.5-7.1k cal yr
BP (Barton et al., 2009). By the time of the Chinese Neolithic and Early Bronze Age,
distinct breeds of dogs had already developed in northern China (Shigehara et al., 1998).
Dogs appear to have also been important to Siberian hunter-gatherer groups. Dog and
human burials at the Ushki Lake site in northeastern Siberia date the use of domesticated
dogs to at least 11.9k cal yr BP (Kuzmin et al., 2008). In the Lake Baikal region dog
interments in burial contexts date dogs to about 7.8-7.0k cal yr BP (Losey et al., 2011).
These dates are not surprising considering evidence that domestic dogs were probably
used by Palaeoindian hunter-gatherers, having presumably crossed into North America
with humans migrating from Siberia (Olsen and Olsen, 1977; Fiedel, 2005). The oldest
dog remains in Japan date to approximately 10.5k cal yr BP (9,300 BP). These Jomon
dogs appear to have one phylogenetic origin, and it is suggested that they may have
accompanied humans migrating from the Asian mainland (Shigehara and Hongo, 2000).
55
Conversely, some of the earliest archaeological evidence for the use of domestic dogs in
Northeast China is clay figurines from Upper Yinggeling, dating to about 3.2-3.1k cal yr
BP (3025 + 90BP, 2985 + 120 BP) (Tan et al., 1995a). In Korea, dogs appear first in the
Early or Middle Chulmun (7.5-4.7k cal yr BP), but it is difficult to constrain the dates for
faunal remains as they do not appear to have been separated by stratigraphic level
(Nelson, 1993). The domestic dog probably spread through most of Northeast Asia by
7.5k cal yr BP, though inhabitants of Northeast China and Korea may have adopted the
animals somewhat later.
Considering the early dates for dog domestication in Europe and the lack of
evidence for anatomically modern humans in Northeast Asia until the late Pleistocene, it
is possible that dogs accompanied some of the first anatomically modern humans into
East Asia as they expanded out of western Eurasia. At the same time, evidence of an
important domestication event in southern China after 16.3 kya (Pang et al., 2009) could
also support the theory that domestic dogs spread north amongst hunter-gatherer bands
over several thousand years, reaching the Central Plains by about 13.0, Siberia by 11.9,
and Japan by 10.5k cal yr BP (Shigehara and Hondo, 2000; Cohen, 2003; Liu, 2004;
Kuzmin et al., 2008). Dogs seem to have been an important resource for sedentary
farming communities in the Central Plains, semi-sedentary Jomon peoples, and mobile
hunter-gatherers in Siberia. They may have variously provided food, protection from
other predators, traction, and/or aid in hunting. The frequent use of dog skeletons in
mortuary and ceremonial contexts suggests some ritual significance of the animal that
56
may have been associated with their importance in hunting, the symbolic ability to
control nature or simply companionship (see Losey et al., 2011).
Pigs
Pigs appear to have been the primary animal domesticated for food in the early Neolithic
of China. Analyses of mitochondrial DNA (mtDNA) from both European and Asian
domesticated pigs indicate that at least two distinct centers of early domestication existed,
one of which was in East Asia (Watanabe et al., 1999; Giuffra et al., 2000; Larson et al.,
2005). Tentative evidence for early domestication of pig, along with dog and chicken,
appear in the Central Plains sites of Nanzhuangtou (12.4-11.0k cal yr BP [10,500-9,700
BP]) and Hutouliang (12.9k cal yr BP [11,000 BP]) (Cohen, 2003; Liu, 2004). Mortality
profiles (60% between the ages of 0.5 to 1 year), tooth measurements, and context of
deposition have all been used to assert that domesticated pigs were being exploited at the
Cishan site in the Central Plains by about 8.9k cal yr BP (Yuan and Flad, 2002). Isotopic
signatures amongst pigs from the Dadiwan site in northwest China further indicate
feeding of pigs with millet beginning sometime between 7.2-5.8k cal yr BP. Other
individuals from the same time period are thought to represent wild animals raiding
millet fields or free-range animals that were only occasionally provisioned (Barton et al.,
2009). By 9.0-7.0 kya sedentary life in farming villages included not only the use of
57
millet as a common staple, but domesticated pigs, and possibly chicken2 (Yuan and Flad,
2002; Liu, 2004).
In Northeast Asia, beyond the Central Plains, pigs were widely exploited in both
domestic and hunted contexts. Pig and deer are common animals species represented in
early Neolithic sites in Northeast Asia such as Xinglongwa at 7.2k cal yr BP (5290 + 95
BC) (Wa, 1992); along with birds these two animals were also represented in art work on
ceramic vessels at Xiaoshan (4.7-4.6k cal yr BP [4110 + 85, 4200 + 85]) (Wa, 1992).
The mortality profile of pigs at Xiaoshan indicates a pattern (mostly aged 1-2 years)
consistent with that of domesticated animals, but it is not accompanied by clear evidence
of skeletal changes associated with domestication (Shelach, 2000). The later Neolithic
group referred to collectively as Hongshan, (5.5k cal yr BP [3535 + 110 cal BC, 2180
BC]) relied heavily on domesticated pigs and may have practiced pig sacrifice in ritual
activities (Wa, 1992).
Wild boar was a key hunted resource on the Korean Peninsula and may have been
domesticated during the Chulmun period (8.0-3.3k cal yr BP) (Nelson, 1993).
Domesticated pigs might also have arrived from North China by 2.0 kya (Kim and Choi,
2002). More careful analysis of faunal remains is necessary to determine the status of
early Sus remains. Pigs, along with dogs, were some of the first domesticated animals in
the Russian Far East. In the Primorye coastal region, bordering Northeast China, Bronze
Age occupations were typified by the exploitation of pig and dog around 4.5-2.5 kya
2
Considering the zoogeographic distribution of Gallus gallus gallus or red junglefowl, the wild ancestor of
domesticated chickens, it is clear that chickens were domesticated farther south than the Central Plains of
China. As such, the earliest evidence of chicken domestication should be sought further south in sites predating Cishan and Peiligang at 7.5k cal yr BP and about 5.5k cal yr BP respectively based on
dendrochronologically calibrated radiocarbon dates (West and Zhou, 1988; Akishinonomiya et al., 1994).
58
(Kuzmin, 1995). Japanese wild boar (Sus scrofa leucomystax) was also extensively
exploited on Hokkaido and Izu Islands in northern Japan, which are outside the record of
natural wild distribution, suggesting that beginning in the Early Jomon (7.0-5.7k cal BP),
wild pigs were transported, and possibly kept and bred, by humans in Japan (Yamazaki et
al., 2005). Though not yet domesticated, some sort of early management has been
suggested for the Okinawan Islands by at least 7.8 kya (7000 BP) (Matsui, 2005).
Considering the evidence, it is probable that pigs were domesticated first in the Central
Plains prior to 8.9 kya and then spread across Northeast China, Korea, and Japan, perhaps
being locally domesticated in one or more of these other regions.
Cattle
Cattle, sheep, and horses form the basis of nomadic pastoralist economies across Central
and East Asia. Cattle (Bos taurus) are exploited by both sedentary agriculturalists and
mobile pastoralists in East Asia. While water buffalo (Bubalus bubalis) and yak (Bos
[Poephagus] grunniens) were domesticated locally, B. taurus is often considered to be a
western domesticate. A regionally specific haplogroup among Northeast Asian cattle
indicates that local aurochs (Bos primigenius) were incorporated into the gene pool of
domestic cattle either through a separate domestication episode or incorporation of wild
female animals into existing herds (Mannen et al., 2004). In India, zebu (Bos indicus)
was domesticated separately from west Asian breeds before 7.5 kya (Loftus et al., 1994;
Meadow, 1996; Chen, S. et al., 2010), but frequencies of zebu haplotypes are low
amongst Northeast Asian cattle breeds (Mannen et al., 2004).
59
The earliest evidence of domestic cattle use in Northeast Asia is from the Zhongri
sites in Qinghai3 at about 5.6-4.0 kya (Flad et al., 2007) and Hongshan sites in Northeast
China at 5.7-5.4k cal yr BP (Guo, 1995a). The Tamsagbulag site in eastern Mongolia,
probably dating to sometime between 6.4-4.5k cal yr BP, also shows evidence of a
reliance on cattle, though no research has been conducted on the status of these
individuals as domesticated or wild. Similar dates can be assigned to the introduction of
domesticated cattle in the eastern steppes of Siberia, where Afanasievo groups (5.5-4.0
kya [3500-2000 BC]) raised cattle, sheep and horses (Frachetti, 2002).
Although it is likely that domesticated cattle were utilized in Northwest China by
5.0ka BP, the Qijia (Ch’i chia) culture (4.2-3.8ka – after Debaine-Francfort, 1995)
appears to have been the first in northwest China to regularly incorporate cattle, sheep,
and horses into their subsistence economy (Flad et al., 2007). In the Russian Far East,
cattle raising was a common element of subsistence beginning at the end of the Bronze
Age or the early Iron Age (3.0-2.0 kya), but may have begun as early as 8.5-6.7k cal yr
BP (Kuzmin, 1995, 1997). “Ox” and “water buffalo” bones have been found in Korean
sites dating to 5.0-4.7k cal yr BP, but their stratigraphic association and taxonomic
designations are unclear (see Nelson, 1993).
Cattle could have been used for traction as well as meat and milk. The possibility
of use in dairying is especially compelling in regions where cattle were most likely
adopted directly from the western Eurasia, where dairying was established and utilized
3
These remains are attributed to Bos sp., but it is not entirely clear that adequate faunal analysis was
conducted to verify attribution to B. taurus or B. (Poephagus) grunniens (see Olsen, 1990, 1991). This lack
of detail should be noted since Qinghai may have been a center of yak domestication (Rhode et al., 2007).
60
extensively as early as 8.0 kya (Evershed et al., 2008). Organic residue analysis of
ceramics should be employed at sites where cattle are reported in order to assess the
possibility of this function, particularly within the geographic limits of modern milk
consumption.
Ovicaprids
Neither sheep nor goats seem to have been domesticated locally in Northeast Asia.
Genetic evidence does indicate more than one domestication episode for both species.
Domesticated goats (Capra hircus) are thought to have had three separate geographic
areas of original domestication, one of which was in South Asia as genetic markers for
this episode are found only in Pakistan, India, Malaysia and Mongolia (Luikart et al.,
2001; MacHugh and Bradley, 2001). The importance of goats in South Asia is
underscored by their initial prominence in faunal assemblages from excavations at
Mehrgarh, along the North Kachi Plain in the Indus region (Meadow, 1996). Evidence of
goats in Northeast Asia is less common and they may have been introduced from the west
or south much later than other herd species.
Domesticated sheep (Ovis aries) show genetic evidence of domestication from
two separate wild subspecies (Wang et al., 1990; Wood and Phua, 1996; Hiendleder et
al., 1998; Hiendleder et al., 2002). One group is considered to be descended from the
mouflon (Ovis musimon/Ovis orientalis), with wild ancestors from the region of modern
Turkey and western Iran (O. orientalis anatolica and O. orientalis gmelini). The roots of
61
the other major domestic sheep branch (O. aries) are still unknown, but derivation from
either the wild urial of the Aralo-Caspian Basin (O. vignei bochariensis) or argali species
(O. ammon spp.) of Central Asia and Mongolia has been discounted (Hiendleder et al.,
1998; Hiendleder et al., 2002). Considering these lines of evidence, it can be inferred
that both sheep and goats were domesticated elsewhere and introduced to Northeast Asia.
Archaeological evidence supports an introduction of fully domesticated species.
The earliest reported evidence of sheep exploitation comes from Shaanxi Province at
Lingkoucun (7.3-6.2 kya) and from Banpo (6.9-5.8 kya), but these individuals may have
been from local wild populations (Flad et al., 2007). More likely candidates for the first
domesticated sheep, from Longshan sites, are older than 4.0 kya (Flad et al., 2007). It is
proposed that domestic sheep, along with cattle and horses, were part of a suite of
domesticates used by the agro-pastoralist Qijia peoples and may have been introduced
along with wheat sometime between 5.0-4.0 kya (Debaine-Francfort, 1995; Flad et al.,
2007); however, Hongshan sites in Northeast China contain evidence of cattle, sheep and
pig, dating to between 5.7-5.4k cal yr BP (Guo, 1995a). The status of Hongshan sheep is
not entirely clear, but such an early age for domesticated individuals might lend credence
to the domestic status of sheep at Lingkoucun and Banpo. Later sites of sheep
exploitation in Northeast China include the Lower Xiajiadian archaeological culture sites
(> 4.0 to about 3.5k cal yr BP), which are broadly contemporaneous with the Qijia in
northwest China and typically include faunal remains of dogs, pigs, cattle, sheep and deer
as well as broomcorn and foxtail millet (Guo, 1995b).
62
By 3.3k cal yr BP, sheep were of special economic importance for both sedentary
agriculturalists and nomadic pastoralists in Northeast Asia. Among early Bronze Age
pastoralist sites in the steppe and mountain zones of southeastern Siberia and Mongolia,
sheep remains are sometimes found in association with horse bones in burial or
ceremonial contexts. Horses and sheep have complementary grazing patterns
(Honeychurch and Amartuvshin, 2006) and were probably introduced as part of an
established herding strategy that developed farther west in Central Asia. The earliest
evidence for the pastoralist suite of animals in Northeast Asia comes from Afanasievo
burial remains (5.5-4.0 kya) in the eastern steppes of Siberia (Frachetti, 2002).
Horses
While sheep are characteristic of pastoralist and agro-pastoralist communities in
Northeast Asia, the enhanced mobility that horses (Equus caballus caballus) enabled was
integral to the development of nomadic pastoralist economies across Eurasia during the
later Bronze Age. Wild horses (E. ferus spp.) were widely distributed across the Eurasian
steppe during the Late Pleistocene, but disappeared from the fossil record of many
regions by about 10.0 kya (Bennett and Hoffman, 1999; Vilà et al., 2001). Domestication
occurred later than that of the other species described here and in further contrast, genetic
analysis indicates that domestication from different wild herds occurred numerous times
(Vilà et al., 2001). Dispersal of techniques in capture and taming could have led to the
spread of horse domestication as a practice, as opposed to the spread of pre-domesticated
individuals (Vilà et al., 2001; Jansen et al., 2002).
Archaeological evidence for local
63
domestication episodes at both Botai (Kazakhstan) and Dereivka (Ukraine) date primary
domestication to about 6.5-5.5 kya (Anthony and Brown, 1991, 2000; Levine, 1999a,
1999b; Olsen, 2003; Outram et al., 2009).
Evidence of domesticated horses in Northeast Asia is restricted until the early
Bronze Age. Occasional finds of horse remains have been made in Neolithic contexts
dating to between 5.0-3.7 kya, but are within the natural range of wild species and cannot
be definitively attributed to domestic animals. Horse remains at Qijia sites in northwest
China date to about 3.7 kya and provide some of the first secure evidence of possible
horse domestication in Northeast Asia (Yuan and Flad, 2005). By 3.4-3.0 kya, horse
breeding was an important activity for Shang aristocracy in China and included symbolic
participation by the king (Yuan and Flad, 2003). Horses were a common domesticate in
northern China by 2.5 kya (Yuan and Flad, 2005; Flad et al., 2007). In Mongolia and
southern Siberia, horse remains were often incorporated into early Bronze Age
monuments called khirigsuurs (by at least 3.2-2.8 kya) (Fitzhugh, 2009). These
monuments are often associated with the first incursion of nomadic pastoralists into the
region (Honeychurch and Amartuvshin, 2006). Similarly, horse bones were included,
along with other domestic animals, in Bronze Age burials (< 3.0 kya) from the Song Nen
(Songjiang-Nenjiang) Plain region of Heilongjiang Province in Northeast China (Tan et
al., 1995b). Horses were not introduced until much later in Korea and Japan and there is
no direct evidence in those regions of domesticated horses in the Neolithic or early
Bronze Age.
64
There is earlier evidence for the exploitation of equids in early Holocene Gobi
Desert archaeological sites, but may relate to hunting (see artefact lists in Fairservis,
1993). Of the few faunal remains recovered from sites in this desert region, it is notable
that the vast majority are equids (presumably E. ferus przewalski or E. hemionus
hemionus as both species were widespread in the Gobi Desert until the late Holocene).
Despite this, there is no evidence that the animals were domesticated. Hunting of equids
as a key food source in the early Holocene would not be surprising since horse bones are
also common in Pleistocene Chinese sites (Han and Xu, 1985; Deng, 2005; Yuan and
Flad, 2005). As such, we might expect that Holocene hunter-gatherers regularly hunted
local wild equids, although there is no reason to reject the possibility that some
individuals from later sites were domesticated.
Camels
Bactrian or two-humped camels have been important herd animals throughout arid
regions of Central, South, and East Asia. The wild ancestor of Camelus bactrianus is C.
ferus, which was once distributed from the great bend of the Yellow River in China,
through Mongolia and central Kazakhstan. Two-humped camels are well-represented in
the rock art of the Altai, Tul-Kun, Tamurasche, Uryankhai, Turgai, and Minusinsk
regions between Inner Asia and Siberia (Peters and von den Driesch, 1997; Potts, 2004).
Although two-humped camels are thought to have been first domesticated in the Gobi
Desert, phylogenetic analysis of eighteen domestic and three wild two-humped camels
indicates that the existing wild individuals from northwestern China and southwestern
65
Mongolia do not share a recent common ancestor with domesticated breeds (Ji et al.,
2009). The same study suggests a single domestication episode for the Bactrian camel.
A clearer understanding of Pleistocene and early Holocene East Asian C. ferus
distribution and phylogeny is required in order to gain a better sense of this singular
event, including the relationship between modern wild camels and the subspecies
ancestral to domesticated breeds.
It is more difficult to ascertain the timing of camel domestication than for other
domestic species because there are currently no osteological criteria for distinguishing
domestic from wild animals (Olsen, 1988). As such, domesticated status is usually
accorded to animal remains outside the natural home range (Peters and von den Driesch,
1997). The most convincing early evidence of camel domestication is older than 5.0 kya
and comes from the Kopet Dagh foothills of southern Turkmenistan. Not only were
camel bones identified, but clay figurines of two-humped camels were also present and
even the earliest levels contained some clay representations of camels harnessed to carts
(Masson and Sarianidi, 1972: 109; Peters and von den Driesch, 1997). There are also
claims for the use of camel by peoples of the Indus Valley civilization (about 4.0 kya)
and artistic representations of two-humped camels were found in early Pakistani and
Indian sites dating from 3.0-2.0 kya (Köhler-Rollefson, 1996; Peters and von den
Driesch, 1997). Andronovo peoples in Xinjiang are also known to have bred camels and
contemporary sites suggest that their use of domesticated camel dates to about 4.0k cal yr
BP (Kuzmina and Mallory, 2007: 252). By 3.0 kya Bactrian camels are found in
66
archaeological contexts from northern China to Bactria (northern Afghanistan/southern
Uzbekistan), and were even more widespread by 2.5 kya (Olsen, 1988; Potts, 2004).
Since these regions bear no evidence for the Holocene existence of C. ferus in
either local palaeontological or archaeological contexts, it seems that domestication must
have occurred earlier farther east (Peters and von den Driesch, 1997). Although an
ancestral relationship between existing wild populations and domesticated Bactrian
camels remains unproven, prehistoric zoogeographic distribution of wild camels have led
most researchers to conclude that camels were domesticated in the desert regions of
southern Mongolia and northern China. Considering the earliest evidence for use of
domestic camels in Turkmenistan, one may suggest that camels were first domesticated
in the arid regions of southern Northeast Asia, perhaps Xinjiang, by 5.5kya. In this case,
it is notable that evidence of the animal appears in Turkmenistan before other parts of
Northeast Asia. This could indicate some level of reciprocal contact between desert
groups in Central Asia at a very early stage in prehistory. Conversely, it could be
suggested that C. ferus existed farther west than current evidence suggests.
67
Animal
Region
Dog
North China
Northeast China
Korea
Japan
Siberia
Mongolia
North China
Northeast China
Korea
Japan
Siberia
Mongolia
North China
Northeast China
Korea
Japan
Siberia
Mongolia
North China
Northeast China
Korea
Japan
Siberia
Mongolia
North China
Northeast China
Korea
Japan
Siberia
Mongolia
North China
Northeast China
Korea
Japan
Siberia
Mongolia
Pig
Cattle (B. taurus)
Sheep
Horse
Camel (C. bactrianus)
Earliest Evidence (k cal yr
BP)
12.9
3.2-3.1
7.5-4.7
10.5
11.9, 4.5-2.5 (Primorye)
Unknown
12.9-8.9
7.2
8.0-2.0
7.0-5.7?, after 2.5
4.5-2.5 (Primorye)
Unknown
5.6-4.0
5.7-5.4
5.0-4.7?
after 2.5
5.5-4.0 (Eastern steppe)
6.4-4.5?
7.3- 4.0
5.7-5.4
Historic
after 2.5
5.5-4.0, Historic
by 3.0
3.7
3.0
after 2.5
after 2.5
5.5-4.0 (Eastern steppe)
4.5-3.5
4.0, 3.0
2.9
N/A
N/A
4.0
before 4.0?
Table 2.1 Earliest evidence for use of major domesticated animal species in Northeast
Asia.
68
2.2.3. Animal domestication in the Gobi Desert
Regional suites of economically important domesticated animals in Neolithic Northeast
Asia were derived from a variety of contexts and should be considered to reflect local
ecology, community needs, and interaction spheres. The types of animals domesticated
within a society indicate several things, including the suitability of each species for
domestication, a pre-existing relationship between local people and the wild progenitor,
and the needs of the group. Pig breeding is most closely associated with higher levels of
sedentism or village communities. Unlike sheep, goats, cattle, and horses, pigs are not
suited to mobile pastoralism, but can be easily contained and will do well feeding on the
concentrated refuse produced by sedentary communities. On the other hand, animals
associated with nomadic pastoralism are grassland-adapted and need ample pasturage.
While sedentary humans with plenty of access to grazing land can quite successfully
manage small herds of these species, especially if winter fodder is grown, large herds are
less labour intensive for mobile people who can move with the animals to find seasonal
pasture and offer regular protection from predators. Considering the sedentary nature of
many early communities that exploited domesticates it is not surprising that herd animals
are rare in Northeast Asia prior to the spread of nomadic pastoralism.
When we consider possible early domesticates in the Gobi Desert, it is important
to consider local plant and animal communities, as well as existing patterns of land-use.
High residential mobility and a varying seasonal focus on upland plateaux and lowland
dune-field/wetland environments are suggested for early to middle Holocene huntergatherers in arid Northeast Asia (Janz, 2006; Bettinger et al., 2007). Equids and
69
ovicaprines were common species across the Gobi Desert and wild camel would also
have been included in the local fauna of the western regions. Wild boars were probably
present in parts of the East Gobi, as well as wild millet, but the lack of evidence for
sedentism makes early domestication of such species unlikely. Geographical proximity
to Afanasievo stock-raising peoples west of the Gobi Desert makes the possibility of
hunter-gatherer adoption of species like sheep, cattle, and horses intriguing, but
compatibility with existing land-use and subsistence strategies must also be considered.
Domestication can be seen as a type of technological change. As with any
technological change, the structure of a society must be modified in order to
accommodate new elements. If dune-field/wetland environments were consistently
productive it seems less likely that hunter-gatherers would adapt organizational strategies
to incorporate sources of subsistence that would require vigilant protection from
predators and constant access to forage. In this case, people would have needed to accept
major organizational changes, such as devoting less time to other regular tasks, in order
to benefit from herd-raising (see Lechtman, 1977; Lemonnier, 1992; Schiffer, 1992,
2001). This implies that certain tasks and benefits would be differentially valued in
various societies (Schiffer, 2005). The perception of herd animals as a luxury item, or the
security of having regular access to meat and/or milk are two types of benefits that might
have justified the adoption of smaller herd animals like sheep or goats.
Consequently, if high residential mobility was typical of the society and
consistent with the needs of proto-domesticates, it would have been easier to incorporate
them into existing economic regimes. High residential mobility would also have
70
contributed to perceived benefits for the use of animals like horses and camels, which
would have made long-distance moves easier (Hämäläinen, 2003). Established patterns
of high residential mobility would have reduced new pressures associated with constantly
providing new forage, particularly if traditional moves coincided with localized seasonal
appearance of pasture. Ease in following and hunting highly mobile prey species would
also have benefited hunter-gatherers. In order to assess the relationship between Gobi
Desert hunter-gatherers and the emergence of pastoralist economies in Northeast Asia, it
is necessary to examine the chronology of such developments in the region.
2.2.4. Chronology of nomadic pastoralism in Northeast Asia
Pastoralism is considered to have evolved first in the Near East. Here, agricultural
communities of fully or semi‐sedentary farmers began managing wild herds of animals,
such as sheep and goats, in a way that produced long‐term changes in behaviour and
morphology (Garrod et al., 1996; Hole, 1996; Zeder, 2001, 2005). Similarly, herding
practices in East Asia are also thought to have originated in agricultural villages (Chang,
1987). Scholars in both regions have speculated that an increased focus on herding arose
when agricultural populations expanded and marginal farmers were forced to adopt
strategies that allowed them to survive in more arid or colder environments (Christian,
1998; Chang, 1987; Garrod et al., 1996).
Pastoralism as an independent subsistence economy was originally thought to
have become possible only with the “Secondary Products Revolution,” when the use of
products like wool and milk allowed herders to be more independent of agriculturalists
71
(Krader, 1959; Sherratt, 1983). On the Eurasian steppes, pastoralists appear to have
become more mobile by about 5.5k cal yr BP. Considering that horses were being
domesticated between 6.5-5.5 kya, it is possible that the domestication of this species
contributed greatly to the potential mobility of pastoralists; however, Kohl (2007: 137144) argues that the introduction of heavy wheeled carts and wagons pulled by oxen was
associated first with increased mobility. Ox and carts were probably widely used by 5.0
kya (end of the fourth millennium BC). This theory deserves careful consideration since
the earliest evidences for horse domestication and riding was localized in only two
centres – the Ukraine and Kazakhstan. Both carts and horses would have facilitated
mobility, but in much of Central Asia cattle were the primary focus of pastoralist
activities in the early period.
Milk and wool are also important to the modern Eurasian pastoralist economy and
presumably would also have been regularly exploited. Despite this, it is unlikely that the
development of dairying played a pivotal role in allowing herders the freedom to become
less sedentary as organic residue analysis of pottery vessels has recently revealed that
milk was used early as 8.0 kya (seventh millennium BC) (Evershed et al., 2008). It is not
known when wool production began, but wool from sheep would have provided access to
an extremely versatile raw material that has since proven central to mobile pastoralist
economies in the Eurasian steppes, including in the construction of portable houses.
Large scale production of woollen textiles is known to have been practiced in
Mesopotamia by about 5.0 kya (late fourth or third millennium BC). Woollen rugs and
72
carpets have also been found in Early Bronze Age burial monuments (Kohl, 2007: 164166).
Nevertheless, the use of wool rather than bast fibres (e.g., linen, nettle, hemp) and
the development of wagons, which were roughly contemporaneous innovations, do not
explain why pastoralism became viable enough for a widespread shift in economic focus.
The emergence of arsenical copper/bronze metallurgy might also be an important
motivating force since primary access to the highly localized raw materials, as well as
distribution, would have been easier for those with improved forms of transportation.
Early Bronze Age metal artefacts from the western Eurasian steppes were probably
produced locally and with the most basic techniques on ingots from the Caucasus (Kohl,
2007:166-180). The role of developing trade networks as an aspect of the “Secondary
Products Revolution,” for controlling trade of both bronze and “secondary” pastoralist
products, has not been investigated.
Domesticated herbivores such as cattle and sheep predated the emergence of
mobile pastoralism. Prior to the institutionalization of nomadic pastoralism, subsistence
strategies were fairly flexible as peoples on the western Eurasian steppes were known to
have relied on fishing and hunting, as well as stockbreeding and agriculture (Christian,
1998; Benecke and von den Dreisch, 2003). At the beginning of the Eurasian Bronze
Age (5.5-5.0k cal yr BP), evidence of sedentary settlements disappear and raised burial
monuments are common markers of the new nomadic pastoralist society (Kohl, 2007:
144-145). In some regions, settlements do become more common again in later periods.
73
Evidence for intensive gathering of wild plant foods such as goosefoot
(Chenopodium) and amaranth (Amaranthus) was found in Late Bronze Age levels in the
Samara Valley, along the middle Volga River in western Russia. This system of
subsistence was later augmented with barley (Hordeum vulgare), wheat (Triticum
dicoccum, T. compactum), and millet (Panicum miliaceum) in the final stages of the
Bronze Age (Kohl, 2007: 157). Cereal grains were probably either acquired through
trade or locally cultivated, but this aspect of subsistence is not well studied. Hunting and
fishing remained important economic activities that contributed to a subsistence base now
including pastoralist products (Frachetti, 2008: 46). Neighbouring hunter-gatherers
groups apparently continued to rely primarily on various species of deer, which were
probably also of economic importance to early pastoralist communities who maintained
flexible subsistence strategies that included hunting, fishing, and gathering (Kislenko and
Tatarintseva, 1999; Honeychurch and Amartuvshin, 2006). In general, subsistence
economies were mixed and varied from region to region, with a seemingly greater
reliance on exclusive pastoralism in Kazakhstan and western Siberia. Considering the
continued importance of wild foods, it seems that most early pastoralist economies on the
Eurasian steppes were at least as flexible their predecessors. Distinguishing markers of
early pastoralists are their increased residential mobility and the incorporation of
domesticated species.
Pastoral strategies in the Early Bronze Age were dominated by the use of cattle,
but also included sheep, goats and horses. Communities in the western Eurasian steppe
are generally assigned to the “Yamnaya culture,” while the term “Afanasievo culture” is
74
reserved for those groups in the eastern steppes of Kazakhstan and southwestern Siberia.
Similarities in material culture and the “Caucasian” physiological type of both groups
have been used as evidence to support the eastward migration of newly mobile western
pastoralists (Kuzmina, 1998; Hemphill and Mallory, 2004). However, calibrated
radiocarbon dates of Afanasievo material tend to be slightly earlier than those of similar
cultures in the western steppe (Rassamakin, 1999).
The predominance of horse remains in the Botai “culture” sites in the north
central steppes of Kazakhstan (99% of fauna) is markedly different from the Yamnaya
and Afanasievo sites where the most common large-bodied domesticated species are
cattle (Frachetti, 2008: 47). Botai peoples used horses for dairy, but are thought to have
primarily relied on domesticated horses in order to aid them in hunting herds of wild
horses (Olsen, 2003; Outram et al., 2009). This pattern of exploitation shifted rapidly to
one more similar to neighbouring areas around 4.5k cal yr BP (2500 BC), when the
exploitation of cattle, sheep and goats overtook that of horses (Benecke and von den
Dreisch, 2003).
The Afanasievo (5.5-4.5 kya) are generally recognized as the first group of
pastoralists in both Northeast Asia and more broadly across the Eurasian steppes. Central
areas of interaction included the Yenisei River Valley and the Minusinsk Basin
(Frachetti, 2002). Subsistence was likely not based on herd animals alone, but also
included hunted animals, gathered plants and fish (Frachetti, 2002). Subtle changes in
ceramic form and burial are indicative of a change in local communities, including a
more complex material culture, and are considered to mark the emergence of the Okunev
75
peoples (4.6-4.0 kya), who were physiologically more similar to Asiatic peoples
(Gryaznov, 1969; Mallory, 1989; Hemphill and Mallory, 2004). This physiological
change may be indicative of the incorporation of more eastern populations into pastoralist
communities and speaks to the increasing influence of locally evolving pastoralist
communities in Northeast Asia.
From about 4.5-3.0k cal yr BP, the nomadic pastoralist peoples who inhabited the
mountain-steppe zone just west of the Gobi Desert are archaeologically recognized by
their exploitation of domestic animals (particularly cattle and sheep), metallurgy, and
distinct material culture, all of which are reflected in their burial complexes and rock art.
The term “Andronovo” (4.0-3.3 kya) has been widely applied to such archaeological
complexes, which post-date Afanasievo and Okunev deposits. In addition to the previous
range of domesticated animals, Andronovo groups were also known to engage in camel
husbandry (Kuzmina and Mallory, 2007). Settlement data from the Dzhungar Mountains
in southeastern Kazakhstan (near the border of China and bounded by the Altai
Mountains; see Figure 2.1) suggest seasonal occupation of lowland and midland zones
along ravines and south-facing cliffs. The occurrence of rectilinear, semi-subterranean
houses with substantial stone foundations, as well as small camps with similar types of
house forms, suggest a pattern of long-term continuity and seasonal reoccupation of the
same environments (Frachetti, 2008: 132). It has been suggested that these groups
planted seed crops around their winter settlements and returned to harvest them in the
fall, but it is also clear that pastoralists might have more easily taken advantage of the
diversity of wild plants available across the array of local biomes that they inhabited
76
seasonally (Gryaznov, 1969; Frachetti, 2008). This pattern of structured and invested
seasonal habitation is quite different from that recognized at contemporaneous huntergatherer habitations in Mongolia and China.
The Karasuk period (3.3-3.0 kya) follows the Andronovo in the Minusinsk Basin
and is contemporaneous with developments in the Late Bronze Age of Mongolia (Houle,
2010). Sheep may have been of more economic importance than cattle at this time, and
their different pasturage requirements could suggest an increase in transhumance. Horse
bones also become more frequent. Horses may have been most important as a means of
transportation, as there is clear evidence of horse-back riding (Legrand, 2006). Burial
and habitation sites are more widespread than in the preceding period and it is suggested
that population increased. Metallurgy became progressively more developed and
important, eventually culminating in the Iron Age Tagar period (Legrand, 2006). In
northern Mongolia, many late Bronze Age ritual and burial monuments have been dated
to between about 3.2 and 2.8k cal yr BP (Fitzhugh, 2009). These monuments include
deer imagery and intensive ritual use of horses (rather than sheep). Such a pattern is
divergent from the Karasuk culture and suggests that nearby Bronze Age groups in
northern Mongolia were developing their own traditions.
Early agropastoralist communities in northwestern China are contemporaneous
with Afanasievo settlements. By this time, sedentary groups using agricultural tools and
domestic species had settled along watercourses in the more humid and temperate climate
zones. Most early cultivators were ancestors of later agricultural specialists, though at
this early period they practiced a more mixed economy of herding, hunting, gathering,
77
and farming. Little is known about subsistence economies in the arid Gobi Desert during
this period. Based on the occurrence of grinding stones and macrotools, some scholars
have suggested that cultivation may have been practiced around dune-fields and shallow
lake basins (Cybiktarov, 2002), but beyond the southeastern-most regions the low
frequency of such artefacts is inconsistent with intensive cultivation and processing.
It is commonly thought that nomadic pastoralism was introduced to East Asia by
migrations from Central Asia, but agrarian cultures in western China may have engaged
in herding practices prior to the florescence of nearby nomadic pastoralist cultures
(Mallory, 1989; An, 1992a; Christian, 1998; Flad et al., 2007). Cultural connections
between agricultural China and northern pastoralist groups during the late Neolithic and
early Bronze Age have been suggested (Jacobson, 1988; Li, 2002). Earlier direct
interactions are unlikely, as there is little evidence of pastoralist activity in Xinjiang,
Gansu, and southern Mongolia, until after 4.5k cal yr BP.
The best known early pastoralists from northern China are copper and bronzeusing groups, known as the Qijia, who practiced both agriculture and herding.
Physiologically Asiatic peoples, the Qijia were centred in eastern and central Gansu, but
were also present in eastern Qinghai, southern Ningxia, and western Shaanxi (DebaineFrancfort, 1995: 13). Dates on Qijia sites fall in the period of roughly 4.8-3.8k cal yr BP
(4260 + 80 at Changshan, 3555 + 95 at Dahezhuang) and the sites often overlie those of
the Late Neolithic Majiayao (Machiayao) type, or in some regions appear to emerge from
other local predecessors (Debaine-Francfort, 1995: 362-367). The Qijia are characterized
by the presence of new western domesticates and increased social stratification (An,
78
1992b; Liu, 2004). Hemp was used for textiles and people subsisted on a variety of
western and eastern domesticates. Plants included foxtail millet (Setarica italica), wheat
(Triticum sp.), barley, and rye (Secale montanum). Pigs were of primary importance,
followed by cattle, sheep, and goats. Horses and dogs were also raised. Deer were still
an important hunted species (Debaine-Francfort, 1995; Flad et al., 2007). In succeeding
cultures, sheep herding gained primary dominance amongst agropastoral people in the
Gansu region (Debaine-Francfort, 1995: 347). Developments in pottery production,
metallurgy, and social stratification at Qijia sites parallel developments in the Central
Plains, but bear distinct impressions of unique trajectories that are more closely aligned
with Central Asian pastoralist groups.
Qijia peoples appear to have had contact with those groups associated with the
Longshan (5.0-4.0 kya) culture further east on the Central Plains. Similarities in tool and
ceramic traditions with the Shaanxi variant of Longshan culture, as typified by
Kexingzhuang II, have been interpreted to show this society as ancestral to the Qijia.
Newer radiocarbon dates for Longshan and Qijia type sites now indicate that this
similarity is more likely the result of interaction and trade between the groups settled on
the periphery of their respective culture zones (Debaine-Francfort, 1995: 327-328). The
Longshan were highly stratified, institutionalized agriculturalists, exhibiting evidence of
intense intraregional conflict (Shao, 2000). Settlements were clustered in centers of
abundant arable land with settlement focused along rivers (Liu and Chen, 2006).
Although copper and bronze were an important aspect of Qijia culture, metal artefacts are
limited in the Longshan archaeological record and were not treated as prestige items until
79
the emergence of the later Erlitou and Shang entities (Shao, 2000). Following the Qijia
period, contacts with eastern China continued. Debaine-Francfort (1995: 347-348)
contends that an increased focus on pastoralism and affiliations with outlying pastoralist
societies dramatically shifted the trajectory of cultures between northwestern China and
the Central Plains.
As Gobi Desert sites north of the Qijia range have yet to be clearly interpreted, it
is not known what type of interaction these horse-riding agropastoralists would have had
with hunter-gatherers groups to the north. Likewise, the relationship between
contemporary Neolithic agriculturalists and hunter-gatherers in the East Gobi has been
largely ignored. Only after the rise of horse-centred nomadic pastoralism is the
relationship between agricultural China and less sedentary neighbours addressed in the
literature. By that time, Gobi Desert groups almost certainly had knowledge of their
neighbours and may have engaged in trade, possibly exporting such products as furs and
stone. By at least 4.0k cal yr BP, the full suite of both western and eastern domesticates
(perhaps with the exception of chickens) would have been familiar to inhabitants of
southern mainland Northeast Asia. By no later than 3.0k cal yr BP, nomadic pastoralism
was well established across the Gobi Desert.
One notable trend among early pastoralists in northern China is the incorporation
of pastoral strategies into established economies already reliant on domesticated animals
and cultivation. Although modern communities in Northeast China are traditionally
associated with nomadic pastoralism, prehistoric groups were practicing a very different
type of production economy prior to the introduction of nomadic pastoralism. Nomadic
80
pastoralism entered the region quite late, perhaps as a consequence of increasing political
and military power of pastoralist communities on the Northeast Asian steppes. This
increased military power, which became especially notable by 2.5 kya, was related to
both the military advantage of an expert equestrian fighting force and the political
consolidation of steppic groups.
Later Bronze Age mortuary complexes, artworks and monuments in Northeast
Asia appear to have evolved locally (Jacobson, 1988, 2002), but share numerous
attributes with their contemporaries farther west (Christian, 1998). In China, Central
Asian pastoralists may have mixed with agricultural “barbarians” such as the Qijia, who
were probably in frequent contact with people practicing irrigation agriculture along the
fertile Yellow River valley and its tributaries (Chang 1987; Christian 1998: 106). This
relationship became increasingly fraught with tension during the rise of military powers
within both China and arid Northeast Asia. We can expect that due to the mobility and
low population density of Gobi Desert hunter-gatherers, their relationship with Neolithic
and early Bronze Age Chinese agriculturalists was different than that developing between
agriculturalists and agropastoralists along the Yellow River and its western tributaries.
Nevertheless, it is clear that with the rise of equestrianism, increasing contact, and
increasing population densities, mobile communities throughout the northern and central
reaches of mainland Northeast Asia were steadily more at odds with urban, agriculture
China.
81
2.3. Current knowledge of the Neolithic transition in Mongolia
Although we are making inroads into understanding of the complexity of relationships
between early agricultural, pastoral, and hunter-gatherer groups in the steppes of
prehistoric Northeast Asia, our knowledge remains rudimentary. And, athough we are
gaining a deeper understanding of individual societies and economies across mainland
Northeast Asia (e.g., Frachetti, 2008; Weber et al. [eds.], 2010), the relationship between
these diverse groups remains ambiguous. Positioning Mongolian prehistory with the
larger framework of contemporary groups is even more difficult due to a lack of research
in this time period. Archaeological studies in the country focus on the more visually
impressive remains of nomadic pastoralist groups, particularly those most closely
associated with modern Mongolian identity (e.g., Xiongnu [Hunnu], Mongol Empire).
Knowledge of the unique trajectory of early Mongolian prehistory would almost certainly
confirm the importance of this geographic region which was central in the development
of Northeast, Central, and East Asian cultures since the earliest record of human
habitation. Numerous post-glacial assemblages both remaining in situ and known from
museum and university collections are largely unpublished since their discoveries. This
situation is unfortunate, since trajectories of development and interaction would
undoubtedly be revealing considering that the Gobi Desert exists as a transitional zone,
separating both western Central Asia from developed agricultural and later pastoralist
communities in Northeast China, and sedentary agriculturalists in southern Northeast
Asia from dedicated nomadic pastoralists in the north.
82
As noted in the introduction, our knowledge of Mongolian and Gobi Desert
(including parts of the Inner Mongolia Autonomous Region, PRC) prehistory comes
largely from two sources – the collections of major early 20th century scientific
expeditions and joint Soviet-Mongolian expeditions. Following the dissolution of the
Union of Soviet Socialist Republics (USSR) in 1991, foreign researchers became more
involved in archaeological research in Mongolia. One of the most notable of these early
collaborations is the Joint Mongolian-Russian-American Archaeological Expedition
(JMRAAE) (Derevianko et al., 1996, 1998, 2000).
A diverse array of Stone Age assemblages have been documented from across
Mongolia, dating from the first hominid occupations of the region to the early Bronze
Age. Dating of the early sites is not well constrained, although it is probable that the
earliest ones in Mongolia are as old as or older than those in Siberia since the two regions
are very similar environmentally and topographically along their common borders.
However, the current study is concerned only with post-LGM sites.
Assemblages containing microblade core technology are often considered to date
to the post-glacial period, though it is now clear that such technology, in the form of
wedge-shaped cores, began as early as the beginning of MIS 2 (ca. 25.0-11.5k cal yr BP)
(Gladyshev et al., 2010). The post-LGM period is typified by the use of percussion and
pressure-flaked microblade cores in a variety of prismatic and sub-prismatic forms, along
with tools made from retouched microblades. The existing literature mainly contains
descriptive summaries of artefacts associated with different chronological periods;
83
however, very few Mongolian sites are dated, making it difficult to trace the region’s
developmental trajectory.
2.3.1. Post-LGM settlement and subsistence in Mongolia
Nels C. Nelson was the first archaeologist to discover, collect, and publish prehistoric
assemblages from Mongolia. His publications primarily outline his findings and draw
attention to the fact that many of the finds were made in dune-field environments (Berkey
and Nelson, 1926; Nelson, 1926a, 1926b, 1939). Nelson did indicate that these extensive
wind-blown sediments appeared to have accumulated under more humid conditions,
when “living” dune-fields were gradually being enlarged from the lee shores of filled
lake basins (Nelson, 1926a). He states that deposits from the primary find site of
Shabarakh-usu (Bayan-dzak) were blown from the walls of adjoining promontory
escarpments within the dune-fields, where flints, pottery and bone occurred in “distinctly
stratified order” (Nelson, 1926a: 251). Since Nelson’s time, continuing archaeological
work has shown that dune-fields are the primary source for Gobi Desert Neolithic-type
assemblages (containing pottery and ground-stone), suggesting that a specific adaptation
to such environments was an important aspect of Neolithic land-use.
Johannes Maringer
The first archaeologist to seriously address the issue of land-use and site location in the
Gobi Desert was Johannes Maringer (1963), who analysed and catalogued the artefacts
84
recovered by Folke Bergman during the Sino-Swedish Expeditions in Inner Mongolia
(Maringer, 1950). He also drew on Nels C. Nelson’s publications and in 1956 he
examined the Central Asiatic Expedition collections housed at the American Museum of
Natural History. He also studied collections in Denmark made by H. HaslundChristensen in Inner Mongolia (Chahar) (Jacobsen, 1940), and consulted Elisseeff’s
(1950) summary of materials from Altan Bulag (Altan Boulaq/Altan Bulaq) along the
Mongolian-Siberian border, as well as Okladnikov’s (1951) monograph on the Soviet
Kiselev Expeditions. Maringer (1963) asserted that, as previously suggested by Nelson
(1926b), Neolithic groups identified by the use of pottery, polished stone tools, grinding
stones, and arrowheads expanded beyond the geographic range of their ancestral
Mesolithic predecessors, covering a greater expanse of steppe and desert areas.
Drawing connections between changes in human land-use and climate change, he
recognized that the early Holocene of Mongolia would have offered very different
environments than in modern times. He pointed out that following the LGM, with the
melting of glaciers, the Gobi Desert would have been much wetter and that all of
Mongolia would have been typified by richer vegetation and the infilling of lowland
basins and river beds. He hypothesized that by the time of the Mesolithic, aridity was
again increasing and people were forced into shrinking areas of remaining water, where
they relied heavily on fishing. The decline of large game would have forced them to rely
on small and medium-sized animals, perhaps leading to the invention of the bow and
arrow, which allowed them to hunt more mobile animals and birds. Neolithic habitations
85
were mostly found around springs, lakes, playas, and river beds, suggesting that the
period was typified by the use of loosely connected oases.
Maringer saw no evidence for agriculture among Neolithic peoples in Mongolia,
though in some regions they may have been influenced by agricultural communities in
China, perhaps through trade. He believed the absence of other agricultural tools
indicated that grinding stones were probably used to grind wild varieties of grain. While
Egami and Mizuno (1935: 61) had suggested that Stone Age peoples along the ChinaMongolia border were involved in the herding of horses, camels, cattle, and sheep, and
Teilhard de Chardin and Pei (1944: 38) interpreted material remains as suggestive of
agricultural endeavours, Maringer rejected such possibilities for the Stone Age of
Mongolia proper. He also pointed out that the few finds of bronze artefacts in these
regions are insignificant and that Mongolia seems to have maintained lithic technology
until the beginning of the Iron Age around 200-100 BCE (2.2-2.1 kya) when a “separate
cultural and ethnic stratum” began to move into the region. As such, Maringer saw the
Mongolian hunter-gatherers as a completely separate entity from the pastoralist peoples
who later inhabited the region and displaced them.
Maringer’s interest in the relationship between Mongolian hunter-gatherers and
adjacent cultures is evident in his discussion of the origins of Mongolian Stone Age
peoples and their contacts with contemporary neighbouring groups. Based on
comparisons with assemblages described from Altan Bulag (Elisseeff, 1950) and material
from the Lake Baikal region in Siberia (Maringer, 1950: 183-185), he linked original
Palaeolithic hunters in Mongolia with the cultures of southern Siberia. Although he
86
noted similarities between Palaeolithic assemblages on the Ordos plain and at
Zhoukoudian (Choukoutien), he saw those as derived from southward-moving groups of
Palaeolithic hunters (see also Maringer, 1950). Also noted is Okladnikov’s assertion
(1951, 1962) that Shabarakh-usu pottery bore remarkable similarities to the Isakovo and
Serovo ceramics from the Lake Baikal region (ca. 6.2k cal yr BP, after Weber, 1995). As
such, Maringer saw Mongolian and Gobi Desert groups as closely allied with their
northern neighbours throughout prehistory. He believed that contact between sedentary
agriculturalists and northern hunter-gatherers was also supported by finds of “northern
Chinese type tools and shards of painted pottery” in the steppe and desert regions of
Mongolia (see also Okladnikov, 1962; Cybiktarov, 2002), but did not explore the nature
of such hypothetical connections.
Mongolian and Russian research (1940s to present)
Later work of Soviet and Mongolian archaeologists largely supported and expanded
Maringer’s original hypotheses about Mesolithic and Neolithic inhabitants of Mongolia.
The relationship between cultures of the Lake Baikal region and Mongolia were
emphasized and seemed to researchers strongest at sites in eastern Mongolia (Larichev,
1962; Okladnikov and Derevianko, 1970; Dorj and Derevianko, 1970; Dorj, 1971).
Overall, local development was emphasized, but the relationship between the material
culture of Mongolia and the Gobi Desert and that of Manchuria (Northeast China) and
Siberia was thought to indicate an important and long-term developmental relationship
between these regions.
87
Finds made by both Soviet and Mongolian archaeologists later in the twentieth
century added great depth to an understanding of the diversity of adaptations existing in
prehistoric Mongolia, particularly at the end of the Neolithic and beginning of the Metal
Ages. The most striking contradiction to Maringer’s model is the discovery of a very rich
Bronze Age horizon. Burial and ceremonial monuments link Mongolia with
contemporaneous developments in greater Central Asia, but suggest a local evolution of
both Bronze and Iron Age cultures (see summaries in Cybiktarov, 2002, 2003). The
Mongolian Mesolithic (post-LGM aceramic horizon) was thought to have continued until
about 6.0-5.0 kya (fifth or fourth millennium BC) (Derevianko and Dorj, 1992: 171, 172)
with dates presumably based on inferences from stratigraphy and contiguous
developments. The Neolithic, which marked the introduction of pottery, followed the
Mesolithic and was associated with population expansion.
Several phases of development are now thought to characterize the Neolithic. By
the middle Neolithic, agriculture is believed to have become an important subsistence
strategy in eastern and southern Mongolia, complemented by hunting and fishing. The
Neolithic-to-Bronze Age transition or the early Bronze Age is thought to date to between
about 4.0-3.0 kya and was typified by more temporary campsites than in the middle
Neolithic. Tool kits included the increased frequency of bifacially retouched tools,
primarily of chalcedony, including finely pressure-flaked projectile points. Flake tools
such as scrapers and microblade inserts are still common in such sites and copper slag is
occasionally found. This pattern, though identified at sites in north-eastern Mongolia, is
also considered to be typical of sites in the western Gobi Desert. Animal husbandry is
88
thought to have begun during the same period as plant cultivation, but was not
widespread until the Bronze and early Iron Ages. The adoption of animal husbandry is
usually associated with increasing aridity and the arrival of nomadic pastoralists from the
west. A more nomadic lifestyle would have begun at this time.
One of the most striking discoveries from the Neolithic or Neolithic-to-Bronze
Age transition is Tamsagbulag, a site near the Chinese border in eastern Mongolia.
Organization of subsistence and settlement appear to be much more consistent with
patterns in adjacent regions of Northeast China than contemporaneous finds elsewhere in
Mongolia. The Tamsagbulag site consists of several rectangular wattle-and-daub and
wooden semi-subterranean houses with roof entrances, located on the edge of an elevated
terrace near a spring in the vicinity of palaeolake Buir (Buiir) Nuur (Okladnikov and
Derevianko, 1970; Dorj, 1971; Derevianko and Dorj, 1992; Séfériadès, 2006). An
economy based both on hunting-gathering-fishing and food production is suggested by
faunal and plant remains, which include evidence of millet and cattle, as well as bird, pig
and horse (Derevianko and Dorj, 1992; Séfériadès, 2006). The domesticated status of
these faunal remains has not yet been demonstrated, although they have commonly been
accepted as such. The assertion that this group practiced millet agriculture is supported
by the presence of numerous grinding stones, hoes, and millstones (Derevianko and Dorj,
1992). Derevianko and Dorj (1992: 174-175) saw the Tamsagbulag site as evidence of
an independent origin for agriculture, which emerged from a base of intensive hunting
and gathering. No parallels in material culture or economy were thought to have existed
in adjacent regions of Central, North or East Asia.
89
Burials under house floors indicate a continuation of the Palaeolithic tradition of
using maral (red deer/elk) (Cervus elaphus) incisors in bead making. Mother-of-pearl
and bone were also used for ornaments. Bodies at Tamsagbulag and other sites in the
Kerulan valley were positioned in a contracted sitting posture, which is rare in Northeast
Asia, but has analogies to some late Neolithic/Eneolithic burials in the southeastern
Trans-Baikal region (Cybiktarov, 2002). Microblade insets were hafted in composite
knives.
Having conducted his own excavations at Tamsagbulag, beginning in 1996, the
French scholar Michel L. Séfériadès proposed that the Tamsagbulag site excavated by
Dorj and Okladnikov (Tamsagbulag 1) dated to about 6.5 kya (5th millennium BC).
Radiocarbon dates were derived from organics in a trench yielding chipped stone tools
and pottery, as well as charcoal and ash, which was located at the foot of the terrace not
far from the Tamsagbulag spring. This material was dated to 6.4k cal yr BP (5590 + 120
BP) and the nearby habitation sites were inferred to have been of about the same age.
Another site was uncovered northeast of Tamsagbulag 1 on the eastern side of the lake
within a small area of sand dunes on the eastern bank of a smaller, nearly dry lake. Lithic
and ceramic materials collected from the windward slope of the dune were also
associated with the long bone of a gazelle or antelope. An unreported radiocarbon date
places this site at around 4.0 kya.
Séfériadès interprets the difference in site types as evidence that a period of
sedentism occurred at Tamsagbulag 1 around 6.5 kya and was followed by increased
mobility after 4.0 kya. He proposes that when the lake began to vanish, people would
90
have moved to the dune site along the residual lake. Eventually, a return to year-round
nomadism would have occurred sometime after 4.0 kya. Aside from the early date
assigned to sedentary occupations at Tamsagbulag, this model is consistent with earlier
Soviet and Mongolian interpretations of the Neolithic-to-Bronze Age transition.
In the Gobi Desert of southern Mongolia, Shabarakh-usu (the key archaeological
locale excavated by Nelson during the Central Asiatic Expeditions, also known as Bayandzak [Baindzak/Bayn-dzak]), was further studied and identified as an early Gobi Desert
Neolithic type site. An aceramic horizon recognized by Nelson in his earlier excavations
was not observed by Okladnikov, but was later confirmed through excavation by Polish
researcher Kozłowski and Mongolian colleague Huhnbator (Kozłowski, 1972), and by
the joint Soviet-Mongolian team (Okladnikov, 1951), who found two distinct layers at
other stratified sites. The earliest levels were found beneath the dune base layer.
Ceramic decoration, arrowpoints, knives, and microblades led to the association of these
assemblages with the Early Neolithic of the Lake Baikal region (or the Serovo period,
which is now considered to belong to the late Neolithic – see Weber, 1995). Pottery was
decorated with textile and woven net impressions and had a pointed bottom. This earliest
level was found in a two-meter deposit of reddish-brown loam overlain by one meter of
light grey buried soil and over 3.5 metres of sand in which the later archaeological
sequence was situated (Chard, 1974). This later stage of development was characterized
by similar microblade reduction sequences, but assemblages are dominated by bifacially
retouched artefacts. The ceramics are thin-walled with flat bottoms. Painted ceramics
91
were also reported from these layers, bearing traces of black ornamentation on a red or
yellow background4 (Okladnikov, 1962; Chard, 1974).
The southern Gobi Desert Neolithic is thought to be characterized by small-scale
plant cultivation. Large numbers of grinding stones are claimed to have been discovered
along with “hoe-like tools” (Derevianko and Dorj, 1992; Cybiktarov, 2002). Reexamination of Gobi Desert assemblages collected during Central Asiatic Expeditions
and the Sino-Swedish Expeditions suggest that the collections – with the exception of
those from the southeastern Gobi Desert – contain few grinding stones and almost no
formal or large types (Fairservis, 1993; Marginer, 1950; Janz, 2006; also see Appendix
C). Nor do field notes of expedition archaeologists indicate that such artefacts were
found and left behind due to constraints on transportation (Nelson, 1925; Pond, 1928,
n.d.). It is not clear if variation between American and Soviet reports is due actual
differences in recovered artefact assemblages, or simply in their interpretations.
By the beginning of the Bronze Age, several distinct regional subsistence
strategies are thought to have been in place in Mongolia. Analyses of Soviet-Mongolian
collections led to the conclusion that there were distinct differences between the western,
eastern and southern regions during the late Neolithic and early Bronze Age (Volkov,
1967, 1981; Novgoroda, 1989; Cybiktarov, 2002). Cybiktarov (2002) divides
archaeological remains from the early Bronze Age into four distinct areas with varying
cultural traditions evidenced by material remains: 1) eastern steppe (eastern Mongolia
4
It should be noted that when I viewed these collections in February 2010 at the Institute of Archaeology
and Ethnology, Novosibirsk and the Institute of Archaeology, Mongolia, no such artefacts were found or
known to be included in the collections.
92
and southern Trans-Baikal); 2) eastern forest (forest zone within the same region); 3)
western Mongolia; 4) southern Mongolia (Mongolian and Chinese Gobi Desert).
Séfériadès (2006) published a nearly identical division of socio-economic and cultural
entities in Mongolia based on the Soviet literature. His four regions included: 1) western
Mongolia (including the Altai regions and west of the Khangai Mountains); 2) northcentral region (south of Lake Baikal); 3) southern region (southern Mongolia and
northern China); and 4) eastern region (northern Mongolia, west of Manchuria).
The characteristics of each of Cybiktarov’s regions are summarized as follows:
The first zone, in eastern Mongolia is considered to have been characterized by sedentism
and a mixed economy based on hunting, gathering, fishing, and stock raising (horses,
cattle, and sheep). Hoe-like implements and ring-shaped “counter-weights” are thought
to indicate low-level agricultural production. Copper smelting is evidenced by pieces of
slag and a few metal artefacts from sites in Siberia and along the border of Mongolia.
Tamsagbulag is included in this eastern steppe group.
Forest-dwelling groups in the Khentei Mountains of eastern Mongolia are thought
to have been separate from steppe and forest-steppe groups. They appear to have
maintained a hunter-gatherer economy without the addition of domesticates, though they
occasionally appear to have engaged in bronze metallurgy using local raw materials after
4.0 kya (early 2nd millennium BC) (Cybiktarov, 2002). Stone continued to be the most
widely used raw material, but emphasis was placed on flake tools rather than prepared
microblade core technology.
93
The early Bronze Age in western Mongolia is known mostly from rock art and
burials and is considered to be the result of a local variant of the Afanasievo and
Okunevo5 groups found throughout eastern Central Asia. A connection with the early
pastoralists of eastern Central Asia suggests a mixed economy that relied heavily on
hunting and gathering, but incorporated elements of food production. No settlement or
habitation sites have been excavated in this region.
Southern Mongolian archaeological sites have been found mostly in dune-field
environments near dried-up rivers and lakes. Past environments are proposed to have
been more semi-arid than arid, with forest-steppe landscapes. Fragmented grinding
stones and “hoe-like” tools are thought to attest to a partial reliance on agriculture.
Reports of painted pottery suggested a relationship with the early agricultural societies of
China. The Gobi Desert is considered to have been a region of unique cultural
developments. Remains from these sites were undated and their relationship to Bronze
Age developments remains unclear.
More recent exploration has been undertaken by the Joint Mongolian-RussianAmerican Archaeology Expedition (JMRAAE), which is in many ways an extension of
the joint Soviet-Mongolian Palaeolithic Gobi Desert and Mongolian Altai expeditions of
the 1980s. Several interesting Neolithic sites were discovered or revisited during this
project, although the primary goals of the expedition included the systematization and
standardization of techno-typological classifications, elaboration of chronostratigraphic
sequences, and reconstruction of palaeoenvironmental conditions and palaeoeconomy in
5
Considering the dates proposed by Cybiktarov (2002), these groups would necessarily be more closely
contemporaneous with the Andronovo entity.
94
Palaeolithic sequences (Derevianko et al., 1996). The cave site Chikhen Agui is the most
notable post-glacial period discovery and offers a new perspective on the organizational
strategies of post-LGM Gobi Desert hunter-gatherers.
Chikhen Agui is a rockshelter located near a spring at about 1970 m a.s.l. at the
eastern extent of the Gobi-Altai mountain range. It was discovered by Derevianko and
Petrin in 1988 and excavated by the JMRAAE team in 1996-1998 and 2000 (Derevianko
et al., 2003). Site structure was indicative of a cold season short-term camp that was
continuously reoccupied over a few millennia (Derevianko et al., 2008). Concentrations
of grass may have served as bedding, while a large hearth near the entrance would have
provided heat to the entire shelter. Microblade insets for composite tools were the most
common artefact type and a wooden haft was also found. Ostrich eggshell beads and
fragments of what may have been a vessel of the same material were also discovered.
Faunal remains from Chikhen Agui include Lepus capensis6 (hare), Ochotona cf. O.
Alpine (pika), Marmota sp. indet. (marmot), Spermophilus sp. indet. (ground squirrel),
Dipodidae gen. et sp. indet. (jerboa), Equus hemionus (wild ass/khulan), Procapra
gutturosa (Mongolian gazelle), and Capra sibirica sibirica (Siberian ibex) (Derevianko
et al., 2008). Excavators interpreted this site as a seasonal hunting camp used for
processing game: animals may have been ambushed as they came to drink at the spring
below the mouth of rockshelter (Derevianko et al., 2003). The addition of non-utilitarian
artefacts such as beads, a serpentine-antigorite pendant, and a wooden post or pole near
the west wall of the cave suggest to researchers that some type of ritual activity might
6
Lepus tolai is common in the Gobi Desert and throughout much of Mongolia and was previously included
in Lepus capensis. It is not clear if this identification takes this new nomenclature into consideration.
95
also have occurred (Derevianko et al., 2008). Notably, the ostrich eggshell bowl
contained unidentified grass seeds and was situated near the wooden pole.
Ten radiocarbon dates bracket the site between 13.4-6.4k cal yr BP (5630 + 220
and 11,545 + 75 yrs BP) (Derevianko et al., 2003), though the two earliest reported dates
were later discounted and a date range of 13.4 to 8.7k cal yr BP (7850 + 100 BP) was
proposed as a more reliable estimate (Derevianko et al., 2008). Recently, additional
pieces of ostrich eggshell from the site were dated to 12.2 and 11.6k cal yr BP (10,060
50 and 10,330
55 BP) (Kurochkin et al., 2009). Unfortunately, soil development was
weak and heavily disturbed by rodent burrows, which made for a challenging
stratigraphic profile. As such, it was not possible to draw many conclusions about
changes in artefact assemblages over time. There do appear to have been changes in raw
material use: lithics in the lower horizon were made primarily on dark siliceous
sandstone, while jasper-like rocks and chalcedony were more common in the upper layer.
The upper layer also yielded geometric microliths which are uncommon in Mongolian
lithic assemblages and were not present in earlier strata (Derevianko et al., 2003).
Chikhen Agui is significant for a number of reasons. It was the first post-LGM
site in the Gobi Desert to be dated chronometrically. The majority of dates are between
10.0 and 9.0k cal yr BP, identifying it as one of few known Epipalaeolithic habitation
sites. Though only representative of a singular aspect of contemporary land-use, it offers
a distinct contrast to the “dune-dweller” sites which dominate the archaeological record,
particularly in the Gobi Desert. Whether or not ritual activity was practiced here, it
represents a high elevation, short-term special purpose site, presumably occupied during
96
the cold season (based on hibernation patterns of represented rodent fauna, the most
likely time is the end of summer or early fall). Both large and small faunal remains
indicate a range of species typical of both mountain environments (ibex, pika) and of the
open plains (wild ass, marmot, ground squirrel, gazelle), suggesting a varied diet. It is
also important to note that the rockshelter seems to have been abandoned after about 8.7k
cal yr BP until modern or historic times.
Inferences of post-LGM Gobi Desert settlement and subsistence based on investigations
in North China
Since 1989, a group of scholars from China and the United States have been investigating
environmental constraints and settlement patterns in the terminal Pleistocene/early
Holocene of North China, including the connection between the domestication of
broomcorn millet (Panicum miliaceum) in northwestern China (as evidenced by the
Cishan site) and the subsistence economy of post-glacial hunter-gatherers (Bettinger et
al., 1994; Madsen et al., 1996; Elston et al., 1997; Richerson et al., 2001; Bettinger et al.,
2007; Bettinger et al., 2010a; Bettinger et al., 2010b; Elston et al., 2011). Investigations
included an analysis of the Sino-Swedish collections from the Alashan Gobi Desert based
on Maringer’s (1950) catalogue of the remains (Bettinger et al., 1994), survey around the
Helan Mountains (Madsen et al., 1996), excavations at Pigeon Mountain Basin in the
Helan Mountains foothills (Elston et al., 1997), and survey of the region around and north
of the Neolithic Dadiwan site on the upper Wei River in the western Loess Plateau.
97
Analysis of the Sino-Swedish expedition materials was based on the presence or
absence of particular artefact categories: cores, adzes/axes, bifaces, and unifaces
(Bettinger et al., 1994). Sites from the Alashan Gobi Desert were divided into six groups,
including residential bases, seasonal base camps, temporary camps, and procurement
locales. Based on the assumption that almost all sites were Neolithic in age and were part
of the same type of land-use and economic strategy, inter-assemblage variability was
thought to have been related to site function and chronology was not considered. Two
types of seasonal camps were identified based on the mutually exclusive presence of
either bifaces or axes (adze/axes). Projectile points were considered separately from
bifaces, although it is clear from my own study of the collections that most artefacts
Maringer classified as projectile points were bifacially retouched. As such, the biface
category includes a range of biface types, including both projectile points and knives.
Bifacially flaked adze/axes were not included in this category.
Residential bases were indentified primarily in areas rich in lithic raw material
such as around the Ukh-tokhoi (Ukh Tohoy) and Khara-dzag (Hara-dzag) plateaux, and
along the dune-field/wetland region of the Mongolian border. Residential bases were
interpreted as winter occupations, when increased sedentism would have allowed for craft
production and more concentrated episodes of processing. Seasonal base camps with
adze/axes but no bifaces were primarily located around the Lang Shan foothills, far from
residential bases, and were proposed to have represented a separate kind of settlement
system or satellites of residential bases along the border of Ukh-tohkoi/Khara-dzag.
98
Several important points emerge in this analysis of the Sino-Swedish artefact lists.
The pattern of land-use postulated is similar to the collector-type model outlined by
Binford (1980) and was believed to represent a more complex intensive and seasonally
differentiated strategy that was a response to increasingly abundant resources at the end
of the Pleistocene. Possible evidence of woodworking tools (adze/axes) was thought to
represent specialized tasks and the exploitation of new resources. Use of ceramics
inferred a decrease in residential mobility, functionally related to an increase in diet
breadth that included more plant species. Microblades were similarly thought to be
connected to subsistence intensification (see also Bettinger et al., 2006). Finally, the
Neolithic occupations represented by these collections were thought to represent a
colonization, rather than in situ development. This colonization was first proposed to
have been driven by increased availability of resources at the end of the Pleistocene or
the beginning of the Holocene.
The general picture drawn was one of post-LGM hunter-gatherers extensively
reorganizing subsistence, mobility, and tool use as increased precipitation as many new
opportunities for food procurement emerged during a period of increased precipitation
and warming (see also Elston et al., 2011). Climatic amelioration has been suggested
elsewhere as a possible motivating factor in broad spectrum foraging or domestication
(Richerson et al., 2001; Elston and Zeanah, 2002), but this model is at odds with the more
common models of intensification as symptoms of either increased population density
(Binford, 1968; Flannery, 1969; Keeley, 1988; Redding, 1988; Stiner et al., 2000; Stiner,
2001) and/or unpredictability/scarcity of resources (Redding, 1988; Bar-Yosef, 1996;
99
Hillman, 1996; Winterhalder and Goland, 1993; Bar-Yosef, 2002; Marshall and
Hildebrand, 2002).
Farther south, survey of sand dunes, alluvial fans, stream margins and lake-marsh
shorelines along the foothills of the Helan Mountains suggested a trend towards reduced
residential mobility over time and a progressive intensification in the use of the lower to
middle reaches of alluvial fans with springs (Madsen et al., 1996). The researchers
interpreted the use of such environments as an indication of early agricultural
endeavours. As in the Alashan Gobi Desert, they identified both residential bases and
seasonal base camps. Chronologically ambiguous sites along dune margins were thought
to represent seasonal occupations focused on the intensive processing of wild seed crops
and small animals (Madsen et al., 1996). Later excavation at Pigeon Mountain basin, a
shallow drainage basin with springs in the Helan Shan foothills, indicates that the use of
microblade technology was in place by about 15.1k cal yr BP (12,710 + 70) and grinding
stones by at least 13.5k cal yr BP (11,620 + 70) (Elston et al., 1997). Unlike in the
Alashan Gobi, possible house structures were sometimes associated with microblades and
microblade cores, as well as other artefacts associated with food processing. Madsen and
colleagues (1996) point out that while Late Palaeolithic sites suggest numerous shortterm camps, a wider range of Neolithic-type sites are noted. They interpret similarities in
core reduction strategies between periods as a sign of local development rather than
colonization.
Based on their analysis of the Sino-Swedish Expedition catalogues and
approximately twenty years of experience in the Helan Shan region, Bettinger et al.
100
(2007) proposed a tentative chronology of post-LGM land-use and agricultural
developments for northwestern China. Although few radiocarbon dates are available for
Helan Shan sites, we can expect that extensive survey of the region should have provided
the researchers with some level of implicit understanding about local chronology. At the
same time, the proposed chronology is hypothetical and should be supported by
additional direct dating and excavation.
Dates from Pigeon Mountain Basin provide the first clearly dated expression of
post-LGM hunter-gatherer occupation in arid North China. These sites belong to what is
characterized as the Helan Period, a time when the frequent use of macrotools and a
diverse array of tool forms suitable for plant processing suggest broader spectrum
subsistence than would have occurred in earlier periods. Large archaeological site
assemblages imply decreased residential mobility compared with the earlier period. The
use of dune field margins near upland environments would have allowed inhabitants to
take advantage of a wide range of resources. Upland sites are fewer in number but are
generally more diverse, suggesting permanent base camps in higher elevation
environments, perhaps occupied seasonally.
The Tengger Period represents terminal Pleistocene and early Holocene
occupation (including the Younger Dryas period of increased aridity – see Madsen et al.,
1998). An increased reliance on microlithic technology is characteristic (see Elston et al.,
1997). A scarcity of milling stones and more evidence of faunal remains are assumed to
result from sampling error, but hunter-gatherers during this period are expected to have
been more focused on hunting relative to gathering due to the decreased abundance of
101
plant species during the Younger Dryas. Dune-field resources became more important,
but the pattern of short-term low-land occupations and upland base camps seems to
persist.
In the early Holocene, it is suggested that increased humidity and interdunal lake
and marsh infilling may have led to longer term habitation in dune-fields. Microblade
technology is thought to have become more important. Wild millet may have been
exploited more intensively during this period. Despite the attractiveness of this model,
the researchers admit that no archaeological sites in the study area date to the early
Holocene. In the southern study area, at the early agricultural site of Dadiwan, the postPleistocene record does not resume until 7.8k cal yr BP (6,950 + 90) (Bettinger et al.,
2007).
The lack of archaeological sites dating to the Early Holocene is a recurrent theme
in North China, with Nanzhuangtou and Yujiagou being the only two dated sites with
occupations from this time period (Xia, 2001; Cohen, 2003; Liu, 2004). The researchers
do not see Neolithic habitation in North China as emerging from a local base (Bettinger
et al., 2007; Bettinger et al., 2010a; Bettinger et al., 2010b). Instead, they suggest that
desert hunter-gatherers familiar with the exploitation of millet were driven southward
during successive intervals of drought in the early Holocene.
The model that they offer proposes that the Alashan Gobi was colonized by
hunter-gatherers from the south, who were becoming more sedentary and expanding their
populations into the more marginal northern desert regions. Geographic expansion would
have been driven by progressively increasing population density along the less arid
102
stretches of northern China, where the better availability of low-ranked resources during
post-glacial climatic amelioration would have increased resource reliability, thereby
stimulating a decrease in residential mobility, and a broadening in diet breadth. Such
developments would have presumably involved decreased child mortality (Handwerker,
1983) and perhaps increased fertility (Hassan, 1973; Hassan, 1981).
Considering the geographic distance between the Helan Shan region and those
regions of the Gobi Desert discussed in this work, the Gobi Desert chronology is not
necessarily expected to conform to that proposed here. However, three important
inferences about local hunt-gatherers can be taken from this analysis: a relationship
between climatic amelioration and increasing dune-field use; the complementary seasonal
use of upland and lowland habitats; and the possibility that the intensive exploitation of
grass seeds by desert hunter-gatherers may be directly related to the development of
sedentary agriculture in central China. Analysis of Gobi Desert assemblages will
consider both these points.
Recent interpretations of the transition to nomadic pastoralism
Contrary to Maringer’s (1963) suggestion that the Metal Ages in Mongolia were the
result of colonization by iron-using foreigners around 2.2 kya, it is now clear that there
was a locally developed Bronze Age. Research by Mongolian and Soviet scholars
supports the likelihood that local hunter-gatherers exploited copper, possibly bronze, and
domestic herd animals by at least 4.0 kya and probably earlier. Despite clear evidence
refuting cultural and economic “stagnation” during the Neolithic, the possibility that true
103
nomadic pastoralism was introduced through the expansion of western groups of different
ethnic affiliations is still debatable. Recent doctoral work by North American scholars
working in Mongolia has addressed this issue (Wright, 2006; Houle, 2010).
The late Bronze Age in Mongolia is considered to be the first period when fully
formed nomadic pastoralism can be recognized. Burial/ceremonial monuments called
khirigsuurs (kheregsuurs) are found across western and central Mongolia, including the
Gobi Desert (see Eredenbaatar, 2004). Dates from both khirigsuurs and deer-engraved
stelae (deer stones) indicate that the culmination of Bronze Age ritual culture would have
occurred between about 3.2 and 2.8k cal yr BP (Fitzhugh, 2009), broadly
contemporaneous with the Karasuk period in the Minusinsk Basin of Siberia (Legrand,
2006; Houle, 2010).
Bronze Age habitation assemblages from the Khanuy Valley in northern
Mongolia suggest that as with the Karasuk, sheep and goats were most important for
subsistence, followed by horse, then cattle, and finally musk deer (Moschus sp.) (Houle,
2010: 122-131). The intensive ritual use of horse in Mongolia is notably divergent from
the ritual use of sheep and cattle most common in the Karasuk period burials (see
Legrand, 2006). Deer imagery is also unique to Mongolia, and perhaps the Trans-Baikal
region, until the Iron Age Scythian period (see Fitzhugh, 2009; Houle, 2010: 5-6). These
unique ritual traditions illustrate a trajectory separate from Siberia and influential to later
Central Asian pastoralist tradition.
Despite evidence of intensified ritual practices including the construction of major
architectural structures, the late Bronze Age in Mongolia is distinct from what is found to
104
the north and west. There is little evidence for the well-defined social stratification and
differentiation that is apparent among pastoralists in the Minusinsk Basin. Specialized
site function and potentially some level of social differentiation were noted by Houle
(2010: 143-176) for the Khanuy Valley (see Figure 2.1). Earlier surveys in the Egiin Gol
Valley (see Figure 2.1) further indicate little difference in land-use between nomadic
pastoralist communities and earlier hunter-gatherers (Wright, 2006). In Mongolia, the
occurrence of slab burials at Bronze Age pastoralist sites is a distinctive form of burial
that can be associated with Okunev and Karasuk periods further west and is relatively late
in comparison to neighbouring regions (the earliest may be 3.6k cal yr BP, with the
majority of dates falling between 2.8 and 2.5k cal yr BP) (Wright, 2006: 279). Evidence
of microlithic scatters around khirigsuurs and chipped stone projectile points within slab
burials indicate that earlier local populations may have been ancestral or otherwise
associated with Bronze Age pastoralists. Wright (2006: 199-263) argues that some
khirigsuurs may have been constructed and used by pre-pastoralist hunter-gatherers and
that slab burials were markers of the first true nomadic pastoralist groups, who may either
have been derived from a local base or have been immigrants incorporated into the
existing population.
Whether or not early monuments were built by microlith using hunter-gatherers,
pastoralists, or hunter-gatherer pastoralists, it is clear that the trajectory and chronology
of developments in Mongolia is different from that observed in the rest of Central Asia.
Nomadic pastoralism may have been adopted at a variable rate across the region, but does
not appear to have formed the basis of economies until sometime between 3.5 and 3.0k
105
cal yr BP. This chronology places the rise of nomadic pastoralism in Mongolia slightly
earlier or contemporaneous with parts of Northeast China. Here, cultures such as
Baijinbao on the Song Nen Plain (3.3 kya) began to rely more heavily on herd animals,
and the Upper Xiajiadian in the Chifeng region (3.0-2.3 kya) exhibited more extensive
ties with Northeast Asian nomadic pastoralists (Guo, 1995b; Tan et al., 1995b; CICARP,
2003). The Xindian peoples (3.2-2.5 kya) in Gansu and Qinghai provinces also appear to
have been nomadic pastoralists practicing limited agriculture (An, 1992b; DebaineFrancfort, 1995). It is also widely believed that the Zhou (3.0-2.2 kya) in the Central
Plains of China were not related to the preceding Shang, but originated north and west of
the Central Plains (Guo, 1995b; Barnes, 1999).
Pastoralism appears to have reached its initial height between 3.2 and 3.0 kya
across Northeast Asia. Bronze Age developments such as metallurgy and equestrianism
reached new heights during the Iron Age only a few hundred years later. Most
importantly, we should consider that just as 4.0 to 3.0 kya is the span that includes the
end of the Neolithic and the beginning of the Bronze Age, it is the period between 3.5
and 3.0 kya that would have marked the true decline of Neolithic hunter-gatherer
societies and the rise of nomadic pastoralist economies.
2.4. Summary
The transition to food production in Northeast Asia was characterized by a series of long
term technological and economic changes that had variable outcomes. Microblades,
pottery, grinding stones, and stone polishing are hallmarks of terminal Palaeolithic
106
adaptations, but these technological elements are never ubiquitious and do not seem to be
definitively associated with specific developments in subsistence. Northeast Asia is
notable for the fact that both civilizations based on sedentary agriculture and powerful
nomadic pastoralist states were dominant in the same region. Having outlined the
transition to agriculture and pastoralism, as well as agropastoralism, in Northeast Asia, it
is clear that both local environments and cultural associations probably contributed to the
formation of historical economic structures. This point should be remembered as we
address the role of Gobi Desert hunter-gatherers within the context of neolithization in
Northeast Asia.
Throughout the terminal Pleistocene and Holocene, Mongolia and the Gobi
Desert region are characterized by a unique set of developmental trajectories.
Neighbouring agriculturalist, agropastoralist, and pastoralist neighbours may have
interacted with and influenced technological and subsistence change in the region, but the
archaeological record prior to the Bronze Age attests to the long-term importance of
hunter-gatherer economies as opposed to the steadily increasing prominence of
domesticated species that is found elsewhere in Central Asia and China. The emergence
of a post-LGM strategy in the Gobi Desert, which focused on the use of dunefield/wetland environments and the occasional use of Neolithic type tools such as pottery,
grinding stones and polished stone tools, suggests that local hunter-gatherers may not
have as regularly exploited domesticated species as in China and Central Asia, but were
similarly developing new methods of adapting to changing local environments.
107
By about 5.0 kya, there was a definite shift in the broader context of subsistence
economies across Northeast Asia as elements of production, or at the very least
intensification, emerged. The wide dispersal of both foreign and local domesticated plant
and animal species is a clear indication of this trend. Food production economies were
firmly established in Northeast Asia by 3.0 kya, except in regions where extreme climates
prevented the adoption of existing domesticates. Tentative evidence for earlier limited
use of domesticated animals in Mongolia is compelling, and while current evidence is
insufficient to support inferences about the status of domestication processes in the Gobi
Desert, it is likely that hunter-gatherers would probably have had knowledge of domestic
herd animals by 5.0-4.0 kya. Shifts in Gobi Desert land-use and artefact assemblages
around this time must be taken into account and given some consideration as they relate
to the possible introduction of domestic species.
This research focuses on modes of land-use and subsistence during what are
traditionally thought of as the Mesolithic, Neolithic, and early Bronze Age periods. Since
land-use is inextricably tied to environmental and cultural constraints, it is important to
establish a more detailed chronology for local post-LGM developments which can then
be compared to contextual data. By considering local artefact assemblages,
palaeoenvironment, and the cultural and economic milieu of neighbouring regions, this
study will build a series of hypotheses about late Stone Age groups in the Gobi Desert
that can be further investigated through excavation and materials analyses. Due to
distinct geographic differences across the Gobi Desert in environment and possible
cultural contacts, three Gobi Desert regions were considered: the East Gobi, the Gobi-
108
Altai and the Alashan Gobi. This study contributes to understanding two major changes
in the cultural trajectory of Northeast Asia: the nature of complex hunter-gatherer
systems prior to the transition to food production economies; and the role that groups
living on the peripheries of both Central Asian pastoralist complexes and agricultural
villages played in the development of the dichotomous production economies that
characterize modern Northeast Asia.
109
CHAPTER 3 – CHRONOLOGY OF GOBI DESERT
ARCHAEOLOGICAL SITES
The absence of good chronological control limits our understanding of prehistoric
land-use, technological developments, and subsistence in the Gobi Desert. Previous
research on existing assemblages has taken various approaches to interpretation of the
regional chronology (Maringer 1950; Okladnikov, 1951; Kozłowski, 1972; Bettinger et
al., 1994), but few clear results have emerged. Although an artefact-based chronology
built on surface associations and stratigraphic evidence has been impressed upon the
consciousness of local archaeologists there is no consensus system for chronologically
ordering prehistoric assemblages. Without temporal controls, it is impossible to relate
shifts in settlement patterning to either ecological shifts (i.e., increased/decreased
precipitation, dune stabilization, enhanced productivity of certain favoured plants or
animals, longer winters), or cultural stimulus (i.e., technological innovation, increased
cultural complexity, trade with agricultural or pastoralist neighbours, population
expansion from Central Asia or China).
Estimated ages of archaeological sites outlined in Chapter 2 may be accurate, but
have not been clearly justified. Chronological resolution has been too coarse to confirm
or deny existing interpretations of temporal variability in technology, land-use,
subsistence, or the role of climate change in influencing these trajectories. In order to
move our understanding of this culturally and environmentally important geographic
region forward it is necessary to begin disseminating a firm artefact-based chronology by
which to recognize chronological ordering in the archaeological record.
110
The first step in this endeavour is to consolidate existing knowledge of artefactbased chronology for the terminal Pleistocene and early Holocene in Mongolia and
adjoining regions. This task is daunting due to both the wealth of regional publications
and the number of languages used in disseminating the data. The goal of this chapter is
to synthesize the literature most accessible to English-speaking audiences, also drawing
on available texts in other languages according to the author’s ability in order to identify
the most reliable temporal indicators in artefact assemblages. It is hoped that in the
future Mongolian, Chinese, Russian, and other scholars might bring additional
knowledge to bear in the chronology of the post-LGM Gobi Desert in order to collaborate
on research that will be accessible to an international community.
It has been difficult to obtain secure chronometric dates for Gobi Desert sites
since most are surface assemblages with few organic remains. Issues of preservation
aside, because the majority of excavated assemblages were collected prior to the advent
of radiocarbon dating, easily dated organic remains like charcoal were not recovered
when present. With the advent of modern chronometric dating methods, excavation can
now provide temporal data, as exemplified by research at the Chikhen Agui site
(Derevianko et al., 2003; Derevianko et al., 2008); however, excavation is costly and this
fact limits large scale recovery and necessarily constrains data to one component of a
larger settlement system.
Existing collections hold a great deal of potential for building interpretations
based on inter- and intra-regional data sets. The broader geographic perspective that they
offer is especially promising. Archaeological collections used in this research are housed
111
at the American Museum of Natural History (AMNH) and the Museum of Far Eastern
Antiquities (MFEA). They are derived from large areas of the Gobi Desert region and
represent a wide range of environmental and site-use contexts. The vast majority of these
museum assemblages derive from recently deflated dune surfaces. Others were still
partially embedded in the matrix. Detailed catalogues, journals and archival records
allow for an understanding of possible collection biases (Nelson, 1925; Pond, 1928, n.d.;
Fairservis, 1993; Maringer, 1950).
For this study, chronometric dating was applied to collections in order to
constrain temporal range of assemblages, test assumptions about possible intermixing,
and offer approximate dates for various diagnostics that might be used for indirect dating
of additional sites. Some researchers dismiss the usefulness of surface collections due to
the perception that they are necessarily chronologically incoherent. In the case of the
Gobi Desert collections we can be confident that there are at least some “single
component” assemblages. Detailed descriptions of excavation and collection clearly
indicate the common phenomenon of small, localized site clusters (often surrounding
hearths or accumulations of fire-cracked rock) within major find localities (Nelson, 1925;
Pond, n.d.; Fairservis, 1993). Likely cases of intermixing were often noted by the
original investigators. If we then concede that such sites are temporally coherent, we can
address the issue of using chronometric dating techniques.
112
3.1. Chronometric dating
A total of 23 separate Gobi Desert archaeological sites were chronometrically dated
(Appendix A) and provide indirect dates for additional sites based on comparative
artefact typology. The most effective method of dating Late Pleistocene and Holocene
archaeological sites is typically accelerator mass spectrometry radiocarbon (AMS) dating
analysis. AMS can be used effectively on both organic remains and carbonates (though
dating of aquatic mollusc shell raises issues of marine reservoir and other effects – see
Ascough and Cook, 2005). Modified ostrich eggshell and pottery are often found in Gobi
Desert assemblages and offer the best potential for AMS dating. Alashan Gobi sites less
often contain ostrich eggshell and the pottery is usually heavily sand-tempered and
without carbonized organics, so the sites in this part of the study area were most difficult
to date. A second approach to chronometry, luminescence dating of pottery shards, was
carried out on Alashan Gobi samples. Luminescence dating was carried out at the
University of Washington Luminescence Laboratory
Many archaeological sites in the study collections contained fragments, decorated
pieces, and disc beads of ostrich eggshell (Maringer, 1950; Okladnikov, 1962; Fairservis,
1993; Janz et al., 2009). Ostrich eggshell fragments and bead fragments from the AMNH
collections were dated using AMS in 2007 (Janz et al., 2009). Additional dates were
subsequently obtained from samples derived from collections at the MFEA. All ostrich
eggshell fragments were dated using the selective dissolution process outlined in earlier
published studies on dating ostrich eggshell (Freundlich et al., 1989; Vogel et al., 2001;
Bird et al., 2003).
113
Pottery is more widespread in Gobi Desert archaeological assemblages than
ostrich eggshell and has been used for direct dating in other Northeast Asian
archaeological assemblages (O’Malley et al., 1998; Keally et al., 2003; Kuzmin and
Shewkomud, 2003; Yoshida et al., 2004). Samples were derived primarily from
collections housed at the AMNH, but one sample was taken from the MFEA collections.
All shards were processed using low temperature combustion (400oC) (O’Malley et al.,
1998).
Obtaining the age of a site using dates on pottery is clearly limited to the period
following the adoption of ceramic technology. It was hoped that ostrich eggshell dates
would provide a source for dates on earlier archaeological sites, but there was concern
over the possible use of fossil eggshell, as the practice of older shell use has already been
documented in the Gobi Desert and other regions (Potts, 2001; Aseyev, 2008; Janz et al.,
2009). On the other hand, radiocarbon dates on ostrich eggshell fragments associated
with post-LGM assemblages (as recognized by the use of developed microblade
technology) have shown that eggshell fragments might be consistent with archaeological
assemblages even if they are occasionally be older than associated artefacts (Jaubert et
al., 2004; Kurochkin, 2009; Janz et al., 2009).
3.1.1. Sample selection
A range of samples were taken from each of the three Gobi Desert target regions – the
East Gobi, the Gobi-Altai, and the Alashan Gobi. Selection of particular specimens for
dating was based on the suitability of materials for dating, association with other
114
relatively common diagnostic artefacts, apparent temporal cohesiveness of associated
assemblages, and the ability to date other assemblage artefacts to control for error (see
Appendix A.2). Site selection was also based on how representative each site was of the
respective region. Many sites outside dune deposits did not fit all the criteria outlined
above, but artefacts from these localities were selected whenever possible in order to
broaden the sample. Sampling was intended to minimize damage to assemblage integrity
– unique artefacts were avoided in favour of more abundant types. Most samples came
from site assemblages systematically analyzed for this study, although exceptions were
made in the case of certain sites (e.g., Chilian Hotoga) because they fit the other criteria.
Multiple samples were taken from several sites.
3.1.2. Dating methods
A number of studies have shown that ratite eggshell can provide reliable radiocarbon
dates (Freundlich et al., 1989; Miller et al., 1999; Vogel et al., 2001; Bird et al., 2003).
Samples of ostrich eggshell, especially those associated with pottery, were prepared by
myself and Dr. George Burr (University of Arizona) and analyzed at the NSF - Arizona
Accelerator Mass Spectrometry (AMS) Laboratory using a selective dissolution
procedure designed to remove the outer layer of carbonate and avoid potential
contamination (Burr et al., 1992), a procedure successfully applied to ratite eggshell
samples from Australia by Bird et al. (2003). In a test sample, the radiocarbon content of
successively dissolved eggshell fractions was measured to assess possible post-
115
depositional carbon exchange. It was found that the multiple dates produced overlapped
at the 2σ level and were therefore statistically indistinguishable from one another,
suggesting that ostrich eggshell is relatively inert and should provide reliable results (Janz
et al., 2009).
The primary concern in using ostrich eggshell to date archaeological sites is the
possibility that fossil shell might have been used. Recent AMS dates from Shabarakhusu, Mongolia (Janz et al., 2009), suggested that although fossil shell does occur in
Holocene archaeological sites, dating multiple artefacts using complementary dating
methods helps to establish the contemporaneity of eggshell with archaeological
assemblages. Chronometric dates on ostrich eggshell have been used in northern and
southern Africa to infer the age of archaeological assemblages with minimal
consideration of the “old eggshell” problem (Freundlich et al., 1989; Vogel et al., 2001;
Halkett et al., 2003). In those regions eggshell for bead-making is thought to have been
obtained from broken water carriers or from eggs collected for consumption
(Sandelowsky, 1971; Orton, 2008), implying temporal association between ostrich
eggshell and archaeological occupations. In our original study, none of the samples
yielding anomalously old dates displayed conclusive evidence of human modification.
While it has not yet been proven that hunter-gatherers used only fresh eggshell, AMS
dates do provide at least an upper limit to the age of sites (Janz et al., 2009).
Dating pottery using AMS is less commonly done, but has been employed
successfully at other Northeast Asian archaeological sites (O’Malley et al., 1998; Keally
et al., 2003; Kuzmin and Shewkomud, 2003). If viable, direct dating of ceramics offers a
116
wealth of opportunities for dating both museum collections and site assemblages in arid
regions of Northeast and Central Asia, where surface assemblages are most common and
preservation of organics rare. All dated ceramic samples underwent a standard acidalkali-acid (AAA)7 pretreatment and low temperature combustion. Some samples needed
to be treated two or three times with acid until all carbonates in the paste could be
neutralized, whereas for others a single treatment sufficed. Similarly, samples with
higher levels of humic acid and were treated with a base up to three or four times in order
to remove all contaminants. All samples were combusted on a vacuum line with CuO at
approximately 400oC. Previous studies suggest that low temperature combustion is most
reliable for AMS dating on pottery as it releases carbon from the temper, but is not hot
enough to release old carbon from the clay (O’Malley et al., 1998). In these earlier
studies, interior portions of the pottery were sampled in order to avoid contamination
from the exterior surface. Bulk samples were not combusted at low temperatures, but
exterior and interior portions were dated separately using low temperature combustion.
Interior subsamples generally provided older ages than the exterior counterparts
(O’Malley et al., 1998).
The source of carbon in the selected samples varied. In some cases the most
obvious source of carbon was residue from burning on the vessel surfaces. Some of the
shards had abundant fibre temper or other carbonated organics in the temper, as
evidenced by the charred and blackened paste. Other shards did not have any obvious
traces of such blackening. Many of the samples with no obvious traces of organic temper
7
This pretreatment is also known as an acid-base-acid (ABA) pretreatment.
117
or blackening produced very low carbon yields (e.g., 73/2796 B produced only 0.08 mg C
or 0.07% of the sample, compared with 73/2797 A with a relatively high yield of ~1.0 mg
C/0.87%). Samples with carbon yields under 0.10% tend to yield unreliable dates
(George Burr, personal communication, July 7, 2011) and are marked with an asterisk in
Table 3.1. As expected, the dates returned from the low-yield samples were anomalous
with the archaeological assemblages, with the exception of AA89873/AMNH #73/887A
(Shabarakh-usu 4) and AA89887/AMNH #73/2231C (Baron Shabaka, Site 19), which
were consistent with other dates from the same localities.
Due to variation in the presumed origin of carbon amongst the samples, bulk
portions were used for dating. Considering the finding that exterior portions of the shards
tend to be younger and more variable (O’Malley et al., 1998), it is reasonable to assume
that bulk samples may have included some younger carbon. As such, if inaccuracies do
exist, we can expect that the AMS dates on pottery are minimum age estimates.
Comparing AMS dates on pottery with ostrich eggshell dates and luminescence dates
should contribute to a clearer understanding of such concerns.
Luminescence dating on pottery complements ostrich eggshell dating for Gobi
Desert sites. Luminescence dating is more destructive than radiocarbon (a fragment of
pottery must be at least 5 mm thick and 30 mm in diameter), but is more reliable because
it provides a direct age range for the firing or use of the pot (Aiken, 1985; Feathers,
2003). Since ceramics are often more fragile and less likely to be used after breakage
than ostrich eggshell, it is also less likely that shards were recycled or re-used. Although
there is a larger degree of uncertainty in luminescence dating, it is effective in building a
118
relative chronology for pottery types (Godfrey-Smith et al., 1997; Herbert et al., 2002)
and for testing the applicability of AMS dates from eggshell. The utility of dating surface
ceramics by luminescence has been demonstrated in several cases (Dunnell and Feathers,
1994; Sampson et al., 1997); the technique is especially useful in circumstances where
multiple occupation episodes may have been intermixed (Feathers, 2003).
The main problem with using luminescence to date these samples is uncertainty in
determining the external dose rate, which includes both gamma and cosmic contributions.
For ceramics, an associated sediment sample is often collected for this purpose where in
situ measurements cannot be made. Since the dated specimens were collected decades
ago no such sediments are available. The problem was diminished by employing finegrained dating (Feathers, 2003: 1496), which is less reliant on the external dose rate. The
fact that the museum-curated samples come from the surface is advantageous, because
the atmosphere contains little radioactivity, thus reducing the gamma contribution.
Uncertainty in the cosmic dose rate and the potential for radon fall-out have been cited as
problems in surface dating, but these concerns are often over-stated (Feathers, 2003) and
at any rate minor compared to the uncertainty with the gamma dose rate. As such, range
of error for samples dated using luminescence was much higher than those dated using
AMS.
AMS and luminescence dating are complementary techniques. AMS provides
dates with a low margin of error, while luminescence dates on pottery can be used in the
absence of organic temper or surfical carbonization. Luminescence dating on pottery can
also be used as a control on AMS from ostrich eggshell and as a method of establishing a
119
relative chronology for diagnostic pottery styles. AMS determinations on humanmodified ostrich eggshell in combination with dates on stylistically distinct pottery from
key archaeological sites will help refine and test assumptions about chronology.
3.1.3. Results
Results of AMS and luminescence dates are summarized in Appendix A.3. Dates that
could be reliably associated with archaeological remains are summarized in Table 3.1.
Direct dates on pottery indicate that the majority of these archaeological assemblages
date to the middle to late middle Holocene. Combined with radiocarbon dates from
Chikhen Agui (Figure 3.1), we can record the earliest known post-LGM occupations in
each macro-region as follows: East Gobi, 9.5k cal yr BP from Shara KataWell; GobiAltai, 13.4k cal yr BP from Chikhen Agui; Alashan Gobi, 5.6 kya from Yingen-khuduk.
Since many aceramic archaeological assemblages have yet to be dated, we expect that the
respective regions were inhabited prior to these dates. Dates for the earliest use of
pottery north of the North China Plain and south of the Lake Baikal region can currently
be assigned to 9.6k cal yr BP and are indicated by sample AA89868 (AMNH #73/466A)
from Shara KataWell in the southeastern Gobi Desert.
120
Site
14C age BP + 1 σ
or KA + 1 σ
KYA (cal. to
68% range)
UW2361
Cat. No.
(K.13 or
73/)
203: 5
Jabochin-khure
3500 + 300
3.20-3.80
AA91693
207: 1
Gashun Well
3385 + 40
3.59-3.68
UW2358
212: 6
Yingen-khuduk
3910 + 300
3.61-4.21
UW2357
212: 123
5690 + 350
5.34- 6.04
UW2360
212: 128
3910 + 230
3.68-4.14
UW2856
248: 5
3540 + 1060
2.48-4.60
UW2355
248: 6
2740 + 200
2.54-2.94
UW2362
298: 15
6460 + 700
5.76-7.16
UW2359
298: 25
3840 +340
3.50-4.18
AA89872
655 A
4308 + 40
4.85-4.95
AA89873
887 A*
Shabarakh-usu 4
3680 + 76
3.92-4.13
AA89877
1189 A
Shabarakh-usu 10
3595 + 41
3.86-3.96
AA89878
1194 A
3246 + 39
3.42-3.54
AA89879
1609 A
5116 + 41
5.78-5.91
AA89880
1609 C
5061 + 49
5.75-5.88
AA89881
1702 A
Barun Daban
1661 + 42
1.53-1.62
AA89868
466A
Shara KataWell
8604 + 51
9.54-9.63
AA89885
2229 A
Baron Shabaka
Well
5609 + 47
6.34-6.44
AA89886
2231 A
5954 + 52
6.73-6.86
AA89889
2237 B
3115 + 47
3.28-3.38
AA89887
2231 C*
5825 + 85
6.53-6.73
AA89892
2526 A
Spring Camp
866 + 51
0.74-0.88
AA89897
2797 A
Chilian Hotoga
6728 + 45
7.56-7.64
Lab. No.
Dottore-namak
Mantissar 12
Ulan Nor Plain
Table 3.1 Results of chronometric dating. * indicates that the date may be unreliable due
to a low carbon yield of under 0.10 mg C (AA89873, AA89887). For site locations, see
Figure 3.3.
121
Alashan Gobi
Gobi-Altai
East Gobi
14000
12000
Oasis 1
10000
8000
Oasis 2
6000
Oasis 3
4000
2000
Metal Ages and Historic
J-k
G
Y-k a
Y-k b
Y-k c
D-n a
D-n b
M 12 a
M 12 b
S-u 1
S-u 4
S-u 10 a
S-u 10 b
UNP a
UNP b
BD
Ch-a a
Ch-a b
Ch-a c
Ch-a d
SKW
BS 19 a
BS 19 b
BS 19 c
BS 19 d
CH 35
0
Figure 3.1 Dates (cal yr BP/ka [2 σ]) plotted for Gobi Desert sites according to target
region and chronological period. Boxed site initials indicate dune-field/wetland sites.
The pattern of dune-field/wetland use is extremely interesting when compared
across regions. The first East Gobi dune-field/wetland sites are markedly earlier (7.6k cal
yr BP) than in the Gobi-Altai (4.9k cal yr BP) and Alashan Gobi (> 6.0 ka) and suggest
that intensive use of dune-field/wetland environments in the more verdant eastern Gobi
Desert may have occurred earlier than in the west. Alashan Gobi sites also tended to
return dates that were in a slightly later range than those from the Gobi-Altai. The series
of dates from all macroregions shows that the specialized use of dune-field/wetlands
environments continued as late as the early Bronze Age, and that these environments
were still occasionally exploited in historic times (Barun Daban, 374 + 49 CE [AA89881,
AMNH #73/1702A]).
122
While the earliest date on pottery from a dune-field/wetland site comes from Orok
Nor, which is in the Gobi-Altai region, the very early date (11.6k cal yr BP [AA89884,
AMNH #73/1792A; see Appendix A.3.) is suspect due to low carbon yields and the
anomalously high-fired quality of the pottery sampled. The carbon yield was only 0.19
mg C/0.04%, which is too low to be considered reliable (see above) and should be treated
with extreme caution. Unlike at other Gobi Desert sites, the ostrich eggshell at Orok Nor
produced younger dates (9.4 and 9.3k cal yr BP [AA89882, AMNH #73/1790A and
AA89883, AMNH #73/1790B]) than did the pottery. Due to the uncertainty of its
accuracy, the Orok Nor pottery date is not reported in Table 3.1. Two samples, AA89873
(AMNH #73/887A) from Shabarakh-usu 4 and AA89887 (AMNH #73/2231C) from
Baron Shabaka Well, did produce dates consistent with others from the same locality,
despite yields of 0.04% and 0.06% respectively. These dates are reported in Table 3.1.
AMS dates on ostrich eggshell suggest that the material is often not
contemporaneous with the Holocene archaeological assemblages with which it is found.
Table 3.2 compares chronometric dates on ostrich eggshell with dates on pottery from the
same site. The data suggest that middle Holocene peoples were indeed using ostrich
eggshell from earlier contexts. As such, we can no longer state that the ostrich eggshell
accurately dates associated human activities in the Gobi Desert (contra Janz et al., 2009).
It should still be noted that ostrich eggshell from Pleistocene contexts in Northeast Asia
has proven to be temporally consistent with associated artefact assemblages in some
cases (Jaubert et al, 2004; Madsen et al., 2001). Decreased availability of ostrich
eggshell following the extinction of ostriches sometime after 8.3k cal yr BP (according to
123
eggshell dates from Shabarakh-usu 1; see Appendix A.3) would explain the use of fossil
shell. In the middle Holocene the long-term importance of ostrich eggshell artefacts may
have encouraged Gobi Desert peoples to continue using the material for bead-making
thousands of years after the extinction of local ostriches. Clear knowledge of the timing
of ostrich extinction would allow us to predict the use of older shell.
Despite discontinuity between eggshell and pottery dates at most sites, numerous
eggshell dates from Shabarakh-usu are notably consistent. Fifteen out of 17 are within
1000 years of each other even though they are from many separate sites of a later age
(Table 3.3). This led to the original assertion that they must correctly date the
archaeological occupation (Janz et al., 2009). The dates are similar, but not close enough
to have been from the same egg or clutch of eggs. One explanation is that ostrich may
have been present in the region only for a brief interval of about a millennium (ca. 9.58.2k cal yr BP). Alternately, eggshells from that period may have been most accessible
to hunter-gatherers due to middle Holocene human use of lake environments that were
previously frequented by ostriches. Human populations contemporary with early
Holocene ostriches might also have exploited fresh eggs, the shells of which were
scavenged from middens by later inhabitants for bead-making. There have been verbal
reports of eggshell middens associated with microliths at this locality and possible early
Holocene occupations were reported by Soviet archaeologist A. P. Okladnikov (described
in Chard, 1974: 82). So far, few terminal Pleistocene or early Holocene aceramic
assemblages have been dated due to the lack of dateable materials.
124
Site
Yingen-khuduk
Ostrich Eggshell Date
(cal yr BP)
Pottery Date
(cal yr BP or ka)
45,646 + 1605
3910 + 300
5690 + 350
3910 + 230
Mantissar 12
> 49,900
6,460 + 700
> 48,500
3,840 + 340
> 48,800
Shabarakh-usu 1
8,295 + 64
4,900 + 46
9,515 + 22
Shabarakh-usu 4
8,399 + 25
4,027 + 105
9,484 + 39
9,244 + 91
Baron Shabaka
Well (Site 19)
14,829 + 298
6,388 + 52
14,714 + 333
6,795 + 67
6,632 + 102
3,332 + 53
Chilian Hotoga
(Site 35)
12,549 + 131
7,601 + 36
11,696 + 202
Table 3.2 Comparison of ostrich eggshell and pottery dates from Gobi Desert sites.
125
Shabarakhusu, Site #
1
2
4
7
14
C yr BP
Cal yr BP
648-01
δ13C
value
-10.4
7,483 + 47
8231-8359
AA89870
648-02
-8.4
8,522 + 50
9492-9537
AA76420
763-01
-10.3
8,159 + 59
9045-9223
AA76421
763-02
-9.6
8,184 + 44
9061-9229
AA76419
764-01
-9.1
7,969 + 37
8754-8950
AA76416
790-01
-9.0
8,396 + 52
9335-9467
AA76417
790-02
-11.1
8,268 + 44
9169-9367
AA76418
790-03
-10.7
30,490 + 780
33,323-43,775
AA89874
894-01
-10.0
7,589 + 47
8373-8424
AA89875
984-01
-10.0
8,473 + 64
9444-9523
AA89876
998-01
-10.0
8,254 + 47
9153-9335
AA76422
1034-01
-11.3
8,054 + 43
8823-9015
AA76423
1034-02
-11.6
38,600 + 1000
42,204-43,850
AA76424
1034-03
-10.7
8,439 + 60
9407-9509
AA76425
1035-01
-11.0
8,081 + 49
8888-9080
AA
Sample ID
AA89869
Table 3.3. Ostrich eggshell dates from Shabarakh-usu.
Although ostrich eggshell dates cannot currently be used to date Holocene
archaeological sites, chronometric dates on pottery can be used to help create a
chronology for site assemblages. Understanding technological and economic
developments in other regions of Northeast Asia are important in contextualizing Gobi
Desert assemblages, especially in the earliest periods when little data is available in any
one region. Without chronometric dates, it had been difficult to make appropriate
associations with data from neighbouring regions, which could be used to identify
temporally diagnostic artefacts and support a chronological framework. New dates for
Gobi Desert sites further allow us to compare recognizable changes in technology and
126
land-use with local palaeoecology in order to build hypotheses about subsistence and
economy.
3.2. Artefact-based chronologies
Artefact-based chronologies in neighbouring regions of Siberia and China are better
understood than those in Mongolia. Understanding the timing of cross-regional
developments such as the introduction of various technologies, decorative styles, or lithic
reduction strategies can help contextual Gobi Desert archaeology. Using better dated
chronologies from surrounding regions in conjunction with new chronometric dates on
Gobi Desert assemblages allows for the recognition of temporally diagnostic artefacts
and can provide a general age for undated assemblages.
3.2.1. Definition of “Neolithic” and issues in terminology
As outlined in Chapter 1, the majority of the archaeological sites discussed here are
described in the literature as either Mesolithic or Neolithic. The term “Neolithic” has
been employed differently across geographic regions. Etymologically, the word means
“New Stone Age”, and was originally used by John Lubbock in 1865 to describe a period
in technological development marked by the use of polished stone tools (Trigger, 1989:
94-95). In much of the world the term carries with it connotations of agricultural
production, sedentary village life, and high levels of social complexity, as characterized
by developments in the Middle East and Europe (Childe, 1953; Hodder, 1990; Renfrew
127
and Bahn, 1996; Liu, 2004). Russian and earlier Soviet scholars have defined the
Neolithic simply based on the first appearance of pottery (Chard, 1974; Kuzmin, 2003).
Due to the prevalence of Soviet and Russian research in Mongolia, the terms
“Mesolithic” and “Neolithic” have been used in the Gobi Desert to differentiate sites with
ceramics from those without ceramics. Since ceramic bearing sites are primarily found in
distinct environmental contexts, the term Neolithic may improperly represent the actual
chronological distribution of archaeological sites in the region (Janz, 2006). As such,
categorizations based solely on occurrences of pottery are misleading and impede
recognition of qualitative measurements by which to distinguish periodicity.
Based on its original use the term Neolithic is appropriately used to describe the
New Stone Age of Mongolia, replete with polished stone tools, and other examples of
technological development – small bifacially flaked points and a more varied tool kit. In
order to promote consistency in terminology and recognition of contemporaneous shifts
in human adaptation, the earliest post-LGM microblade-based assemblages in Northeast
Asia should be referred to as Epipalaeolithic, regardless of the presence or absence of
pottery. The period evidenced by a florescence of new flake tools, the more regular use
of pottery, and a reliance on ground and polished stone technologies should be referred to
as Neolithic. Finally, the following period is evidenced by incipient use and knowledge
of copper and bronze working and can be referred to as the Eneolithic (currently referred
to as Late Neolithic or Early Bronze Age). This system of nomenclature relies on a series
of technological indicators other than pottery and is most appropriate for the
128
developmental trajectory of Northeast Asia, as well as being more consistent with the
original etymology of existing terminology.
Within this framework, it is possible to introduce local terminological variants for
regional chronologies. In order to capture the Gobi Desert regional chronology more
precisely, a local terminology is proposed here. Although the use of regional
terminologies can be confusing due to differential recognition of archaeological cultures
across international borders, the use of a local terminological framework allows for
necessary distinctions between different subsistence economies or material cultures.
Accordingly, I have sought to construct a clear and descriptive terminology for the Gobi
Desert region by which we can define and categorize three chronologically distinct
periods of related organizational strategies in the Gobi Desert region of Mongolia and
China. Nomenclature for the three periods takes note of dune-field/wetland exploitation
across the Gobi Desert and beyond, and represents a span of time stretching from the Late
Epipalaeolithic to the Eneolithic: Oasis 1, spanning the Late Epipalaeolithic from 13.5 to
8.0 kya, this period represents incipient task-specific dune-field/wetland use; Oasis 2,
coinciding with the Neolithic and dating from 8.0 to 5.0 kya, this is a period of intensive
habitation of dune-field/wetland environments; and Oasis 3, spanning the Neolithic-toBronze Age transition or the Eneolithic, dating from 5.0 to 3.0 kya, this period is also
represented by continued habitation of dune-field/wetland environments, and represents
the transition from hunting and gathering to pastoralism. This terminology was designed
to eventually be applied to other arid regions of Central Asia should it prove suitable.
129
3.2.2. Chronological variation in technology, subsistence, and land-use
3.2.2.1. Early Epipalaeolithic (Upper Palaeolithic/Late Palaeolothic) – 19.0 to 13.5k cal
yr BP
The Early Epipalaeolithic is typified by the use of formal microlithic technology. Landuse seems to have been characterized by high mobility and utilization of a wide range of
environments. In Mongolia, low density scatters of both chipped macrotools and rough
microblade cores and flakes are thought to belong to this phase, more often referred to as
the Late Palaeolithic (see Derevianko [Ed.], 2000: 263-271). The earliest post-LGM
chronometric dates are 18.2k cal yr BP in northern Mongolia (Gladyshev et al., 2010) and
13.0k cal yr BP in the Gobi-Altai region (Derevianko et al., 2003). Although an overall
lack of radiocarbon dates for Mongolian sites may contribute to a perceived lack of early
post-LGM habitation, it is also possible that the region of modern Mongolia was very
sparsely populated until the terminal Pleistocene or early Holocene to which the majority
of microlithic sites appear to date.
Climate and subsistence
Post-LGM climate was highly variable in Northeast Asia, but showed increases in
average annual precipitation and temperature. Summer monsoon circulation probably
recommenced between 19.0-17.0k cal yr BP and lake levels stabilized across the region
(Herzschuh and Liu, 2007; Wünnemann et al., 2007). Large ungulate populations may
have declined during this period, as Pleistocene “megafauna” became extinct; however,
130
there is now evidence that some large Pleistocene fauna like mammoth and ostrich
survived into the terminal Pleistocene and early Holocene (Janz et al., 2009; Kurochkin et
al., 2009; Kuzmin, 2010). Increased precipitation, enlarged desert oases, lakes, and
stabilization of other diverse ecosystems are thought to have encouraged a more broadspectrum subsistence, in concert with a decline of large game (Lu, 1999: 16-17; Bettinger
et al., 2007).
Figure 3.2 Map of geographic locales mentioned in Chapter 3. Base map copyright of
maps.com, used by permission.
The sporadic use of milling stones and pottery, florescence of microblade
technology, and decline in the use of large points in various regions like China and Korea
have been cited in support of this theory (Lu, 1998; Madsen and Elston, 2007; Seong,
2008; Elston et al., 2011). In the Upper Yenisei/Western Sayan Mountains region (see
Figure 3.2) of southern Siberia zooarchaeological remains from Ui (Uy) I (see Figure
3.3), an LGM period habitation site, included remains of Bos primigenius (aurochs),
131
Equus hemionus (Asiatic wild ass/onager), Cervus elaphus (red deer) and Capra sibirica
(Siberian wild goat), while the similarly aged site of Tarachikha contained both large and
small species such as Mammuthus primigenius (woolly mammoth), Rangifer tarandus
(reindeer), Ovis ammon (Argali sheep), Bison priscus (steppe bison), Alopex sp. (fox),
Lagopus sp. (grouse), Marmota sp. (marmot), and Citellus sp. (squirrel) (Markin, 1998).
Early post-LGM period sites indicate a similar reliance on large game such as Alces alces
(moose), aurochs, and Equus ferus (horse), but also contain grinding stones and a range
of other resources such as eggs, Gulo gulo (wolverine), Lepus sp. (hare), fox, birds, and
fish (Vasil’ev and Semenov, 1993). In North China, faunal remains from Xiaonanhai,
dated to between about 15.7-12.8k cal yr BP (13,075 + 500, 11,000 + 500 BP), indicate
continued exploitation of large-bodied species such as Rhinoceros tichorhinus (extinct
rhinoceros), Equus hemionus (khulan), Cervus canadensis (elk or wapiti), and Bubalus
wansjocki (extinct water buffalo) (Tang and Gai, 1986). During the LGM, edible plant
species would have included mostly herbs and grasses with edible seeds (Lu, 1999).
Post-LGM adaptations in North China are thought to include high mobility and varied
land-use strategies that included dune-fields, wetlands, highland plateaux, and desertsteppes (Bettinger et al., 2007).
Post-LGM population expansion may have been facilitated by the florescence of
microblade/core technology (Goebel, 2002; but see also Barton et al., 2007). Scattered
evidence of microblade core technology has been found to pre-date 20.0 kya (see below),
but these examples are rare and occur in lithic assemblages containing mostly large tools.
Although the first evidence for the use of microblade cores appears just prior to 25.0k cal
132
yr BP in Northeast Asia, it is not until after about 17.5k cal yr BP (15.0k cal yr BP) that
microblade technology became widespread (Vasil’ev and Semenov, 1993; Chen, 2007;
Kuzmin, 2007; Norton et al., 2007; Seong, 2008; Gladyshev, et al., 2010). Microblade
core technology became a dominant core reduction technique in Northeast Asia by 15.411.5k cal yr BP (13.0-10.0k yr BP) and assemblages from this period are characterized by
a diversity of microblade core types (Aikens and Akazawa, 1996; Ackerman, 2007;
Elston et al., 1997; Lu, 1998; Xia et al., 2001; Cohen, 2003; Barton et al., 2007).
Figure 3.3 Map of archaeological sites mentioned in Chapter 3. Base map by maps.com.
1. Maina; 2. Ui; 3. Ust’-Khemchik; 4. Toora Dash; 5. Cheremushnik; 6. Ust’-Belaya; 7. Saganzaba; 8. Ulan Khada; 9. Oshkurovo; 10. Ust’-Kyakhta; 11. Studenoe; 12. Altan Bulag; 13. Ust’Karenga; 14. Tolbor; 15. Moil’tyn am; 16. Orok Nor; 17. Chikhen Agui; 18. Barun Daban; 19.
Ulan Nor Plain; 20. Shabarakh-usu; 21. Mandal Gobi/Ulan-khovor; 22. Yingen-khuduk; 23.
Dottore-namak; 24. Ukh Tokhoi sites; 25. Mantissar sites; 26. Jabochin-khure; 27. Gashun Well;
28. Camp Ruined Lamasary (Site 11/11A); 29. Baron Shabaka sites (sites 19, 20, 21); 30. Jira
Galuntu (Site 18); 31. Shara KataWell; 32. Spring Camp (Site 16); 33. Alkali Wells (Site 26); 34.
Paoling Miao Southeast (Site 31); 35. Chilian Hotoga (Site 35); 36. Dulaani Gobi; 37.
Tamsagbulag; 38. Khutyn-bulag; 39. Ovoot; 40. Hail’er; 41. Daxingtun; 42. Tengjiagang; 43.
Ang’angxi; 44. Xinkailiu; 45. Zuojiashan; 46. Yaojingzi; 47. Fuhegoumen; 48. Chahai; 49. Xinle;
50. Houwa; 51. Xibajianfang; 52. Xinglongwa; 53. Zhaobaogou; 54. Menjiaquan; 55. Hutouliang;
56. Yujiagou; 57. Youmafang; 58. Nanzhuangtou; 59. Cishan; 60. Xiaonanhai; 61. Peiligang; 62.
Xiachuan; 63. Chaisi; 64. Xueguan; 65. Dadiwan; 66. Pigeon Mountain Basin.
133
Early Epipalaeolithic technology in adjacent Northeast Asia
The archaeological record of Northeast Asia suggests regional continuity in lithic
assemblages throughout the duration of the LGM and early post LGM. Microblade core
reduction had been widely adopted at this time, but the use of expedient flake core and
blade core reduction sequences was more common (Vasil’ev and Semenov, 1993;
Markin, 1998; Lu, 1998; Norton et al., 2007; Sano, 2007; Ikawa-Smith, 2008; Chen,
2010). A trio of tool types – microblades, flake tools and heavy-duty tools – is identified
with the LGM and early post-LGM, as recovered at Mengjiaquan in the North China
Plain 20.9k cal yr BP (17,500 + 250) (Lu, 1999). The pre-LGM period in Japan (Late
Palaeolithic II) is characterized predominantly by the use of formal blade cores and
retouched blade tools (Ikawa-Smith, 2008). During the LGM points, including backed
points, were typical (Yuichiro, 2005), while later sites containing pottery (16.7-15.7k cal
yr BP) are variously associated with microblade core technology, partially ground and
chipped axe/adzes, large leaf-shaped points, gravers, scrapers and blades (Keally et al.,
2003). In Korea, tanged points were common in pre-LGM and LGM assemblages, but
became rare in the post-LGM period (Seong, 2007, 2008). At the pre-LGM Chaisi site
(ca. 30.0k cal yr BP) in North China, microblade cores were associated with large tools
such as choppers, scrapers and bolas, while post-LGM sites contained large cores, blade
cores, microblade cores, unifacial points, small flake tools, and pebble tools like
choppers, hammers, and chipped adzes (Chen and Wang, 1989; Lu, 1998).
Following the LGM core reduction strategies included many of the same elements
found in preceding periods, but an increased focus on the use of microblade cores and a
134
variety of new types are often noted. Several Japanese early post-LGM sites, dated to
around 18.6k cal yr BP (15,470 + 190), show a shift in microblade core types from subconical, to boat-shaped, to wedge-shaped (Sano, 2007). In Korea, wedge-shaped cores
are considered to be the oldest type of microblade core technology, while smaller cores
and more simplified reduction strategies are considered to be later developments (Norton
et al., 2007). Chinese early post-LGM assemblages offer a contrast to the more common
pattern of increasing variety in core forms. At the Xiachuan site (ca. 25.0-20.0k cal yr
BP [21.0-16.5k BP]) microblade cores were mostly conical with a few wedge- and boatshaped types (Tang and Gai, 1986; Chen and Wang, 1989; Lu, 1998; Tang, 2000; Chen,
2007), while the later Xueguan assemblage (16.5k cal yr BP [13,550 + 150 BP])
primarily contained wedge-shaped microblade cores and only a few of the conical form
(Lu, 1998). Early microblade core reduction techniques appear to have been more
diverse in early North China sites (e.g., Xiachuan and similar undated assemblages from
the lower Yellow River Valley – see Lu, 1998) than in other regions. The increasing
frequency of microblade cores in post-LGM assemblages indicates a corresponding
reliance on the use of microblade insets for organic hafts.
Large unifacial points are found in assemblages spanning the LGM and early
post-LGM in several regions. The lack of reliable dates limits our interpretation of supraregional patterns in the use of such points, but they are often associated with the LGM.
At Xiachuan they are found with both microblade cores and backed knives made on blade
tools (Lu, 1998). Points were similarly associated with microblade and blade industries
at the pre-LGM Ui I in southern Siberia (Vasil’ev and Semenov, 1993). Large tanged
135
points are indicative of late Pleistocene tool kits in Korea until the end of the LGM
(Seong, 2007, 2008). Judging by the age of those sites, it is probable that large unifacial
points are diagnostic of the LGM and very early post-LGM, but were not commonly used
in later periods. Following the LGM, the increasing reliance on microblades as insets in
composite points may be related to a decline in the use of unifacial points.
Microblade core technology
Methods of microblade core reduction in Northeast Asia developed over time and the
earliest manifestation of the technology differs from later post-LGM developments.
Despite the scarcity of reliable chronometric dates, several authors have attempted to use
microblade core morphology and reduction sequences to create an artefact-based
chronology within Northeast Asia (Tang and Gai, 1986; Chen and Wang, 1989; Seong,
1998). It is clear that wedge-shaped8 microblade core technology is the most widespread
in early assemblages, and the only type of microblade core present in pre- and early postLGM Siberian and Mongolian assemblages (Vasil’ev and Semenov, 1993; Slobodin,
1999; Gladyshev et al., 2010).
Tang and Gai (1986) outline a progressive developmental technology for North
China based on the relative frequency of certain cores types from directly and indirectly
8
Unfortunately, the term “wedge-shaped” has been used to describe a morphology that is found supraregionally during all time periods and can include several types of very different reduction strategies.
Seong (1998) outlines several reduction strategies recognized by Japanese and Chinese archaeologists that
can produce wedge-shaped cores (see also Morlan, 1976), but argues that analysis of lithic assemblages
should focus on flexible typologies based on perform formation, platform preparation and blade production
techniques. Pleistocene-type wedge-shaped cores made on heavily prepared nodules are herein referred to
as “formal wedge-shaped cores.”
136
dated archaeological sites (for alternate nomenclatures see Morlan, 1976; Seong, 1998).
The earliest incarnation of microblade core technology in China was found at Xiachuan,
where the most prevalent core forms were conical and boat-shaped. As described by
Chen and Wang (1989), “conical cores” show variable shapes at discard. They are
described as being made on small chunks with platforms trimmed from various angles.
Cylindrical cores are similar, but with opposed platforms. Boat-shaped cores were
similar to wedge-shaped, but were prepared from platform to distal end and lack an
intentionally worked keel edge. The technique is reminiscent of the Horoka type from
Hokkaido (Morlan, 1967; Seong, 1998).
Microblade cores from Xiachuan are considered to represent some of the most
ancient examples of the technology in China, but the relative frequency of core types
from the site is atypical of other LGM sites in Northeast Asia. In post-LGM Chinese
sites, more formal wedge-shaped microblade cores are increasingly dominant, while
conical cores are much less common or absent (see Chen and Wang, 1989). Conical
cores are usually more common in Northeast Asian Holocene sites, where wedge-shaped
cores are comparatively rare. Holocene wedge-shaped cores also vary in form from their
Pleistocene predecessors. They are usually made on minimally prepared cobbles or
rectangular nodules, and form thick-bodied performs with a keel edge in the back
(Morlan, 1976; Chen and Wang, 1989). Aside from the keel, the exhausted form is quite
similar to Xiachuan conical cores. Considering the range of other artefact types
discovered at Xiachuan, including axes and grinding tools, the pre-LGM dates are
extremely precocious.
137
Early Epipalaeolithic of Mongolia
In northern Mongolia, flake-based industries replaced blade-based ones around the same
time that pressure flaked microblade cores were first used. Recent evidence from the
open-air Tolbor sites suggests that this transition occurred by about 25.0k cal yr BP
(Gladyshev et al., 2010). Pressure and percussion-flaked microcores became common
during the post-LGM period, along with increasing numbers of microblades. Retouched
points on flakes were also more common, a pattern that is repeated elsewhere in
Northeast Asia. Little is known about the timing of this transition (Gladyshev et al.,
2010).
Moil’tyn am is an important stratified site on the bank of the Okhotsk River near
the ruins of the medieval Mongol city of Karakorum in central Mongolia. The site is
undated, but lithic assemblages appear to date from the Middle Palaeolithic to
Epipalaeolithic (Okladnikov, 1981). Layer 3 may correspond to the early Upper
Palaeolithic (after Gladyshev et al., 2010) and is notable in the occurrence of
subprismatic and irregularly shaped microcores with elongated flake scars reminiscent of
microblade technology (Okladnikov, 1981: 289). Heavy-duty tools made on pebbles,
prepared and expedient blade and flake cores, points and other retouched tools make up
the majority of the finds. A prepared core strategy reminiscent of the Levallois technique
is common. Pebble, flake and blade tools were found in Layer 2, but blades and
elongated flakes are most characteristic.
Layer 1, which might be tentatively assigned to the Early Epipalaeolithic,
indicates the continued use of heavy-duty pebble tools, along with more extensive
138
retouch of flakes and a few true microblades and microblade cores. Microblade cores are
mainly wedge-shaped and sub-conical, and are most consistent with forms described for
the Xiachuan assemblage. The Dno Gobi (Дно Гоби9) locality, Sites 2 and 3 contain a
range of tool types, including points, bifacially worked macrotools, flake tools and many
blade and blade tools, as well as both conical and formal wedge-shaped cores
(Okladnikov, 1986: 168, 200-201). Some elements of these sites appear to pre-date the
LGM, but may date to as late as the early post-LGM.
Summary of Early Epipalaeolithic
Early Epipalaeolithic assemblages in Northeast Asia tend to focus on the production of
large retouched flake and blade tools. Microblades were largely unretouched and were
probably used primarily as knife blades or projectile inserts for organic hafts. Microblade
cores were of a formal wedge-shaped variety throughout southern Siberia and northern
Mongolia, while boat-shaped and conical cores were also used in northern China and in
early post-LGM Japanese assemblages. Large unifacially retouched points are found in
LGM and early post-LGM assemblages in some regions of Northeaast Asia, but do not
appear to have been common throughout the Early Epipalaeolithic. Mongolian Early
Epipalaeolithic assemblages are characterized by the use of both microblade core
technology and retouched flake tools.
9
For ease of recognition, names are given in Cyrillic when they are transliterated directly from the Russian
or Mongolian literature and have not been previously used in English publications.
139
Considering the evidence from northern Mongolia and northern China, we expect
that Early Epipalaeolithic lithic assemblages in the Gobi Desert would be typified by
retouched flake tools, and both microblade and blade core technology. Large unifacial
points could also be present in LGM or very early post-LGM assemblages. Microblade
core technology might be represented by wedge-shaped, boat-shaped and conical core
technology, with an emphasis on wedge- and boat-shaped forms. So far, no
archaeological sites have been reliably dated to this period and none of the undated sites
are typical of the Early Epipalaeolithic. Ostrich eggshell from the East Gobi has been
dated to the Early Epipalaeolithic (see Appendix A.3), but the temporal association
between dates and the artefacts is dubious. The region may have been quite sparsely
populated during and immediately following the LGM, with increased population only in
the terminal Pleistocene.
The tendency towards increased use of microblade core technology is typical of
the Early Epipalaeolithic. A decline in the use of earlier lithic types like large unifacial
points and, in northern Mongolia and Siberia, blade cores, is also notable within postLGM tool kits. Technological shifts underlying changing subsistence strategies are not
as clear as suggested for the Russian Far East and Japan by the use of pottery, but the
more widespread use and importance of microblade core reduction sequences within
Early Epipalaeolithic industries is an important characteristic that may reflect changes in
early post-LGM foraging strategies. It has been suggested that the consistency with
which microblade technology gradually became dominant over flake and blade tool
technologies during and following the LGM, was related to the extremely cold seasonal
140
temperatures and uneven distribution of resources that would have characterized the
period in northern latitudes (e.g., Elston and Brantingham, 2002; Seong, 2007, 2008). As
a reliable and efficient reduction strategy (sensu Elston and Brantingham, 2002),
microblades could have served foragers as an important reliable hunting technology in
circumstances where food resources were less reliable during the post-LGM period with
fluctuating climatic conditions and increasing seasonality (Seong, 2007).
3.2.2.2. Late Epipalaeolithic/Oasis 1 (Mesolithic/Early Neolithic) – 13.5 to 8.0k cal yr
BP
After about 13.5k cal yr BP a distinct shift organizational strategies, typified by
diversification in subsistence and technology, occurs widely across Northeast Asia. A
reliance on microblade core reduction strategies is representative of the Late
Epipalaeolithic. Pottery becomes more widespread and is nearly ubiquitous by the end of
the Late Epipalaeolithic. Regional developments in Northeast Asia appear to be more
complex and diverse during this period. Stylistic and technological relationships between
assemblages from the Gobi Desert and southern Siberia and northern China have been
consistently emphasized in the archaeological literature and, along with their geographic
proximity, make them important regions for contextualizing Mongolian and Gobi Desert
post-LGM archaeology.
141
Climate and subsistence
The Late Epipalaeolithic began during a period of extreme low moisture availability
corresponding to the Younger Dryas, lasting from 13.0 to 11.6 kya in monsoonal Central
Asia and 12.0 to 10.0 kya on the Loess Plateau (Madsen et al., 1998; Herzschuh, 2006).
The continuing post-LGM trend towards warmer, wetter conditions recommenced
following the Younger Dryas with strong intensifications of both the Indian and
Southeast Asian Monsoons at about 11.5k cal yr BP, along with an increase in seasonality
that continued until after 9.6k cal yr BP (Herzschuh, 2006). Increases in monsoonal
precipitation caused the boundary of arid lands in Northeast Asia to retreat dramatically
(Starkel, 1998; Feng et al., 2007). Increased seasonality during the Late Epipalaeolithic
may have limited the range of some species that were previously widespread. At the
same time increased effective moisture would have resulted in the infilling of lake basins
and river channels as well as the stabilization of alluvial and aeolian deposits formed
during the LGM and terminal Pleistocene (see Owen et al., 1997; Hülle et al., 2009).
Archaeological faunal assemblages from the Late Epipalaeolithic are more varied
than in the Early Epipalaeolithic, with a decreased emphasis on large-bodied game. The
increasingly widespread use of pottery and grinding stones also indicates a shift in
subsistence practices and land-use. In southern Siberia, faunal remains indicate that the
exploitation of cervids and fish were increasingly important during the Late
Epipalaeolithic (see Michael, 1984; Kirillov and Derevianko, 1998). Studenoe, a
stratified site near the Mongolian border, exemplifies the transition in subsistence
economy that occurred in much of Northeast Asia during the Late Epipalaeolithic. The
142
site was used from about 21.6k cal yr BP to about 2.3k cal yr BP and was probably
primarily inhabited throughout that time as a seasonal hunting camp (Buvit et al., 2003).
Fauna from early layers (13.2-13.7k cal yr BP [11,300 + 100 to 11,400 + 200 BP, 11,40011,800 years ago]) included forest, montane and steppe species such as Cervus elaphus
(red deer), Capreolus capreolus (roe deer), Poephagus baikalensis (Baikal yak), Bovinae
(bovines; aurochs or bison), and Capra sibirica (Siberian mountain goat) (Kirillov and
Derevianko, 1998). Later Epipalaeolithic layers probably date until about 9.0k cal yr BP
and contain mostly forest species such as Alces alces (moose), Cervus elaphus (red deer),
Capreolus capreolus (roe deer), and fish species such as Siberian elets, Lota lota
(burbot), and Esox sp. (pike) (Kirillov and Derevianko, 1998). The lowest cultural layer
(XI) of the Ulan Khada site near Lake Baikal is thought to date to the very end of the
Late Epipalaeolithic (Mesolithic). Here, the use of fishing equipment and seal bone
(Pusa sibirica) further support the increasing importance of aquatic resources.
A different series of adaptations are represented by faunal assemblages from sites
in North China. At Hutouliang the incorporation of small- to medium- bodied fauna –
Rana sp. (frog), Citellus citellus (ground squirrel), Cricetulus varians (hamster), Vulpini
(fox), Sus scrofa (wild boar), Canis lupus (wolf), Cervus sp. (deer), Procapra
picticaudata (goa or Tibetan gazelle), Gazella subgutturosa (goitered gazelle) – into a
diet already rich in large ungulate species such as Equus hemionus (khulan), Equus
przewalskyi (wild horse), Bos sp. (cattle) suggests diversification of the existing diet (Gai
and Wei, 1977; Chen and Wang, 1989; Lu, 1999). Faunal assemblage from the terminal
Pleistocene to historic period site of Yujiagou (based on TL and 14C, estimated to date
143
from 13.7 to 2.1 ka), also in the Sanggan River region, contains an almost identical range
of species but with a special focus on gazelle/antelope exploitation (Xia et al., 2001; Wu
and Zhao, 2003). Possible early evidence of animal domestication also occurs during this
period in the lake and marsh deposit of Nanzhuangtou (12.8-9.9k cal yr BP [10,815 + 140
and 8800 + 108 BP]) (Lu, 1999; Cohen, 2003). The site is thought to have been used for
processing and cooking meat. Faunal remains included Canis lupus (wolf), Sus scrofa
(wild boar or possibly domesticated pig), Canis familiaris (dog), Gallus sp. (possible
domesticated fowl), and high frequencies of deer (Cervus elaphus [red deer], Elaphurus
davidianus [Pere David’s deer/Milu], Capreolus capreolus [roe deer], Cervus nippon
[sika deer]) (Lu, 1999; Cohen, 2003). Fish, bird, soft-shelled turtle and shellfish remains
were also recovered (Cohen, 2003; Wu and Zhao, 2003).
A change in the focus of subsistence is probably related to local shift in vegetation
during the terminal Pleistocene, which would have influenced species composition (see
Lu, 1999: 16-17). Megafauna and steppe species like mammoth, rhinoceros and equids
were present in Early Epipalaeolithic sites in southern Siberia, but were probably no
longer available in the more heavily forested terminal Pleistocene environments
(Michael, 1984; Derevianko and Markin, 1998; Kuzmin, 2010). Hunters would have
begun to target large- to medium-bodied cervids as an alternative. Small fur-bearing
species appear to have been exploited prior to the LGM and in the Early Epipalaeolithic,
but are less often noted in Late Epipalaeolithic faunal assemblages (see Vasil’ev and
Semenov, 1993; Markin, 1998). Stratified sites in Siberia suggest that hunter-gatherers
144
may have maintained a set pattern of land-use, but adjusted species selection according to
local faunal diversity.
Likewise, data from North China indicate the continued use of steppe-dwelling
ungulate species with the addition of small- and medium-bodied species from a variety of
new environments like forests and marshlands. Foragers in the arid Gobi Desert and
northwestern China (Bettinger et al., 2007) are expected to have relied on a comparable
range of species from various environments, such as newly formed marshlands around
stabilized lake and dune-fields. Thus, the Late Epipalaeolithic is suggested to have
represented the incipient stages of an oasis adaptation that became central to Holocene
human land-use in arid Northeast and Central Asia.
Late Epipalaeolithic technology in Northeast Asia
Microblade-based tool kits dominate Late Epipalaeolithic assemblages across Northeast
Asia. Expedient flake technology continued to be used, but blade cores were no longer
common in lithic assemblages. The importance of microblade technology coincides with
an overall emphasis on the production of small flakes and tools and a corresponding
decline in the use of macrotools. Pressure-flaking was a common method of retouch.
Homogeneous cryptocrystalline stone such as jasper and chalcedony were most highly
favoured and were probably more suited for producing the small formal microblade cores
and finely retouched tools than more coarse-grained stone.
Assemblages from southern Siberia represent a series of technological changes
that are broadly consistent with those in Mongolia. The Late Epipalaeolithic (locally
145
referred to as the Mesolithic) of the Lake Baikal region is best known and is
characterized by the increasing frequency of microblade cores and a diversity of
microcore types, including less formal prismatic cores. Other markers of the period
include the disappearance of prepared flake core technology (“Levallois”), declines in the
number of large points, less varied scraper types, retouched microblades, and the
occasional use of new technologies such as polished stone, bifacially worked tools, and
fish hooks (see Michael, 1984; Kirillov and Derevianko, 1998). As in the rest of
Northeast Asia, polished stone, pottery, and grinding stones became more common. The
earliest evidence of pottery in the region comes from around Lake Baikal, where
ceramics were found at several Late Epipalaeolithic sites10: Studenoe 1, layers 6 and 7;
Ust’-Kyakhta; and Ust’-Karenga, Layer 7 (Kuzmin and Orlova, 2000). The first use of
polished stone, especially in the manufacture of fishing gear, also dates to the end of the
Late Epipalaeolithic.
The earliest stage of Late Epipalaeolithic technology is represented at
Cheremushnik, where lithic assemblages indicate the continued use of Early
Epipalaeolithic tool kits with the introduction of bifacially worked tools, projectile points
(“arrowheads”), and axes (Michael, 1984). The stratified Oshkurovo site near Ulan-Ude
is thought to date to between about 15.1-12.6k cal yr BP, the later layers exhibiting an
increase in the number of worked small and medium sized pebbles, the addition of
prismatic cores, and microblades with lateral retouch (Kirillov and Derevianko, 1998).
10
All ceramic bearing sites are traditionally referred to as “Neolithic” in the Soviet/Russian literature.
Such sites should be re-evaluated for consistency of technological associations and subsistence before
being grouped with other Neolithic sites.
146
Upper layers of stratified assemblages from Ust’-Belaya post-date 10.1k cal y r BP (8960
+ 60 BP) and likewise indicate increasingly diverse microblade core forms, along with
new technological developments like polished stone, small bifacial projectile points
(“arrowpoints”) and fishing hooks (Michael, 1984). Pre-dating 8.5k cal yr BP, the lowest
level of Ulan Khada represents the Epipalaeolithic to Neolithic transition and indicates a
heavy dependence on microblade core technology, fishing technology, polished
adze/axes, and temporally diagnostic core tools from which burin spalls were taken to
make “drills”. Sinkers and blanks for hooks underlie evidence of the importance of
fishing and a piece of seal bone indicates exploitation of the Baikal seal or nerpa (Pusa
sibirica) (Khlobystin, 1969; Kuzmin and Orlova, 2000).
Lithic assemblages from the Late Epipalaeolithic of North China are similarly
characterized by the increased use of microblade reduction strategies, but with the
continued use of flake and heavy-duty tools. Wedge-shaped and boat-shaped microblade
cores were dominant. The earliest type of formal wedge-shaped core is identified by the
Hutouliang technique (Saikai/Fukui)11. Other wedge-shaped core types were also found
in Late Epipalaeolithic sites along with boat-shaped and the more rare conical types. The
He Tai technique (Yubetsu) may have developed in northern China following the
Hutouliang type and was characterized by the preparation of a bifacial core, from which
ski-shaped spalls were longitudinally detached to form a platform that was generally not
rejuvenated during the process of reduction. The third variant was the Sanggan technique
11
The technique is thought to have been represented in an early form at Xiachuan (25.0-20.0k cal yr BP)
and then later at Hutouliang (12.9k cal yr BP). In this variant, the platform was shaped by transverse blows
on a unifacially flaked blank, with microblades detached from one end of the platform (Tang and Gai,
1986).
147
(Oshoroko), where small spalls were removed from the tip of a bifacially prepared core to
create a successively rejuvenated platform (for illustration, see Seong, 1998) (Tang and
Gai, 1986). All three types were found at the Hutouliang localities, near the Sanggan
River, Hebei Province12.
New dates from Hutouliang suggest that pottery may have been used as early as
16.0k cal yr BP (13,080 + 200) (Yasuda, 2002), but early pottery is more often
recognized at Late Epipalaeolithic sites. Three Late Epipalaeolithic pottery sites are
Yujiagou, Toumafang, and Nanzhuangtou. Pottery at Yujiagou, in the Hutouliang area of
the Nihewan Basin, was associated with a thermoluminescence (TL) date of 11.6 ka
(11,870 + 1720) (Tang, 1997; Xia et al., 2001). East of Yujiagou, near Nihewan village,
the Toumafang site has been dated to 8.3k cal yr BP (7530 + 100 BP). Artefact
assemblages include not only microblades and microblade cores, but small flaked tools, a
chipped adze, pottery, grinding slab fragments, a few polished stone tools, and a large
number of flaked and polished bone and antler tools (Lu, 1999). Likewise, the
Nanzhuangtou site contains not only stone hammers and cores/flakes, but an array of new
artefact types that are most often recognized in later periods, including variously
decorated pottery shards (cord-marked, punctuate, and applied decorations), grinding
stones, a roller or pestle, worked wood, a bone awl, and an antler drill (Lu, 1999; Cohen,
2003; Wu and Zhao, 2003). Interestingly, there are no microblades.
12
The best known date for the assemblages is 12.9k cal yr BP (or 12.5k cal yr BP based on 10,690 + 210 in
Lu, 1998, 1999), but more recent dates on pottery from Hutouliang indicate an even earlier age of about
16.0k cal yr BP (13,080 + 200) for at least one of the sites (Yasuda, 2002).
148
Lithic assemblages in northwest China are distinct, but follow a similar pattern to
that farther east. At Pigeon Mountain Basin in the Helan Shan foothills a milling stone
was found above sediments dated to 13.5k cal yr BP (11,620 + 70 BP) (Elston et al.,
1997). Microlithic technology was also dominant after about 13.5k cal yr BP,
corresponding with a decline in the use of macrotools (Bettinger et al., 2007). Retouched
microblades occurred only in the later layers, dated to at least 11.7k cal yr BP (10,130 +
70 BP), along with an “arrowpoint,” which appears from the illustration to have been a
small unifacially retouched flake (see Elston et al., 1997). Macrotools included gouges
on cobbles, scrapers, flake tools, blade tools, spheroids, debitage, and roughly chipped
Helan points resembling artefacts from Hutouliang, Xiachuan, and Xueguan (Elston et
al., 1997). A shift in raw material preferences was also noted, in that the rise of
microlithic technology appears to have corresponded with a shift from the use of
quartzites and coarse-grained metavolcanics to crypocrystalline cherts and chalcedonies
(Bettinger et al., 2007).
In Northeast China (Manchuria/Dongbei), the first use of microblade technology
is later than in southern Siberia and North China. Few sites pre-date the Neolithic period,
when evidence for sedentary habitation is found along with developed Neolithic stone
tools and pottery. The two earliest post-LGM sites are Xibajianfang (Bajianfang),
Liaoning Province, and Daxingtun, Heilongjiang Province. Xibajianfang contains mostly
extant species with the exception of Bos primigenius (aurochs) and is thought to date to
the terminal Pleistocene or early Holocene. There were small tools, but no microblades
(Jia, 2007). At Daxingtun (13.7k cal yr BP [11,800 + 150 BP]), only one prismatic
149
microblade core and 14 truncated microblade fragments were reported (Lu, 1998) and the
status of these artefacts as true microblades has been questioned (Jia, 2007). Small
flakes, pebble tools, and microblades are thought to be typical of terminal Pleistocene
assemblages, although the latter appear to have been less widely distributed. Direct
percussion was used for detaching flakes, while indirect percussion was used for retouch.
Flaked adzes, knives, and grinding slabs are not included until the early Holocene (Lu,
1998). As in southern Siberia, pottery was first introduced during the Late
Epipalaeolithic and is associated with small tool/microlithic technology. Core typology
appears to have been more diverse than in the Yellow River and North China Plain
regions (Lu, 1998), which may be related to the use of less standardized core forms.
Considering the late occurrence of microblade technology and the developed nature of its
first appearance, it can be suggested that microblade technology diffused to Northeast
China from neighbouring regions in the terminal Pleistocene.
Late Epipalaeolithic of Mongolia and Oasis 1 technology
Evidence of Late Epipalaeolithic occupation has been found throughout Mongolia,
although chronological determinations have been based largely on typology rather than
chronometric dating. As in Northeast China, there is a scarcity of Early Epipalaeolithic
archaeological sites, but available dates indicate that Epipalaeolithic microblade core
technology does pre-date developments farther east. The earliest dated post-LGM
assemblage is from the Chikhen Agui cave site (13.4 to 8.7k cal yr BP [11,545 + 75 to
150
7850 + 100 BP]) (Derevianko et al., 2003). New radiocarbon dates on pottery also place
the Shara KataWell site at the end of the Late Epipalaeolithic (9.6k cal yr BP, Table 3.1).
Other Mongolian Late Epipalaeolithic sites are recognized, but undated. Ostrich
eggshell dates from Shabarakh-usu (Janz et al., 2009), Orok Nuur, Chilian Hotoga (Site
35), and Alkali Wells (Site 26) fall into the Late Epipalaeolithic, but ostrich eggshell
dates are often problematic due to the use of sub-fossil shells. More probable Late
Epipalaeolithic sites in Mongolia include Bygat-2 (possibly an early or transitional late
Epipalaeolithic site, see Gladyshev, 1987), the lower levels of Dulaani gobi (Tseveendorj
and Khosbayar, 1982), a slightly weathered artefact group from Ulan-khovor/Mandal
Gobi (Govï) (Gábori, 1963; Kozłowski, 1972), Kerulen, Site 9 (Dorj, 1971: 29-30, 111),
and Altan Bulag, Horizon I (Gábori, 1963).
Late Epipalaeolithic assemblages in Mongolia include highly developed
microblade core technology based on a diverse array of core types, as evidenced by the
Chikhen Agui assemblage (see Derevianko et al., 2003). Classic Epipalaeolithic wedgeshaped microblade cores typified by the Yubetsu (He Tao), Hutouliang (Saikai), and
Togeshita (Yangyuan) techniques are not found in Late Epipalaeolithic assemblages, and
wedge-shaped cores seem to conform more closely with the Yadegawa (Nodake)
technique described by Seong (1998), where little specific platform preparation was
conducted on the flat surface of the blanks. The technique employed might also be
compared to that used in the manufacture of conical cores at Xiachuan. Notably, many of
the Gobi Desert wedge-shaped cores are bifacially flaked on the opposite end of the
flaking surface (or flute), allowing them to have been used as cutting tools (though many
151
do not show clear evidence of usewear) (see Maringer, 1950; Morlan, 1976; Fairservis,
1993). The same technique is also evidenced at the probable Early Epipalaeolithic
locality Dno Gobi, Sites 2 and 3 (Okladnikov, 1986), suggesting the long-term local
utilization of both formally prepared wedge-shaped and flexible conical core reduction
strategies in Gobi Desert assemblages.
The Shara KataWell site in the East Gobi yielded two cores, both of which are
wedge-shaped. These specimens appear to have been made on flat cobbles of chalcedony
with the exterior surface removed transversally from the sides and on the edges to create
a rough D-shaped blank. Aside from the removal of short spalls, there was little or no
platform preparation before microblades were detached from one end of the short axis.
The technique bears similarities to the Togeshita (Yangyuan) technique reported from
Hutouliang (Gai, 1984; Seong, 1998). Associated pottery (fibre-tempered with quartzite
inclusions and light cord-marking) dates this site to about 9.5k cal yr BP. Reduction
strategies and the more easterly locale of Shara KataWell may suggest technological
influences from developed industries farther south.
Late Epipalaeolithic Mongolian assemblages are more typically characterized by
smaller cores and less standardized platform preparation than other Northeast Asian
Epipalaeolithic assemblages. Flexibility in manufacture appears to have been typical for
later post-LGM Mongolian archaeological sites. The use of more flexible reduction
strategies was probably aimed at exploiting a range of widely available, but differently
shaped, raw material packages such as small cobbles of high quality stone characteristic
of the Gobi Desert. The result is a wide diversity of core types. Such a situation is not
152
unique, as assemblages in southern Siberia also seem to evidence a greater diversity of
core forms by this time.
Finely retouched microblade tools also became common in the Late
Epipalaeolithic. Microblades were not only used as inset blades in composite tools, but
were retouched using indirect percussion to create multiple tools types. Assemblages
recovered from Chikhen Agui Horizons 2 and 3 show that 26% and 18%, respectively, of
all used and/or retouched flakes were made on microblades retouched using pressureflaking techniques (Derevianko et al., 2003: Table 2). Microblades were retouched along
one or more lateral edges or ends. Of the total artefacts from Horizon 2, unused
microblades (< 10 mm wide) and bladelets (11-14 mm) made up 32% of the entire
assemblage, while in underlying Horizon 3, unused microblades and bladelets made up
23% of the entire assemblage (Derevianko et al., 2003). While both layers indicate the
regular use of microblades, the increased frequencies of both retouched and unretouched
microblades in Horizon 2 underscores the growing importance of microblades within
post-LGM tool kits. A greater reliance on microblade technology is in contrast to preLGM and Early Epipalaeolithic assemblages, which were based primarily on blade and
large flake technology, including production from prepared Levallois-like cores.
Refinement of existing technological traditions and increasing reliance on
microblade core reduction strategies at the end of the Mongolian Early Epipalaeolithic
suggest continuity between those assemblages and the more widespread Late
Epipalaeolithic sites. A refinement in existing methods of microblade core production is
evident in Horizon I of Altan Bulag, in the Selenga River Basin of northern Mongolia
153
(see Gábori, 1963: Planche III). A widespread pattern of increasing post-LGM
microblade use, resulting in more efficient methods of manufacture, partially explains the
technological similarities between Neolithic assemblages in the Gobi Desert and Lake
Baikal region that have been widely noted in the literature on Mongolia and Northeast
Asia (i.e., Gábori, 1963; Maringer, 1963; Chard, 1974). Due to the continuity of foreststeppe and river valleys south of the boreal forests between southern Siberia and
Mongolia, a natural path of early expansion and later contact and trade may have existed
between Siberia and northern Mongolia, possibly extending south across the open steppes
and river valleys into the western Gobi Desert. Likewise, such contact between the more
eastern parts of Mongolia and southern Siberia would explain similarities noted between
those regions late in the Neolithic (Cybiktarov, 2002).
Summary of Late Epipalaeolithic developments and Gobi Desert Oasis 1
The term “Oasis 1” is used in this thesis to refer to the Late Epipalaeolithic. According to
evidence from northwestern China, during this period dune fields began to be targeted as
camp sites from which a range of ecozones could be exploited (Bettinger et al., 2007).
Faunal remains consistently indicate the incorporation of medium- and small-bodied
game, especially cervids. Although faunal assemblages are not as well studied in
Northeast China, diversification of prey types seems to have been common across
Northeast Asia as aquatic species and a wider range of mammalian fauna are recorded in
many Late Epipalaeolithic sites. Evidence from Nanzhuangtou in northern China further
suggests that wetland environments may have become important procurement locales
154
during the Late Epipalaeolithic. Large game would have continued to be targeted and
high residential mobility would have persisted. The persistent exploitation of larger
game like equids in northern China is not surprising due to the relative prevalence of
open steppe environments. It is proposed that the margins of sand dunes in proximity to
upland regions were beginning to be important to foragers in arid northwest China at this
time as they provided access to resources within both the dune-fields and sand-free
piedmonts (Bettinger et al., 2007).
A similar pattern could be proposed for the Gobi Desert, based on environmental
similarities. Possible Late Epipalaeolithic dune-field sites in basins or valleys include
Shara Murun Crossing (Site 3) in the East Gobi, and in the Gobi-Altai Barongi Usu
Valley, and Shabarakh-usu Site 2 components 2a and 2b (see Table 3.7). However, many
Gobi Desert Oasis 1 sites were situated in high elevation environments, near springs or
rivers (see Chapter 4). The lack of confirmed dune-field sites may be related to the
difficulty in identifying and dating early aceramic sites, but might also represent a unique
pattern of land-use focused on upland environments.
Based on the increased exploitation of cervids and fish in southern Siberia and the
use of a more diverse array of small species in northern China, the Late Epipalaeolithic in
Mongolia may also have been characterized by a reliance on medium- and small-bodied
species. The grass seeds and a wealth of small-bodied animals at Chikhen Agui (though
it is not clear if all the faunal remains can be associated with human activity), as well as
the occasional use of pottery exemplified at Shara KataWell, may support such a model
for the Gobi Desert. Unfortunately, comparative data for earlier sites is entirely lacking.
155
Technological trends during the Late Epipalaeolithic of Mongolia and the Gobi
Desert are most consistent with those in southern Siberia and northwestern China.
Microlithic tools, including retouched microblades, were more frequent than macrotools
and were often shaped by pressure-flaking. Later Epipalaeolithic peoples tended to
increasingly favour homogeneous cryptocrystalline stones (jasper and chalcedony were
the most popular raw materials in the Gobi Desert) over coarser-grained materials such as
siliceous sandstone, basalt and quartzites. By the end of the Late Epipalaeolithic, fibretempered cord-marked pottery had been incorporated into tool kits, though ceramics are
still rare. Ostrich eggshell beads, which were also found in pre-LGM Northeast Asian
sites, occur with increasing regularity during the Late Epipalaeolithic, including at
Chikhen Agui, Yujiagou, Hutouliang, and other Nihewan Basin sites. Microblade core
technology indicates a more varied and flexible approach to reduction strategies,
frequently focused on prepared cores made from small nodules, and was organized
around the exploitation of homogeneous raw materials producing reliable conchoidal
fractures.
3.2.2.3. Neolithic/Oasis 2 (Early to Middle Neolithic) – 8.0 to 5.0k cal yr BP
The beginning of the Neolithic period in Northeast Asia, defined by widespread
diversification of new lithic technologies, took place in the early Holocene.
Approximately 8.0-5.0k cal yr BP is suggested as a broadly encompassing date for Oasis
2, or the early to middle Neolithic of Northeast Asia. The persistence of aceramic
microlithic assemblages in the Gobi Desert has resulted in continued use of the term
156
“Mesolithic” for sites that are more closely allied to the broader trend in Neolithic
technology, subsistence, and land-use. This includes specialized task sites outside of the
base camps where ceramics were more commonly deposited. Due to our current reliance
on relatively rare diagnostic technologies like pottery, grinding stones, and adze/axes, the
introduction of a chronology able to distinguish the relative age of archaeological
assemblages based on ubiquitous artefact types and specific elements of core reduction
sequences is of central importance. Assigning known archaeological sites to a specific
period is fundamental in identifying key shifts in land-use and patterns of resource
exploitation.
Archaeological sites from the Neolithic are much more numerous across
Northeast Asia and the comparative data are richer than for the preceding period. As
outlined in Chapter 2, developments in subsistence and technology varied from region to
region despite the general trend in diversification that sometimes included food
production. Neolithic archaeological assemblages are more abundant in Mongolia than
Epipalaeolithic sites. Many Neolithic sites are from the Gobi Desert and a dramatic
increase in the number of sites may be related to increases in population density.
Regional chronologies from the borderlands of northern China and southern Siberia will
be used as a comparison, but primarily new radiocarbon dates from Gobi Desert sites will
be used to identify local temporally diagnostic technologies.
157
Climate and subsistence
A period of high effective moisture associated with soil stabilization and higher
vegetative biomass in arid environments took place in the early to middle Holocene and
is known as the Holocene Climatic Optimum. Soil formation and the infilling of lakes
are typical markers, but there is a great deal of regional diversity in both the timing of
increased moisture and the environmental shifts that resulted. In the Gobi Desert,
differences in the local dominance of circulation systems controlled temporal variability
in Holocene climatic amelioration. Optimally moist conditions occurred variously
between about 9.0 and 4.5k cal yr BP (Tarasov et al., 2000; Mischke et al., 2005; Jiang et
al., 2006; Herzschuh, 2006; Rudaya et al., 2008). A widespread decline in precipitation
and expansion of permafrost are evidence of climatic degradation that occurred across
Northeast Asia between 5.8 and 4.5k cal yr BP (Starkel, 1998). All regions appear to
have been effected by an overall decrease in effective moisture after ca. 3.0k cal yr BP
(Herzschuh, 2006).
The Neolithic tends to be typified by the adoption of domesticated plant and
animal species, and sedentary agricultural communities did emerge in North China and
parts of Northeast China at this time. However, despite the prominent role of these new
agricultural economies in studies of the Neolithic they are not typical in much of
Northeast Asia. Hunter-gatherer groups continued to focus on wild resources. At the
same time, the emergence or spread of new technologies like grinding stones, pottery,
and specialized fishing gear often do suggest intensified exploitation of certain foods.
158
In the Lake Baikal region, hunter-gatherer economic strategies persisted into the
Bronze Age and aside from dogs (Bazaliiskiy and Savelyev, 2003; Losey et al., 2011),
plant and animal domesticates were not a regular component of economic strategies. One
exception is the Sagan-Zaba site on the west coast of Lake Baikal, spanning the
Mesolithic to Bronze Age, which is recorded as containing “frequent” remains of horse
and sheep or goat (Weber et al., 1998). Fishing technology (composite fishing tools and
harpoons) suggests the increasing importance of aquatic resources, while stable isotope
studies indicate that ungulate meat continued to be of key importance in the diet
throughout the Neolithic and Eneolithic (Weber et al., 1998; Weber and Bettinger, 2010).
Faunal remains from sites in the Yenisei River region included red deer, Siberian goat
and sheep, and molluscs (Vasil’ev and Semenov, 1993). Southern Siberia is typified by
hunter-gatherer-fisher cultures. Domesticated species do not typically appear until about
5.5k cal yr BP in the Yenisei region and even later in the Lake Baikal region.
The Neolithic of Northeast Asia is best known by developments in the Central
Plains and Yellow River Basin. Faunal remains from Nanzhuangtou suggest that
domesticated animals may have already been adopted in some areas by the early
Holocene. Bones of domesticated dog, pig, cow and possibly chicken from Peiligang
culture sites (8.5-7.5k cal yr BP) attest to the role of animal husbandry in Early Neolithic
economies (Underhill and Habu, 2006; Lu, 1999). The importance of millet agriculture
during the Neolithic is evidenced by large scale storage of millet (Panicum miliaceum)
dating to between 10.3-8.7k cal yr BP at Cishan in the North China Plain, provisioning of
animals at Dadiwan by 7.9-7.2k cal yr BP, and indications of “slash-and-burn”
159
cultivation beginning around 7.7 ka (Li et al., 2009; Lu et al., 2009; Barton et al., 2010).
At the same time, a diverse array of floral and faunal remains from Cishan13 and the
slightly later Peiligang site indicates that hunted and gathered resources were still an
important part of subsistence. By 5.0k cal yr BP, domesticated dogs, pigs, cows and
sheep were all a part of many local agricultural economies, alongside cultivated millet,
rice, soybeans and hemp (Crawford et al., 2005; Lee et al., 2007; Li et al., 2009).
Sedentism and cultivation also became common in Northeast China during the
Early Neolithic, but such agricultural sites are mainly distributed over the alluvial plains
of southern Northeast China and parts of Inner Mongolia. The earliest of these sites are
Xinglongwa and Chahai, dated respectively to 8.1 and 7.8k cal yr BP (7240 + 95 BP,
6925 + 95 BP) (Guo, 1995a). No evidence of domesticated plants was recovered from
these sites, but the site structure and tool kits indicate a reliance on cultivation and plant
foods in conjunction with hunting (Guo, 1995a). Better evidence for millet domestication
comes from Xinle I (Lowe Xinle), where village settlements in the Lower Liao River area
were dated to 7.5-7.0k cal yr BP (6620 + 150 to 6145 + 120) (Jia, 2007). By the time of
the Hongshan culture group at about 6.0-5.0 kya, agriculture included domesticated millet
and pigs (Guo, 1995a). Faunal remains from Yaojingzi in Jilin Province (5.5k cal yr BP
[4726 + 79 BP]) included cattle, horse, and dog, and are taken as evidence of animal
13
Faunal remains from Cishan included Ctenopharyngodon idellus (grass carp), Lamportula sp. (mollusc),
Cuora sp. (Asian box turtle; originally described in the literature as Emydidae, recent changes in taxonomy
indicate that the genus of the animal referred to is probably Cuora), Macaca mulatta (rhesus monkey),
Meles meles (badger), Paguma larvata (masked palm civet), Panthera pardus (leopard), Sus scrofa (wild
boar), Anser jabalis (bean goose), Gallus gallus (fowl, possibly domesticated chicken), and several species
of cervids (Lu, 1999). Plant remains included Juglans regia (common walnut), Geetis bunseana
(hackberry seeds? – from Lu, 1999: 36, taxonomic designations are confused), and Corylus leterapluylea
(hazelnut) (Lu, 1999).
160
husbandry (Liu, 1995). Highly ritualized behaviour is noted for this period. Many
hunter-gatherer sites are still associated with the Neolithic period in Northeast China and
are distributed primarily in the northernmost part of the region and on the steppe to the
east of the Daxing’anling (Da Khinggan) mountains (Lu, 1998). The Xinkailiu site in
Heilongjiang Province (6.2k cal yr BP [5430 + 90 BP]) was associated with lake
environments; fish storage pits and fishing tools, in conjunction with evidence of hunting,
indicate a heavy reliance on fishing (Tan et al., 1995a; Lu, 1998).
The emergence of sedentary agricultural communities represents an important
divergence from mobile hunter-gatherer subsistence economies, but dedicated huntergatherers were most common across Northeast Asia. Fishing, hunting, and gathering
appear to have remained important throughout the Neolithic, with varying degrees of
economic utility (Lu, 1998). No evidence of domesticated species has been recovered
from Gobi Desert sites, though this may be due to poor preservation. The use of grinding
stones and pottery has been cited as possible evidence for cultivation (see Chapter 2), but
might also be related to the intensified processing of certain wild foods.
Neolithic technology in Northeast Asia
Although certain tool types associated with the Neolithic were used much earlier in some
areas of Northeast Asia (see Chapter 2), it was not until about 9.0-8.0k cal yr BP that the
use of pottery, polished stone, grinding stones, and bifacial flake technology became
widespread. Regional differences in subsistence economies are notable during the Early
161
Neolithic and technological trajectories also begin to diverge more clearly during this
time. Microblade technology continued to be important among hunter-gatherer groups,
but was less dominant in the tool kits of sedentary agriculturalists.
Comparisons have been made most often between assemblages from Gobi Desert
and the Lake Baikal region (Maringer, 1950; Okladnikov, 1962). Unfortunately, there
are few studies of lithic chronology in the Lake Baikal region (see Weber, 1995). The
stratified Ulan-Khada site provides some of the best data for building such a chronology
and spans the terminal Late Epipalaeolithic to Eneolithic. The earliest dates on pottery
from the Lake Baikal region come from this site and are dated to 8.5k cal yr BP (Kuzmin
and Orlova, 2000). The “net-impressed” pottery is associated with small bifacial
projectile points with straight bases, and new forms of composite fish hook shanks.
“Pseudo wedge-shaped” and prismatic microblade cores, along with endscrapers on
spalls, retouched microblades, and burins on microblades were found in all levels of the
site. Polished stone appears earliest in the form of slate shanks for composite fishhooks
and an adze, but is more common in younger layers (Goriunova and Khlobystin, 1991;
Kuzmin and Orlova, 2000; Weber, 1995). By 5.0k cal yr BP, arrowshaft straighteners,
bifacial knives and polished nephrite (jade) adze/axes were also present in the tool kit,
along with a diversity of pottery types.
Due to a lack of information on lithic assemblages, ceramics have been
considered the best markers to distinguish between the earlier Kitoi (8.8-6.9k cal yr BP)
and the later Serovo/Glazkovo (6.2-3.0k cal yr BP). Early Kitoi pottery includes oval and
mitre-shaped vessels with net impressions and later designs include cord-impressions
162
near or below the rim and/or various incised lines forming geometric motifs. SerovoGlazkovo period pottery exhibits greater variation in pottery styles, including: Ust’Belaya, typified by simple oval pots with stab-and-drag and sometimes comb
impressions; Posol’sk, typified by thick and straight-walled vessels with appliqué about
1-1.5 cm from the rim, featuring dentate impressions, stab-and-drag type lines, and often
a series of small holes located above the appliqué; smooth-walled; cord-impressed
decorations; and hatched decorations (Weber, 1995; McKenzie, 2009).
The Upper and Central Yenisei River region to the west of Lake Baikal is another
area of cultural developments highly relevant to this discussion, though much less well
understood. Neolithic components from stratified sites like Maina indicate the
importance of microblade technology and include “double-platform,” prismatic, and flat
(or tabular) microblade cores (Vasil’ev and Semenov, 1993). Bifacial microlithic points
are triangular with a concave base or a lateral notch, and there are also rhomboid and oval
pieces. Bifacially flaked blades are interpreted as inserts for composite tools and are
similar to Early Neolithic specimens from the Gobi Desert (see below). The Maina
assemblage also included retouched microblades, burins, wedge-like tools, endscrapers,
sidescrapers, retouched flakes, notched and denticulated pieces, and pebble tools
(Vasil’ev and Semenov, 1993). The tool kit is consistent with other aceramic Neolithic
archaeological sites in the region, including Ui II, Ust’-Khemchik 3, and Toora-Dash.
Differences in assemblages included the occurrence at Ui II of a roughly flaked axe,
wedge-shaped microblade cores, knives, and grinding stones. Triangular bifacial points
with straight bases and conical microblade cores were found at Ust’-Khemchik 3.
163
Distinct artefact types from Toora-Dash include bifacial inserts on elongated flakes
(lunate, rectangle, and trapeze shapes), arrowheads with concave bases, knives, and an
antler adze.
Despite the absence of pottery at Early Neolithic sites in the Upper Yenisei River
area, in the Central Yenisei region pottery appeared first near the beginning of the Early
Neolithic (8.1-7.4k cal yr BP [7330 + 35 BP, 6530 + 60 BP]) (Kuzmin and Orlova,
2000). Pottery is thin-walled and tempered with sand and gravel. Rim fragments feature
serrated impressions and a “belt of small pinholes” are often found on the upper side of
the rim. “Back-stepped blade impressions” encircle vessels below the rim, with
perpendicular lines coming from the lowest horizontal lines. These patterns have been
compared to Posol’sk ceramics from the Lake Baikal region (McKenzie, 2009). Slightly
later ceramics from various sites are more diverse and decorations include net
impressions, cord impressions, oval stamps, horizontal rows of small oval serrated
depressions, crescent impressions, zig-zag stamps, and punctate-comb zig-zag
decorations (McKenzie, 2009). Later sites are extremely diverse and isolation of cultural
complexes based on ceramics is problematic.
In northern China, regional tool kits reflect divergences in economic strategies
represented by agriculturalists and hunter-gatherers. Microblade and flake technology
are completely absent at Cishan, while chipped and/or polished axes comprise over 56%
of the lithic assemblage (Lu, 1999). Spades are also common, along with grinding
slabs/querns, rollers, hammers, chisels and anvils. Tools made from bone, antler, animal
tooth, and shell are also present and include knives, chisels, drills, net shuttles, spades,
164
needles, various ornaments, arrowheads for hunting, and harpoons for fishing. Simply
shaped sand-tempered pottery vessels are typical. Most are roughly made, but finer
vessels are also found. Tripod stands were used. Cord marking, comb impressions, and
narrow bands of relief were common finishes (Lu, 1999). A similar lithic assemblage
was found at Peiligang, but fully polished axes and spades are more common and there
are lower frequencies of bone and antler tools. Quartzite and chert flake scrapers and
points were also used (Lu, 1999). Microblade tools were used at Yujiagou, in the
Nihewan Basin, as late as 2.1 ka, although frequencies declined steadily after about 11.6
ka (Xia et al., 2001). Neoliths (polished stone tools? – see Xia et al., 2001) appeared at
about 8.7 ka and were most common between 6.6 and 2.1 ka. At Dadiwan, in
northwestern China, microlithic technology was used by early millet-using huntergatherers, but is rare in later Neolithic agricultural sites (Bettinger et al., 2010b).
Similarly, microblades declined in quantity and variety throughout agricultural
regions of Northeast China while polished stone and organic tools became more
important (Lu, 1998). Lithic assemblages associated with sedentary communities are
characterized by chipped and/or polished macrotools like hoes, spades, adze/axes, and
grinding stones (querns and rollers), although microblades were used in composite tools
(Guo, 1995a; Lu, 1998; Jia, 2007). Later Neolithic site groups like Zhaobaogou (6150 +
85, 5980 + 85), Lower Houwa (5600 + 100 BP), and Fuhegoumen (4735 + 100 BP)
contain highly variable frequencies of microblades, but all indicate declining reliance on
microblade technology (Guo, 1995a; Xu, 1995; Lu, 1998). Pottery was still handmade
and sand-tempered pottery, but features a greater variety of surface designs. Cylindrical
165
pots are the most common vessel type in both agricultural and hunter-gatherer sites (Guo,
1995a).
Groups more reliant on hunting and gathering focused more heavily on
microblade technology (Lu, 1998). Although flake tools are much more common,
microblades are numerous at such sites, and both boat-shaped and conical cores are
recovered. Pressure flaking was heavily used for retouch and in the manufacture of
arrowheads and spearheads made on flakes. Bone and antler tools are often found at
Early Neolithic sites (Lu, 1998).
There are several key Neolithic hunter-gatherer sites in Northeast Asia. The
terminal Epipalaeolithic or Early Neolithic assemblage from Tengjiagang in Heilongjiang
Province (8.4k cal yr BP [7570 + 85]) contains pottery, stone, bone and antler tools (Lu,
1998). Bone and antler tools make up about 30% of this assemblage. The Ang’angxi
(Ang-ang-hsi) site group on the Song Nen Plain is closely allied with Gobi Desert
assemblages. Located in a group of four sand dunes south of Wufu Station, the lithic
assemblage is characterized by a diversity of unifacial points, large and small bifacial
points, microblades for composite tools, scrapers and bifacial knives on blades (Chard,
1974: 106-107 [Figure 2.46]; Tan et al., 1995a: 132-135 [Figure 4.6]). Bone spearheads
and knife blades were also found, and bone harpoons attest to the importance of fishing.
High-fired pottery tempered with fine sand and some shell was mostly in the form of
cylindrical vessels and was decorated with irregular shapes, incised lines, stab and drag
lines, and fingernail marks (Tan et al., 1995a). At Xinkailiu, Heilongjiang Province (6.2k
cal yr BP [5430 + 90 BP]), tool kits and faunal remains indicate a heavy reliance on both
166
fishing and hunting (Tan et al., 1995a; Lu, 1998). Small flake and microblade tools are
most common and associated with polished axes, chisels, and grinding stones. Bifacially
flaked projectile points are common and varied in character (Tan et al., 1995a).
Neolithic of the Gobi Desert and Oasis 2 technology
Prior to this study, there were no chronometric dates for the Early Neolithic of the Gobi
Desert. Five archaeological assemblages are now dated to between 8.0-5.0k cal yr BP
and assigned to the Neolithic or Oasis 2 (Table 3.1). All are from the Gobi Desert and
include: Chilian Hotoga, Site 35 (7.6k cal yr BP [6728 + 45 BP]), components of Baron
Shabaka Well, Site 19 (6.8 and 6.4k cal yr BP [5954 + 52 BP, 5609 + 47 BP]) in the East
Gobi; Ulan Nor Plain (5.8k cal yr BP [5116 + 41 BP, 5061 + 49 BP]) in the Gobi-Altai;
and components of both Yingen-khuduk (5610 + 370 ka), and Mantissar 12 (6460 + 700
ka) in the Alashan Gobi.
The majority of Oasis 2 sites appear to have been longer term habitation sites with
hearths. As at other Northeast Asian sites from this period, there are substantial additions
to existing technology, including pottery, grinding stones, polished stone tools, and a
proliferation of pressure-flaked points and bifaces. Most habitation sites contain pottery
and chipped or lightly polished adze/axes. East Gobi sites are characterized by the use of
large, formally-made grinding stones, but these artefacts are rare in more western Gobi
Desert sites. Finely made points, presumably used as arrowheads, were found in all sites,
including unifacially flaked microblade points from Chikhen Agui. Aside from the
presence of small unifacial points at Chikhen Agui, Horizon 1 is relatively consistent
167
with underlying horizons, although it may date to the early stages of Oasis 2 (after 8.7k
cal yr BP) (Derevianko et al., 2008). This may be related to the task-specific nature of
the site, but the impression of coherency could also result from a lack of stratigraphic
integrity.
One difficulty in reconstructing an artefact-based chronology for Oasis 2
assemblages from museums is that Oasis 2 and Oasis 3 sites were often recovered from
the same locality and collected as one group. Baron Shabaka Well, Yingen-khuduk, and
Mantissar 12 all contain artefacts directly dated to both periods. The Chilian Hotoga and
Ulan Nor Plain sites are more likely to represent coherent Oasis 2 assemblages (see
Appendix B). Although historic pottery was included in the Ulan Nor Plain assemblage,
this is not evidence of post-depositional intermingling of sediments, but rather a result of
the collection by Nelson and Pond of artefacts scattered across the dune surface (Nelson,
1925; Pond, 1928, n.d.). It is common for the AMNH collections to include such
artefacts (personal observation, August 2009).
168
a.
c.
b.
d.
Figure 3.4 Dated Oasis 2 pottery: a. “net-impressed” pottery from Yingen-khuduk, 5.6
ka; b. “net-impressed” pottery from Baron Shabaka, 6.4k cal yr BP; c. textile-impressed
pottery from Chilian Hotoga, 7.6k cal yr BP; cord-impressed pottery from Mantissar 12,
6.5 ka.
169
Dated decorated pottery shards are the best diagnostic artefact type. Samples
from Yingen-khuduk (5.3-6.0 ka; UW2357, #K.13212: 123) and Baron Shabaka (6.4k cal
yr BP; AA89885, #73/2229A) both bear a distinctive “net-impressed” surface design that
can be considered characteristic of Oasis 2 (Figure 3.4a, b). Chilian Hotoga textileimpressed pottery, dated to 7.6k cal yr BP, is a coarse sand- and organic-tempered greyware blackened on the interior surface (Figure 3.4c). Oasis 2 pottery is typified by the
use of heavy sand temper, thick walls, and a low-fired paste. The interior paste of most
samples is blackened, which is typical of low-fired pottery with high organic content.
The shard from Yingen-khuduk is exceptional in that the interior paste is not blackened.
This is probably due to the fact that it was only tempered with sand rather than additional
organics; the clay that was used might also have been less rich in organic matter. The
lack of organic temper in Alashan Gobi pottery is typical of all periods. Some of the
spongy-textured grey-ware from the East Gobi may also date to this early period (Figure
3.9). Pottery use became more common during Oasis 2 and is typified by “netimpressed,” corded, or textile impressed low-fired grey or brown-ware.
Additional diagnostic artefacts can be recognized from sites that have been
chronometrically dated or that contain pottery typical of Oasis 2. Unifacially retouched
points made on microblades are found in several Neolithic sites in Northeast Asia and can
be considered diagnostic of early Oasis 2 (Table 3.4). They are similar to some examples
of perforators on microblades, but retouch on perforators is more localized on the distal
end. Retouch on small unifacial points is more consistently executed and less steep than
with the production of other microblade tools. Such artefacts were found in East Gobi
170
sites and Horizon 1 of Chikhen Agui. According to the youngest accepted dates from
Horizon 2 (8.7k cal yr BP; see Derevianko et al., 2008), Horizon 1 should post-date 8.7k
cal yr BP (Derevianko et al., 2008). Such points are also found at Chilian Hotoga and
Baron Shabaka, dating to 7.6k cal yr BP and 6.8k cal yr BP respectively. According to
illustrations of typical assemblage artefact types, similar points were recovered in the
Lower Liao River region of Northeast China from Xinle I (Lower Xinle) (7.5-7.0k cal yr
BP [6620 + 150 to 6145 + 120]) and the similarly aged Zuojiashan I (Yaojingzi and
Yuanbaogou) in the Changchun (Ji-Chang/Jilin-Chang) region of central Northeast China
(Jia, 2007: 74-75, 125). Additional undated Neolithic sites also contain such points,
including Haila’er and Ang’angxi (Chard, 1974: Figure 2.46). Based on the absence of
unifacial points in later dated archaeological sites, it is reasonable to suggest that they are
temporally constrained to approximately 8.0-6.5k cal yr BP.
Assemblages from eastern Mongolia collected by Mongolian and Soviet
archaeologists also indicate the widespread use of unifacial points in the eastern regions.
Examples of such finds include surface assemblages from Dornogovi aimag (or Eastern
Gobi province) (Dorj, 1971: 170), surface assemblages from Sükhbaatar aimag (Dorj,
1971: 156), the Khutyn-bulag (Хуйтэн-Булаг) lake site (Dorj, 1971: 39-40, 136), and at
Locale 9 (Стоянка 9) (including shouldered points) (Dorj, 1971: 29-30; 111). Based on
Dorj’s (1971) descriptions and illustrations, such sites probably date to about 8.0-6.5 kya.
The upper layers of Dulaani Gobi (Дулааны Гобь), Dornogovi aimag, contained both
unifacial and bifacial points and may be attributed to about 7.5-6.5 kya (see Tseveendorj
and Khosbayar, 1982). I suggest that Munkh-tolgoi (Мунх-толгой) also dates to about
171
7.5-6.5 kya (Dorj, 1971: 30-31, 112-113). The Ovoot (Овоот) assemblage, excavated
from a sand deposit on the Kerulen River near Ovoot Mountains, contains pottery, bone
harpoons, and composite hafts. Based on the frequency of retouched microblades and the
early forms of pressure and rough bifacial flaking on macrotools, the site probably dates
to the Epipalaeolithic-to-Neolithic transition (Dorj, 1971: 33-36, 119-125).
In the East Gobi, unifacial micropoints were recovered at Jira Galuntu (Site 18)
(Figure 3.5b, c), Baron Shabaka Well (Site 19), and Baron Shabaka South (Site 21)
(Figure 3.5d). Chilian Hotoga (Site 35) also contained a retouched microblade that was
probably used as a hafted projectile, but it is associated with bifacial flake points and is
not finely finished (Figure 3.5a). Some unifacial points from Jira Galuntu were
shouldered (Figure 3.5b), which is also observed in Zuojiashan I assemblage unifacial
points (Jia, 2007: 125). Similar artefacts are depicted for Lake Baikal (Trans-Baikal)
region sites (Chard 1974: 86 [Figure 2.27]). Shouldered points are a distinctive trait of
the Early Neolithic in southern Siberian, including the Maina site, attributed to the early
Holocene climatic optimum (Vasil’ev and Semenov, 1993: Figure 3). The presence of
unifacial points at Chikhen Agui indicates that the technology was distributed across the
Gobi Desert region, although possibly more common in Eastern Mongolia and Northeast
China (Table 3.4).
172
a.
c.
b.
d.
Figure 3.5 Unifacial and bifacial points from Oasis 2 sites: a. Chilian Hotoga, bifaces
and uniface (second from right); b. Jira Galuntu, unifacially retouched shouldered
microblade points; c. Jira Galuntu, unifacially retouched flakes and microblade segment
(middle); d. Baron Shabaka South, bifacial and unifacial points.
173
Region
Site Name
East Gobi
Baron Shabaka
Baron Shabaka
South
Chilian Hotoga
Jira Galuntu
Dulaani gobi,
upper layers
Dornogovi
aimag (surface)
Chikhen Agui
Sükhbaatar
aimag (surface)
Khutyn-bulag
Locale 9
Munkh-tolgoi
Xinle I
Zuojiashan I
Haila’er
Ang’anxi
Gobi-Altai
East Mongolia
Northeast
China
Chronometric
dates
6.8, 6.4 kya
N/A
Unifacial
points
yes
yes
Bifacial points
yes
yes
Shouldered
points
no
no
7.6 kya
N/A
N/A
yes
yes
yes
yes
yes
yes
no
yes
N/A
N/A
yes
yes
N/A
After 8.7 kya
N/A
yes
yes
no
yes
no
N/A
N/A
N/A
N/A
7.5-7.0 kya
N/A
N/A
N/A
yes
yes
yes
yes
yes
yes
yes
yes
no
yes
N/A
N/A
yes
yes
N/A
yes
N/A
N/A
yes
N/A
N/A
Table 3.4 Summary of sites containing small unifacial points in the Gobi Desert, East
Mongolia, and Northeast China.
The method of manufacturing of small bifacial points may have been derived
from techniques used to retouch microblades, particularly the manufacture of unificial
points. Narrow bodied bifacially flaked points have been found associated with unifacial
points at Jira Galuntu, Baron Shabaka Well, Baron Shabaka South (Figure 3.5d) and
Chilian Hotoga, possibly indicating that bifacial points were derived from experiments in
forming points from minimally retouched blades and flakes. Bifacial points from Baron
Shabaka South appear to have a slightly concave base and occasional light fluting (see
Figure 3.5d). Small bifacial blades are probably contemporaneous with bifacial points,
both having been recovered from the Chilian Hotoga Well site (Figure 3.5a). Bifacially
flaked points and blade-like pieces were recovered from the upper levels of Maina in the
174
Upper Yensei region (Vasil’ev and Semenov, 1993: Figure 3). This assemblage was
excavated from a “buried double soil” that was ascribed to the optimal early Holocene
phase. Small bifacial points date to as early as 8.5k cal yr BP in the Lake Baikal region,
but the use of bifacial techniques for the manufacture of both points, blades and knives
was more widespread in this region of southern Siberia sometime after 7.2k cal yr BP
(6310 + 70; see Kuzmin and Orlova, 2000; Weber, 1995). According to data from the
Chilian Hotoga Well assemblage, bifacial pressure-flaking appears to have been used in
the Gobi Desert by at least 7.6 kya.
Dated Oasis 2 assemblages, along with the Jira Galuntu and Baron Shabaka South
sites, are detailed in Appendix B. These site assemblages indicate a number of
characteristics that distinguish them from the Late Epipalaeolithic/Oasis 1. Formal
grinding stones are one of the most notable technological developments in this period,
particularly in the East Gobi, and are consistent with similar technologies in northern
China. Likewise, partially polished stone adze/axes are found in East Gobi sites.
As discussed above, unifacial blade and bifacial flake points are a hallmark of the
Early Neolithic across the Gobi Desert and eastern Mongolia. Pressure-flaked microtools
are a hallmark of the Mongolian Neolithic and can be used to distinguish Neolithic from
Epipalaeolithic sites. A range of bifacial flaking techniques was employed in the
manufacture of microblade cores, chipped macrotools like adzes and axes, and fine
bifacially pressure-flaked points, blades, and knives. Small chipped adzes are also
representative of Oasis 2 and were often made on high quality cryptocrystalline stones
(Figure 3.6). Semi-lunar knives originate during Oasis 2, as evidenced at Baron Shabaka
175
Well and Ulan Nor Plain. Bifacial flaking was also used in the preparation of wedgeshaped microblade cores, which are characterized by a knife-like edge opposite the
flaking face.
Figure 3.6 Small chipped adze from Baron Shabaka Well, on high quality jasper.
Microblade core manufacture during Oasis 2 continues to be characterized by
flexibility and a number of cores forms are recognized. Many microblade cores do not
fall into distinctive shapes and were classified during data collection as “unknown” or
“informal”. More carefully executed and heavily reduced cores are less common, but
generally fall into three categories: wedge-shaped, conical, and cylindrical. Conical and
176
cylindrical types sometimes have a wedge-shaped back, which would have facilitated
core reduction if used to grip the core in a vice. A wedge-shaped back might also have
been useful as a multipurpose tool. The Epipalaeolithic layers at Chikhen Agui contain
well-made and heavily used microblade cores, though they tend to be blockier and
squatter than the slender cylindrical and conical cores typical of later periods.
Numerous test pieces from the Ulan Nor Plain site help to construct the primary
method of microblade core preparation at the site (see Figure 3.12). A platform was first
prepared on a small, roughly oval cobble by removing one narrow end. A rough wedge
or U-shape was then formed by transverse or sometimes longitudinal chipping from the
edges or platform. Microblades were sometimes struck from one of the wide, flat faces
of the core, but more often from one narrow edge, progressing around onto the adjacent
sides. The narrow edge was sometimes bifacially prepared, as evidenced by test pieces
and numerous “keel flakes.” The exact method of core preparation appears to have been
primarily determined by the shape, cortical structure, and striking quality of the raw
material. Simple elongated flakes were also struck from cores prepared in this way. The
edge opposite the surface of microblade removals was often bifacially flaked into a knifelike edge, which was probably related the process of core reduction (Flenniken, 1987;
Tabarev, 1997), but could also have been used as a tool (Morlan, 1976).
The most heavily reduced cores generally tended to take a conical or cylindrical
form (the latter was usually formed by the use of two opposing striking platforms). There
are many examples of intermediate specimens discarded at the transition from wedgeshaped to small conical core. Other heavily reduced cores retained the wedge-shaped
177
back at discard and were most heavily reduced from the narrow face. Flexibility of core
form throughout the artefact’s use life indicates a variability in core reduction strategies
that was probably partially controlled by the original nodule shape and size. The lack of
formal core preparation prior to microblade removal probably contributed to different
choices during the reduction sequence based on variation in the location and nature of
remaining cortical surfaces and evolving core shape.
Other significant core types include expedient cores, biface cores, and informal
blade/bladelet/elongated flake cores. Some cores are also classified as “core tools” and
these are informal, amorphous cores with evidence of use and/or retouch and numerous
flake removal scars that do not intentionally contribute to overall core morphology.
Expedient cores are amorphous cores with numerous flake removal scars and no distinct
platform. Biface cores are also found in Oasis 2 sites and are typically thick bifaces
covered in rounded, rather than elongate, flake scars. They are often U-shaped. When
evidence of use and/or retouch suggests that the cores were used as tools, they are
classified as “biface core tools”. Biface cores and biface core tools are less common than
microblade cores. The category of “informal blade/bladelet/elongated flake cores” is
used to classify cores with a roughly prepared platform, and elongate flake scars,
indicating the removal of elongate flakes wider than microblades. Such cores are present
in the majority of Gobi Desert assemblages, including those from Epipalaeolithic sites.
Near the end of Oasis 2 the use of high quality jaspers probably became more
widespread. As exemplified at Chikhen Agui, raw materials were locally obtained in the
Late Epipalaeolithic. More selective use of various high quality cryptocrystalline stones
178
in the manufacture of unifacial and bifacial points, bifacial knives and some heavily
reduced microblade cores suggests that high quality raw materials were procured for the
production of certain tool types, even if they were not local. Since knowledge of raw
material sources is scanty, it is currently not possible to define “local” and “exotic” (but
see Kulik et al., 2006 for the Gobi-Altai region). Petrographic studies of lithic
assemblages and raw material sources would greatly improve our knowledge of land-use
and possibility of trade-routes from source locales like the Arts Bogd-Ulan Nor Plain and
the Ukh-tokhoi/Khara Dzag plateaux regions.
Summary of Neolithic developments and Gobi Desert Oasis 2
One of the most commonly cited inferences about Neolithic Gobi Desert peoples is that
they were engaged in some form of incipient agriculture during the middle Holocene (e.g.
Cybiktarov, 2002). Although the dates of millet domestication in Northeast Asia make it
possible that Gobi Desert people were familiar with such developments, the contents of
Gobi Desert archaeological assemblages do not support the conclusion that Gobi Desert
peoples were agriculturalists. In the western Gobi Desert, grinding stones are rare in both
Oasis 2 and Oasis 3 assemblages. More extensive use of grinding stones in the East Gobi
is suggestive, but aside from one possible “hoe-like tool” at Baron Shabaka Well there is
little evidence of the digging and sowing tools, or storage facilities characteristic in North
and Northeast China agricultural sites. Evidence of permanent settlements, subterranean
houses, and the small villages typical of early sedentary agriculturalists is likewise
missing. Hunter-gatherers may have heavily exploited wild grass seeds and/or tubers in
179
the dune field/wetland environments where grinding stones are found. They might also
have aided in the abundance of seasonal resources by scattering seed. But there is
currently no convincing evidence of full-fledged agricultural endeavours in the Gobi
Desert proper.
Likewise, there is little of evidence for domesticated animals during the
Mongolian Neolithic; however, domesticated dogs were probably widespread amongst
hunter-gatherers in Northeast Asia by the beginning of the Neolithic and it is possible that
they were used by hunter-gatherers in Mongolia. Other domesticated species like pigs,
cows, and sheep were all present in the Early Neolithic of North China. Chickens may
also have been kept. Pigs and chickens are not adapted to high residential mobility and
are unlikely to have been adopted by Gobi Desert groups. The Tamsagbulag site in
eastern Mongolia is thought to represent an economy reliant on hunting, fishing,
gathering, but complemented by millet cultivation and cattle husbandry as early as 6.5
kya (Derevianko and Dorj, 1992; Séfériadès, 2006). Both the dates and evidence for
domestication are provisional and must be investigated further. Middle Holocene plant
and animal domestication in eastern Mongolia would suggest continuity with
contemporaneous economic developments in neighbouring Northeast China.
Tools characteristic of Oasis 2 are summarized in Table 3.6 and include large
formal grinding stones, polished stone, chipped and/or partially polished adze/axes,
pressure-flaked unifacial microblade points, and a variety of bifacial tools such as small
pressure-flaked points, knives and blades. Microblade cores include wedge-shaped,
conical, and cylindrical types. Biface cores are also typical of Oasis 2. While expedient
180
flake cores and amorphous flakes were found in Oasis 2 assemblages, many tools were
still based on microblades. The most notable exception is the production of microlithic
endscrapers, which were often made on thick semi-circular flakes or microblade core
platform reduction spalls (i.e., thumbnail scrapers). Bifacial preparation on flakes and
microblades was also important. By the end of Oasis 2, lithic assemblages were typified
by microblade cores, tools on microblades, and a range of bifacially flaked tools
manufactured on high quality cryptocrystalline stone. Pottery use was more widespread
than during Oasis 1 and was characterized by the use of low-fired brown-wares with
simple surface treatments (Table 3.5).
The environmental distribution of Gobi Desert sites is discussed in more detail in
Chapter 4; however, all dated Oasis 2 sites are associated with dune-field environments
near streams or former lakes. According to site distribution and the proliferation of
grinding technology and pottery, it is probable that Oasis 2 represents the first intensive
use of dune-field resources. Faunal remains from Chilian Hotoga indicate the utilization
of a range of resources that included both small- and large-bodied animals, dunefield/wetland (e.g., bird, frog), and steppic species (e.g., equids). Gazelle remains are
common in many sites from North China and we can predict that they were also regular
prey for Gobi Desert inhabitants. Grinding stones suggest the processing of small plants
like grass seeds and/or tubers. The appearance of less portable tools like grinding stones,
pottery, and large adze/axes suggests a shift in residential mobility, perhaps related to
both increased investment in dune field/wetland resources and decreased seasonal
181
mobility. The absence of permanent dwellings indicates a continuance of relatively high
residential mobility.
Timing in the introduction of Neolithic tool kits is similar across Northeast Asia.
Between 9.0-8.0k cal yr BP, the use of pottery, polished stone, grinding stones, and
bifacial flake technology became widespread. It has long been observed that
assemblages from southern Siberia, much of Northeast China, and Mongolia bear
remarkable similarities in material culture, particularly in the development of bifacial
flake technology and the early diversification of microblade core types. The occurrence
of broadly similar developments in material cultures, despite highly variable
environmental conditions, makes this situation particularly interesting. Although the idea
that the Gobi Desert was peopled by migrations from the Lake Baikal region is untenable
based on differences in pottery chronologies and subsistence patterns, the Early Neolithic
of the Gobi Desert suggests a much closer circle of interaction with and influence from
northern and eastern cultures than with southern agricultural neighbours of the Yellow
River Valley.
3.2.2.4. Eneolithic/Oasis 3 (Late Neolithic/Early Bronze Age) – 5.0 to 3.0k cal yr BP
The Late Neolithic, Eneolithic, or Early Bronze Age in Northeast Asia is here dated to
the period between about 5.0 to 3.0k cal yr BP. Technology and land-use is more
consistent with the preceding period than during the Epipalaeolithic-to-Neolithic
transition. However, changes in social networks and economic endeavours are
represented by subtle shifts in surviving material culture, and mark a divergence that
182
often prefigured later historic economies in the area. By 5.0 kya, pastoralism was
widespread across the Eurasian steppe. Many domesticated animals and crops of
Western Asian origin were already being incorporated into Northeast Asian subsistence
strategies (see Chapter 2). In many cases, the geographic domains of historic pastoralist
and agriculturalist societies were already clearly defined by 3.0 kya, although hunting and
gathering still contributed to economic activities.
Despite the shorter length of time encompassed by Oasis 3, evidence of human
activity is more visible than in earlier periods. This situation is not restricted to the Gobi
Desert. One need only make a brief survey of regional literatures in order to see that our
knowledge of this period is much more detailed than for the preceding one due to a
greater number of identified and excavated sites (e.g., Svyatko et al., 2009). Increased
site visibility in the Gobi Desert, particularly in the Alashan Gobi and the Gobi-Altai,
might reflect either a peak in population density or a shift in residential mobility that
affected site distribution and visibility.
However, the difficulty in distinguishing late Oasis 2 from Oasis 3 assemblages
might also contribute to a perceived overabundance of Oasis 3 sites. In order to gain a
better understanding of shifts in Holocene land-use, clear chronological markers must be
established. Distinct differences between Oasis 1 and Oasis 2 assemblages make it fairly
simple to separate Epipalaeolithic from Neolithic sites, but fewer diagnostic markers
distinguish early from late Neolithic assemblages. Primary chronological markers like
polished stone, chipped adze/axes, formal grinding stones, and bifacial points were all
introduced early in the Neolithic and may have been used until the end of the Eneolithic.
183
Increasing regionalization during the Eneolithic hinders inferences about
technological change based on comparative assemblages from neighbouring regions.
Gobi Desert material culture has been compared to that of contemporary groups in
Central Asia, the Lake Baikal region of southern Siberia, northwestern China, and the rest
of Northeast Asia (Maringer, 1950; Formozov, 1961; Larichev, 1962; Okladnikov, 1962;
Derevianko and Dorj, 1992). Little evidence of influence in material culture is
recognized from the Central Plains and Lower Yellow River Valley, where the trend
towards tool kits focused on polished stone and bone tools continued from the early
Neolithic. Despite occasional similarities to assemblages from other regions, Gobi
Desert lithics and ceramics are distinct and well-represented in dated assemblages. As
such, the focus of this section is to identify evidence for temporal markers within local
assemblages, drawing comparisons to neighbouring regions only when most applicable.
Chronological interpretations are based on new dates and detailed site descriptions.
Climate and subsistence
The Eneolithic broadly coincides with the end of the Holocene warm/wet phase. Due to
differences in the dominance of various circulation systems across the Gobi Desert, the
three target regions (i.e., East Gobi, Gobi-Altai, Alashan Gobi) were affected differently
by Holocene trends in precipitation and effective moisture. Despite a general pattern of
increasing aridity, all regions maintained higher effective moisture than prevails today.
There is evidence for the expansion of arid steppe environments after 6.3k cal yr BP in
the East Gobi, and more widespread aridification between 4.5-2.9k cal yr BP (Herzschuh,
184
2006). Beginning around 5.2k cal yr BP, aridity increases in the Gobi-Altai and
intensifies after 4.0k cal yr BP (Starkel, 1998; Tarasov et al., 2000). In the Alashan Gobi,
desert expansion recommenced after about 4.0k cal yr BP, but local vegetation may not
have been strongly affected in areas of increased moisture availability (e.g., around lakes
and rivers) until after about 3.2k cal yr BP (Herzscuh et al., 2004; Mischke et al., 2005).
An overall decrease in mean effective moisture after 3.0k cal yr BP is apparent across the
entire Gobi Desert (Herzschuh, 2006).
The economic strategies and material culture in neighbouring regions were much
different than in the Early Neolithic. The Eurasian Bronze Age began between 5.5-5.0
kya, when nomadic pastoralist communities began to replace sedentary settlements. By
5.0 kya, heavy wheeled carts and wagons pulled by oxen were used throughout the
Eurasian steppe and large scale wool production was practiced in the Middle East (Kohl,
2007). By between 5.5-5.0 kya, the use of cattle and/or sheep was firmly established in
agricultural regions of North China and Northeast China.
By the middle of the Eneolithic, a suite of both West and East Asian domesticates
would have been familiar to many inhabitants of northern China and southern Siberia.
Beginning around 4.5 kya, various established agricultural communities along the Yellow
River and its tributaries were growing crops like soybean, adzuki bean, hemp, Chinese
cabbage, buckwheat, canola/rapeseed, and wheat (Crawford et al., 2005; Lee et al., 2007;
Li et al., 2009). Rice was cultivated as far north as Korea sometime between 5.0 and 3.5
kya. Camels were bred and used by Andronovo peoples in Xinjiang by about 4.0 kya
(Kuzmina and Mallory, 2007: 252). Some of the first secure evidences of possible horse
185
domestication in Northeast Asia come from Qijia sites in northwest China and date to
about 3.7 kya (Yuan and Flad, 2005). The presence of painted pottery in sites from the
Alashan Gobi suggests contact between Gobi Desert groups and neighbouring
agropastoralists who had domesticated herd animals such as cattle, sheep, and horses.
From about 4.5 to 3.0k cal yr BP, nomadic pastoralist peoples who inhabited the
mountain-steppe zone just west of the Gobi Desert are archaeologically recognized by
their use of domestic animals, metallurgy, elaborate burial complexes, and rock art. The
period between 3.5 and 3.0 kya probably marked the true decline of Neolithic huntergatherer societies and the rise of nomadic pastoralist economies, as pastoralism reached
its height across Northeast Asia between 3.2 and 3.0 kya (see Chapter 2). Agriculture
had also intensified in Northeast China and Korea, where sedentary communities
cultivated millet, rice, wheat, barley, and sorghum. In northern Mongolia, Bronze Age
ritual and burial monuments based on the intensive ritual use of horses have been dated to
between about 3.2-2.8k cal yr BP (Fitzhugh, 2009).
Based on the continuity of lithic assemblages during this period, there is little
evidence to contradict the continuance of hunter-gatherer subsistence strategies in the
Gobi Desert. Occupation of dune-field/wetland sites remains extensive and indicates the
importance of associated resources. A decline in formal grinding technology is notable.
Different methods of use or a focus on those species requiring less extensive
technological investment is more likely related to this shift in technology than the
decreased importance of plant foods. A change in residential mobility and the need for
more portable tool kits might also have contributed to such a trend. Likewise, an increase
186
in the diversity and abundance of pottery could be indicative of either a shift in
processing methods (i.e., boiling instead of grinding), or a new investment in decorative
elements less related to basic subsistence. More detailed analyses of possible shifts in
land-use and subsistence during Oasis 3 are contained in chapters 4 and 6.
Eneolithic of the Gobi Desert and Oasis 3 technology
The majority of dated Oasis 3 sites are from the Gobi-Altai and the Alashan Gobi (see
Figure 3.1). There are five archaeological sites dated to between 5.0-3.0k cal yr BP and
assigned to the Eneolithic or Oasis 3, including: components of Baron Shabaka Well, Site
19 (3.3k cal yr BP [3115 + 47 BP]) in the East Gobi; Shabarakh-usu 1 (4.9k cal yr BP
[4308 + 40 BP]), Shabarakh-usu 4 (4.0k cal yr BP [3680 + 76 BP]), Shabarakh-usu 10
(3.9k cal yr BP [3595 + 41 BP], 3.5k cal yr BP [3246 + 39 BP]), in the Gobi Altai; and
Jabochin-khure (3500 + 300 ka), Gashun Well (3.6k cal yr BP [3385 + 40 BP]),
components of Yingen-khuduk (3910 + 300 ka, 3910 + 230 ka), and components of
Mantissar 12 (3840 + 340 ka), in the Alashan Gobi. Several Oasis 3 assemblages are
considered temporally coherent and offer clear examples of typical tool kits. Detailed
descriptions of dated sites are offered in Appendix B.
One of the most important post-LGM localities in the Gobi Desert is Shabarakhusu. Many separate sites were collected from the same locality during the 1925 Central
Asiatic Expedition. However, unlike at similar localities such as Yingen-khuduk and
Baron Shabaka Well, each assemblage was separately curated at the AMNH and fully
described by Nelson in his field notes (1925). For this reason, sites at the Shabarakh-usu
187
locality were most heavily sampled for chronometric dating. All the sites from this group
have been dated to the Oasis 3 phase, but two distinct periods of Oasis 3 habitation are
represented (see Figure 3.1). Shabarakh-usu 1 is dated to the Oasis 2-Oasis 3 transition
(4.9k cal yr BP [4308 + 40 BP]), while the two remaining sites span the period from
about 4.0-3.5k cal yr BP. Despite low carbon yields for Shabarakh-usu 4, the date
produced for this site (4.0k cal yr BP [3680 + 76 BP]) was similar to that from the surface
of Shabarakh-usu 10 (3.9k cal yr BP [3595 + 41 BP]).
Differences between Oasis 2 and Oasis 3 assemblages are present in both the
ceramic and lithic components, although changes in ceramic technology are most
recognizable. Ceramics are more common than in Oasis 2 and more intensive
manufacturing processes were involved in the production of certain pottery types. Highfired ceramics began to be produced. Homogeneous red-ware, including painted pottery
(Figure 3.7), from the Alashan Gobi and Baron Shabaka Well in the East Gobi appear to
have been fired at temperatures suggestive of formal kiln firing. Low-fired ceramics
were still regularly used. In addition to the brown and grey-wares typical of Oasis 2, red
and reddish-brown-wares become common and may indicate the effects of higher firing
temperatures on local clays.
188
a.
b.
c.
d.
Figure 3.7 Painted pottery from the Gurnai Depression. From sites: a. Mantissar 8, K.
13294:1; b. Mantissar 12, K. 13298: 1-4; c. Mantissar 7, K. 13293: 2; and d. Mantissar
12, K. 13298: 5.
A greater diversity of surface treatments is recognized from Oasis 3 assemblages
(see Figure 3.8, Table 3.5). Common surface finishes include string-paddled, incised
geometric designs, stamps (including a shard from Baron Shabaka Well that was
impressed using a roller stamp with evenly spaced rows of square punctates or “toothed”
impressions Figure 3.8e), channelled ware, moulded rims, and raised clay bands that were
moulded or incised. The majority of pottery from Shabarakh-usu 1 was string-paddled,
but there is also evidence of moulded rims and incised geometric patterns. Shabarakhusu 4 and Shabarakh-usu 10 both contain a wide range of other diagnostic of Oasis 3
pottery types (see Appendix B). Handles and miniature lugs are new features that are
189
found at both early and late Oasis 3 sites. A variety of vessel shapes are noted (see Table
3.5), although comparative data on Oasis 2 vessel shapes is lacking. One partially
reconstructed string-paddled bowl from Gashun Well is dated to 3.6k cal yr BP, but flatbottomed pots appear to have been widely used. Cylindrical flat-bottomed vessels were
found in East Gobi sites.
Oasis 1
Firing
low
Oasis 2
Firing
low
Temper
coarse sand and organic
temper
Temper
sand and organic temper
fine sand temper
coarse sand temper
Finish
thin cord markings
Form
unknown
Finish
Net-impressed
Textile
Cord markings
Form
straight-walled
unknown
Oasis 3
Firing
low
high
Metal Ages
Firing
low
high
Temper
none
sand and organic
fibre
sand
mica or nacre
Temper
sand
sand and fibre
unknown
Finish
Paddled (string/cord)
Geometric-incised
Channelled
Painted
Smeared basket
Burnished
“Toothed” roller stamp
Raised/incised clay
bands
Moulded rim
Finish
Stamped
Moulded rim
Raised clay bands
unknown
Form
unknown
Form
Straight-walled
flat-bottomed
cylindrical
globular
bowls
handles
lugs
Table 3.5 Characteristics of pottery associated with each period. Based on direct dates
and associated assemblages.
190
a.
b.
c.
f.
d.
e.
g.
h.
Figure 3.8 Examples of Oasis 3 pottery from dated sites: a. channelled-ware, Shabarakhusu 10; b. “checker-stamped”, Shabarakh-usu 10; c. “geometric-incised”, Yingenkhuduk; d. “geometric-incised”, surface collection of Shabarakh-usu 1 and 2; e. rollerstamped punctate design, Baron Shabaka Well; f. incised raised clay band, Shabarakh-usu
1; g. incised raised band, Shabarakh-usu 10; and h. moulded rim, Yingen-khuduk.
191
There appears to have been little change in the types of temper used; sand or a
combination of sand and organic temper is most common. As observed for Oasis 2 sites
(see above), regional variation in temper types continues during Oasis 3. Sand-temper is
most common and is sometimes mixed with organics, shell, or unidentified mediums.
The majority of shards from the Alashan are heavily tempered with sand, though grainsize varies. Evidence of organic temper is rare at Alashan sites. Untempered-wares are
more common here than in the rest of the Gobi Desert. Many East Gobi ceramics are
distinct from those in western Gobi Desert sites. One common tempering medium in the
East Gobi occurs only occasionally in pottery from the Gobi-Altai or the Alashan; it is a
shiny material that might be mica or nacre (“mother-of-pearl”). Another distinct type of
temper common in East Gobi sites (16% of shards from Baron Shabaka Well) is
unidentified; when broken, the interior paste is characterized by a spongy, porous texture
that is usually blackened (Figure 3.9). Some sort of organic temper is suggested.
Inclusions of mica/nacre are usually associated with this type of fabric. Similar
specimens were also recovered with less frequency from Gobi-Altai sites.
While pottery manufacture appears to have been more important during Oasis 3,
the use of large grinding tools declined. Large, formal grinding stones are most common
in the East Gobi at Oasis 2 sites. Such tools are rare farther west and there little evidence
for use of the ground pestles (one exception is a possible fragmentary specimen from
Abdertungtei in the Alashan Gobi; see Maringer, 1950: Plate XXXIX), rollers, and saddle
querns that have been found in East Gobi sites. Heavier and more formal types of
grinding equipment from western sites include large, rather flat hand stones with one
192
angled edge from Abderungtei (K. 13209: 139), Yingen-khuduk, and Ulan Nor Plain, and
a large, flat grinding slab from the Shabarakh-usu (Bayan-dzak) collections housed at the
Institute of Archaeology, Mongolian Academy of Science. Oasis 3 assemblages
sometimes include grinding stones or “rubbing” stones (see Fairservis, 1993), but they
are smaller and presumably more portable.
Figure 3.9 Example of “spongy-textured” paste common in East Gobi sites.
Lithic assemblages represent a clear continuation of flintknapping techniques with
only minor variations. Microblades are frequently retouched into perforators and other
tools. Endscrapers on microblades are rare in Oasis 2 sites, but more common during
Oasis 3. Drills with expanded bases made on microblades are a new artefact type.
193
Although adze/axes were identified in Oasis 2 sites, fully polished specimens are rarer
and associated with Oasis 3 habitations. Chipped adze/axes are found in sites from all
regions beginning in Oasis 2, but seem to be more widespread during the later phase.
Larger knife blades were also added to the tool kit (Figure 3.10). In comparison
with smaller Oasis 2 bifacial blades, the longer size, curved shape, and single rounded
end of Oasis 3 specimens suggest that these artefacts were intended to function as fullsize hafted blades. Such tools are also recognized in upper levels of Ulan-khada in the
Lake Baikal region, dating to 4.0k cal yr BP or slightly later (Khlobystin, 1969;
Goriunova and Khlobystin, 1991). The earliest example of curved knife blades comes
from Ulan Nor Plain, dated to about 5.8k cal yr BP. New types of bifacial projectile
points were also introduced and included forms with stemmed and convex bases (Figure
3.11). Such projectile points were found at the following sites: Camp Ruined Lamasery
Obo (Site 11/11A), Baron Shabaka West (Site 20), and Paoling Miao Southeast (Site 31),
in the East Gobi; Shabarakh-usu 1 and Shabarakh-usu 2, in the Gobi-Altai; and Mantissar
7 (K. 13293), in the Gurnai Depression of the Alashan Gobi. The extensive use of
bifacially flaked tools is typical of late Oasis 2 and Oasis 3 assemblages, and they appear
to have played an increasingly important role in tool kits throughout the Neolithic and
Eneolithic.
194
Figure 3.10 Example of large curved bifacial blade from Shabarakh-usu 4.
Figure 3.11 Examples of Oasis 3-type bifaces from Shabarakh-usu 2. On the right are
examples of biface blades, including one small, but slightly curved specimen and one
specimen typical of Oasis 2 “bifacial inset blades.”
195
Microblade core reduction strategies underwent subtle regional shifts during
Oasis 3. Wedge-shaped cores are less common and conical and cylindrical forms are
dominant. Notably, wedge-shaped cores are found more regularly in Alashan Gobi sites
during this period than in the East Gobi or Gobi-Altai. Cylindrical cores are sometimes
characterized by use of opposing ends as striking platforms. Numerous stubby
cylindrical cores indicate heavy reduction by successive platform rejuvenation. Such
core types are less common during Oasis 3 in the East Gobi than in the Gobi-Altai and
the Alashan Gobi. In the western Gobi Desert barrel-shaped cores, or massive cylindrical
microblade cores, are probably diagnostic of the Oasis 3 period, though such cores were
present during Oasis 2 in East Gobi sites.
A series of partially prepared microblade cores from Yingen-khuduk indicates the
use of roughly cylindrical preforms with a heavily prepared platform (Figure 3.12a).
Cortex removal and nodule shaping on the Yingen-khuduk specimens appears to focus on
the production of thick, round preforms with a predetermined cylindrical form that is
divergent from the flatter morphology exemplified at Ulan Nor Plain (Figure 3.12b). At
Ulan Nor Plain, preforms were made on flat nodules or nodules with one side removed
for thinning. One or more of the sides were retouched in order to create a wedge. The
two reduction strategies are not necessarily temporally distinct, since cores produced
from both types of preforms are present in the assemblages, but rather represent different
strategies favouring either the production of cylindrical or wedge-shaped cores.
196
a.
b.
Figure 3.12 Microblade core preforms from Yingen-khuduk (a) and Ulan Nor Plain (b).
197
Summary Eneolithic developments and Gobi Desert Oasis 3
Many Oasis 3 archaeological sites have been dated to between 4.0-3.5 kya, the period
immediately preceding the archaeologically visible rise of nomadic pastoralism.
According to existing research, the Neolithic-to-Bronze Age transition or the early
Bronze Age in Mongolia is thought to have taken place between about 4.0-3.0 kya. The
characterization by Soviet and Mongolian archaeologists of tool kits from this period
indicates the continued use of flake tools and microblade inserts along with an increase in
bifacially retouched tools such as finely pressure-flaked projectile points. These earlier
observations are consistent with those recorded here. In contrast, the Eneolithic upper
levels of the Yenesei River region site Toora-Dash indicate the continued use of flake
technology, but an absence of arrowheads and bifacial inserts (Vasil’ev and Semenov,
1993).
Soviet literature suggests that animal husbandry and plant cultivation were of
some importance to local economies in the steppe and forest zones of eastern Mongolia,
but no concrete evidence of domestication has been recognized in the Gobi Desert.
Evidence of agriculture includes hoe-like implements and ring-shaped “counter-weights,”
which are thought to be indicative of low-level agricultural production (Cybiktarov,
2002). Several tools from the Baron Shabaka Well locality might be similarly
categorized (see Appendix B), but are inconclusive evidence of cultivation. Based on
excavations at Tamsagbulag and other archaeological sites, it has also been proposed that
the Neolithic-to-Bronze Age transition was typified by more temporary campsites than in
earlier phases of the Neolithic (Séfériadès, 2006). Thus far, all post-glacial Gobi Desert
198
archaeological sites suggest high residential mobility and there is little evidence of
permanent or semi-permanent occupations. The possibility of higher residential mobility
at this time contradicts the likelihood of agricultural production, which is typically
associated with increased sedentism. Examples of probable spindle whorls might
similarly be taken as evidence of the use of spun wool, but are not dated could have
served for spinning other types of fibres, such as bast.
Copper smelting is indicated by pieces of slag and a few metal artefacts from sites
in Siberia and along the border of Mongolia. In the Lake Baikal region of southern
Siberia, copper and/or bronze artefacts first appear during the Glazkovo period (5.2-3.4k
cal yr BP after Weber and Bettinger, 2010). Limited evidence of smelting suggests lowlevel local production in the forest region of northeastern Mongolia (Cybiktarov, 2002;
Séfériadès, 2006). Several microlithic assemblages from the Gobi Desert contain
evidence of copper or bronze slag. Dottore-namak (K. 13248) is one such site in the
Alashan Gobi, but ceramics were dated to about 2.5 ka (3540 + 1060 ka, 2740 + 200 ka)
(Table 3.1), indicating that the assemblage dates to the late Bronze Age or early Iron Age.
Oasis 3 represents a period of continued intensive occupation of dunefield/wetland environments in the Gobi Desert. Extensive remains in the dune-field
marshland environment of the Gurnai Depression (Maringer, 1950: 151-163) exemplify
the importance of such habitats to Gobi Desert groups in this later phase. Hunting and
gathering were most certainly the basis of local economies. Although fishing is of
particular importance in much of Northeast Asia, Gobi Desert tool kits show little
evidence of intensive fish exploitation. Expedition field notes make frequent references
199
to the diversity of bird life in the Gobi Desert, particularly around lakes, and birds may
have been an important food source. Plant foods would have been especially abundant in
dune-field/wetland environments and would continue to have been important during this
time. Some sites were found in steppe environments, suggesting that hunters continued
to take advantage of abundant wild ungulate herds like gazelle, horse, and khulan.
A possible shift in the relative importance of formal grinding technology is
represented by the use of smaller informal grinding tools and “rubbing stones” in GobiAltai and some East Gobi Oasis 3 sites. All types of grinding stones are rare in the
Alashan Gobi. Larger grinding tools may have continued to be used in Oasis 3 sites from
the East Gobi, but they are rare and may have been derived by later inhabitants from
Oasis 2 occupations. Decreased investment in milling technology suggests a shift in the
use and processing of local plant foods. A desire for more lightweight grinding stones is
implied and is intriguing when combined with an increased investment in pottery
production. While enhanced portability is suggested by the adoption of more lightweight
technology, the more frequent use of pottery could be considered contradictory.
Nevertheless, pottery is not as heavy as grinding stones, and can be attached to thongs for
transport. Unlike large grinding stones, ceramic vessels are frequently used in mobile
societies (Hoopes, 1995; Bollong et al., 1997; Pavlů, 1997; Bright and Ugan, 1999; Rice,
1999; Eerkens et al., 2002). At the same time, the use of high-fired and more decorative
pottery (e.g., stamped, painted, incised designs rather than simple paddled pots) indicates
that some vessels were probably manufactured with the intention of longer-term curation.
200
Grinding stones are usually assumed to indicate the production of flour, which
hunter-gatherers could then mix with grease and other foods and store for delayed
consumption (for examples see Manne, 2012). In other cases, grinding stones are thought
to have been related to macerating foods for supplemental feeding or weaning of infants
(Hillman, 1989). Many ethnographic cases of pottery use among hunter-gatherers
indicate that the technology was frequently used to prepare stews (Bollong et al., 1997;
Eerkens, 2004; Malainey et al., 1999; Mercader et al., 2000; Pratt, 1989), the broth of
which can also provide substantial nourishment for feeding infants. Storage of dried
seeds or plants might be an alternate function of curated pottery vessels, which is a
possibility that can be explored partially through studies of usewear (i.e., relative
frequency of evidence for cooking or lack thereof on shards). A shifting emphasis
between these two technologies may suggest different trends in the relative importance of
food processing and/or transport methods.
The exploitation of ostrich eggshell for beads seems to have flourished during
Oasis 3. Many dune-field sites from this period contain older eggshell scavenged for the
manufacture of beads. Heightened evidence for the production of ostrich eggshell beads
along with marked diversity in ceramic styles shows the increased importance of
ornamental elements.
Oasis 3 is broadly characterized by increased investment in the manufacture of
pottery vessels and adze/axes, decreased investment in formal grinding technology, and a
greater variety of bifacially flaked tools.
201
3.2.3. Estimates of ages for undated sites
Chronometric dates on pottery and associated artefacts have been used to identify several
key diagnostic artefact types and technological trends for each period. These findings are
summarized below in Table 3.6. Diagnostic pottery types are summarized above in Table
3.5. In order to make a more reliable assessment of land-use and mobility from the
Epipalaeolithic to Eneolithic, it is important to determine the relative age of as many sites
as possible. By using initial findings based on chronometric dates, a number of undated
sites can be assigned a relative age. Expanding the site sample size for each period will
also contribute to preliminary assumptions about artefact chronology.
Estimated ages are assigned based on data collected during analysis of collections
and are limited primarily to those sites for which detailed quantitative and qualitative data
was obtained. The majority of data is presented in the form of tables in order to ease
presentation of extensive data. Table 3.7 lists sites for which data was collected in 2008
and 2009, with limited representation of sites analyzed in 2004. Estimated ages are based
on the presence of diagnostic artefacts (e.g., unifacial points, biface blades, small chipped
adze/axes, polished adze/axes, distinct pottery types) or the relative frequency of certain
artefact types (e.g., wedge-shaped cores, cylindrical/conical cores, endscrapers on
microblades, adze/axes). Artefact lists for each site are published in Maringer (1950) and
Fairservis (1993). The distribution of artefact types in each assemblage was a key
consideration in establishing a relative age for each site (see Appendix C); however, the
use of photographic images was also extremely important and allowed for the recognition
of less easily quantified traits.
202
Pre-LGM
Palaeolithic
Large pebble
tools
Cores
“Levallois”
type
Blade
Microblade
(rare)
Flake
Microblade
cores
Pressureflaked
Sub-prismatic
Flake tools
Elongated
flakes
Various
retouched
Early
Epipalaeolithic
Large pebble
tools
Bone/antler
tools
Cores
“Levallois”
type (rare)
Blade
Microblade
Flake
Microblade
cores
Boat-shaped
Wedge-shaped
Conical?
Flake tools
Sidescrapers
Endscrapers
Large unifacial
points
Various
retouched
Oasis 1/ Late
Epipalaeolithic
Large and small
pebble tools
Informal
grinding stones
Pottery
Bone/antler
tools
Cores
Microblade
Flake
Microblade
cores
Boat-shaped
Wedge-shaped
Conical
Cylindrical
Microblade
tools
Insets
Various
retouched
Flake tools
Endscrapers
Various
retouched
Oasis 2
Oasis 3
Metal Ages
Large formal
tools
Formal grinding
stones
Pottery
Bone/antler
Large formal
tools
Informal
grinding stones
Pottery
Bone/antler
Slag
Bronze
projectiles
Pottery
Cores
Microblade
Biface
Flake
Cores
Microblade
Biface
Flake
Microblade
cores
Wedge-shaped
Conical
Cylindrical
Microblade
cores
Wedge-shaped
Conical
Cylindrical
Barrel
Microblade
tools
Unifacial points
Perforators
Awls
Various
retouched
Flake tools
Bifacial points
Endscrapers
Perforators
Bifaces
Projectile points
Shouldered
points
Large points
Inset blades
(parallel-sided)
Large knives
Large tools
Adze/axes
(ground, edgeground,chipped,
micro)
“Hoes”
Microblade
tools
Endscrapers
Perforators
Expanded base
drills
Flake tools
Bifacial points
Endscrapers
Perforators
Various
retouched
Bifaces
Projectile points
(straight,
stemmed,
convex, fishtailed bases)
Blade knives
Large tools
Adze/axes
(polished, edgeground,chipped)
Grooved slabs
Table 3.6 Summary of artefact chronologies. Based on dated assemblages and
comparative regional data.
Cores
Microblade
Flake
Microblade
cores
Conical
Cylindrical
Flake tools
Various
retouched
203
It is difficult to distinguish Oasis 2 from Oasis 3 sites in the absence of diagnostic
pottery types. Pottery is mostly found in sites near dune-fields and larger water sources,
making it necessary to use other types of artefacts to recognize different phases within
Neolithic/Eneolithic assemblages. Several Oasis 3 chronological markers have already
been noted in the discussion of dated Gobi Desert assemblages: expanded base drills,
curved bifacial blade knives, projectile points with stemmed, fish-tail and convex bases,
barrel-shaped cores, endscrapers on microblades or elongated flakes, and finely polished
adze/axes. Some of these artefact types (including endscrapers on microblades) have
been found in limited numbers in late Oasis 2 sites, but are primarily restricted to Oasis 3
sites. The association of several kinds of Oasis 3-type artefacts is most definitive of age.
Whenever possible, estimating the approximate age of a site was more reliant on the
range of associated artefacts, rather than one or two diagnostics (see Appendix C).
204
SITE
SKW
3
6D
7
9
9B
9C
9D
10/10A/
10B/10C
11/11A
12
12/12A/
12B
13
13A
14
15
18
ENVIRONMENT
Mountains/river
Dunes/valley/river/ba
dlands
Hillslope/river
Hillside/river
Mesa/river
Hilltop
Base hill/river
Hills
Hills/stream
EAST GOBI
PERIOD
Oasis 1
Oasis 1
Oasis 1
Oasis 3
unknown
Palaeolithic?
Oasis 2 or 3
Oasis 2 or 3
Oasis 2 or 3
Knoll/dune/stream
Plains/stream
Mountains/stream/so
me on plains (12)
Mountains/stream
Mountains/stream
Mountains/stream
Hill at head of wash
Dune/basin/lake
Oasis 3
Oasis 2
Epi/Palaeolithic,
Oasis 2
Late Oasis 2?
Oasis 2 or 3
Oasis 1
Oasis 2
Early Oasis 2
19
20
Dune/basin/lake
Valley/dunes/lake
Oasis 2, Oasis 3
Oasis 2 and/or 3
20A
21
Dune/basin/lake
Hillside/dunes/lake
Various
Early Oasis 2
23/23A
28
29
Mesa/river
Hillside/lake/sand
Valley slope/sand
30/30A
Valley bottom/river
Oasis 3
Oasis 3
Late Oasis 2 or
early Oasis 3
Late Oasis 3
31
Valley/dune/river
34
35
36
Mesa /lake
Dune/basin/lake
Dune/basin/lake
Late Oasis 2 and
early Oasis 3
Oasis 3?
Early Oasis 2
Oasis 1
QUALIFICATIONS
Dates
Core/scraper types, large flakes,
microblades, thick bifacial knife
Core/scraper types
Backed microblades, cowrie, pottery
1 core preform only
Debitage, rough macrotool
Scraper types
Partially retouched point, scrapers
Mixed sites, few diagnostics, biface
fragments
Mixed?, core types, biface types
Mixed?, core/scraper types, material
Mixed, core/scraper types, material,
debitage
Core/scraper types, raw materials
Core/scraper types, biface, material
Core/scraper types, raw materials
Biface, core/scraper types, material
Unifacial points, grinding stones, polished
stone, pottery; also Historic
Dates
Pottery, bifaces, drill, grinding stones, core
and scraper types
Mixed
Unifacial points, cores, grinding stones,
chipped adze/axes
Pottery, grinding stones, material
Macrotools, bifaces, material, cores
Core/scraper types, bifaces, drills, polished
macrotool, raw materials
Macrotool, grinding stone, core/scraper,
bifaces, possibly mixed
Mixed, core/scraper types, macrotools,
bifaces, grinding stones
Based on chalcedony biface blanks
Dates
Microcore/scraper types/material
Table 3.7a Age estimates for studied archaeological sites, East Gobi. SKW = Shara
KataWell.
205
GOBI-ALTAI
PERIOD
Epipalaeolithic
and Oasis 3
Early Oasis 2
Oasis 1?
Palaeolithic
Late Oasis 2 and
some Oasis 3?
Early Oasis 2
Late Oasis 2
Epipalaeolithic
Oasis 2
Metal Ages?
SITE
CM
ENVIRONMENT
Mesa/stream
GBW
BUV
DH
UNP
Basin/well
Mtn valley/dunes
Foothills/well
Dune/basin/stream
JW
JW s-s
SG
AB
KhO
BD
Dune/basin/stream
Dune/basin/stream
Foothills
Foothills/spring
Mountain
meadow/streams
Dune/basin/lakes
ON
SC
Su 1
Su 1A
Su 2
Su 2a
Plains/dune/lake
Plains/stream
Dune/basin/lake
Dune/basin/lake
Dune/basin/lake
Dune/basin/lake
Su 2b
Su 4
Su 7
Dune/basin/lake
Dune/basin/lake
Dune/basin/lake
Oasis 2 and Oasis
3
Oasis 2 and 3
Oasis 3
Early Oasis 3
Oasis 2
Early Oasis 3
Epipalaeolithic or
Oasis 1
Oasis 1
Oasis 3
Early Oasis 3
Su 8
Su 10
Su 11
Su 13
Dune/basin/lake
Dune/basin/lake
Dune/basin/lake
Dune/basin/lake
Late Oasis 2
Oasis 3
Late Oasis 2
Oasis 3
QUALIFICATIONS
Differences in surface abrasion,
core/scraper types, debitage
Pottery, cores, mortar
Core/scraper types, weathering
Debitage, large flakes
Dates, some pottery types might be Oasis
3, curved biface; also Historic
Bifaces, core/scraper types
Bifaces, core/scraper types/material
Core/scraper types, no microblades
Biface, core/scraper types, blade tools
Raw material, core/scraper types, presence
of Metal Age monuments
Core/scraper types, pottery, polished
macrotool; also Metal or Historic
Pottery, core/scraper types; Historic
Core/scraper types
Dates
Pottery, core/scraper types
Pottery, macrotools, biface types
Raw material, core types, microblades
Core types, eggshell – 9.4k cal yr BP
Dates and comparison to Su 10
Core/scraper/biface types, pottery,
shouldered drills and perforators
Pottery, bifaces
Dates
Bifaces, raw material
Pottery, bifaces, drills
Table 3.7b Age estimates for studied archaeological sites, Gobi-Altai. Shabarakh-usu
sites 2, 8, 11 and 13 were analyzed in 2004 (Janz, 2006) and quantitative data is not as
comprehensive, although additional photographs were taken in 2008. CM = Cemetery
Mesa, GBW = Gashuin Bologai Well, BUV = Barongi Usu Valley, DH = Dubshi Hills,
UNP = Ulan Nor Plain, JW = Jichirun Wells, JW s-s = sub-surface component of Jichirun
Wells, SG = Sairim Gashato, AB = Artsa Bogdo, KhO = Khunkhur Ola, BD = Barun
Daban, ON = Orok Nor.
206
SITE
176
ENVIRONMENT
Dune/lake
ALASHAN
PERIOD
Oasis 3
179
183
186
188
Dune/lake
Plains/dunes
Plains/dunes
Plains/basin
Oasis 3
Oasis 3
Oasis 2?
unknown
202
203
204
Plains/basin
Plains/basin
Plains/basin
Epipalaeolithic?
Oasis 3
Epipalaeolithic
207
208
209
210
212
213
216
218
219
220
222
223
Sand/basin/well
Dune/basin/river
Basin/stream/well
Basin/stream/well
Dune/lake
Dune/basin
Foothill/spring
Mountain/stream
Mountains
Mountain/stream
Mountain/well
Mountain/well
226
229
230
Valley/river
Valley/river/sand
Mountain/cave
Oasis 3
Oasis 3
Oasis 3
Oasis 3
Oasis 2 and 3
Oasis 2
Oasis 3
Oasis 3
Palaeolithic
Oasis 2
Oasis 3
Late Oasis 2 or
Early Oasis 3
Oasis 1
unknown
Oasis 3
231
237
240
Mountain/river
Valley/river/sand
Hill plains
Oasis 2
Palaeolithic
Oasis 3
241
247
248
251
258
259
Hills/swamp
Dune/basin
Dune/basin/spring
Plains/sand
Basin/lake/sand
Basin/lake/sand
Oasis 2?
Oasis 2
Metal Ages
unknown
Oasis 2 or 3
Oasis 2?
277
280
287
289
Dune/lake
Dune/lake
Dune/wetland
Dune/wetland
Oasis 3?
Oasis 3
Oasis 3?
Oasis 3
QUALIFICATIONS
Pottery, core types, scraper types, partially
polished axe
Pottery like 176
Painted pottery, polished axe, cores
Large microblades, scraper types
Microblade debitage and large
flakes/cobbles
See 204, intrusive pottery?
Dates
Macrotools, few microlithics, no
microblades
Dates
Bifaces, macrotools, core types
Biface/pottery types
Biface/pottery types
Dates
Partially polished axe, various cores
Core types, scraper types, bifaces –drills?
Core types, scraper types, biface fragment
Scraper on macroflake
Inset blade knife, “stray finds” = mixed?
Biface frag., scraper types, biface core
Spearpoint, bifaces, core/scraper types,
pottery
Core types
Odd undiagnostic macrotool or core tool
Pottery, chalcedony bead making, one
scraper on microlithic flake
Large bifaces, blade core, late core types
Core types, lack of microliths
Spindlewhorls, pottery, polished axe,
grooved slab (also grinding stone)
Core types, chipped adze/axes
Core types, scraper types, drills, bifaces
Dates
1 scraper only
1 cylindrical core only
Aceramic, biface types, hoe-like tools,
awl, unifacial perforator, adze/axe types
Slag, pottery, scraper types, Metal Age?
Painted pottery
Pottery, cores, scraper types, Metal Age?
Pottery, biface
Table 3.7c Age estimates for studied archaeological sites, Alashan Gobi.
207
ALASHAN
PERIOD
Oasis 3
SITE
290
ENVIRONMENT
Dune/wetland
293
294
296
298
Dune/wetland
Dune/wetland
Dune/wetland
Dune/wetland
303
307
311
316
321
322
Dune/wetland
Dune/wetland
Dune/wetland
Dune/wetland
Plains
Plains/spring
Oasis 3
Oasis 3
Oasis 3
Oasis 2 and
Oasis 3
Oasis 3
Oasis 3?
Oasis 3
Oasis 3
Early Oasis 2
Palaeolithic
323
324
Mountain/spring
Mountain/spring
Epipalaeolithic
unknown
(continued)
QUALIFICATIONS
Pottery, core/scraper types, bifaces,
polished stone frag., turquoise frag., drills
Painted pottery, biface, drill, core/scrapers
Painted pottery, cylindrical core
Painted pottery
Dates
Scraper types
Slag, pottery, core/scraper types, mixed
Scraper types
Similar raw materials to other sites
Core/scraper types, blade knife, adze/axe
Large flake tools, thumbnail scrapers
(intrusive?)
Core/scraper types, blade
1 amorphous core on cobble
Table 3.7c (continued) Age estimates for studied archaeological sites, Alashan Gobi.
Sites 209, 210, 240, 241, 280, 289, and 296 were not analyzed, but assigned to a period
based on presence of diagnostic artefacts. Shara-khulus (K. 13241) also includes the
fragment of a pottery spindlewhorl, which were also recovered from Altat (K. 13240).
One challenge with respect to East Gobi archaeological sites was the presence of
formal grinding stones. In general, formal, often finely polished pestles and rollers are
characteristic of Oasis 2 assemblages like Chilian Hotoga and Baron Shabaka Well. The
occasional occurrence of such artefacts in sites with a typical range of Oasis 3 core,
scraper, and biface types is problematic. While the presence of grinding stones is rare in
the Alashan Gobi, most Oasis 3 sites in both the East Gobi and the Gobi-Altai contain
only small grinding slabs or what are sometimes referred to in AMNH catalogues as
“rubbing stones”. Use of formalized grinding stones may have continued in the East
Gobi during Oasis 3, or they may have been scavenged from Oasis 2 sites as is done by
modern herders (see Pond, n.d.: 90-91). Baron Shabaka West (Site 20) is the best
208
example of this situation, where large formal grinding stones are associated with Oasis 3
style pottery (i.e., pottery with rolled cord finish, and high-fired stamped grey-ware
similar to that dated to 3.3k cal yr BP from Baron Shabaka), a blade knife, an expanded
base drill, and core/scraper types. While the grinding stones may have been scavenged or
artefacts from two different periods of occupation mixed, the most important point is that
the presence of formal grinding stones does not definitively distinguish an Oasis 2 period
occupation.
Despite the difficulty in properly assigning sites to a particular period or phase,
temporal associations between certain artefact types were strengthened in the process.
Considering assemblage inventories, quantitative data, and photographic images from
numerous sites allowed for more comprehensive recognition of possible diagnostic traits
in lithic assemblages. Just as expanded base drills are associated with dated Oasis 3 sites,
artefact associations within additional undated sites suggested that fully retouched awls
(recognized by long triangular shape as illustrated in Maringer, 1950: Plate XXIX, 14)
are more closely related to late Oasis 2 occupations. Large thin bifacial flake knives are
similarly assigned to Oasis 2, as are the largest formal tools like “hoes” or the massive
pick found at Baron Shabaka Well (see Appendix B). An abundance of bifacial tools are
more often associated with Oasis 3. While bifaces began to be used during Oasis 2, small
bifacial projectile points are most common in Oasis 3 assemblages. Intensive production
of small bifacial points, as exemplified by Shabarakh-usu 7 and 11 (see Fairservis, 1993;
Janz, 2006), is associated with Oasis 3. Bifacial points appear to be less common in late
Oasis 3 sites and are so far absent in sites dated to the Metal Ages.
209
Use of a limited range of high quality raw materials is also characteristic of Oasis
2 and early Oasis 3 assemblages. The increasing reliance on a select range of
homogeneous cryptocrystalline raw materials began in the late Epipalaeolithic and is a
trend that has been noted at Chikhen Agui (Derevianko et al., 2003). A limited selection
of reliably high quality raw materials is particularly noticeable in the Gobi-Altai, where
homogeneous cryptocrystalline stone is easily obtainable. At the same time, many
microlithic assemblages across Mongolia contain artefacts made on red or purple jasper
characteristic of the Gobi-Altai region (Joshua Wright, personal communication, May
2011). Assemblages characterized by the use of more roughly flaked projectile points on
a range of local, less homogeneous crytocrystallines, including chalcedony pebbles, may
date to the end of Oasis 3. Chalcedony appears to have been preferred for the
manufacture of small bifacial tools, and the increased predominance of those tools may
be reflected in a decline reliance on exotic jaspers. These preliminary observations must
be tested through the analysis of raw material frequencies amongst the debitage of
directly and indirectly dated sites, but this data was not collected in the course of my
study. A better understanding of tool stone distributions across the Gobi Desert is
necessary.
In some instances, sites can be identified only as Oasis 2/Oasis 3. Late Oasis 2
sites (e.g., the late Oasis 2 site of Ulan Nor Plain, which contains blade knives and
endscrapers on microblades/elongated flakes; see Appendix B), and early Oasis 3 sites
(e.g., Shabarakh-usu 1; see Appendix B) are most difficult to distinguish. Late Oasis 2
assemblages show the initial incorporation of Oasis 3-type technologies. Assigning a
210
relative age to smaller assemblages is most problematic because they often appear to
have been related to raw material procurement or short-term task sites where finished
diagnostic artefacts, including cores, were rarely discarded. The situation is especially
notable in the East Gobi. Late Oasis 2 and early Oasis 3 assemblages are recognized by
the use of only a few high quality raw materials, as opposed to a wider range of local,
often poor quality stones. The consistent focus on high quality raw materials at type sites
like Baron Shabaka Well and Yingen-khuduk support this categorization. Core reduction
strategies were also considered and artefacts like keel flakes indicate increased core
preparation typical of Oasis 1 and 2.
Future refinements of Neolithic/Eneolithic artefact chronologies will allow for
more precise categorizations, but similarities in artefact types across Oasis 2/Oasis 3 are a
reminder that although period designations are useful for grouping like sites they are
incapable of capturing the transitional nature of local technological change. A
designation of “late Oasis 2 or early Oasis 3” encompasses the period between 6.0-4.0k
cal yr BP. After this, only slight shifts in technological strategies appear to have
occurred, including a decline in the frequency of finely finished projectile points, more
regular use adze/axes, another increase in the variety and relative frequency of pottery,
and a decline in the importance of high quality exotic stone. A decline in the relative
importance of microblade production might also have occurred, but the possibility should
be explored in more detail when additional late Oasis 3 sites have been dated.
211
Several other significant observations should be made:
a.
Pottery first appears in Oasis 1 sites, but is not common until Oasis 3.
b.
Wedge-shaped cores appear to be more typical of Oasis 1 and 2 assemblages
in the East Gobi and the Gobi-Altai, but are more representative of late Oasis
2 and/or early Oasis 3 assemblages in the Alashan Gobi.
c. Alashan Gobi occupations are notable for the high frequency of macrotools or
large formal tools such as adzes, axes, gouges, and chisels. Currently, sites
with large, thin bifaces and awls are assigned primarily to Oasis 2. Most of
these sites are aceramic, even the large lakeshore site of Gashun-nor (K.
13259).
d. Two types of Oasis 3 sites are distinguishable – those with few microblades
and lower quality raw materials, and those with more microblades and high
quality raw materials. It is difficult to assess whether this is related to
chronology or simply due to situational constraints. For example, highly
localized raw material constraints such as those related to seasonal mobility
could have created such a pattern. Oasis 2 sites do often contain very finely
finished projectile points and microblade tools, along with more uniform and
slender microblades made on a narrow range of raw materials. However, the
later Oasis 3 site of Gashun Well (3.4k cal yr BP) shows the continued use of
high quality raw materials and intensive microblades manufacture. Additional
research on this issue is required.
212
e. Ta Sur Heigh (Site 7) in the East Gobi contains evidence for the use of exotic
cowry (Cypraeidae) shells, which are also found in Neolithic and Early
Bronze Age sites in northern China and are associated with long distance trade
(Peng and Zhu, 1995). Other cowry finds in the East Gobi are not as clearly
associated with Neolithic/Eneolithic habitation sites. Pottery and stone tool
types from Ta Sur Heigh are consistent with other Oasis 3 assemblages.
Northwest China is identified as the main centre of cowry use and cowries
have been associated with Majiayao (~5.3-4.0 kya) cultural sites. More
tentative data had also suggested that their use in burial contexts may date to
the Early Neolithic in North China (Peng and Zhu, 1995). By about 4.0 kya,
cowries had begun to spread east, being used at lower Xiajiadian sites around
the Yanshan Mountains and in eastern Inner Mongolia. Cowry use was most
widespread between about 3.6-2.6 kya. The association of cowries with Oasis
3 occupations is chronologically appropriate and indicates relationships with
contemporaneous Neolithic cultures and some access to Western-Eastern trade
goods.
213
Based on direct dates and periodization of additional sites, the following set of
chronological indicators is proposed for the Gobi Desert:
1. Early Epipalaeolithic, 19.0-13.5k cal yr BP – There is no clear evidence of
archaeological sites from this time period in the Gobi Desert, although some
possible examples have been noted. Judging from trends in neighbouring regions,
we can expect that lithic assemblages were focused on the production of large
unifacially retouched flake and blade tools. If microblade technology was
present, it can be expected that microblade flakes were largely unretouched and
used as inserts for organic hafts. Microblade core forms differed between
southern Siberia/northern Mongolia and northern China; therefore, we cannot
predict which forms would have been adopted in various Gobi Desert target
regions.
2.
Oasis 1 or Late Epipalaeolithic, 13.5-8.0k cal yr BP – Microlithic tools largely
replaced flake and blade tools and were sometimes retouched using pressureflaking. Fibre-tempered cord-marked pottery was adopted by 9.5k cal yr BP in
the East Gobi. Ostrich eggshell bead production becomes more common.
Microblade core technology frequently focused on prepared cores made from
smaller nodules and organized around the exploitation of cryptocrystalline stones
producing reliable conchoidal fractures. By the end this phase, cryptocrystalline
stone was favoured for knapping over coarser-grained materials like siliceous
sandstone, basalt and quartzite.
214
3. Oasis 2 or Neolithic, 8.0-5.0k cal yr BP – New technologies like formal grinding
stones, chipped and/or ground adze/axes and bifacially flaked tools were added to
microlithic assemblages. Pottery became increasingly common. String-paddled,
net-impressed and undecorated plain-wares are most typical. The production of
microblades is the dominant reduction strategy, but formal biface cores and
expedient flake cores were also used. Unifacial projectile points made on
microblades are typical of early Oasis 2, with bifacial forms emerging slightly
later. Small endscrapers were made on thick semi-circular flakes or microblade
core platform reduction spalls. Elongated endscrapers (tongue-shaped) on
bladelets (10-14 mm in width) are occasionally recovered, especially in the
Alashan Gobi. Endscrapers on amorphous flakes are most common. Beginning
around 6.0k cal yr BP, there was a diversification of bifacially flaked tools types.
Large, thin bifaces are also associated with the late Oasis 2 phase and may have
served as large knives. Late Oasis 2 assemblages are also characterized by a trend
towards the use of a few select high quality raw materials.
4. Oasis 3 or Eneolithic, 5.0-3.0k cal yr BP – Pottery use was more extensive during
the final stage of the Neolithic. Both low- and high-fired ceramics are found and
the variety of treatments includes geometric-incised, channelled, moulded rims,
and raised clay bands. Handles and miniature lugs were used. Red and reddishbrown-wares are common. Lithic assemblages are similar to those of late Oasis 2,
but new tool types include fully polished adze/axes, expanded base drills, and
projectile points with stemmed, fish-tail, and convex bases. Small informal
215
grinding stones are more typical of Oasis 3 than the large formal types used
during Oasis 2 (primarily in the East Gobi). Wedge-shaped cores are less
common in the East Gobi and the Gobi-Altai and more common in the Alashan
Gobi. Barrel-shaped cores are typical of tools kits in Alashan Gobi and the GobiAltai.
3.3. Discussion
Chronometric dates on several post-LGM archaeological sites from three Gobi Desert
regions provide a starting point from which to begin ordering numerous existing
microlithic assemblages. For early periods a comparison with technological
developments elsewhere in Northeast Asia is instructive. These comparisons illustrate
not only larger trends in technology, but also in economic developments. The temporal
continuity of widespread shifts in technology also suggests some level of
interconnectedness amongst hunter-gatherers in mainland Northeast Asia. As population,
or at least the visibility of archaeological sites, increased during the Early Holocene,
technological assemblages are more regionally divergent and stylistic comparisons
between regions are much less useful.
By the Late Neolithic or Eneolithic, the widespread rise of sedentism and
production economies based on the exploitation of domesticated plants and animals
contrast markedly with evidence of mobile foraging economies in the Gobi Desert.
However, mobile foraging economies without evidence of domesticated plants and
animals persisted in many other parts of Northeast Asia, including the Lake Baikal region
216
and much of the higher latitude far eastern regions of Russia and Northeast China. The
use of microblade and biface technology flourished in many of these regions, as it did in
the Gobi Desert, though in diverse forms. The technological diversity that emerged
during the Neolithic might be associated with increasingly localized interaction spheres
and group identification, as geographically discontinuous developments punctuated a
previously more homogeneous economic landscape (e.g., compare the archaeological
records of the Minusinsk Basin and the Lake Baikal region, or the Central Plains and
Northeast China, during the Late Epipalaeolithic and the early Bronze Age).
The artefact-based chronology that is presented here represents a preliminary
method of classification for the Gobi Desert based on current knowledge of assemblage
composition and diagnostics. Future excavation and chronometric dating will allow for a
more refined understanding of chronology, including relative dependence on various raw
material types, changes in microblade core reduction sequences, and trends in biface
production. Distinct technological continuity between the late Oasis 2 and early Oasis 3
phases is notable and it is hoped that the relationship between these two phases will be
more clearly elucidated by future research.
217
CHAPTER 4 – PREHISTORIC HUMAN LAND-USE IN THE GOBI
DESERT
The goal of this study is to not only to build a preliminary technological
chronology, but to elaborate temporal and spatial characteristics of land-use in the Gobi
Desert during the terminal Pleistocene to late middle Holocene in order to better
understand modes of subsistence and settlement prior to the emergence of nomadic
pastoralism. I have used chronometric dating and an artefact-based chronology to assign
approximate ages to almost 100 archaeological sites (Table 3.7) from three
environmentally distinct regions of the Gobi Desert. Reconstructions of land-use in this
chapter are based on information from original site descriptions (i.e., environment,
topography, assemblage composition) and quantitative analysis of lithic assemblages.
Since each of the three regions is environmentally distinct, they will be considered
separately in order to account for possible differences in land-use related to variable
resource distribution, and the timing of climate-mediated environmental shifts (see
Chapter 5).
Examination of shifts in organizational strategies begins with a series of
hypotheses about expected changes in land-use based on current knowledge of post-LGM
archaeology in the Gobi Desert and other parts of Northeast Asia, as outlined in Chapter
2. The following hypotheses are based on previous interpretations of Gobi Desert
archaeology as well as the new chronometric dates detailed in Chapter 3.
1. By 13.5k cal yr BP, post-LGM climatic amelioration had allowed highly
mobile desert-adapted hunter-gatherers to expand across the Gobi Desert,
218
exploiting a wide range of environments, including dune-fields. A period of
increased aridity and resulting resource stress during the terminal Pleistocene
and early Holocene may have encouraged the introduction of more specialized
processing technologies like grinding stones and pottery for extracting
additional nutrients from arid-adapted plant species like grass seeds (Bettinger
et al., 2007; Elston et al., 2011). Residential mobility was high and organized
in a circulating pattern more typical of Binford’s (1980) foraging system.
2. Optimally warm and moist conditions following 8.0k cal yr BP led to the
stabilization of dune-fields and the creation of diverse new habitats around
rivers, lakes and larger interdunal marshes. Since local hunter-gatherers had
begun using dune-field environments in previous periods, the new abundance
of resources, along with the infilling of associated depressions, rivers and lake
basins led to more regular and/or prolonged occupation of such environments.
Hunter-gatherers continued to exploit a variety of environments using a
radiating pattern of land-use similar to a collector system (Binford, 1980),
with longer term seasonal base camps centred on dune-field environments.
3. After about 5.0k cal yr BP, widespread deterioration of steppe environments
and the contraction of lakes and wetlands outside of dune-fields encouraged
hunter-gatherers to intensify seasonal use of dune-field/wetland environments,
which were oases of productivity. Diminished returns encouraged huntergatherers to make more frequent moves, resulting in the geographic expansion
of smaller field camps and task sites. Herd animals were first incorporated
219
into the hunter-gatherer system at this time, providing new resources for
subsistence and raw materials.
This chapter uses quantitative lithic analysis of assemblages assigned in the
previous chapter to the Late Epipalaeolithic (Oasis 1), Neolithic (Oasis 2), and Eneolithic
(Oasis 3) in order to test the proposed patterns of land-use. In the following chapter,
local and supra-regional palaeoenvironmental data is synthesized in order to
contextualize reconstructions of land-use with notable shifts in local effective moisture,
vegetation, and hydrology. According to the resulting data, each hypothesis is reassessed
in the concluding chapter.
4.1. Analysis of land-use
Human exploitation of various ecozones is expected to have shifted with post-LGM
environmental change. Key ecozones would have become more or less productive in
meeting the needs of local hunter-gatherers. The internally-drained basin-range and
steppe topography of the Gobi Desert is expected to have provided discontinuous access
to patchy resources, particularly when increased precipitation resulted in the capture of
moisture from run-off within lowland basins, supporting the formation of lakes and
denser vegetation. Lowland basins, open desert or desert-steppes, and mountainous
highlands would have offered a range of seasonal resources. Wide swaths of low
elevation land covered by dune-fields interspersed with wetlands and small lakes must
have supported higher plant and animal diversity and productivity (Nicholas, 1998).
220
Ungulate populations noted by expedition scientists and explorers in the early 20th
century (Allen, 1938) should have been more numerous under less arid conditions and
would have provided more reliable access to large- and medium-bodied prey than in
modern times. Numerous accounts of seasonally abundant avian fauna also suggest a
former richness in small game (Nelson, 1925; Pond, n.d.; Hedin, 1943).
Raw material resources, particularly high quality tool stone, would also have had
important affects on land-use preferences and overall mobility. However, very little is
known about the distribution of such resources in the Gobi Desert. According to
expedition archaeologists the richest sources of raw materials in the western study areas
were in the Arts Bogd-Ulan Nor Plains (Gobi-Altai) and the Ukh-Tokhoi/Khara Dzag
plateaux region (Alashan Gobi) (Nelson, 1925; Maringer, 1950; Fairservis, 1993; see also
Kulik et al., 2006). Pond (n.d.) also notes the presence of silicified volcanic ash and
chalcedony sources in the Southwest subregion of the East Gobi (see Figure 1.2), but the
extent of these resources is not clear. Some inferences about the availability of lithic raw
material can be based on regional variability in frequencies of primary reduction and
relative nodule size, inferred from percentage of remnant cortex and core volume. A
more reliable assessment of the influence of raw material availability upon lithic
assemblage variability will only be possible, however, with additional research on
material sources.
Differences in subsistence and settlement can be inferred for each phase by
examining the distribution of archaeological sites within each target region during each
time period (Oasis 1, 2, and 3). Palaeoenvironmental data can then contribute to our
221
understanding of resource availability. Large-scale changes in land-use across the Gobi
Desert can also be recognized. Lithic assemblage composition and lithic reduction
strategies are used to assess relative levels of residential mobility between regions and
time periods. Formal cores and tools tend to be associated with situations of high
residential mobility and raw material conservation, while informal cores and tools are
more common in situations of decreased mobility. Examining qualitative and
quantitative differences between lithic assemblages will allow us to ascertain whether
certain environments with large site assemblages were associated with longer term
occupations, or whether they were simply frequently reoccupied. While all environments
might yield proof of human habitation and exploitation, the goal is to recognize changes
in their function and relative importance throughout time.
4.1.1. Categorization of sites and environments
In addition to period assignments, each archaeological assemblage is categorized
according to relative assemblage size and environmental context. Artefact assemblage
composition and the dominant modes of lithic reduction are used as evidence for site
function and relative duration of occupation. The relative density of each site is based on
orders of magnitude for artefact counts (< 10, 11 to 100, 101 to 1000, 1001 to 5000, >
5000). This approach is considered a more reliable estimate of site size due to variation
in surface collection methods. Interpretations of raw material use are based on core
dimensions, relative percentage of remnant cortical surface, reduction strategies, and
extent and intensity of retouch. Resulting interpretations of relative length of occupation
222
within different ecozones can then be used to interpret overall mobility and aspects of
subsistence.
Based on knowledge of Gobi Desert landforms and previous interpretations of
land-use in arid Northwest China (e.g., Bettinger et al., 2007) we should expect
significant variation in the use of five distinct terrains: dune-field, lake, steppe, midelevation (e.g., mesas, hillslopes, plateaux), and mountainous terrain. Four separate
environmental factors should be important in delineating ecozone categories based on
variation in modern floral and faunal distribution: elevation, topography, water source,
and presence/absence of sand or dune-fields (Allen, 1938; Jigjidsuren and Johnson, 2003;
Batsaikhan et al., 2010). Data on elevation (metres a.s.l.) was sometimes recorded in
field journals, but I more often estimated elevation by locating each archaeological site
on expedition base maps (Hill, n.d.; Hill and Roberts, n.d.; Roberts et al., n.d.;) and recent
topographic maps (Army Map Services, Corps of Engineers, U.S. Army, Washington,
D.C., 1949, 1950, 1954; Norin, 1978; Bureau of Geological Investigation, Geological
Survey of Mongolia, 2003a, 2003b, 2003c, 2003d) according to site names and
descriptions (Nelson, 1925; Pond, 1928, n.d.; Maringer, 1950). Expedition maps were all
at the 1:200,000 scale. Regional maps were at the 1:500,000 scale for the Gobi-Altai
region, at a 1:1,000,000 scale for the Alashan Gobi and East Gobi (Norin and Montell,
1969), and at the 1:500,000 for the Gurnai Depression (Norin, 1978).
I derived information about topography, water sources, and presence or absence
of sand primarily from site descriptions, complemented by cartographic data where site
descriptions are lacking. For greater accuracy, information on nearby water sources was
223
referenced with maps, since changes in the surficial hydrology of the study area have
been dramatic over the past 3000 years. Many sites were recorded as being discovered
near wells, but maps clarify the hydrological context by indicating the proximity of
“intermittent streams” or drainage channels. In such circumstances, the site is considered
to be found near a stream. Wetlands and lakes were grouped together since the
distinction may have been less relevant in wetter periods where what are today wetlands
or marshes would have been extant lakes. Moreover, wetland/marsh margins of lake
habitats were probably just as, or more, important to hunter-gatherers as the lakes
themselves. Table 4.1 outlines the categories used for classifying various localities.
Parameter
Elevation
metres a.s.l.
Topography
Water
source
Sand
Site size
# artefacts
1 = < 1000
Categorization
2 = 1000-1200
groupings
3 = 1201-1400
4 = > 1400
1 = basin
(basin/
valley)
none
2 = steppe
(plains/basin
plains)
well/spring
3 = promontory
(mesa/hillslope/
hilltop/foothills)
river/stream
4 = upland
(mountains/
foothills)
lake/wetland
1 = none
1 = < 10
2 = present
2 = 10-100
3 = 101-1000
4 = 1001-5000
Table 4.1 List of categories used for analysis of site context.
5 = > 5000
224
In order to determine which environmental parameters were most relevant to
archaeological locations, Pearson’s chi-square was used to evaluate the natural
juxtaposition of environmental parameters (Table 4.2). The relationships between water
source and both elevations and topography were difficult to test due to low cell counts,
but individual cells showed a strong significance between the association of rivers and
streams and camps in upland locales (especially 1201-1400 m a.s.l. and mountainous
terrain), while camps in basins or valley lowlands suggested significant association with
lakes and/or wetlands. Sites associated with sand dunes showed a strong correlation with
basin/valleys, lake/wetlands, and elevations between 1000 and 1200 m a.s.l. (p = <
0.0001). These relationships are not surprising considering natural geological processes;
however, the distribution of sites within a specific intersection of environmental zones is
worth noting.
225
ELEVATION/
WATER
Count
Expected
None
Well/spring
River/stream
Lake/wetland
TOPOGRAPHY/
WATER
None
Well/spring
River/stream
Lake/wetland
ELEVATION/
SAND
No sand
Sand
TOPOGRAPHY/
SAND
No sand
Sand
WATER/ SAND
No sand
Sand
< 1000
1000-1200
1201-1400
> 1400
5
1.9
2
1.2
4
4.2
2
5.7
Basin/valley
8
10.3
5
6.4
17
23.2
41
31.0
Steppe
2
2.6
1
1.6
10
5.9
5
7.8
Promontory
1
1.7
2
0.7
5
2.6
0
3.5
Upland
7
9.9
3
6.4
14
20.9
41
27.9
< 1000
5
2.4
4
1.6
3
5.1
4
6.9
1000-1200
2
2.0
0
1.3
8
4.2
3
5.6
1201-1400
3
2.7
4
1.8
11
5.8
0
7.7
> 1400
5
5.5
8
7.4
Basin/valley
21
30.3
50
40.7
Steppe
13
7.7
5
10.3
Promontory
8
3.4
0
4.6
Upland
9
5.6
4
7.4
River/stream
26
15.4
10
20.6
18
7.7
0
10.3
Lake/wetland
3
20.6
45
27.4
10
27.9
55
37.1
None
10
7.3
7
9.7
11
6.9
5
9.1
Well/spring
9
4.7
2
6.3
Table 4.2 Relationship between environmental parameters, showing natural juxtaposition
of variables. Elevation/water: X2 = 26.04, p = 0.0020. Topography/water: X2 = 40.88, p
= <0.0001. Elevation/sand: X2 = 22.24, p = <0.0001. Topography/sand: X2 = 52.10, p =
<0.0001. Water/sand: X2 = 47.53, p = <0.0001.
226
Due to the low number of Palaeolithic and Epipalaeolithic sites and a sample
biased (either through collection practises or natural occurrence) towards sites in lowland
dune-field environments, individual cell counts are too low to test the significance of
distribution; however, visual consideration of sites according to period (i.e.,
Palaeolithic/Early Epipalaeolithic, Oasis 1, Oasis 2, and Oasis 3) suggests that certain
environmental zones were probably favoured during different periods (Figure 4.1, Table
4.3). Oasis 1, Oasis 2, and Oasis 3 sites show a strong distribution at elevations between
1000-1200 m a.s.l. High elevation sites (> 1400 m a.s.l.) are most rare. Distribution of
sites according to elevation is probably closely related to natural local topography and
collection biases – the majority of landmass in the Gobi Desert rests at elevations of
between 1000 and 1200 m a.s.l. and high elevation locales were not as extensively
explored during collecting expeditions. Despite this probable bias, individual cell counts
in Table 4.3 show that Palaeolithic/Epipalaeolithic sites are more commonly associated
with elevations between 1201 and 1400 m a.s.l. than would be expected for a random
sample.
Considering the lack of sites at elevations above 1400 m a.s.l., and the potential
for an underrepresentation of sites at elevations above 1200 m a.s.l., topographic
parameters are considered to be more informative (i.e., basin/valley, steppe, promontory,
upland). An entire range of topographic environments were exploited during all periods,
but Oasis 1, 2 and 3 sites were most common in basins, depressions, or valleys (Figure
4.1, Table 4.3).
227
100%
75%
>1400
50%
1201-1400
25%
1000-1200
< 1000
0%
P/E
O1
O2
O3
a.
100%
75%
Upland
50%
Promontory
25%
Steppe
Basin/Valley
0%
P/E
O1
O2
O3
b.
100%
75%
Lake/Wetland
50%
River/Stream
25%
Well/Spring
None
0%
P/E
O1
O2
O3
c.
100%
75%
50%
Sand
25%
No Sand
0%
P/E
O1
O2
O3
d.
Figure 4.1 Distribution of Gobi Desert sites according to each environmental parameter:
a) elevation; b) topography; c) water source; d) sand. Palaeolithic/Early Epipalaeolithic,
N= 13; Oasis 1, N = 8; Oasis 2, N= 30; Oasis 3, N = 47.
228
ELEVATION
< 1000
1000-1200
1201-1400
> 1400
TOPOGRAPHY
Basin/valley
Steppe
Promontory
Mountains
WATER SOURCE
none
Well/spring
River/stream
Lake/wetland
SAND
absent
present
Palaeolithic/
Epipalaeolithic
2
1.4
5
8.1
4
2.0
1
0.5
Oasis 1
Oasis 2
Oasis 3
0
0.9
6
5.4
0
1.3
2
0.3
3
3.4
20
20.3
7
5.0
5
7.7
6
5.3
34
31.1
5
7.7
1
1.9
4
8.0
3
1.9
2
1.5
4
1.8
5
4.9
0
1.1
1
0.9
2
1.1
18
18.4
6
4.3
2
3.4
4
4.0
33
28.8
5
6.7
6
5.3
3
6.2
5
1.7
3
1.1
4
4.0
1
6.2
1
1.1
0
0.6
5
2.4
2
3.8
3
4.0
2
2.4
10
9.2
15
14.4
4
6.2
3
3.8
11
14.4
29
22.5
11
5.2
2
7.8
4
3.2
4
4.8
10
11.9
20
18.1
14
18.7
33
28.3
Table 4.3 Actual and expected distribution of sites according to each environmental
parameter. Elevation: X2 = 18.47, p = 0.0301. Topography: X2 = 11.95, p = 0.2163.
Water source: X2 = 22.41, p = 0.0077. Sand: X2 = 13.74, p = 0.0033.
229
One of the most distinct patterns is found for water sources. Palaeolithic and
Epipalaeolithic sites are almost entirely absent from wetland/lake environments, while
Oasis 2 and Oasis 3 sites cluster around major water sources (river/stream, wetland/lake).
Oasis 1 sites are most common around rivers or streams, while Oasis 3 sites are most
commonly found around lakes or wetlands. Oasis 2 sites are primarily distributed near
rivers/streams and wetland/lakes.
Other notable patterns include the association of Palaeolithic and Early
Epipalaeolithic sites with elevations between 1201-1400 m a.s.l. and upland topography,
the association of Oasis 3 sites with basin/valley topography, and the preferential
association of Oasis 3 sites with sandy environments. Dune-fields appear to have been
important to Gobi Desert groups during Oasis 2 (20 out of 30 sites) and Oasis 3 (33 out of
47 sites), but are underrepresented in earlier periods. While cell counts indicate
significant associations with basin/valleys, lake/wetlands, and sand only in Oasis 3, the
numbers suggest that such environments were avoided in early periods and that habitation
became progressively more common beginning in Oasis 1. Stabilization of dune-fields
activated during the LGM was probably not complete until after the Younger Dryas and
early Holocene (Grunert and Lehmkuhl, 2004). Prior to the stabilization of dune-field
environments in the early Holocene, they would not have offered rich ecosystems for
foraging. Similarly, the wetland habitats that were apparently favoured in later periods
may not have been well-established and productive until Oasis 2.
Site distribution according to period can also be investigated for each of the three
regions. Figures 4.2-4.4 indicate consistency in site distribution between the three target
230
regions. The East Gobi group (Figure 4.2) is notable in that Oasis 3 sites are equally
distributed between basin/valley and promontory locales. This contradicts the general
finding that lowland environments were uniformly favoured during both Oasis 2 and
Oasis 3. East Gobi sites are also distributed mostly around rivers or streams, including
about 50% of Oasis 2 and Oasis 3 sites. This condition may be attributed to either
increased archaeological survey in 1928 around large rivers such as the Shara Murun and
its tributaries, or to local hydrology, which is characterized by many rivers and drainage
channels bounded by marshes and wetlands.
The Gobi-Altai group (Figure 4.3) shows a relatively high frequency of
Palaeolithic/Early Epipalaeolithic (~25%) sites around lakes – a situation entirely lacking
in other regions. The apparent distribution of these early sites around lakes might be due
to more intensive sampling of dune-field/lake sites in the Gobi-Altai, where several
distinct sites were recorded from the same locality. Better integrity of individual
Shabarakh-usu sites might also contribute to enhanced recognition of early sites. In the
East Gobi and Alashan Gobi, distinct sites were not maintained during collection, which
confuses the recognition of early components. As noted in Appendix B, some artefacts
from the Baron Shabaka Well locality in the East Gobi are most typical of the
Epipalaeolithic, but the lack of site structure limits interpretation. Excavation of subdune layers around dune-field/wetland sites is probably needed in order to better assess
this circumstance.
231
100%
75%
>1400
50%
1201-1400
25%
1000-1200
0%
P/E
O1
O2
O3
a.
100%
75%
Upland
50%
Promontory
25%
Steppe
Basin/Valley
0%
P/E
O1
O2
O3
b.
100%
75%
Lake/Wetland
50%
River/Stream
25%
None
0%
P/E
O1
O2
O3
c.
100%
75%
50%
Sand
25%
No Sand
0%
P/E
O1
O2
O3
d.
Figure 4.2 Distribution of East Gobi sites according to each environmental parameter: a)
elevation; b) topography; c) water source; d) sand. Palaeolithic/Early Epipalaeolithic, N
= 3; Oasis 1, N = 5; Oasis 2, N = 11; Oasis 3, N = 10.
232
ELEVATION
1000-1200
1201-1400
> 1400
TOPOGRAPHY
Basin/valley
Steppe
Promontory
Mountains
WATER SOURCE
none
River/stream
Lake/wetland
SAND
absent
present
Palaeolithic/
Epipalaeolithic
1
2.1
2
0.7
0
0.2
Oasis 1
Oasis 2
Oasis 3
3
3.4
0
1.2
2
0.3
6
7.6
5
2.6
0
0.8
10
6.9
0
2.4
0
0.7
0
1.3
0
0.1
1
0.9
2
0.6
2
2.2
0
0.2
1
1.5
2
1.0
6
4.9
1
0.4
2
3.4
2
2.3
5
4.5
0
0.3
5
3.1
0
2.1
1
0.1
2
1.6
0
1.2
0
0.2
4
2.8
1
2.1
0
0.4
5
6.1
6
4.5
0
0.3
5
5.5
5
4.1
3
1.4
0
1.5
3
2.4
2
2.6
4
5.3
7
5.7
4
4.8
6
5.2
Table 4.4 Actual and expected distribution of sites in the East Gobi according to each
environmental parameter. Elevation: X2 = 19.88, p = 0.0029. Topography: X2 = 11.31, p
= 0.2549. Water source: X2 = 11.97, p = 0.0627. Sand: X2 = 4.39, p = 0.2224.
233
100%
75%
>1400
50%
1201-1400
25%
1000-1200
0%
P/E
O1
O2
O3
a.
100%
75%
Promontory
50%
Steppe
25%
Basin/Valley
0%
P/E
O1
O2
O3
b.
100%
75%
Lake/Wetland
50%
River/Stream
25%
Well/Spring
None
0%
P/E
O1
O2
O3
c.
100%
75%
50%
Sand
25%
No Sand
0%
P/E
O1
O2
O3
d.
Figure 4.3 Distribution of Gobi-Altai sites according to each environmental parameter:
a) elevation; b) topography; c) water source; d) sand. Palaeolithic/Early Epipalaeolithic,
N = 4; Oasis 1, N = 2; Oasis 2, N = 10; Oasis 3, N = 10.
234
ELEVATION
1000-1200
1201-1400
> 1400
TOPOGRAPHY
Basin/valley
Steppe
Promontory
WATER SOURCE
None
Well/spring
River/stream
Lake/wetland
SAND
absent
present
Palaeolithic/
Epipalaeolithic
2
2.9
2
0.9
0
0.1
Oasis 1
Oasis 2
Oasis 3
2
1.5
0
0.5
0
0.1
8
7.3
2
2.3
0
0.4
7
7.3
2
2.3
1
0.4
1
2.8
2
0.9
1
0.3
2
1.4
0
0.5
0
0.1
8
6.9
2
2.3
0
0.8
7
6.9
2
2.3
1
0.8
1
0.3
1
0.5
1
0.9
1
2.3
1
0.1
0
0.2
0
0.5
1
1.1
0
0.8
2
1.1
3
2.3
5
5.8
0
0.8
0
1.1
2
2.3
8
5.8
3
1.1
1
2.9
0
0.5
2
1.5
2
2.7
8
7.3
2
2.7
8
7.3
Table 4.5 Actual and expected distribution of sites in the Gobi-Altai according to each
environmental parameter. Elevation: X2 = 3.97, p = 0.6810. Topography: X2 = 5.92, p =
0.4320. Water source: X2 = 12.83, p = 0.1706. Sand: X2 = 5.92, p = 0.1154.
235
100%
75%
>1400
50%
1201-1400
25%
1000-1200
<1000
0%
P/E
O1
O2
O3
a.
100%
75%
Uplands
50%
Steppe
25%
Basin/Valley
0%
P/E
O1
O2
O3
b.
100%
75%
Lake/Wetland
50%
River/Stream
25%
Well/Spring
None
0%
P/E
O1
O2
O3
c.
100%
75%
50%
Sand
25%
No Sand
0%
P/E
O1
O2
O3
d.
Figure 4.4 Distribution of Alashan Gobi sites according to each environmental
parameter: a) elevation; b) topography; c) water source; d) sand. Palaeolithic/Early
Epipalaeolithic, N = 6; Oasis 1, N = 1; Oasis 2, N = 9; Oasis 3, N = 27.
236
ELEVATION
< 1000
1000-1200
1201-1400
> 1400
TOPOGRAPHY
Basin/valley
Steppe
Mountains
WATER SOURCE
none
Well/spring
River/stream
Lake/wetland
SAND
absent
present
Palaeolithic/
Epipalaeolithic
2
1.3
2
3.2
0
0.4
1
0.1
Oasis 1
Oasis 2
Oasis 3
0
0.3
1
0.6
0
0.1
0
0.0
3
2.4
6
5.7
0
0.7
0
0.2
6
7.0
17
16.5
3
1.9
0
0.6
3
4.0
1
1.0
2
1.0
1
0.7
0
0.2
0
0.2
4
6.1
3
1.5
2
1.5
21
18.2
3
4.4
3
4.4
3
1.4
2
0.7
1
1.1
0
2.8
0
0.2
0
0.1
1
0.2
0
0.5
3
2.1
0
1.0
2
1.7
4
4.2
4
6.3
3
3.1
4
5.0
16
12.5
5
2.5
1
3.5
1
0.4
0
0.6
4
3.8
5
5.2
8
11.3
19
15.7
Table 4.6 Actual and expected distribution of sites in the Alashan Gobi according to each
environmental parameter. Elevation: X2 = 10.48, p = 0.3134. Topography: X2 = 5.65, p =
0.4637. Water source: X2 = 14.95, p = 0.0923. Sand: X2 = 7.31, p = 0.0625.
237
In the Alashan Gobi (Figure 4.4), Oasis 3 sites are most common around lakes or
wetlands, but appear to have been more evenly distributed across different types of water
sources than in other regions. The majority of Oasis 2 and Oasis 3 sites from
mountainous terrain or elevations over 1400 m a.s.l. were from the Ukh-tokhoi/Kharadzag plateaux region of the Alashan Gobi. Most of these sites are attributed to Oasis 3.
Occupations identifiable as Oasis 2 or only Neolithic/Eneolithic (Oasis 2 or Oasis 3) were
also recovered from mountainous terrain in the East Gobi, but these sites appear to have
been related to raw material procurement. Only one site in the Gobi-Altai, associated
with Oasis 3, was collected at an elevation of more than 1400 m a.s.l.
Broadly similar trends in the distribution of sites during different periods are
characteristic of the Gobi Desert, despite above noted divergences between regions. The
most notable trend is the use of dune-fields and lake/wetland environments during Oasis
2 and Oasis 3. It is probable that the abundance of Oasis 2 and Oasis 3 sites within dunefield/wetland environments is related to factors other than sampling bias. Dunefield/wetland environments are closely linked with topography and elevation; based on
the distribution of sites within various topographic, hydrological, and depositional
environments, five distinct ecozones can be distinguished: 1) lowland dune-field/wetland
( lowland [< 1200 m a.s.l.] dune accumulations around rivers, marshes and lakes); 2)
lowland river (sites situated on plains or wide valleys near rivers); 3) lowland dry
(lowland sites with spring/well or no clear source of water); 4) upland (sites on mesas,
mountains or other higher elevation regions near major water source); and 5) upland dry
238
(upland [> 1200 m a.s.l.] sites with no apparent nearby water source or a minor water
source such as a spring or well14).
Figures 4.5 and 4.6 illustrate the distribution of sites for each region and for the
Gobi Desert as a whole. Distribution of sites across ecozones in the entire Gobi Desert
sample further supports the trends noted above. Individual counts are too low for some
variables to permit the use of chi-square, but the cell counts suggest a high probability
that the distribution of sites is not a random effect of sampling (Table 4.7).
Several trends in land-use can be identified. The lack of early sites in dunefield/wetland environments is consistent for all regions. Although dune-field/wetlands
appear to have been favoured, a range of environments were exploited during Oasis 2 and
Oasis 3. Ecozones without access to major water sources appear to have been avoided
during Oasis 1 (coinciding with the Younger Dryas; see Madsen et al., 1998), though the
small sample size may contribute to this effect. According to site distribution, the use of
high elevation (> 1200 m a.s.l.) or upland environments appears to have continued as an
important component of Oasis 1 organizational strategies. Most clearly, the distribution
of Gobi Desert sites indicates a gradual trend towards increasing exploitation of sandy
lowland environments around major water sources, most of which would have been
surrounded by marshland. While the increasing use of dune-field/wetland environments
is evident in this sample, it is not clear how the use of other environments may have
changed with time as new resources and technologies were incorporated into subsistence
14
Although springs probably functioned in earlier periods, and wells may have been dug into former
springs, these sources are considered in a different category because they provide water but none of the
floral, faunal or material resources typically associated with major bodies of water.
239
strategies. Recognizing differences in how various ecozones were used will further
enhance our understanding of changes in land-use over time. Appendix D lists the
ecozone grouping of each studied site.
100%
75%
Upland dry
Upland
50%
Lowland dry
Lowland river
25%
Lowland dune-field/wetland
0%
P/E
O1
O2
O3
Figure 4.5 Distribution of Gobi Desert sites according to ecozone. Palaeolithic/Early
Epipalaeolithic, N = 13; Oasis 1, N = 7; Oasis 2, N = 30; Oasis 3, N = 47).
Ecozone
Lowland dune-field/wetland
Lowland river
Lowland dry
Upland
Upland dry
Palaeolithic/
Epipalaeolithic
1
7.7
2
1.2
3
0.8
3
2.4
4
0.9
Oasis 1
Oasis 2
Oasis 3
4
4.7
1
0.6
0
0.4
3
1.3
0
0.5
20
17.8
2
2.8
1
1.8
7
5.6
0
2.2
33
27.8
4
4.4
2
2.9
5
8.7
3
3.4
Table 4.7 Actual and expected distribution of Gobi Desert sites according to each
ecozone. X2 = 32.544, p = 0.0011.
240
East Gobi
100%
Upland dry
75%
Upland
50%
Lowland river
25%
Lowland dunefield/wetland
0%
P/E
O1
O2
O3
Gobi-Altai
100%
Upland dry
75%
Upland
50%
Lowland river
25%
Lowland dunefield/wetland
0%
P/E
O1
O2
O3
Alashan Gobi
100%
Upland dry
75%
Upland
50%
Lowland dry
25%
Lowland river
0%
P/E
O1
O2
O3
Lowland dunefield/wetland
Figure 4.6 Distribution of sites in each target region according to ecozone. East Gobi:
Palaeolithic/Early Epipalaeolithic, N = 3; Oasis 1, N = 5; Oasis 2, N = 11; Oasis 3, N =
10. Gobi-Altai: Palaeolithic/Early Epipalaeolithic, N = 4; Oasis 1, N = 2; Oasis 2, N =
10; Oasis 3, N = 10. Alashan Gobi: Palaeolithic/Early Epipalaeolithic, N = 6; Oasis 1, N
= 1; Oasis 2, N = 9; Oasis 3, N = 27.
241
East Gobi
Ecozone
Lowland dune-field/wetland
Lowland river
Upland
Upland dry
Palaeolithic/
Epipalaeolithic
0
1.5
0
0.2
2
1.1
1
0.1
Oasis 1
Oasis 2
Oasis 3
2
2.6
0
0.3
3
1.9
0
0.2
7
5.7
1
0.8
3
4.2
0
0. 4
6
5.2
1
0.7
3
3.8
0
0.3
Oasis 1
Oasis 2
Oasis 3
2
1.5
0
0.2
0
0.2
0
0.1
8
7.3
1
1.1
1
1.1
0
0.4
8
7.3
1
1.1
1
1.1
0
0.4
Oasis 1
Oasis 2
Oasis 3
0
0.6
1
0.1
0
0.1
0
0.1
0
0.1
5
5.0
0
0.8
1
1.2
3
0.8
0
1.0
19
15.1
2
2.5
2
3.8
1
2.5
3
3.1
X2 = 13.34, p = 0.1476
Gobi-Altai
Ecozone
Lowland dune-field/wetland
Lowland river
Upland
Upland dry
Palaeolithic/
Epipalaeolithic
1
2.9
1
0.5
1
0.5
1
0.1
X2 = 8.90, p = 0.4470
Alashan Gobi
Ecozone
Lowland dune-field/wetland
Lowland river
Lowland dry
Upland
Upland dry
Palaeolithic/
Epipalaeolithic
0
3.3
1
0.6
3
0.8
0
0.6
2
0.7
X2 = 32.42, p = 0.0012
Table 4.8 Actual and expected distribution of sites in each region according to ecozone.
242
4.1.2. Assemblage composition
Variation in assemblage composition can help decipher the use of different ecozones in
which sites were found. A variety of information can suggest site function. Relative site
size and assemblage composition are the most important data used for this purpose.
Artefact assemblage composition, based on the presence or absence of certain artefact
types, gives a general sense of tasks carried out at each site, as well as an overall
impression of technological organization within the broader organizational strategy.
Aspects of subsistence practices, residential mobility, transport limitations, scheduling,
perceived risk, and raw material availability all contribute in different ways to the
composition of tools kits (Binford, 1979; Torrence, 1983; Bleed, 1986; Parry and Kelly,
1987; Nelson, 1991; Bousman, 1993; Kuhn, 1994; 1995; Bamforth and Bleed, 1997).
Specialized technologies are especially useful for the recognition of particular activities
within each environment.
Misrepresentation of site function can occur due to factors such as differences in
the rates of discard, variation in occupation span, and post-depositional collection of
artefacts for reuse, or as souvenirs (Schiffer, 1987: 47-50; 114-119). Rare artefacts that
are very sensitive to site function and chronology, such as decorated pottery, and finished
tools such as projectile points, are those most frequently collected by archaeologists in
past decades (Schiffer, 1987: 116). Comparisons between large sites where researchers
were only able to recover and curate a small incomplete sample, and very small sites
from which an entire assemblage could be collected, are expected to be especially
problematic. In the Gobi Desert, as Pond (n.d.) observed, it is probable that grinding
243
stones were often scavenged over the millennia. The practice of scavenging heavy
ground stone tools for reuse may have increased with the introduction of animals for
transport (see Schiffer, 1987: 114).
Differences in depositional contexts and collection practices can influence
interpretations of intra-assemblage variability. Due to the possibility of such biases, site
categorizations are based on presence/absence rather than artefact counts. However, I
believe the same indices, such as the ratio of formal to informal cores in each
assemblage, is less biased by collection practices. High variation among sites, including
a suggestive pattern of high levels of expedient technologies in the larger assemblages
(see section 4.1.3) suggests that collection biases were minimal. Variation due to
collection biases is also expected to be more significant in comparisions between regions
than in comparisons within them. Assemblages within each region were collected, if not
by the same researcher, by the same expedition group and are expected to be relatively
internally cohesive.
High density of occupation debris within Gobi Desert dune-field/wetland
environments are a key characteristic of the Neolithic/Eneolithic archaeological record.
However, it is not known if these occupations resulted from relatively long term
habitation in this particular environment or from successive short-term reoccupations.
Discrete sites or artefact clusters were clearly noted at Shabarakh-usu and Baron Shabaka
Well (Nelson, 1925; Pond, n.d.; Fairservis, 1993; Janz, 2006). Such clusters could have
resulted from sequential reoccupations or from the activities of contemporaneous task
groups associated with a single longer term occupation site. Identifiable differences in
244
site function are expected to exist between ecozones, with the use of certain environments
changing over time. The increased focus on dune-field/wetlands during Oasis 2
anticipates distinct differences in site function during that period when compared to
earlier Oasis 1 occupations.
4.1.2.1. Concepts and methods
Two types of archetypical hunter-gatherer settlement systems were used by Binford
(1980) to represent extremes on a continuum of residential mobility strategies. Situating
Gobi Desert groups along this continuum is helpful in understanding these arid lands
high-latitude hunter-gatherers. At either end of the continuum are “foragers” and
“collectors.” Foragers are identified with high residential mobility organized in a
circulating pattern of land-use, while collectors are characterized by a pattern of logistical
radiating land-use centred on a more sedentary home base. Both foragers and collectors
hunt and gather from aggregation sites known respectively as residential bases or a base
camps. Short-term procurement locales or task sites were used by both groups and are
referred to by Binford (1980) as locations. Field camps are sites where long-ranging task
groups are maintained away from the collector base camps. Collector caches and stations
were not considered here due to issues of survival and recognition in the archaeological
record. Based on the archaeological visibility of site types, a general designation of
residential sites is used to encompass residential bases, base camps, and field camps,
while task sites can account for the remainder of site functions.
245
Foragers tend to employ a circular pattern of mobility based on “moving
consumers to goods” (Binford, 1980). Foragers move residential bases when the risk of
procuring sufficient resources from the local environment is considered to outweigh the
risk of moving to another camp (after MacArthur and Pianka, 1966; Kelly, 1995: 132148). As such, short-term residential sites are regularly created across the entire
landscape and resulting artefact assemblages show low variability. Task groups hunt and
gather food within a daily walkable radius of the camp site, but their activities leave few
material remains. In highly patchy environments, where key resources are abundant in
only a few places, foragers are expected to fall into a pattern referred to as tethered
nomadism, wherein extreme redundancy is exhibited in the reuse of key environments
(Taylor, 1964; Yellen and Harpending, 1972; Binford, 1980). The pattern of tethered
foraging is expected in deserts where water is scarce, but even within the same desert
environment the level of seasonal residential mobility varies greatly based on access to
local resources (Kelly, 1995: 126-128).
Collectors exhibit lower residential mobility. Base camps are situated near key
resources and occupations often extend over one or more seasons. Goods are moved to
consumers through procurement of resources by different types of task groups (Binford,
1980). Field camps serve as temporary home bases for task groups operating far from the
residential base and are expected to vary according to the nature of targeted resources.
Task sites or locations are also typical of collector groups, but larger consumer group
sizes may result in higher site visibility than those produced by foragers. As noted by
Habu (1996) for the Early Jomon Moroiso phase in Japan, hunter-gatherer habitation sites
246
with evidence of semi-subterranean houses and high inter-assemblage variability are
good candidates for typical collector-type systems. Collector-type systems are thought to
be most common in temperate, arctic, and subarctic environments because the seasonal
constraints of mobility and availability of food resources play an important role in
determining mobility and resource procurement (Binford, 1980; Lieberman et al., 1993;
Kelly, 1995: 117).
In principal, archaeological evidence of residential and task sites should be quite
different at extreme ends of the collector/forager continuum. However, extrapolation
from the archaeological record is not so straight forward. Hunter-gatherer land-use
strategies are not naturally dichotomous and can seldom be neatly categorized as such
(Binford, 1980, 1982; Lieberman et al., 1993). Seasonal variation in acquisition
strategies may typify a group as foragers in the summer months and collectors in the
winter months, or vice versa. Interpretations of site function might vary according to
differences in the original organization of habitation sites or methods of discard: carrying
out certain tasks farther from the main group; provisioning of sites for subsequent
reoccupation (caching); cleaning practices such as the removal of debitage and debris
from the primary habitation area; or failure to discard valuable, highly curated tools
(Schiffer, 1972, 1987: 58-72, 89-97). Palimpsest accumulations from multiple
occupations of various types can also confuse interpretations of land-use patterning
(Binford, 1982). Dense accumulations of archaeological remains are most prone to
misinterpretation. While inter-assemblage variability should increase with length of
247
occupation and be indicative of residential sites, overlapping task sites might also
simulate a pattern of longer-term habitation (e.g., Schiffer, 1975).
However, distinguishing between residential and task specific sites should be
easier than recognizing more detailed aspects of site function. Cooking activities would
seldom occur outside of field camps, base camps, or residential bases. Similarly, the
large grinding stones found in East Gobi assemblages are heavy and so were probably
limited to residential rather than task sites. The fact that they are associated with denser
accumulations of cultural remains confirms this point. In view of the cost of transporting
grinding equipment and the laborious task of processing, it is likely that large formal
grinding stones would only be found at longer-term residential bases where intensive
processing was being undertaken. Grooved slabs used for finishing beads or used as
shaft straighteners are likewise less portable. Despite frequent evidence for the use of
pottery by hunter-gatherers, ceramic vessels are also somewhat less portable. This is
especially true when considering less fragile technologies like baskets or more expedient
ways of cooking such as with open-fires or earthen pits (Nelson, 1991; Sassaman, 1993).
It is expected that both the friable nature of ceramics, particularly the low-fired ones used
in Oasis 2, and spatial limitations on food preparation activities should limit their use to
the base camp or forager residential base. Less portable tool types such as grinding
stones, grooved slabs, and pottery are recorded for each site and are expected to give an
indication of both site-specific activities and site type.
Site categorizations can be made based on the range of activities represented by
tool types. Habitation sites should include evidence for a range of activities, while task
248
sites are expected to be focused on one or possibly two. Cooking, manufacturing,
hunting, bead-making, and lithic reduction are recognized as key activity types.
Woodworking is tentatively assigned based on the presence of adze/axes (Hayden, 1989),
though it is not clear that this was their primary or sole utility. The presence of bone is
noted as it suggests butchering or processing meat was one on-site activity, but is not an
especially significant distinction due to high variation in the survival rates of bone in such
assemblages.
In this study, residential sites are primarily defined on the presence of pottery,
grinding stones, hearths, or fire-cracked rocks, which are attributed to extensive
“cooking.” Evidence for cooking is not considered definitive since pottery and hearths
may be underrepresented as they are artefacts more subject to destructive postdepositional processes than lithics. “Manufacturing” includes tools such as scrapers,
drills, grooved slabs, whetstones, knives, and slag, which are associated with production,
processing, and repair activities. “Hunting” is based on the occurrence of large and small
unifacial or bifacial points morphologically consistent with arrows or spears. The latter
category is also expected in sites where manufacture and maintenance of hunting
equipment was carried out. Microblades were probably multifunctional and are not used
to infer site function. Lithic reduction is recognized by the presence of cores,
hammerstones, and unused/unmodified debitage flakes. Ornament production or beadmaking is recognized based on the presence of unfinished beads and pendants.
The presence of specialized tools and formal generalized tools is informative not
only about site function, but also about site specific approaches to raw material use and
249
the relative importance of different tasks. Tools manufactured explicitly for one function
are expected to be more efficient. They should be most common at sites produced under
time constraints, where only a limited range of activities took place (Torrence, 1983). A
specialized tool kit includes a diverse range of highly efficient, but less multi-functional
implements. An array of specialized tools is likely to be more common when there is
ample time for increased investment in manufacture, but actual tool use occurs under
situations of frequent repetition and limited time. In contrast, when many different tasks
are being undertaken, a more generalized tool kit is expected. Generalized tools are those
that serve a multitude of functions, but are not designed to fill any one particular role.
For this reason, long-term residential sites are expected to contain a select range of
formal specialized tools and numerous informal generalized types. Both specialized and
generalized tools can be either expediently or formally made, although specialized tools
are frequently types that require higher investment in labour.
Residential and individual mobility have a great affect on differences in access to
high quality raw materials. Access to raw materials is expected to most heavily influence
technological choices, since decisions of curation and transport are based on the
frequency and predictability of resource distribution (Andrefsky, 1994). Highly mobile
individuals are should produce assemblages consistent with the need to either transport
required tools to each new location or make use of whatever raw material resources are
immediately available (Shott, 1986; Kuhn, 1994; Brantingham, 2003; Barton et al.,
2007). Formal tools are likely to have been favoured over more expediently produced
ones in circumstances where the reliability of a tool is important and there is intermittent
250
but predictable access to the quality of raw materials needed to produce standardized
forms. When highly mobile individuals choose to transport formal tools, the constraints
associated with transport or the carrying costs support the use of multi-functional rather
than specialized tool types (Bleed, 1986; Torrence, 1989; Kuhn, 1994).
Composite hafts with microblade insets, as attested to in Gobi Desert assemblages
(Dorj, 1971; Derevianko and Dorj, 1992; Derevianko et al., 2003), are an excellent
example of multi-functional formal tools that are portable and easily maintained (Bleed,
1986). Some researchers have also suggested that they were favoured as being more
reliable in big game hunting than previous technologies (Myers, 1989; Elston and
Brantingham, 2002). These characteristics of portability, versatility, and reliability are
thought to indicate both an increased investment in the aquisition of large game and high
mobility among microblade-using post-LGM hunter-gatherers in Northeast Asia (Elston
and Brantingham, 2002). The large bifaces used by Palaeoindians in North America are
another example of a technology that is portable, versatile and reliable (Kelly, 1988;
Surovell, 2003: 229-236). Such technologies can be thought of as generalized formal
tools.
Conversely, higher levels of sedentism are usually associated with expedient tools
and cores (Parry and Kelly, 1987). Expedient technologies are associated with a lower
degree of raw material conservation and lessened emphasis on reliability (Wallace and
Shea, 2006). Expedient tools are more common when raw material is plentiful. In such
circumstances, generalized expedient technology should be favoured when a wide range
of tasks will be undertaken, and there is plenty of time to complete specific tasks.
251
Specialized expedient technology would then be indicative of a situation where there is
ample raw material, but constraints on other aspects of production/processing such as
limited time or the need for a higher level of performance than a multi-functional tool
could provide (i.e., reliability).
Therefore, within each assemblage tool characteristics such as expedient, formal,
specialized, and generalized are indicative of both site function and raw material access.
Raw material access is a function of both local distribution and relative levels of
residential or individual mobility. Specialized tools in Gobi Desert assemblages include
drills and awls (drilling – formal), perforators (drilling – expedient/informal), projectile
points (hunting – formal), grinding stones (food processing – both formal and informal
types), grooved grinding slabs (bead-making or hunting – formal15), pottery (cooking or
storing – formal), edge-ground and possibly chipped adze/axes (woodworking – formal).
Generalized tools include used flakes (informal), expedient scrapers (informal),
used and/or retouched microblades (formal), and bifacial knives (formal). Composite
points with microblade inserts are also formal specialized tools, but can not be identified
solely on the basis of microblades, which were used individually or as insets for a variety
of tools. Microblades are considered a type of generalized formal technology. Wedgeshaped microblade cores are also considered to be generalized formal tools, which
provide both a source of standardized tool blanks, and a knife-like edge for cutting.
Adze/axes are unique formal tools that can serve a multitude of purposes, but
might also be considered specialized in that they are usually associated with
15
Grooved grinding slabs are known to have been used for bead manufacture, but were possibly also used
as shaft straighteners (see Maringer 1950: 109-110; Janz, 2006).
252
woodworking (Hayden, 1989; Mills, 1993; Yerkes et al., 2003). Edge-ground tools are
known to have been used during episodes of intensive, repetitive processing under high
constraints on time (Hayden, 1989). The occurrence of chipped and/or polished
adze/axes in Oasis 2 and Oasis 3 assemblages is particularly notable due both to implied
function and implications for raw material conservation. While the majority of tools in
these collections were made on small flakes, adze/axes are notable because they have the
potential to be resharpened and reused over an extended period. Edge-ground adze/axes
can be resharpened “seemingly indefinitely” (Hayden, 1989). At the same time, edgegrinding is extremely time-consuming and appropriate downtime would have been
required to manufacture and maintain functionality (Hayden, 1989; Owen, 2007). Such
tools are larger and heavier, so theorectically less portable, than the typical microlithic
tools. Woodworking or other highly intensive processing and manufacturing activities
are suggested when adze/axes are present.
The aforementioned characteristics of artefact type and assemblage composition
were used to categorize each site as summarized in Appendix D. The dichotomous label
of foragers versus collectors is not applied here. There is no clear evidence from the
Gobi Desert of either large middens or permanent/semi-permanent prehistoric structures.
Site occupations probably lasted no longer than one season. Collector-type settlement
systems known among other prehistoric Northeast Asian hunter-gatherers like the Early
Jomon in Japan are not typical of this region. Binford’s (1980) emphasis on the
continuum of hunter-gatherer adaptations and his overarching definition of foragers and
collectors – the former focused on moving themselves to resources and the latter as
253
focused on moving resources to themselves – is a more compelling distinction that can be
better applied to the Gobi Desert archaeological record.
Due to variation in the organization of settlement systems and the difficulty in
discerning discrete site functions, the three site type designations are used that transcend
categorizations based on Binford’s organizational systems. Residential A sites should
coincide with either longer term base camps or frequently revisited residential bases. As
artefact assemblages they can be recognized as larger (>1000 artefacts16) multipurpose
sites with evidence of cooking. High variability in activities is attested to by the range of
tool categories, use of less portable processing technologies, the dominance of
generalized tool types, and a diverse array of specialized tools. Due to heightened levels
of discard at longer term or reoccupied sites, highly valued and maintained artefacts such
as adze/axes are more likely to be associated with Residential A type sites. Both caching
behaviour and irreparable breakage can account for their presence.
Residential B sites are generally smaller (<1000) multipurpose sites. Such
residential sites are considered to be shorter-term habitation sites such as singularly or
infrequently occupied forager-type residential bases or satellite field camps of collectortype groups. They should contain a range of artefact types associated with different
activities, but have fewer artefact types than Residential A sites. Fewer artefact types
might also be present in field camps than in forager residential bases. Some less portable
processing technologies could be present (e.g., grinding stones at a milling-intensive field
16
Although artefact counts can be misleading since they do not take into account increased production of
debitage from certain types of activities like intensive lithic reduction, assemblage magnitude is still often
representative of occupation intensity and can be given some consideration. Inferred site size was not
necessarily decisive in site type categorization.
254
camp), but are likely to be more portable versions of formal prototypes. Formal
generalized tools such as bifaces and wedge-shaped microblade cores should dominate
the tool kit.
Those sites referred to as task sites are consistent with locations in Binford’s
terminology. They are expected to occur within all types of organizational systems. Such
sites are expected to produce highly variable artefact assemblages based on their different
functions. In general, they are characterized by small site size (most notably sites with
less than 100 artefacts) and a lack heavy equipment or evidence of cooking. The
presence of specialized or generalized tools and degree of formality in manufacture
should be related to site function and distance from a residential site; however, due to
carrying constraints, generalized tools should be common and specialized tools rare or
absent. When specialized tools do occur in task sites, they may suggest the importance of
a singular activity. Scrapers should be found in notable quantities only when they are
intensively used as part of a site specific activity. Discarded tools should be expedient or
heavily utilized.
4.1.2.2. Results
The environmental distribution of chronologically ordered residential and task sites
allows for interpretations of land-use during the terminal Pleistocene to middle Holocene.
Palaeolithic and Early Epipalaeolithic sites were not included in this analysis since they
are so sparsely represented in the sample. This is partially due to my intentional selection
of probable post-LGM sites. The primary focus is Late Epipalaeolithic and
255
Neolithic/Eneolithic (Oasis 1, 2, and 3) land-use. Oasis 1, 2, and 3 assemblages are most
viable for comparison of inter-assemblage variability due to similarities in the basic
modes of reduction, reliance on microblade core reduction sequences and the use of other
microlithic tools.
Only one Residential B site and seven task sites were assigned to Oasis 1 across
all regions. Since categorization of residential sites was based partially (but not entirely)
on the presence of pottery, it is possible that residential sites were underrepresented;
however, it is likely that Oasis 1 sites were simply more ephemeral than
Neolithic/Eneolithic residential dune-field sites. Smaller site sizes and reduced interassemblage variability among Oasis 1 type sites might also have contributed to the failure
to recognize some residential base camps. Alternately, as suggested by the relative dense
concentration of finds at Chikhen Agui, more intensively occupied Oasis 1 sites may
have been centred on environments that were not heavily sampled in these collections.
Site distribution and reduction strategies for Oasis 1 assemblages were not further tested
due to low sample size. Clear differences in site distribution are sufficient to support the
hypothesized shift in land-use by about 8.0 kya.
The distributions of residential sites across ecozones for the entire Gobi Desert are
represented in Figure 4.7 and Table 4.9. Residential sites are primarily confined to
lowland dune-field/wetland environments. Task sites are situated in a range of different
environments and probably represent lowland-dwelling groups using upland
environments in a pattern reminiscent of collectors – characterized by the strategy of
256
moving select resources from upland environments into lowland environments where
residential camps were situated.
As with the previous samples, counts of some variables were so low as to make
suspect the results of Pearson’s chi-square test. Raw counts (Table 4.9) indicate that for
Oasis 2, Residential A sites are more common in lowland dune-field/wetland
environments and less common in upland environments. Residential B sites are similarly
common in lowland dune-field/wetland environments, with few sites found in other
environments. Individual cell counts show that the distribution of task sites is most
notable. Task sites are fewer than expected in lowland dune-field/wetlands for both
periods and significantly more common in upland environments during Oasis 2. The
majority of Oasis 3 sites were derived from dune-field/wetland environments and there
are too few sites in other environments to accurately test the significance of distribution.
Nevertheless, this situation may also reflect an increasing reliance on dune-field/wetland
habitats, which is itself significant.
257
Oasis 2
100%
75%
Upland
Lowland dry
50%
Lowland river
25%
Lowland dunefield/wetland
0%
Residential A
Residential B
Task site
Oasis 3
100%
Upland dry
75%
Upland
50%
Lowland dry
Lowland river
25%
Lowland dunefield/wetland
0%
Residential A
Residential B
Task site
Figure 4.7 Distribution of site types across Gobi Desert for Oasis 2 and Oasis 3.
258
Oasis 2
Ecozone
Lowland dune-field/wetland
Lowland river
Lowland dry
Upland
Residential A
10
8.0
0
0.9
1
0.4
1
2.7
Residential B
5
4.0
0
0.4
0
0.2
1
1.3
Task site
3
6.0
2
0.7
0
0.3
4
2.0
X2 = 10.62, p = 0.1007
Oasis 3
Ecozone
Lowland dune-field/wetland
Lowland river
Lowland dry
Upland
Upland dry
Residential A
9
7.8
0
0.6
0
0.6
2
1.4
0
0.6
Residential B
12
10.7
1
0.8
0
0.8
1
2.0
1
0.8
Task site
6
8.5
1
0.6
2
0.6
2
1.6
1
0.6
X2 = 7.93, p = 0.4399
Table 4.9 Actual and expected distribution across Gobi Desert ecozones for Oasis 2 and
Oasis 3 sites.
259
Table 4.10 shows the regional distribution site types within each ecozone. The
Alashan Gobi is notable for the high frequency of Residential B sites during Oasis 3.
Otherwise, broadly similar trends in site type distribution are visible.
Ecozone
Site type
Lowland
dune-field/wetland
Res. A
Res. B
Task
Res. A
Res. B
Task
Res. A
Res. B
Task
Res. A
Res. B
Task
Res. A
Res. B
Task
Lowland river
Lowland dry
Upland
Upland dry
East Gobi
O2
5
2
0
0
0
1
0
0
0
0
1
2
0
0
0
O3
4
2
0
0
1
0
0
0
0
1
1
1
0
0
0
Gobi-Altai
O2
3
2
1
0
0
1
0
0
0
0
0
1
0
0
0
O3
3
2
0
0
0
1
0
0
0
0
0
1
0
0
0
Alashan
Gobi
O2 O3
2
2
1
8
2
6
0
0
0
0
0
0
1
0
0
0
0
2
1
1
0
0
1
0
0
0
0
1
0
1
Table 4.10 Regional distribution of Oasis 2 and Oasis 3 sites for each ecozone.
Residential A and Residential B sites are expected to represent two different types
of residential organization. A greater reliance on Residential A sites would suggest either
more long term occupation of residential sites, or more frequent reoccupation of the same
sites. If certain sites are frequently revisited or regularly incorporated into a yearly
round, hunter-gatherers are more likely to cache heavy equipment or stockpile raw
materials at key locations. In contrast, an increase in the number of Residential B sites
should be related to either shorter term habitation or to less frequent visits. Table 4.11
shows the distribution of site types for the entire Gobi Desert and for each region during
260
Oasis 2 and Oasis 3. In contrast to the sharp increase in the relative frequency of
Residential B sites during the Oasis 3 period in the Alashan Gobi, the distribution of site
types in the East Gobi and Gobi-Altai is similar in both periods. This indicates that the
Alashan Gobi sample is unique, with a possible decline in Residential A sites and a
corresponding increase in Residential B sites during the Oasis 3 phase. Statistically, the
difference in distribution of Alashan Gobi residential type sites between periods is
significant (X2 = 4.41, p = 0.0358).
Site Type
Period
Residential A
Oasis 2
Oasis 3
Oasis 2
Oasis 3
Oasis 2
Oasis 3
Residential B
Task site
All Gobi
Desert sites
12
11
6
15
9
12
East Gobi
5
5
3
4
3
1
Gobi-Altai
Alashan Gobi
3
3
2
2
3
2
4
3
1
9
3
9
Table 4.11 Actual and expected distribution of residential site types for Oasis 2 and
Oasis 3 across the entire Gobi Desert and for each region.
A few notable divergences in tool kits, including the greater visibility and
formality of grinding stones in the East Gobi, suggests that regional trajectories may have
been quite different during the Neolithic/Eneolithic despite cross-regional trends in the
transition from Epipalaeolithic to Neolithic. Local environments were also quite
divergent, as is detailed in Chapter 5. For that reason, variation at the regional level is
highly important. East Gobi sites show no significant difference in the frequency of
different site types between Oasis 2 and Oasis 3 (df = 1, X2 = 0.084, p = 0.771). The
261
Gobi-Altai sample (N = 10) is too small to test, but the distribution of residential site
types is identical for Oasis 2 and Oasis 3 (Table 4.11).
Despite a lack of statistically significant temporal variability for the East Gobi,
residential sites may be more widely distributed in Oasis 3 than in Oasis 2. During Oasis
3, both Residential A and Residential B sites were found outside of dune-field/wetland
environments, while only one such Residential B site can be attributed to Oasis 2.
Yeggah Bolagah In Sumu (Site 15) was found in an upland setting near a major water
source and may represent occasional use of upland environments for short-term
habitation, or perhaps misidentification. Residential A sites were confined to dunefield/wetland environments during Oasis 2. Distribution of sites in the Gobi-Altai region
is mostly consistent between periods, with residential sites restricted to lowland dunefields. Upland zones appear to have been more commonly exploited for residential
locales, including Residential A, in the Alashan Gobi during both Oasis 2 and Oasis 3.
These habitations are related to an extensive series of archaeological sites in the UkhTokhoi/Khara Dzag plateaux region, only a small portion of which were sampled. It is
probable that the tool stone rich region was an important locale for seasonal habitation
and raw material procurement (Maringer, 1950: 97-99, 103-111; Bettinger et al., 1994).
The lack of Residential A sites prior to 8.0k cal yr BP suggests that early postLGM hunter-gatherers exhibited high residential mobility, while the site distribution
suggests the utilization of a variety of environments with a focus on high elevations (>
1200 m a.s.l.). After 8.0k cal yr BP, low elevation dune-field/wetlands were extensively
utilized for residential sites, while small sites probably represent task group oriented
262
procurement of additional resources from a wide range of environments. These
indications so far support the original hypotheses; however, the organization of
residential mobility within dune-field/wetland environments is not defined and there is no
clear indication of variation in organizational strategies between the Neolithic (Oasis 2)
and Eneolithic (Oasis 3).
4.1.3. Lithic reduction strategies
Defining residential mobility and exploring possible changes in land-use between Oasis 2
and Oasis 3 are the focus of quantitative lithic analysis. The increased importance of
bifacial technology, decorative elements (ostrich eggshell beads, decorative pottery
finishes), higher investment in pottery, polished adze/axes, and a decline in the
importance of formal milling technology all suggest a change in prioritization of needs
during Oasis 3. A shift in the the relative frequencies of non-portable tool types like
large grinding stones, and the introduction of lugs and handles to pottery manufacture
could be an indication of an underlying increase in residential mobility. Raw material
conservation practices can also attest to this type of strategic reorganization.
Quantifiying certain aspects of lithic raw material conservation can test assumptions
about changes in residential mobility between periods.
The previous section indicates that there are two types of Neolithic/Eneolithic
residential sites – Residential A and Residental B. These two site types functioned
during both Oasis 2 and Oasis 3 A and Residential B sites are distinguished in the
previous section based on site size, inter-assemblage variability, and relative importance
263
of formalized technology. Successive reoccupation, however, can contribute greatly to
the illusion of long-term habitation since the potential is greater for accumulation of a
variety of tool types and the incorporation of less frequently used expedient technologies
as the site is successively reoccupied. However, if all sites are related to the same length
of occupation and function, there should be no difference in the use of tool stone aside
from that imposed by local availability of raw material.
Consistent differences in the use of lithic raw materials between site types would
support the distinction of two separate types of residential occupation, just as temporal
differences would support a shift in land-use over time. In order to identify differences in
the length of occupation related to time period and residential site types, lithic analysis is
used to address issues of raw material procurement, relative intensity of use and retouch
prior to discard, and relationships between patterns of such and degree of residential
mobility. Primary variables of analysis included: core dimensions, relative frequency of
formal to informal core types, remnant core and scraper cortex, and scraper retouch.
4.1.3.1. Modeled expectations and methods
The degree to which artefacts are reduced or retouched before discard signals aspects of
raw material access and conservation that is related in part to individual and residential
mobility. Assemblage-specific reduction strategies reflect choices in raw material use
that give an indication of both transport limitations, and the proximity and/or regularity of
access to raw material sources (Binford, 1979; Bamforth 1986; Bleed 1986; Kuhn 1991,
1994, 1995). When it is impossible or inconvenient to acquire more raw materials for
264
tool manufacture, the knapper will regularly make more efficient use of the materials on
hand. If raw material is readily available he/she will be less inclined to retouch and reuse
tools or to reduce cores in a conservative manner (Shott, 1986). As a result, highly
mobile hunter-gatherers who may not encounter tool stone for long periods tend to
organize technological systems around the curation of artefacts, while decreased
residential mobility is usually related to the use of less standardized and more expedient
tools as raw materials can be more easily stockpiled (Shott 1986; Parry and Kelly 1987;
Torrence 1989; Kuhn 1995). Some tools or cores may be discarded at an early stage due
to raw material flaws or mis-strike, but an assemblage considered as a whole should
indicate a particular pattern of raw material use and tool curation (see Kuhn, 1995;
Surovell, 2003).
Various characteristics were quantified in order to gauge intensity of raw material
use and curation: relative abundance of formal and informal core types, relative
percentage of cortex remaining on cores and scrapers, invasiveness of scraper retouch,
core dimensions, and variability of core size (CV). The analysis focuses on cores and
scrapers. These two artefact types have proven useful in demonstrating a correlation
between length of occupation/level of mobility and intensity of use and reduction because
they tend to be repetitively reduced, reused, and retouched (Kuhn 1990, 1991; Andrefsky,
1998; Blades, 2003; Eren et al., 2005; Wallace and Shea, 2006). Equally important, these
artefact classes are virtually ubiquitous in Gobi Desert sites, occurring in a range of
ecological contexts, independent of site size (see Fairservis, 1993; Maringer, 1950).
265
Flexibility and formality
Quantifying relative frequencies of formal and informal core types is one method of
recognizing more conservative reduction strategies. According to Shott’s (1996)
definition of curation, formal core technology can not necessarily be identified as
exhibiting high levels of curation, since curation is a continuous category and relative to
the maximum utility of the object. At the same time, it is clear that formal prepared cores
are volumetrically and strategically designed to produce high numbers of standardized
blanks, which minimizes risk, and extends the use life of raw materials and potential for
curation (Andrefsky, 1987; Clark, 1987; Wallace and Shea, 2006). As such, formal cores
do represent heightened raw material conservation.
Such “conservative” reduction strategies therefore suggest limited access to raw
materials; level of access generally being a function of specific provisioning strategies
(see Kuhn, 1995, 2004). Provisioning strategies are most closely linked with the
organization of land-use, including mobility. Kuhn (1995) outlines two potential
methods of planned provisioning – provisioning places or provisioning individuals.
Choice of provisioning strategy is expected to reflect the frequency and duration of
residential moves. Provisioning of places, or stockpiling raw materials at frequently
revisited sites, should be most extreme when residential moves are few and occupations
are more long-term. In contrast, when the frequency of residential moves is high and
occupations are more short-term, individuals make sure that they, themselves, are
constantly provisioned with raw materials. Provisioning of highly mobile individuals
requires attention to transport costs. As previously noted, transported tool kits ideally
266
contain flexible multi-functional tools which are resharpened and reused until their edges
wear out (Kuhn, 2004) or until new high quality raw materials can be procured and new
tools made. Exploitation of environments with less even distribution of high quality raw
materials are likely to encourage more conservative patterns of use and discard.
Specialized microblade core reduction sequences in the Gobi Desert suggest the
continuous importance of provisioning individuals with efficiently transportable
equipments from the period immediately following the Last Glacial Maximum (LGM)
until the early Metal Ages. Microblades appear to have been used as inset blades for
organic hafts in many Northeast Asian assemblages (Derevianko and Dorj, 1992; Kirillov
and Derevianko, 1998; Elston and Brantingham, 2002; Derevianko et al., 2003).
Microblade cores offer a number of benefits, including the potential to be heavily
volumetrically reduced before reaching a point of non-utility, a very low failure rate once
microblade production is initiated (Bamforth and Bleed, 1997), and potential to be used
as both cores and tools (sensu Kuhn, 2007). Dual platforms on cylindrical microblade
cores would have further extended use life by allowing for greater reworking
opportunities when mis-strikes and raw material flaws resulted in heavy step-fracturing
around one platform. Wedge-shaped microblade cores may have been used as an ideal
transportable core-tool, providing knife-like edges that offered additional utility to
microblade production, and making them well-suited for use in transported tool kits
(Kuhn, 1994). This is also true of large bifaces and biface cores (see Kelly, 1988;
Hofman, 1992; Ingbar, 1992; Surovell, 2003), which are common in late Oasis 2 and
early Oasis 3 sites.
267
Curation through “conservative” approaches to raw material use can also be
recognized within “flexible” reduction sequences. When raw material is limited and
artefact curation important, flexible and versatile design is important. If an artefact is to
be carried and used over long distances and in diverse settings, it must be able to perform
in a wide range of activities since the timing and specific uses are not consistently
predictable (Nelson, 1991). Microblades make flexible tools because they can be
shortened and retouched as component insets of a larger blade, or left long to be hafted as
individual tools (see Tabarev, 1997). Flexibility of microblades allowed users to adapt
these tools to continually changing needs and resources.
Additionally, changes in Northeast Asian microblade core reduction sequences
following the LGM, specifically an increase in less formally prepared cores and more
flexible core types (see Chapter 3), suggest that flexibility became increasingly valued
during the terminal Pleistocene and early Holocene. It can also be asserted that the
appeal of less formally prepared and more flexible microblade core reduction sequences
was their potential for continuous rejuvenation and adaptability to raw material size and
shape. While some microblade core preforms from Yingen-khuduk exemplified a
reduction strategy based on the production of cylindrical/conical rather than wedgeshaped cores, several examples of Gobi Desert microblade cores clearly indicate the
transformation of a heavily reduced wedge-shaped core into a conical or more rarely a
cylindrical shape. Judging from the dimensions of exhausted cores in Gobi Desert
assemblages, microblade cores tend to take on a conical form when heavily reduced.
Conversely, wedges were often present on the backs of extant conical and cylindrical
268
core types. Such a pattern of core reduction indicates that microblade cores were
designed to be rejuventated and reused until most they were very small and most of the
original nodule was used.
Thinking about microblade core technology as the basis for a highly portable tool
kit can help explain the lasting popularity of this technology among mobile huntergatherers in Northeast Asia, despite the requirement of relatively high quality raw
materials. Ethnographic data from Nunamiut informants suggests a pattern of use for
discoidal transported cores that might be comparable to microblade cores, “as they put it,
you carry a piece that has been worked enough so that all the waste is removed, but that
has not been worked so much that you cannot do different things with it” (Binford, 1979:
262). Flexibly designed reduction strategies, those able to be adjusted according to
immediate and changing needs, are also archaeologically attested in other regions.
Mousterian sites from southwest France indicate a mode of core reduction that could be
redirected at different points of use, transport, and/or reduction, creating a sort of
branched operating sequence (chaîne opératoires ramifiées) (Bourguignon et al., 2004).
Tool kits based on microblade core technology represent an organizational strategy
broadly focused on regular raw material conservation and formal standardized forms.
Wallace and Shea (2006) explored the relationship between expectations of
reduced mobility in the archaeological record and an increase in informal core types
during the Levantine Mousterian. By reclassifying Middle Palaeolithic core types as
“formal” and “expedient,” they found that later assemblages, particularly ones with signs
of more prolonged occupation or occurring in richer environmental regions capable of
269
supporting longer term habitation, emphasized expedient core types. According to
Kuhn’s (1995) detailed analysis of strategic approaches to raw material use, the strategy
of provisioning places in anticipation of future needs should contribute to a decline in the
importance of formal core and tool technologies due to the greater availability of raw
materials (also Kuhn, 2004). Regardless of the underlying motivation for the use of
expedient core technologies, it is clear that such strategies prioritize the importance of
immediate functional needs over long-term raw material conservation (Binford, 1979;
Wallace and Shea, 2006).
The relative importance of formal to expedient core technology relays information
about both group-wide adaptations and site-to-site variation in the production of different
core types. Availability and the quality of raw material play an essential role in the
manufacture of formal versus informal cores. Andrefsky’s (1994) analysis lithics from
western North America suggests that the manufacture of formal cores and tools is more
often mediated by raw material availability than length of site occupation. Formal and
informal tool types are both produced when high quality stone is abundant, but informal
tools are produced when only low quality tool stone is available, irregardless of quantity.
High quality raw materials in low abundance are usually related to formal manufacture.
As such, we can infer that Gobi Desert assemblages indicate infrequent access to
high quality raw materials at the supraregional scale. Such an interpretation is notable
since high quality raw materials are widely available across the Gobi Desert. The
importance of microblade core technology in the Gobi Desert indicates either high
270
residential mobility or regular periods of high individual mobility with sporadic access to
raw materials.
Therefore, site-to-site variation in the relative frequency of formal to informal
tools should give some indication of differences not necessarily in the distribution of high
quality raw material, but in individual access to those resources. Inhabitants at some
locales appear to have relied extensively on one type of low quality raw material, which
we can assume came from a local source. Such circumstances do mirror local raw
material availability. On the other hand, raw material use should be consistent within site
categories and variable between them if site type categorizations (i.e., Residential A,
Residential B, and task sites) reflect real differences in site function and length of
occupation rather than simply resulting from differences in the frequency of
reoccupation. Temporal changes in residential mobility might also be reflected in
frequencies of formal to informal tool types between Oasis 2 and Oasis 3 sites.
Based on the importance of microblade and other formal core types in extending
the life of a nodule, the ratio of formal to informal cores should reflect situational
conservation at the regional level, differentially representing: constant, frequent, and
reliable access to raw material when expedient cores are numerous and formal cores rare;
reliable access to raw materials with anticipated periods of short term or individual
restrictions (e.g., preparation of cores/flakes for use in hunting forays or other tasks); or
at the other extreme, highly restricted access to raw materials. According to Andrefsky’s
(1994) model, the presence of low quality raw materials but restricted access to high
quality tool stone can also influence the frequency of expedient core types. The use of
271
lower quality tool stone in assemblages with a high frequency of expedient types should
be considered. Frequency of informal/expedient cores at each site is recorded in
Appendix E. The ratio of microblade to other cores at each site is recorded in Appendix
D.
Remnant cortical surface
Just as the use of formal rather than informal cores can be a measure of raw material
conservation, the amount of cortical surface remaining on cores and scrapers at discard is
related intensity of reduction. Since high levels of remnant cortical core surface
generally indicate a lack of extensive reduction, principles of raw material conservation
or lack thereof can be applied. As with the measure of formal to informal core types,
mean percentage of remnant cortical surface within an assemblage should vary between
site types and time periods according to raw material access. High levels of remnant
cortex are expected at sites where large amounts of raw materials are stockpiled and
where primary reduction occurs (Dibble et al., 2005).
Extensively prepared formal cores tend to exhibit lower levels of remnant cortex
since much of the original surface is removed during initial preparation (Dibble et al.,
2005). Since informal cores are considered to represent a less conservative approach to
raw material use, a high percentage of remnant cortical surfaces could represent either a
less intensive reduction of cores or a more expedient and less conservative approach to
raw material use. Assemblages with higher numbers of expedient/informal cores are
expected to display higher levels of remnant core cortex. Circumstances of interest are
272
those where cortex percentages are low and informal core types predominate, or where
there are many formal core types and above average distributions of high cortical surface.
Core morphology is an important consideration for explanations of intraassemblage variability as different formal core types exhibit natural divergences relative
to remnant cortical surface. Most core reduction sequences practiced in the Gobi Desert
were based on extensive core preparation prior to flake removal, a method that leaves
very little cortex (see Dibble et al., 2005). Microblade cores are naturally more likely to
have low levels of remnant cortex. Exhausted microblade cores should have little
remaining cortex. Wedge-shaped cores may maintain cortex on the knife-like edge if the
wedge back is not heavily bifacially worked. Similarly, flat-backed microblade cores
might bear evidence of a cortical surface until entirely exhausted. There should be less
remnant cortical surface on conical and cylindrical microblade cores. Expediently
prepared or “informal” microcores have minimal preparation of the nodule prior to
removal of useable flakes and should have high percentages of remnant cortical surface.
Cores most sensitive to measurements of cortical surfaces are amorphous cores, and to a
lesser degree expediently prepared microblade (< 10 mm), bladelet (11-15 mm), and
elongated flake cores.
Remnant cortical surface for scrapers should be correlated with both intensity of
reduction and the types of flakes used. Continuous core reduction following removal of
all cortical surfaces results in increasingly low frequencies of remaining cortical surface
in the assemblage (Dibble et al., 1995; Dibble et al., 2005). Assemblages with high
levels of cortex on scrapers would then be indicative of the types of flakes used: scrapers
273
produced on primary flakes indicating primary reduction, and scrapers produced on
unreduced pebbles or cobbles.
At the same time, scraper cortex will not necessarily fit these expectations.
Cortical flakes are larger and may be selected for transport due to increased utility in
comparison to smaller flakes produced later in reduction (sensu Kuhn, 1994; Douglass et
al., 2008). Adding to the incongruity of this relationship, sites of primary reduction may
also be loci for replacement and discard of transported formal tools (Gramly, 1980),
suggesting that heavily reduced scrapers (and cores) may be found at sites of primary
reduction due to transport from another locale rather than on-site use. Depending upon
the conditions, remnant cortical surface on flakes may be underrepresented in situations
of primary reduction and overrepresented in conditions of high curation (see Douglass et
al., 2008). Heavily used and retouched scrapers with remaining cortex are more likely
products of transport and extended curation than those only minimally used. This
combination of qualities can particularly bias samples within smaller assemblages.
Remnant cortical surface must be considered based on other elements of the assemblage
in order to determine whether they relay useable information about practices of transport
and discard.
Several expectations related to the measure of remnant cortical surface can be
outlined:
a.
In typical circumstances, a reliance on informal core types should be
mirrored by relatively high frequencies of remnant cortical surface.
274
b. Sites where the majority of tools were locally manufactured, used, and
discarded should be associated with higher levels of remnant cortical
surface on both cores and scrapers.
c. Remnant cortical surface should be even higher in conditions where raw
materials were stockpiled in anticipation of recurrent occupations or
longer stays. This is especially true if unreduced or minimally reduced
nodules were transported to the site, as may have been typical of nodule
procurement in microblade core manufacture (Bleed, 2002).
d. When a site is occupied for only a short duration, we expect the relative
percentage of remnant cortical surface to be most sensitive to differences
in distance (whether logistic or geographical) from raw material source,
conditions of transport, and situational use and discard.
Methods of calculating remnant cortical surface and extent of reduction vary. A
scale based on percentages of remant cortex was used in this study (after Dibble et al.,
2005). Remaining cortical surface was measured by categorizing cores and scrapers
according to approximate percentage of entire surface covered by cortex: 1 = 0%, 2 = 125%, 3 = 26-50%, 4 = 51-90%, 5 = 91-100%. Microblade cores require relatively
extensive reduction and most of the cortex is removed prior to use. Cores with cortex
covering more than 25% of the total surface area are considered to have notably high
percentages of remnant cortical surface. The same measure will be used for scrapers,
since the main reduction strategy represented in these assemblages – the production of
microblades – results in high numbers of non-cortical flakes that are used for unifacial
275
tools like scrapers. Assemblages classified as having “high” overall levels of remnant
cortex are those assemblages where at least 50% of a particular artefact type (i.e., cores or
scrapers) has cortex on over 25% of the total surface area. Assemblages with “low”
levels of remnant cortical surface ones where 50% of cores or scrapers exhibit less than
25% remnant cortical surface at discard. The results are summarized in Appendix E and
in section 4.1.3.2, they presented along with relative frequencies of formal versus
informal core types and core dimensions in order to better understand overall patterns in
raw material use.
Dimensions
Extensive reduction prior to discard is related to recurrent use and rejuvenation,
suggesting fuller exploitation of the maximum core utility (Shott, 1996). Core
dimensions can give a good indication of reduction. Core volume at discard is
determined by both the original nodule size and the intensity of reduction. An
assemblage with larger cores may result from consistently less intensive reduction, but
may also represent the regular use of larger nodules. The use of larger nodules might
arise in a number of situations, including the procurement of stone from primary rather
than secondary deposits (Kuhn, 2004), or similarly quarrying large blocks rather than
collecting smaller surface nodules through embedded procurement (see Reher, 1991).
Platform surface area (cm2) is even more sensitive to level of reduction at discard than
core volume because it most closely represents remaining working surface.
276
By comparing core dimensions with remnant cortical surface, a clearer
understanding of how volume and surface area relate to reduction and original nodule
size is possible. For example, smaller cores have a higher cortex-to-volume ratio and
may retain more cortical surface at the final stage of reduction due to size constraints. An
assemblage with a low mean core volume should represent the use of small nodules,
rather than intensive reduction, if remnant cortical surface is high. Likewise, a high mean
core volume in combination with low percentages of cortex suggests that large nodules
were being used. Since larger cores can be more extensively reduced prior to discard
they will exhibit extensive reduction in the form of low remnant cortical surface at a
larger size (and with larger platform surface area) than will cores worked from smaller
cobbles. These measurements draw attention to variation between assemblages in
procurement strategies.
The following interpretations can be made based on different combinations of
core volume/platform dimensions and remnant cortical surface:
a.
large core/platform + low cortex = large raw materials used, more
reduction
b. small core/platform + low cortex = small nodules and more reduction,
or larger nodules and very high reduction
c.
large core/platform + high cortex = large raw materials, little
reduction
d.
small core/platform + high cortex = small raw materials, little
reduction
277
Coefficient of variation (CV = δ/μ) can also contribute to a more accurate
assessment of reduction intensity by measuring variability in core/platform size within
the assemblage. The coefficient of variation (CV) is a standardized measure of
variability that takes into account differences in sample size. High variability of core
dimensions in one assemblage can suggest differences in either reduction or original core
sizes, allowing for inferences about consistency in nodule size and reduction intensity. A
low CV indicates low variation in core dimensions. When associated with a low mean
core volume, low CV should reflect original nodule size, particularly in very large
assemblages where highly extensive reduction of all raw materials is unlikely.
Measurements of core volume, relative remnant cortex, and favoured reduction
strategies can be combined to give a more accurate indication of site specific patterns in
the use of tool stone. If informal core types tend to be more expedient, less extensive
reduction is expected. As such, consistently small core volumes are expected for
informal cores only when raw materials are procured in small packages. In situations of
extremely low raw material conservation, we can expect a combination of high core
volume, a predominance of informal cores, and high remnant cortical surface on cores.
At the other end of the spectrum, extremely high levels of raw material conservation
would result in a hypothetical situation of low core volume, formal core types, and low
levels of remnant cortical surface. In both circumstances, a very low CV can be
attributed to consistency in original nodule size. Among cores made on smaller nodules,
high standardization in size is even less likely because the increased regularity of raw
material flaws and human error can make more cores unworkable at an early stage in
278
reduction. For that reason, a high CV and low mean core volume and platform surface
area are most likely related to the consistent use of smaller raw materials.
Core volumes (cm3) and platform surface areas (cm2) were both recorded. Core
volume was measured for each core based on three dimensions. Microblade cores
dominated the collections and were best measured according to height and two
perpendicular platform measurements including the length and width or shortest and
longest measurements of the platform (“P1” and “P2”). Amorphous cores without an
obvious length, width and thickness were measured based on the two longest
perpendicular measurements and the shortest major edge to major edge measurement
(i.e., shortest length excluding measurements of odd protrusions atypical of overall core
shape). Original measurements were taken in mm, but core volume was recorded in cm3
for ease of analysis. Mean core volume was generated for each site assemblage. Results
are summarized in Appendix E.
Retouch
Scraper retouch was also considered. Invasiveness of retouch is compared to other
variables in order to assess the relationship between intensity of use and remnant cortical
surface, and between core volume and overall patterns of raw material conservation.
Scraper retouch and use is conditioned by numerous variables, including situations of
short-term intensive processing, which are not necessarily related to availability of raw
materials or length of site occupation. Periods of short-term intensive processing should
promote retouch rather than replacement, particularly on hafted endscrapers (which make
279
up the majority of the collection) since we expect that it is easier to retouch than rehaft a
small tool. Despite various possible controlling factors, the extent/intensity of scraper
use and retouch complements interpretations of overall use patterns.
By measuring the proportion of total artefact width lost during retouch, we can
essentially compare relative frequencies of retouch prior to discard. Invasiveness of
retouch is calculated based on measurements used in Kuhn’s (1990) reduction index for
unifacial tools (t/T or vertical thickness of flake at termination line of retouch scars
divided by maximum medial thickness of the flake). Since retouch was accomplished
using pressure-flaking rather than hard-hammer percussion, as on the specimens
measured by Kuhn, initial dramatic changes in the depth of retouch scars are expected to
be minimal and progressive retouch episodes more gradual. A contrasting method for
measuring loss of volume during reduction has also been proposed (Eren et al., 2005), but
Kuhn’s method is considered more appropriate due to the small size of typically hafted
endscrapers which dominate this collection and the high remaining mass and volume at
discard in comparison to original size.
4.1.3.2. Results
It was suggested at the introduction to this chapter that during Oasis 3, sometime after
about 5.0 kya, increasing aridity might have resulted in the reduced availability of
resources outside dune-field/wetland environments, encouraging further reliance on
dune-field/wetland resources with the geographic expansion of small field camps and
task sites in order to supplement declining local resources. Such a shift should be
280
evidenced by either longer-term occupation of residential sites or an increase in the
number of residential sites in dune-field/wetland environments. A decline in the
frequency of Residential A sites and a corresponding increase in Residential B sites
appear to have taken place in the Alashan Gobi during Oasis 3. Since the relationship
between Residential A and Residential B sites is based primarily on site size and interassemblage artefact variability, either an expansion of residential sites due to increased
population (i.e., new residential sites established in Oasis 3 would be smaller and less
diverse because they have been less frequently reoccupied since the beginning of Oasis 2)
or an increase in residential mobility (i.e., sites are occupied for shorter periods of time,
also resulting in smaller, less diverse assemblages) could contribute to this result.
Using the methods outlined in the previous section, we can assess whether
Residential A sites were more frequently reoccupied than Residential B sites or whether
they were occupied for longer durations. Lowered residential mobility is frequently
related to a decline in formal technology, a corresponding increase in informal/expedient
technology, and the provisioning of places rather than people. If Residential A sites were
occupied by for longer periods of time, they should show higher ratios of informal to
formal core types. Provisioning of places suggests that raw materials were stockpiled at
frequently used locations. Such provisioning is most directly associated with planned
reoccupation of certain sites, but transporting raw materials to a locale is most likely
when people return frequently to a residential base after regular short-term foraging trips
over a long period of time. Such locales should be characterized by a relatively higher
ratio of informal to formal core types, a somewhat higher % remnant core cortex,
281
relatively larger platform sizes (less intensive reduction), and evidence of greater variety
in original nodule size (suggesting provisioning from different sources).
Residential A and Residential B sites were compared for the entire sample using
the t-test. Significant differences were evident for measurements of core volume,
platform and length, mean % microblade and expedient core types, and mean % of cores
with remnant cortex over 25% (Table 4.12). Comparisons among sites within the GobiAltai region were not significant for any variables. Variation between core volume,
platform, and length, and remnant core cortex were all significant for East Gobi sites,
indicating that cores from Residential B sites were smaller, more heavily reduced, and
tended to have less remnant cortex than those from Residential A sites. Tests for
differences in mean % expedient core types and mean t/T for scrapers had p-values of
0.0598 and 0.0658 respectively. While these values are slightly below statistical
significance, they indicate a 93-94% probability that variation between samples was not
due to random sampling. Residential A type sites appear to have had higher numbers of
expedient core types and more invasive scraper retouch on average. Alashan Gobi sites
also reflect the overall pattern with significant variation for variables of remnant cortical
surface and expedient core types.
282
Variable
% informal cores
% microblade cores
Mean % cores with
remnant cortical
surface (25%+)
Mean % scrapers
with remnant
cortical surface
(25%+)
Mean core volume
(cm3)
Mean core platform
surface area (cm2)
Mean t/T (scrapers)
Res. A mean
Res. B mean
t=
p=
Res. A mean
Res. B mean
t=
p=
Res. A mean
Res. B mean
t=
p=
Res. A mean
Res. B mean
t=
p=
Res. A mean
Res. B mean
t=
p=
Res. A mean
Res. B mean
t=
p=
Res. A mean
Res. B mean
t=
p=
All sites
33.5
15.9
7.19
0.0105
58.1
75.9
5.40
0.0250
30.5
10.7
18.08
<0.0001
18.0
11.3
1.12
0.3314
190.6
97.0
6.40
0.0152
46.7
27.8
9.38
0.0038
0.571
0.515
1.17
0.2848
East Gobi
45.3
25.1
4.19
0.0598
49.9
63.7
1.13
0.3056
33.6
14.0
6.04
0.0276
14.3
6.9
4.20
0.0595
182.6
74.9
5.35
0.0364
44.7
27.5
5.35
0.0364
0.537
0.361
3.98
0.0658
Gobi-Altai
37.8
28.2
0.255
0.6295
55.0
70.5
0.48
0.5073
37.6
19.0
2.65
0.1423
22.3
17.5
0.25
0.6304
243.2
169.6
0.43
0.5295
55.5
39.5
0.69
0.4293
0.530
0.567
0.36
0.5638
Alashan Gobi
16.9
4.9
6.44
0.0228
69.8
85.2
3.87
0.0680
21.8
5.6
7.14
0.0174
19.3
12.1
0.57
0.5776
160.1
91.1
1.41
0.2539
42.3
24.8
3.35
0.0871
0.641
0.640
0.0003
0.9863
Table 4.12 T-test results for Residential A and B lithic data sets. Values for the entire
sample and individual regions are included.
Despite a lack of statistical significance between residential site types in all
regions, there is a consistent pattern: in comparison with Residential A sites, all
Residential B sites exhibit smaller core volumes, smaller platform and length
measurements, lower ratios of informal to formal core types, higher ratios of microblade
to other core types, and less remnant cortex on cores and scrapers. These data indicate
that Residential A sites had better access to raw materials and were either being
provisioned with stone or habitually situated in areas with better access to raw materials.
283
The latter is unlikely since both Residential A and Residential B type sites were
sometimes collected in the same vicinity (within a 5 km radius). A combination of
smaller remaining platform surface area and lower remnant cortex show that cores from
Residential B sites were more extensively reduced than those from Residential A sites.
An emphasis on transportable core types at Residential B sites also suggests strategic
conservation of raw materials and the importance of reliably standardized blanks.
Coefficient of variation (CV) indicates that cores from Residential B sites show
more overall variation in volume (CV = 1.24) and platform surface area (CV = 0.721)
than those from Residential A sites (respectively, CV = 0.653, CV = 0.444). More
uniformity in core sizes in conjunction with more remnant cortex and larger platform
surface area might indicate a combination of both larger original nodule sizes and less
intensive reduction for Residential A sites. Differences in raw material procurement
strategies are also implied, with higher variation in original core size suggesting a wider
range of exploited raw material sources with a more embedded strategy of procurement
associated with Residential B occupations. Somewhat less invasive scraper retouch is
implied for Residential B sites in the East Gobi, but may be related to less intensive
processing activities at the sampled sites. Residential B sites were not simply reoccupied
less frequently, but were probably occupied for shorter periods than Residential A sites.
An increase in the number of Residential B and decline in the number of
Residential A sites in the Alashan Gobi suggests higher residential mobility, as evidenced
by more short-term habitation sites. An increase in population density could also
contribute to such a pattern by necessitating the use of new habitation sites. Having been
284
less intensively used over the millennia, new habitation sites could potentially mimic
short-term residential sites, but implied variation in raw material use and access does
suggest a difference in site function. Residential A sites assigned to Oasis 3 include (see
Figure 1.2 for subregions): Gashun Well (Site 207) (Galbain Gobi subregion); Yingenkhuduk (Site 212), associated with both Oasis 2 and Oasis 3 occupations (Galbain Gobi);
Site 218 (Ukh-tokhoi/Khara Dzag subregion); and Site 223 (Ukh-tokhoi/Khara Dzag),
thought to belong to either late Oasis 2 or early Oasis 3. Mantissar 7 (Site 293) is
classified as Residential B, but could have been classified as Residential based on the
wide range of artefact types. It is the only possible Residential A site in the Gurnai
Depression. Distribution of Residential A sites in the Alashan Gobi suggests that the
Galbain Gobi and the Ukh-tokhoi/Khara Dzag plateaux subregions may have been the
primary centers for longer-term residential base camps during Oasis 3. Residential A
sites were probably also located in the Juyanze subregion during Oasis 3, but probable
Oasis 3 sites were not studied in this analysis. Further analysis of sites in Juyanze,
Galbain Gobi, and the Ukh-tokhoi/Khara Dzag subregions might help clarify patterns in
regional-wide Alashan Gobi land-use.
East Gobi and Gobi-Altai assemblages show no significant shift in the relative
importance of Residential A and Residential B sites after 8.0k cal yr BP, but we can also
consider possible temporal differences in mobility based on implied access to raw
material. The same set of variables applied to evaluation of residential type sites can be
applied to period categorizations. Table 4.13 summarizes statistical results for the entire
sample, as well as each target region. As a unified sample, similar differences are
285
exhibited between Oasis 2 and 3 as were noted for Residential A and B: in comparison
with Oasis 2 sites, Oasis 3 sites exhibit smaller core volumes, smaller platform and length
measurements, lower ratios of informal to formal core types, higher ratios of microblade
to other core types, and less remnant cortex on cores and scrapers. General trends
suggest decreased access to raw material in the Oasis 3 period, corresponding to a
probable increase in residential mobility. However, at the regional level, the differences
between periods are not significant for the East Gobi or the Gobi-Altai and relationships
for each variable are inconsistent. These results do not support the conclusion that there
is a significant difference in raw material conservation in either the East Gobi or the
Gobi-Altai from Oasis 2 to Oasis 3.
The Alashan Gobi is markedly different than the other regions. In this sample, pvalues for variables related to core type, size, and remnant cortex in the Alashan Gobi
region do show significant variation. Microblade cores are relatively more common
during Oasis 3, while the use of informal core types declines. A reduction in core size
during Oasis 3 is best represented by a significant reduction in platform surface area.
When associated with a possible trend towards decreased remnant cortical surface (p =
0.0633), evidence of more intensive core reduction can be proposed. Calculations for CV
further indicate higher variation in core volume (Table 4.14). Together, a higher CV and
low mean core volume suggest the use of smaller nodules during Oasis 3. Degree of
variation in platform surface area between periods is too small to be significant. The use
of smaller nodules might be related to the use of randomly encountered cobbles rather
than quarrying of large blocks, or simply to the transport of smaller pieces to production
286
sites. Knowledge of raw material sources would give us a better indication of the
situation; however, there is a clear difference in raw material use and procurement.
Alashan Gobi data support the hypothesis of increased residential mobility and more
transitory occupation of residential bases during the Oasis 3 period. However, the higher
frequency of Residential B sites that is implied could also result from an increase in the
number of logistical field camps and tasks sites associated with each residential base,
rather than higher residential mobility. Regardless of the circumstances, the data imply
more conservative use of tool stone in Oasis 3.
Data on invasiveness of retouch and remnant cortical surface for scrapers do not
contribute to interpretations of collection-wide patterns in reduction strategies. The
results are insignificant and variable. Scraper data are expected to yield more useable
results at the level of individual sites they do for regional comparisons. High variation in
the intensity of retouch and the use of primary flakes between sites, particularly smaller
task sites, is likely controlled more by specific site function and incidental availability of
blanks than by region-wide trends in raw material use.
287
Variable
% informal cores
% microblade cores
Mean % cores with
remnant cortical
surface (25%+)
Mean % scrapers
with remnant
cortical surface
(25%+)
Mean core volume
(cm3)
Mean core platform
surface area (cm2)
Mean t/T (scrapers)
Oasis 2 mean
Oasis 3 mean
t=
p=
Oasis 2 mean
Oasis 3 mean
t=
p=
Oasis 2 mean
Oasis 3 mean
t=
p=
Oasis 2 mean
Oasis 3 mean
t=
p=
Oasis 2 mean
Oasis 3 mean
t=
p=
Oasis 2 mean
Oasis 3 mean
t=
p=
Oasis 2 mean
Oasis 3 mean
t=
p=
All sites
34.4
21.9
3.01
0.0877
51.6
71.5
6.75
0.0117
23.6
21.2
0.18
0.6686
17.3
15.0
0.23
0.6362
196.6
142.7
1.41
0.2400
475.8
343.1
3.59
0.0630
0.502
0.529
0.24
0.6247
East Gobi
37.8
44.0
0.27
0.6065
47.3
51.5
0.09
0.7708
22.4
33.1
1.42
0.2504
13.1
10.4
0.34
0.5681
132.7
165.4
0.41
0.5296
37.8
42.8
0.35
0.5593
0.421
0.463
0.20
0.6590
Gobi-Altai
47.2
37.0
0.34
0.5708
45.9
60.8
1.82
0.3765
27.9
35.6
0.32
0.5796
25.7
10.6
3.11
0.1034
267.1
271.3
0.001
0.9741
57.7
50.4
0.20
0.66
0.447
0.523
0.40
0.5383
Alashan Gobi
17.6
4.8
9.74
0.0045
63.6
87.9
10.66
0.0033
20.6
8.9
3.79
0.0633
13.7
19.8
0.34
0.5656
198.1
80.0
3.96
0.0580
48.4
23.3
4.99
0.0350
0.667
0.565
1.65
0.2120
Table 4.13 T-test results for Oasis 2 and Oasis 3 lithic data sets. Values for the entire
sample and individual regions are included.
Variable
Core volume
Platform surface
area
Period
Oasis 2
Oasis 3
Oasis 2
Oasis 3
CV
0.988
1.353
0.813
0.797
Table 4.14 Coefficient of variation for core volumes and platform surface areas of Oasis
2 and Oasis 3 assemblages from the Alashan Gobi.
288
4.2. Discussion
Three hypotheses were detailed at the beginning of this chapter. The first hypothesis is
that Oasis 1 foragers were organized in a circulating pattern of mobility and exploited a
wide range of habitats, including the occasional use of dune-field environments. The
sample size is very small for Oasis 1 sites; only two sites characteristic of a residential
type habitation (based primarily on evidence of cooking) are currently recognized during
this period – Shara KataWell and Chikhen Agui cave. Both were found in mountainous
environments, suggesting a preference for upland environments. This is further
supported by site distribution according to ecozone (Table 4.7). The lack of Residential
A type sites supports the hypothesis that Oasis 1 groups showed a higher level of
residential mobility compared with later periods, but the study of additional Oasis 1 sites
should be conducted to further support this hypothesis. The short-term exploitation of
dune-field/wetland environments is indicated in both the Gobi-Altai and the East Gobi
(Figure 4.6 and Table 4.8). There is some evidence from the Gobi-Altai of early
Epipalaeolithic task sites in dune-field/wetland environments, but the site was overlain
with sand and may pre-date establishment of the dune-fields that were exploited in later
periods. It is expected that additional data will reveal more task sites dating to the end of
the Epipalaeolithic and corresponding with Oasis 1.
A climatic optimum, characterized by high effective moisture, in the early
Holocene would have created an extremely rich ecological niche around recently
stabilized dune-fields. It is hypothesized that the beginning of Oasis 2 coincides with this
ecological shift. Longer term seasonal base camps centred on dune-field/wetland
289
environments were hypothesized, and thought to have been complemented by pattern of
radiating field camps and task sites. Distribution of sampled sites clearly supports the
oft-mentioned claim in existing literature that the occupation of dune-field/wetland
environments was central to hunter-gatherer organizational strategies by this time.
Oasis 2 marks the beginning of a definitive phase in hunter-gatherer
organizational strategies with the establishment of the longer-term multipurpose
Residential A sites, mostly situated in dune-field/wetland environments (Figure 4.7 and
Table 4.9). In the Alashan Gobi, longer-term base camps are not restricted to lowland
dune-field/wetland localities, but are also present in drier lowland plains (one site in the
Black Gobi, west of Juyanze) and among the higher elevation wetlands of the Ukh-tokhoi
plateau. Distribution of high quality tool stone and local hydrology probably contribute
to this situation. Additional shorter term multipurpose Residential B sites were identified
throughout the same low-lying wetland habitats. Task sites were distributed across the
remaining ecozones, from which additional resources would have supplemented those
available at residential locales (Table 4.10).
Residential A sites appear to have been provisioned with raw materials and
inhabited for longer (Table 4.12). The occurrence of large, heavy formal grinding
technology in the East Gobi further suggests that Residential A sites in that region were
provisioned for planned reoccupation during seasons when resources such as tubers and
seeds could have been gathered, dried, and ground for future use. Such provisioning is
frequently associated with locations where heavily utilized foods or other resources are
more abundant (Kuhn, 1995: 22).
290
Late middle Holocene climatic deterioration may have resulted in the contraction
of lakes and wetlands, as well as the desiccation of early Holocene steppe and desertsteppe environments (see Chapter 5). It is hypothesized that Oasis 3 hunter-gatherers
responded by intensifying the seasonal use of dune-field/wetland environments, which
were oases of primary productivity. At the same time, they more often would have
supplemented their diets from a wider range of environments.
Various data support this hypothesis. In the East Gobi, the distribution of Oasis 3
residential sites is consistent with such a shift, including evidence for an increased focus
on the use of upland environments compared with Oasis 2 (Table 4.10). Increased
residential mobility in the Alashan Gobi during Oasis 3 is also supported by differences
in the distribution of Residential A and Residential B sites (Table 4.10) along with
changes in raw material curation practices that includes higher relative frequencies of
formal core types and more intensive reduction of cores prior to discard during Oasis 3
(Table 4.13). A greater reliance on formalized core technology includes the increased
use of highly portable and flexible wedge-shaped core technology during Oasis 3 (see
Chapter 3). The use of dune-field/wetland environments did not decline, but mobility
may have increased either through more frequent moves or by an intensified radiation of
inhabitants into short-term field camps and task sites. This suggests less group
aggregation with a pattern of dune-field/wetland habitation that is less concentrated.
Such a shift in land-use is consistent with an avoidance of over-utilization that might be
tied to lowered productivity. A higher frequency of residential sites in dunefield/wetland habitats during Oasis 3 suggests the creation of additional sites through
291
more frequent moves. It is not clear if this trend is related to an increase in the number of
radiating field camps associated with longer-term Residential A sites, or to fewer
instances of aggregation at larger residential sites in favour of overall higher residential
mobility.
The hypothesis that smaller field camps and task sites were used more extensively
during Oasis 3 is not supported by data from lithic analysis in the East Gobi or the GobiAltai regions, but is supported for the Alashan Gobi. Based on differences in the
distribution of Oasis 2 and Oasis 3 sites, it is probable the East Gobi groups extended the
range of residential camps into upland environments during the Oasis 3 phase.
Additional attention to the distribution of Neolithic/Eneolithic sites in the East Gobi and
Gobi-Altai regions is required in order to assess possible differences in land-use over
time. A larger, more geographically extensive sample of archaeological sites is most
essential in the Gobi-Altai region.
Finally, it is hypothesized that hunter-gatherers began to adopt herd animals
towards the end of Oasis 3 in order to make up for declining productivity in dunefield/wetland environments. The adoption of domesticates is proposed to have
culminated under conditions of relatively extreme desiccation after 3.0k cal yr BP. The
increased exploitation of upland and plains environments is consistent with pasturage
needs and typical of later nomadic pastoralist land-use (Fernandez‐Gimenez, 1999), but is
not definitive. Ceramic spindle whorls in some East Gobi and Alashan Gobi sites are
similarly both suggestive and inconclusive.
292
At the same time, it is clear that by about 3.2-3.0k cal yr BP, nomadic pastoralism
had become established across much of Mongolia. Agropastoralists along the southern
border of the Gobi Desert were long-since acquainted with domesticated herd animals.
Painted pottery from Gurnai Depression sites show stylistic similarities with pottery
belonging to agriculturalists and/or agropastoralists farther east. Although the sample
size is limited and the shards are very small, black-on-red painted shards bear possible
affinities to Majiayao or Qijia (Figure 3.7, Appendix B). A relationship between
neighbouring agropastoralist groups and Gobi Desert hunter-gatherers is particularly
persuasive considering the stylistic similarities evidenced in these pottery traditions. The
apparent termination of a long-standing dune-field/wetland adaptation after 3.0k cal yr
BP further supports the assertion that these ephemeral traces of pastoralist adaptations do
indeed represent the first stirrings of a production economy within the still strong bounds
of hunter-gatherer subsistence.
Data from lithic analysis supports many of the original hypotheses and provides
additional insight into organizational strategies associated with different types of
residential sites. Due to the small sample size, hypotheses about Oasis 1 land-use can be
neither supported nor negated. During Oasis 2 longer-term multipurpose sites appeared
in the lowland dune-field/wetland environments. Cores from such sites, categorized as
“Residential A,” show a tendancy towards less intensive reduction and the locales were
probably provisioned with raw materials in anticipation of regular reoccupation. In
contrast, smaller multipurpose residential sites, categorized as “Residential B,” were
likely inhabited for shorter durations and core data suggests that they are associated with
293
more intensive raw material conservation, and perhaps more embedded procurement
strategies. An increase in the number of Residential B and decline in the number of
Residential A sites in the Alashan Gobi suggests higher residential mobility and more
transitory occupation of residential bases in that region during Oasis 3. Although there is
a clear difference in lithic reduction strategies between Residential A and Residential B
sites, the hypothesized shift in land-use during Oasis 3 is not supported by data in any
region other than the Alashan Gobi. Variation in site distribution for the East Gobi does
support the possibility of increased logistical mobility through an expansion of residential
sites into new ecozones, possibly in response to declining dune-field/wetland resources.
In the following chapter, a synthesis of recorded palaeoenvironmental data offers a unit
of comparison for Holocene organizational strategies across the Gobi Desert.
294
CHAPTER 5 – PALAEOENVIRONMENTAL CONTEXT
Climate in the Gobi Desert is largely controlled by an interplay in the relative
strengths of the Indian and East Asian Summer Monsoon systems, the SiberianMongolian high pressure system (which is the source of the East Asian winter monsoon),
and the Westerlies. When the Siberian-Mongolian high dominates, beginning at the end
of August and continuing until April, strong Westerlies force the East Asian summer and
winter monsoons south, blanketing the region with cold dry air. When this high pressure
system weakens in the summer, strong summer monsoons bring warm wet air into the
region. The strength of the summer monsoon is related to heat transport over an area
stretching from the southern Indian Ocean to the Tibetan Plateau (Winkler and Wang,
1993). Numerous recent palaeoenvironmental studies in this region permit better
consideration of human ecology following the end of the Last Glacial Maximum (LGM).
Palaeoenvironmental studies allow us to infer a great deal about the nature and
timing of climatic shifts, but can be contradictory and otherwise problematic. The recent
wealth of palaeoclimatic studies in northeast Asia has drawn considerable attention to
inconsistencies at both the local and regional level (An et al., 2006; Zhao et al., 2009).
Proxy data indicate a much different timeline for climatic amelioration in the East Gobi
than either the Gobi-Altai or the Alashan Gobi. Due to the influence of the mid-latitude
Westerlies, climate in the latter regions is more directly controlled by topography and
surface temperatures in the North Atlantic Ocean and western Eurasia than by conditions
in southern Asia (Chen et al., 2008). Furthermore, large-scale reconstructions aimed at
interpreting general climate trends across Central or East Asia can give some insight into
295
the timing of periodic shifts in controlling mechanisms like the summer and winter
monsoon systems, but extreme variability in elevation, topography, and soil type result in
each climatic event being expressed differently from region to region. For example, a
strengthening in summer monsoons after the LGM would have positively affected many
semi-arid regions, but irregular precipitation could have stimulated enhanced evaporation
in the extremely arid desert regions like the Alashan Gobi, resulting continued or
increasing aridity until precipitation stabilized (see Broccoli and Manabe, 1992;
Herzschuh, 2006).
Even at the more local level, contradictory results and interpretational
disagreements can occur. Various proxy measures such as pollen, sediment grain-size,
and depositional sequences sometimes yield differing results, and incongruous
interpretations of “relative” humidity and aridity contribute to the confusion. Pollen
studies allow for a more detailed understanding of vegetation which can be used to infer
ecological context of relevant landscapes, but can be difficult to interpret due to the lack
of reliable modern environmental comparisons and differential transport of certain pollen
types. As noted, variability in the relative influence of the East Asian summer and winter
monsoons, and various high and low pressure systems across the Gobi Desert play a role
at the larger-scale, while local vegetation, elevation, groundwater and fluvial hydrology
further influence the effects precipitation and temperature (An et al., 2006; Zhao et al.,
2009). Further limitations exist in this particular study due to a lack of highly localized
data for most key locales.
296
Despite these limitations, the abundance of recent high resolution
palaeoenvironmental studies across Northeast Asia gives some important information
about the nature and timing of important climatic events. In combination with more
detailed studies of how local environments respond to changes in precipitation,
temperature and other climate influenced events, we can infer a great deal about
resources that hunter-gatherers might have exploited or the suitability of certain
environments during periods of occupation. The range of methods used for
palaeoenvironmental reconstruction also contributes to a more nuanced recognition of
environmental conditions. The goal of this chapter is to avoid deterministic causal
interpretations by synthesizing regional palaeoenvironmental records in order to
understand more complex aspects of the relationship between human land-use and
climate change.
5.1. Palaeoenvironmental chronology of the Late Pleistocene and Early Holocene in
Northeast Asia
Late Pleistocene climates in the Gobi Desert and across adjacent regions can be described
in a very general way that relates to the intensification and weakening of the Asian
summer monsoon systems and gives an overall picture of climate change dynamics. Four
broad periods are relevant to understanding the regional evolution of climate and include:
the middle and late stages of Marine Isotope Stage (MIS) 3 (43.0-25.0ka); early to middle
MIS 2 (25.0-19.0ka); middle to late MIS 2 (19.0-11.5ka); and the early to late middle
stages of MIS 1 (11.5-3.0ka).
297
5.1.1. Middle to late MIS 3 (43.0-25.0k cal yr BP)
During middle and late MIS 3, the climate reached a condition of high effective moisture
(Herzschuh, 2006), witnessed by several periods of palaeosol formation and the infilling
of massive palaeolake basins (Shi et al., 2001; Grunert and Lehmkuhl, 2004; Feng et al.,
2007). Desert lakes persisted with only brief periods of lower water, and the Gobi was
reduced by north and south extension of steppe forest and steppe (Feng et al., 2007). To
the south, on the Tibetan Plateau, it is estimated that mean annual temperature was 2-4oC
higher than at present (Shi et al., 2001). Increased humidity in the north is evidenced by
contraction of desert expanse at around 34.0, 31.0, 29.0 and 25.0k yr BP (39.5, 35.0,
33.5, 30.0k cal yr BP), and in the south at 50.0-40.0, 37.0-32.0, and 29.0-23.0k yr BP
(54.0-44.0, 42.0-36.0, and 33.5-27.5k cal yr BP) (Feng et al., 2007). This shift in
effective moisture is probably due to significant increases in precipitation during the
summer. The northward extension of the summer monsoons would probably have greatly
exceeded conditions in the middle Holocene, as suggested by the existence of large lakes
in the Tengger and Badain Jaran Deserts (Shi et al., 2001). As we tend to perceive ideal
environments for human habitation in more arid environments in terms of higher biotic
productivity, the middle to late MIS 3 likely witnessed the most optimal environmental
conditions during the entire period that anatomically modern humans occupied North
Asia.
298
5.1.2. Early to middle MIS 2 (25.0-19.0k cal yr BP)
Mean moisture values decreased noticeably due to a strong intensification of the winter
monsoon and a weakened summer monsoon throughout much of MIS 2 (~25.0-19.0k cal
yr BP) (Herzschuh, 2006). A period of increased humidity at the end of MIS 1 and the
beginning of MIS 2 allowed the northward expansion of semi-arid conditions between
25.0-21.0k yr BP (30.0-25.0k cal yr BP) and a southward expansion between 29.0-23.0k
yr BP (33.5-27.5k cal yr BP) (Feng et al., 2007). As summer monsoons retreated during
a period of reduced temperatures, an overall moisture minimum known as the Last
Glacial Maximum (LGM) caused a decline in lake levels and an increase in aeolian
deposition (Grunert and Lehmkuhl, 2004; Wünnemann et al., 2007). Many Asian lakes
dried out and hyperarid areas reached their maximal extent between 21.3-19.0k cal yr BP
(Pachur, et al. 1995; Lehmkuhl and Haselein, 2000; Herzschuh, 2006; Feng et al., 2007).
Desert vegetation shifted south to the Loess Plateau and eastward between 32o N and 40o
N to the coast line, while taiga extended far south into former desert steppe (Yu et al.,
2000; Feng et al., 2007). This general trend may have been expressed somewhat
differently across the Gobi Desert depending on pre-existing conditions, as there is
evidence that lower rates of evaporation due to lower temperatures actually allowed for
stable or increased effective moisture (Komatsu et al., 2001; Liu et al., 2002b; Herzschuh,
2006).
299
5.1.3. Middle to late MIS 2 (19.0-11.5k cal yr BP)
The immediate post-LGM period can be broadly characterized as one of gradual climatic
amelioration. Between 19.8-17.2k cal yr BP mean moisture values gradually increased
(Herzschuh, 2006). Lake levels were medium-to-high as steppe and forest-steppe
expanded (Wünnemann et al., 2007; Herzschuh and Liu, 2007). A phase of stable and
slightly wetter conditions occurred in Asia between 18.5-17.0k cal yr BP, due either to
glacial melt or the onset of a summer monsoon circulation after the LGM. Moisture
values for monsoonal Central Asia gradually increased until 17.2k cal yr BP, with stable
to somewhat drier values until 15.4k cal yr BP (Herzschuh, 2006). Amelioration is
indicated by the dramatic retreat of the northern Gobi Desert boundary between 19.016.0k cal yr BP (16.0-13.0k yr BP). Intense aridification immediately followed with a reexpansion of desert areas to their maximal extent between 16.0-9.5k cal yr BP (13.0-8.6k
yr BP). Aeolian deposition extended at least as far north as 55o N and south to 33o N
(Feng et al., 2007). Extreme low moisture availability, corresponding to the cold
Younger Dryas event, occurred at 13.0-11.6k cal yr BP across monsoonal Central Asia
and 12-10 ka on the Loess Plateau (Madsen et al., 1998; Herzschuh, 2006).
5.1.4. Early to late middle MIS 1 (11.5-3.0k cal yr BP)
Early to late middle MIS 1 is the period most directly relevant to this study. Following
the cold/dry Younger Dryas, the trend towards warmer and wetter conditions continued.
A strong intensification of both the Indian and East Asian monsoons occurred at the
Pleistocene/Holocene transition ~11.5k cal yr BP, and conditions became significantly
300
wetter in monsoon-influenced regions at this time (Herzschuh, 2006). Corresponding
periods of aridity in regions influenced by the Westerlies may have resulted from
enhanced subsidence of air masses to the lowland areas (Broccoli and Manabe, 1992;
Herzschuh, 2006). Although annual average temperatures and precipitation had both
increased by 9.6k cal yr BP (9.0k yr BP), orbital changes acting on summer and winter
monsoon systems caused an increase in seasonality, with colder winters and warmer
summers, that would have limited the expansion of some species. Seasonality was
temporarily reduced after 9.6k cal yr BP (9.0k yr BP) and even more so after 6.8k cal yr
BP (6.0k yr BP). Heightened seasonality returned after about 3.2k cal yr BP (3.0k yr BP)
when the strength of the winter monsoon declined (Kutzbach, 1981; Winkler and Wang,
1993).
The retreat and advance of arid lands reflects these shifts in weather systems. In
regions less influenced by the monsoonal systems than the Westerlies, the northern
boundary of arid lands retreated most dramatically between 9.6-7.8k cal yr BP (8.6-7.0k
yr BP). In monsoon-influenced regions, the southern boundary of arid lands retreated
northwards between 10.1-4.5k cal yr BP (9.0-4.0k yr BP) due to an increase in
monsoonal precipitation (Starkel, 1998; Feng et al., 2007). Altitudinal zones also reflect
the shift in climate-mediated ecotones as indicated by changes in the upper and lower tree
lines and permafrost limits (Starkel, 1998). During the early Holocene, hyperarid and
arid zones retreated to an area that extended maximally between 38o N and 48o N (Feng
et al., 2007).
301
Divergence in the palaeoclimatic records of separate Gobi Desert regions
influenced by different circulation systems is especially notable during the Holocene
(Table 5.1). Data from the East Gobi, influenced by the East Asian Monsoon system,
indicate wet conditions from about 11.5-1.7k cal yr BP, with optimal moisture between
8.3-5.5k cal yr BP (Herzschuh, 2006). In stark contrast, the Gobi-Altai and Alashan
Gobi, areas under the influence of the Westerlies do not show a pronounced a moisture
maximum, but rather more constant values between 12.1-2.7k cal yr BP with a much
more arid climate prior to 8.0k cal yr BP and a short maximum between 7.5-6.8k cal yr
BP (Herzschuh, 2006; Chen et al., 2008). The end of optimal conditions in the GobiAltai is signalled by the expansion of permafrost and a decline in precipitation between
5.8-4.5k cal yr BP (5.0-4.0k yr BP) (Starkel, 1998), and in the Alashan Gobi by an
overall decrease in mean effective moisture after ~3.0k cal yr BP (Herzschuh, 2006).
5.2. Regional Variability
5.2.1. East Gobi
The East Gobi Desert region, for the purposes of this study encompasses an area between
41o 00’ - 44o 00’ N and 108 o 00’ - 115 o00’ E. Situated primarily within the Nei Mongol
or Inner Mongolia Autonomous Region of the People’s Republic of China (PRC), the
region extends from south to north from just above the northernmost bend in the Yellow
River into the southeastern corner of Mongolia, and from the northernmost bend of the
Yellow River in the west to the Hunshandake Sandy Land in the east (see Figure 5.3). Of
302
the three regions discussed, the East Gobi climate is the least continential and arid, being
more heavily influenced by the East Asian monsoon system. In recent times annual
average precipitation has been approximately 200-400 mm. The climate is considered
middle-temperate sub-dry, a designation that applies to an area extending southwest
across the Ordos Plateau into the southwestern reaches of the Yellow River (Winkler and
Wang, 1993). Today, the northern limit of the summer monsoon divides the target region
diagonally from southeast to northwest.
Most palaeoenvironmental studies have centred on the transitional zone between
semi-humid and semi-arid environments southeast of the study area, where shifts in the
extent of monsoonal precipitation are most noticeable. Here, monsoonal precipitation is a
controlling factor in Holocene moisture availability, which in turn helps to regulate
vegetative cover. According to proxy data from chronostratigraphic biologic and
geomorphic research, the northern extent of the summer monsoon has changed
dramatically throughout the Holocene. At about 10.0k cal yr BP (9.0k yr BP), the
summer monsoon reached no farther than the southernmost boundary, but by 6.8k cal yr
BP (6.0k yr BP), in concert with a northward shift of the Westerlies, it extended into
Mongolia as far as 44o N (Rea and Leinen, 1988; Winkler and Wang, 1993: 224, Figure
10.3.b).
This progressive northerly shift in monsoonal precipitation during post-LGM
warming would have played an important role in Holocene environmental change.
According to studies of sediment cores derived from a series of lakes lying between 40o
30’- 43o 00’ N, 112o 30’-117o 00’ E, temperatures were much cooler and annual
303
precipitation much lower than modern averages prior to 16.5k cal yr BP (13.5k yr BP).
Indicators of climatic amelioration after 16.5k cal yr BP (13.5k yr BP) include a
reduction in aeolian activity (Wang et al., 2001). Slight increases in humidity and
temperature are seen to have led to a decrease in Chenopodiaceae pollen after about
14.0k cal yr BP (12.0k yr BP), although humidity increased again about 12.4k cal yr BP
(10.5k yr BP) (Wang et al., 2001; Liu et al., 2002a; Wang et al., 2010), when the climate
was somewhat warmer and wetter than today (Shi and Song, 2003). Pollen studies show
that prior to the Holocene, Betula (birch), Chenopodiaceae, Artemisia and Ephedra
dominated, suggesting a woodland-steppe mosaic (Wang et al., 2001; Wang et al., 2010).
Palynological records from Diaojiao Lake (41° 18’ N, 112° 21’ E), at the northern foot of
the Daqingshan Mts., indicate that between 10.4-8.8k cal yr BP (9.2-7.9k yr BP) desertsteppe vegetation dominated an environment that was once rich arboreal species (Shi and
Song, 2003). This brief decrease in arboreal species due to lower temperatures and
increased aridity would have interrupted earlier Holocene amelioration.
Temperatures appear to have begun rising around 8.9k cal yr BP (8.0k yr BP).
There were corresponding increases in Picea (spruce) and Ulmus (elm) (Liu et al., 2002a;
Shi and Song, 2003; Peng et al., 2005; Wang et al., 2010; but see Jiang et al., 2006). At
Lake Bayanchagan (41° 38' N, 115° 12' E, 1355 m a.s.l.), this climatic amelioration was
followed by decreases in the length of growing season. Decreases in mean temperature
of the coldest month, mean temperature of the warmest month, and mean annual
temperature around 8.0k cal yr BP (ca. 7.5k yr BP) were also noted (Jiang et al., 2006).
304
Figure 5.1 Map of Gobi Desert showing palaeoenvironmental locales and study regions.
The Holocene Climatic Optimum occurred at different times after 8.9k cal yr BP
(8.0k yr BP) and is recognized by an increase in forest species pollen and decreased
representation of steppe species (Jiang et al., 2006). Increases in Picea (spruce), Pinus
(pine) and Quercus (oak) are attested to in various locales before 7.8k cal yr BP (7.0k yr
BP) (Wang et al., 2001; Wang et al., 2010) and there is evidence of an optimal forest
zone with a more mosaic forest-steppe vegetation in the immediate vicinity of Daihai
Lake (40° 35’ N, 112° 40’ E) at 7.9-6.5k cal yr BP (7.5-5.5k yr BP). Pollen data from
Daihai Lake indicate that arid herbs and shrubs dominated the lake basin during the early
Holocene, but from 7.9k cal yr BP to 4.45k cal yr BP large scale covers of mixed
coniferous and broadleaved forests marked a warm and humid climate in the lake area.
Before the end of the Holocene Climatic Optimum woodlands and woodland-steppes
305
formed with Quercus (oak), and Pinus (pine) at higher elevations, replacing arid steppe
and Betula (birch) -dominated woodland or woodland-steppe in some regions (Wang et
al., 2001; Liu et al., 2002a; Peng at al., 2004; Jiang et al., 2006).
This trend toward increased humidity throughout the semi-humid/semi-arid
transitional zone can be generally seen as peaking around 7.7k cal yr BP (6.9k yr BP), but
there is some variation between studied locations (Wang et al., 2001; Peng et al., 2005;
Jiang et al., 2006). At Anguli-nuur (41° 18-24’ N, 114° 20-27’ E, ~1315 m a.s.l. – Wang
et al., 2010) the climatic optimum occurred much earlier than 7.7k cal yr BP, but at the
more westerly locale of Diaojiao it occurred later (Shi and Song, 2003). The period of
maximum humidity lasted for approximately 1,000 years and was followed by a steady
decline across the entire region, marked by increases in aeolian activity (Wang et al.,
2001).
The decline in overall vegetation, particularly deciduous and coniferous trees,
continued after the Holocene Climatic Optimum as steppe vegetation became
increasingly dominant. At Lake Bayanchagan mosaic forest-steppe vegetation gave way
to a steppe-dominated environment with small forest patches along mountains and river
valleys between 6.5-5.1k cal yr BP (5.5-4.3k yr BP) (Jiang et al., 2006). Decreases in
humidity and ensuing environmental degradation appear to have occurred later in the
south, where climatic conditions may have peaked between 5.0-4.8k cal yr BP (Peng et
al., 2005), and at Diaojiao Lake, where drainage from nearby mountain ranges might
have prolonged the effects of climatic amelioration (Shi and Song, 2003). Palaeosol
formations in the Hulun Buir (Kölün Buyir) (48o N, 120o E) and Hunshandake sandy
306
lands (42-44o N, 112-118o E), estimated by archaeologists to date to about 5.0-3.0 ka,
may be associated with the same climatic event (Winkler and Wang, 1993; Yang et al.,
2008). Post-optimum increases in aridity and declines in temperature are characterised
first by the declining representation of deciduous trees in favour of coniferous species
(Jiang et al., 2006; Wang et al., 2010), and by the formation of sparse coniferous tree
woodland-steppe in the mountains (Shi and Song, 2003).
Increasing regional variation in climatic optima and ensuing aridity by the midHolocene may be related to local vegetational and hydrological feedback. After 4.5k cal
yr BP (4.0k yr BP) pollen profiles at Lake Bayanchagan indicate a notable rise in the
frequency of desert species (Jiang et al., 2006). This transition appears to have been
prolonged at Diaojiao Lake, where drainage from the mountains may have buffered the
effects of decreased precipitation until after 3.2k cal yr BP (3.0k yr BP) (Shi and Song,
2003). Beginning around 2.9k cal yr BP, there is a decline in representation of woody
plants at Daihai Lake, suggesting a transition from warm/humid to cold/dry climatic
conditions (Peng et al., 2005).
In summary, the Pleistocene-Holocene transition in the East Gobi was marked by
an increase in humidity and temperature by 12.4k cal yr BP (10.5k yr BP) that was
followed by a period of optimal climatic conditions between 8.9-6.3k cal yr BP (8.0-5.5k
yr BP), peaking at about 7.7k cal yr BP (6.9k yr BP). Increasing arboreal pollen in the
southeastern reaches of the study region indicate that the early Holocene was typified by
a shift in both the composition of tree species and the gradual expansion in the extent of
woodland environments. The Holocene Climatic Optimum was typified by a peak in the
307
extent of woodlands, with temperate coniferous and broadleaved forests extending across
the East Gobi (see also Figure 10.11., Winkler and Wang, 1993) and some loss of open
steppe environments, though forests dominated only in high elevations and river valleys.
After 6.3k cal yr BP (5.5k yr BP) there was a decline in arboreal species at lower
elevations corresponding to the expansion of steppe environments. Increased aridity
marked by increased aeolian activity and the prevalence of desert species in pollen
profiles differed regionally according to local hydrology between 4.5-2.9k cal yr BP (4.02.8k yr BP).
While these data give a good indication of climatic conditions along the
southeastern fringes of the study area, they lie farther south or east of the actual
archaeological sites studied here. As such, it might be expected that these localities,
which are now located in much more arid environments, experienced slightly later
climatic and environmental amelioration in accordance with the lag in monsoonal
migration. Optically stimulated luminescence (OSL) dates on palaeosol and sand
sequences in two northeastern desert regions of Inner Mongolia, Hunshandake (~43° N,
116 o E) and Hulun Buir Deserts (~49° N, 118 o E) indicate a shift from dune formation to
sand stabilization that began about 10.0 ka and continued until about 3.6 ka (Li et al.,
2002), with optimal conditions (warm/wet) from about 5.0 to 3.0 ka (Yang et al., 2008).
In both regions climatic amelioration appears to be broadly synchronous with
developments in the semi-humid/semi-arid transitional zone. Vegetation composition in
these arid regions would have differed significantly from that in the more verdant
southeastern region, which falls along the modern boundary of semi-arid to arid
308
environments. Corresponding conditions of heightened evapotranspiration and albedo
feedback may have resulted in a less dramatic moderation of temperatures and
groundwater retention, probably decreasing the significance of arboreal species
(Ganopoloski et al., 1998; An et al., 2006; Zhao et al., 2009).
Conditions may have been similar to those across the comparably arid Loess
Plateau, south of the Yellow River. Here the warm/humid Holocene was evidenced by
semi-stabilization of dunes, the formation of steppe and forest-steppe environments and
of lakes or swamps in interdune depressions (Yang et al., 2004). The distribution of
pollen data from the northwestern Loess Plateau (~35o 00’ - 38o 30’ N, 104o 00’ – 109o
30’ E) indicate that at sites in river valleys and terraces the most warm/humid periods
were characterized by forest, forest-steppe, or steppe with sparse trees. Other sites
showed only sparse forest-steppe or humid steppe with some tree pollen, suggesting that
elevation and hydrology play a major role in vegetation composition (Zhao et al., 2009).
The occurrence or absence of forest-steppe environments is essential in understanding
human land-use, as an increase in arboreal species would have dramatically altered the
biotic landscape. We might infer that during the Holocene Climatic Optimum, the
archaeological study area was typified by steppe with forest-steppe mosaic along upland
plateaux and river valleys, while mixed forests may have dominated mountain ranges.
Quercus pollen from Hulun Bair at about 5.0-3.0 ka indicates the presence of nut-bearing
hardwood species in this northeastern desert region.
According to dates from East Gobi archaeological sites (see Chapter 3), it is
probable that intensive use of dune-field/wetland environments may have begun around
309
8.0 kya and ending by 3.0 kya. Such archaeological sites, frequently associated with sand
deposits and lake basins, belonged to hunter-gatherers described as “dune-dwellers” by
Central Asiatic Expeditions members when they were discovered in 1925 (Nelson,
1926a). This period of dune-field/wetland habitation is contemporaneous with palaeosol
formation and the expansion of steppe and forest-steppe in what were previously and are
now desert environments. According to comparative data from neighbouring locales, this
period of increased precipitation and warmer temperatures provided an environment of
many small lakes and swamps in basins and interdunal depressions, expanded steppe, and
mosaic mixed (coniferous and broadleaved) forest-steppe habitats along river valleys,
mesas and upland plateaux. Based on dates for the expansion of forest species farther
south, it is probable that the longer-term habitation in lowland dune-field/wetland habitats
did not commense until forest stands comprised of species such as oak and pine were
well-established.
The distribution of East Gobi archaeological sites dating to the terminal
Pleistocene and early Holocene indicates that habitation shifted from a focus on higher
elevations (> 1200 m a.s.l.) and river/streams during Oasis 1 (~13.5-8.0 kya), to the use
of low (< 1200 m a.s.l.) and mid-altitudinal habitats (1201-1400 m a.s.l.) near
lake/wetlands during Oasis 2 (~8.0-5.0 kya) (see Figure 4.2 and Table 4.4). Residential
sites associated with Oasis 2 and Oasis 3 centred on lowland dune-field/wetland habitats
(Table 4.11). They indicate an intensive type of planned longer term habitation that is
not characteristic of earlier periods. Furthermore, formal grinding stones are typical of
Oasis 2 sites and represent a sort of specialized non-portable processing equipment that is
310
unique in Gobi Desert prehistory. Considering the types of resources available in dunefield/wetland habitats, such technology is expected to have been used for seeds and
tubers, but the pollen data summarized above suggest that nut-bearing species such as oak
and pine might also have been available. Forest expansion might also be represented by
the development of adze/axe technology, frequently associated with woodworking,
during Oasis 2. The shift in the focus of grinding technology towards less formal types
during Oasis 3 suggests a change in diet or processing practices, perhaps related to a
declining representation of deciduous nut-bearing trees such as oak (see Jiang et al.,
2006; Wang et al., 2010). The decline in arboreal species, particularly deciduous types,
would have been most pronounced at the end of Oasis 3, as indicated by decreasing
representation in pollen profiles and the corresponding increase of desert species between
4.5-2.9k cal yr BP (Jiang et al., 2006; Shi and Song, 2003; Peng et al., 2005).
5.2.2. Gobi-Altai
The region referred to in this study as the Gobi-Altai encompasses an area extending
from 43o 00’ – 46o 00’ N, 98o 00’ - 105o 00’ E and is part of the “arid Central Asia”
region dominated by the Westerlies (Chen et al., 2008). Here precipitation depends on
the availability of water vapour transported by the Westerlies from the North Atlantic
Ocean, and inland seas and lakes along the mid-latitude cyclonic storm paths (Böhner,
2006). Analysis of dust storm and blowing sand records from over 680 stations in China
and Mongolia, collected between the years 1961 to 2000, confirms that the Arctic
oscillation (representative of non-seasonal sea-level pressure variations) and the
311
Westerlies control direct dust dispersal in Mongolia, suggesting that they also play an
important role in climate shifts (Han et al., 2008). Despite the lack of direct monsoonal
influence today, the region would have received increased moisture from both run-off
and precipitation when the East Asian summer monsoon reached its northernmost limit
(Figure 5.2; also see Figures 10.3(d)., Winkler and Wang, 1993).
Figure 5.2 Map showing the study region at about 6.0 kya, including northernmost limit
of East Asian summer monsoon. Approximation of forestation and lake extent derived
from data cited in text. Monsoon data derived from Winkler and Wang, 1993.
The Gobi-Altai region is separated from the Alashan Gobi by the southeasternmost foothills of the Altai Mountains and includes the more temperate Valley of the
Lakes, situated between the Gobi Altai and Khangai (Khangay) mountain ranges of
Mongolia. These lakes are fed by rivers running from the Khangai Mountains. Depth
312
fluctuates greatly on annual and decadal scales, but in recent times all are brackish,
shallow and internally drained. Palaeoclimatic studies are more limited here than in the
Chinese Gobi Desert. This is particularly true since much of the earlier Soviet work is
largely inaccessible. Despite the paucity of literature pertaining to local climate change,
several studies have focused directly on the southern Mongolian Gobi Desert.
Data suggests that conditions during the LGM were largely consistent with other
parts of Central and East Asia. Optically Stimulated Luminescence (OSL) dates on
cryoturbated structures around the Arts Bogd and Gurvan Saikhan (Gurvan
Sayhan/Gurvan Sayhany nuruu) ranges (~43o 41’ – 44o 26’ N, 102o 15’ – 102o 22’ E,
1200-2400 m a.s.l.) indicate that permafrost was developed along the desert floors and
alluvial fans during the Last Glacial Maximum (22-15 ka). Luminescence dates of sandy
sediments close to the Late Quaternary ice margin indicate an age of about 21 ka for the
maximum advance of glaciers (Lehmkuhl and Lang, 2001). The formation of alluvial
fans between about 23-9 ka further indicates a period of increased of aridity and flash
flooding (Owen et al., 1997). In support of this finding, OSL dates from sediments in the
Khongoryn Els dune field, which receives drainage from the Gurvan Saikhan just to the
north (44-43o N, 102-104o E), indicate the major accumulation of aeolian layers and
fluvial-lacustrine sediments (probably from local river blockage by sand accumulation)
occurred during a brief episode sometime between 18-11ka (~15ka) (Hülle et al., 2009).
Despite evidence for increased aridity, the physical quality of cryoturbation
structures and the occurrence of ice wedge casts indicate that the region was
characterised by adequate water. Detailed stratigraphic study of Chikhen Agui, an
313
archaeological rockshelter site in the Gobi-Altai Mountains, also indicates a continuation
of depositional processes and water percolation throughout the LGM (Komatsu et al.,
2001). Colder temperatures may have contributed to reduced evapotranspiration and are
indicated by an annual freeze-thaw depth of about 2 m, severe winter cooling with
continuous permafrost, and mean annual air temperatures below approximately -6o C
(Owen et al., 1998). In contrast, the lack of permafrost structures in the southern and
westernmost portion of the Gobi-Altai region (< 43o 00’ N, 100o 30’ E) might indicate
either an insufficient ground water supply or an annual air temperature closer to 0o C
(Owen et al., 1998). Across the Gobi-Altai region, permafrost degraded with the onset of
warmer average annual temperatures after about 13-10 ka (Owen et al., 1998). Divergent
records of effective moisture may result from variation in the ability of local soils to hold
seasonal melt-water and precipitation under regimes of reduced precipitation.
Holocene climate change in northern China, particularly northeast of the Yellow
River, has been related to increased precipitation resulting from shifts in the northern
extent of the summer monsoon, but the Gobi-Altai study region was not as heavily
influenced by the East Asian summer monsoon system (see Figure 5.2). Even at its
northernmost extent monsoonal precipitation would not have reached far beyond the
modern border separating Mongolia and China. Earlier Soviet work, though based
primarily in northwestern Mongolia, suggests that the climate across Mongolia was
relatively stable from 11.5-8.9k cal yr BP (10.0-8.0k yr BP), but cooler and wetter from
8.9-5.8k cal yr BP (8.0-5.0k yr BP) (An et al., 2008). Western Mongolian Altai lakes
Uvs Nuur and Bayan (Bajan) Nuur show evidence of post-glacial infilling around 13.0 ka
314
with high lake levels at about 11.0 ka (Grunert et al., 2000). Pollen and plant macrofossil
records from across Mongolia more precisely indicate a widespread drier climate before
9.5 ka and a distinctly more mild and wet climate during the early to mid-Holocene
(Gunin et al., 1999; Tarasov et al., 2000). At Hoton-nuur, a lake in the far west
Mongolian Altai Mountains, a cool/dry climate is evidenced during the terminal
Pleistocene and early Holocene, until ca. 12.7k cal yr BP (10.7k yr BP), followed by
forest development and a decrease in continentality (Rudaya et al., 2008). Humidity
appears to have peaked between 10.5-7.4k cal yr BP (9.3-6.5k yr BP) with a slight
cooling around 8.9k cal yr BP (8.0k yr BP) (Rudaya et al., 2008). Recent dates on ostrich
eggshell from the Gobi-Altai region indicate that East Asian ostriches were present in the
vicinity of Bayan-dzak (Shabarakh-usu locality) (44o 10’ N, 103o 42’ E) between 12.28.8k cal yr BP (10.3-7.9k yr BP) (Janz et al., 2009; Kurochkin et al., 2009), probably
indicating a more moist, and perhaps less seasonal, environment (Janz et al., 2009).
Luminescence dating of beach strands along the shore of Adagin Tsagaan Nuur
(Tsagaan/Adagin Cagaan Nuur/Nor) (~40o 00’ N, 100o 05’ E, 1331 m a.s.l.) indicate that
by 8.5 ka this lake joined Bon Tsagaan Nuur (Böön/Bon Cagaan Nuur/Nor) to form a
large palaeolake with an area of about 1923 km2 (Lehmkuhl and Lang, 2001) and a depth
of up to 100 m (Komatsu et al., 2001). Covering sediments and geomorphology of a
1303 m a.s.l. beach line at Orok Nuur (Orog Nor/Nuur) (~45o 00’ - 45o 15’ N, 100o 30 –
101o 05’ E, 1303 m a.s.l.) correspond with those at Adagin Tsagaan Nuur, suggesting a
similar date for this 394 km2 palaeolake (Lehmkuhl and Lang, 2001). Palaeoshoreline
landforms around Orok Nuur are complex, large and well-established, implying that the
315
palaeolake was stable for an extended period and did not experience frequent
transgressions and regressions (Komatsu et al., 2001). Evidence from ten lakes across
Mongolia suggests variable periods of Holocene lakeshore expansion with a widespread
trend towards wetter conditions by 8.3k cal yr BP (7.5k yr BP) (Tarasov et al., 2000).
The expansion of palaeolakes during the early Holocene was probably the result of a
number of conditions, including the initial influx of glacial melt, higher seasonal
groundwater availability due to permafrost melt, increased precipitation, and cooler
temperatures (lower levels of evaporation) (see Komatsu et al., 2001).
Aridity and continentality gradually increased after about 7.4k cal yr BP (6.5k yr
BP) (Rudaya et al., 2008). Warm/dry conditions prevailed between 5.8-2.6k cal yr BP
(5.0-2.5k yr BP) (An et al., 2008). Animal bones and feces from caves in the Tsagaan
Bogd Mountains (Cagan Bogdo/Segs Tsagaan Bogd Uul ) indicate a trend towards
increased aridity after 5.2k cal yr BP (4.5k yr BP) with two more humid phases following
this, including one after 3.2k cal yr BP (3.0k yr BP) (Kniaziev, 1986; Starkel, 1998).
Dramatic retreat of boreal forests and significant lowering of lake levels throughout
Mongolia by 4.0-3.0 ka indicate a drastic change in climate, resulting in essentially
modern conditions since 2.0 ka (Tarasov et al., 2000).
Radiocarbon dates on pottery from archaeological sites in the Gobi-Altai region
are later than those from the East Gobi. Sampling may have contributed to this result;
however, human habitation at Chikhen Agui overlaps with the earliest dune-field/wetland
occupations in the East Gobi (see Figure 3.1) and this may indicate that Late
Epipalaeolithic land-use persisted longer in the Gobi-Altai. Based on comparisons of
316
pottery and artefact typology, many dune-field/wetland sites can be assigned to a period
dating to between approximately 6.0-3.0 kya. This period corresponds with greatly
reduced seasonality across Northeast Asia between 6.8-3.2 kya.
If climatic amelioration – particularly higher effective moisture – was the driving
factor behind a shift to dune-field/wetland-centric organizational strategies, we could
predict that this pattern of land-use would have arisen in the Gobi-Altai sometime
between 9.5-7.4k cal yr BP; however, such a pattern is so far not evident in the
archaeological record until the middle Holocene. Instead, human population density and
dune-field/wetland use in the Gobi-Altai region appears to have been most intensive after
the onset of gradually increasing aridity (7.4 kya). At the Shabarakh-usu locality, many
sites were found below high water strandlines and are firmly dated to between 4.9-3.4k
cal yr BP (see Chapter 3). The termination of Oasis-type land-use occurred only after a
more dramatic retreat of lake levels and boreal forests across Mongolia (see Tarasov et
al., 2000). Despite the trend towards increased relative aridity, we can not assume that
middle Holocene environments were typified by the modern desert and desert-steppe
landscapes. According to the data summarized above, the Gobi-Altai would continue to
have been characterized by numerous stable lakes, and probably by arid or desert-steppe
until at least about 4.0 kya.
5.2.3. Alashan Gobi
The Alashan Gobi lies south of the Gobi-Altai mountain ranges and north of the Qilian
Mountains. For the purposes of this study, a territory encompassing 40o 30’ – 43o 00’ N,
317
97o 00’ – 106o 30’ E is considered. Although this represents a larger territory than that of
the other target regions, environments are less variable and archaeological sites more
dispersed. The Alashan Gobi is part of “arid Central Asia,” in which the modern climate
is controlled by the Westerlies. Monsoon activity over the Tibetan Plateau also
contributes to effective moisture due to the importance of drainage from higher elevations
in the south. The northward advance of monsoonal precipitation would have played a
greater role in Holocene precipitation than in the Gobi-Altai, but less of a role than in the
East Gobi.
Two landforms are of particular importance in both palaeoenvironmental and
archaeological studies: the Badain Jaran (Badan Jarang) Desert, which covers much of
region; and the Juyanze palaeolake system. The Badain Jaran Desert is a landscape of
active dune fields that includes some of the highest dunes in the world (up to 460 m). It
is located between 39o 20’ – 42o 00’ N and 99o 48’-104o 14’ E and bounded in the south
by mountains and in the north and west by former lake beds. The Yabulai Mountains
border Badain Jaran in the southeast. Varying levels of precipitation, derived primarily
from the East Asian summer monsoon system, are reported throughout the region, from
118 mm in the southeast to 37 mm in the northwest (Yang et al., 2003).
The Juyanze palaeolake system is located in the northwest portion of the study
area. It is an endorheic or closed basin with a drainage pattern of river channels from the
south and presently dry channel systems from the northern highlands in the Gobi-Altai
region (see Figure 5.1; Hartmann and Wünnemann, 2009). At its southern extent, the
Juyanze landscape is typified by a network of current and former river channels, marshes
318
and oases. The area is known for an abundance of both prehistoric sites and significant
historical remains (Sommerström, 1956). The palaeolake was historically fed by the
Ruoshui/Heihe drainage system (called Heihe in the southern mountainous regions and
Ruoshui [Ejina He/O-Chi-na Ho] in the north), which originates in the Qilian Mountains
on the Tibetan Plateau. The Ruoshui/Heihe forms two primary branches, the western is
called the Xihe (Moren Gol) and the eastern is the Donghe (Dongduer or Edsen/Etsin
Gol). Although both these rivers once drained into the Juyanze palaeolake, they now
terminate before reaching the terminal basin due to heavy use for irrigation (Lu et al.,
1997; Yang and Williams, 2003).
The Juyanze palaeolake system (including Gashun-nuur [Gaxun], Sogho-nuur
[Sogo/Sokho-nor/nur] and Eastern Juyanze) appears to have reached its maximum extent
earlier in the Pleistocene, but between 37-20 ka a large freshwater lake and swamp area
several times larger than present still existed (Wünneman et al., 1998a, 1998b). This
suggests more effective runoff from surrounding mountain ranges and considerably
higher levels of precipitation on the Tibetan Plateau (Wünneman et al., 1998a;
Wünneman et al., 2007). This period of increased precipitation probably occurred under
the influence of the Westerlies, since the East Asian summer monsoon was quite weak
during glacial times (Yang et al., 2003). After about 30.0k cal yr BP (25.0k yr BP)
aeolian deposition increased and cold-dry conditions predominated, although lake levels
did not reach modern lows (Wünneman et al., 2007). A period of complete desiccation
characterised the Eastern Juyanze between 23.0-17.2k cal yr BP (19.0-14.0k yr BP)
(Wünneman et al., 1998b). Increased lake levels are evident at Eastern Juyanze around
319
17.0k cal yr BP (13.8k yr BP). The formation of lake carbonates around 16.0k cal yr BP
(13.0k yr BP) at Baijian Hu, in the Tengger Desert south of the Yabulai Mountains, also
indicates a return to warmer/wetter conditions following the LGM (Wünneman et al.,
1998a, 1998b).
The Badain Jaran is a sand desert characterised by high dunes and numerous
lakes, most of which are concentrated in the southeast. There are over 100 permanent
lakes in the south. The southeastern edge of the Badain Jaran hosts many small shallow
lakes less than 0.2 km2 in extent and 2 m deep. Large, deep lakes are found farther north
(Yang and Williams, 2003). Aside from the Juyanze system, these lakes have no surface
in- or outflow. Recent studies of ion chemistry from water in several of the lakes indicate
that they are mostly charged by precipitation-derived groundwater rather than runoff
from adjacent mountains. Many of the lakes with low salinity were probably formed by
the emergence of freshwater springs in new locations when the movement of the dunes
changed groundwater flows (Yang and Williams, 2003).
At one time, the lakes in this area were much larger and deeper and many were
joined forming single large lakes. Thermoluminescence (TL) and radiocarbon dates on
sediments and peat from dried lake beds indicate that high lake levels occurred between
about 8.8-4.5k cal yr BP (8.0-4.0k yr BP), but that even as late as ca. 800 years ago, lake
levels were much higher than today (Yang and Williams, 2003). A more humid climate
in the region between ca. 9-4 ka probably indicates increased precipitation from the East
Asian summer monsoon, which reached its maximum northern extent during this period
(Yang et al., 2003). Pollen assemblages reveal more abundant middle Holocene
320
vegetation, with tree pollen of Pinus (pine), Picea (spruce), Ulmus (elm) and Salix
(willow) from peat samples dated to 7.8k cal yr BP (7.0k yr BP) and 7.4k cal yr BP (6.5k
yr BP) (Yang and Williams, 2003). Grass and bush pollens from these samples were
typical of desert-steppe to steppe with riparian or wetland conditions (compare to
Herzschuh et al., 2003; Xu et al., 2009; Zhao et al., 2009). Although lake levels were
higher, lakes changed gradually from freshwater to saline toward the end of the middle
Holocene due to increasing aridity (Yang et al., 2003).
In the Yabulai Mountains, southwest of the Badain Jaran Desert, a similar system
of groundwater-fed lakes exists as in the east. This mountainous region is better
vegetated than the sand seas of the Badain Jaran, and permanent lakes exist in closed
depressions (Yang, 2006). Groundwater is partially recharged by melting ground frost in
spring, but is maintained primarily through local precipitation that rapidly infiltrates the
sandy dunes and sediments throughout the region’s mountainous terrain (Yang, 2006).
TL dates on sediments from interbedded lacustrine and aeolian sediments at Shugui Lake
indicate high lake levels at about 128 ka, 16 ka, and intermittently between 7.5-4.5 ka
(Yang, 2006).
Today, as in the Gobi-Altai, precipitation in the Badain Jaran/Yabulai Mountain
region is controlled by the Westerlies with little influence from the East Asian summer
monsoon, which now reaches only the southeastern margin of the Badain Jaran Desert.
During the middle Holocene, however, precipitation would have reached an average of
perhaps about 200 mm/year, probably due to the northward migration of the Asian
summer monsoon (Wünneman et al., 1998a; Yang and Williams, 2003). By about 6.8k
321
cal yr BP (6.0k yr BP), northward migration of the East Asian summer monsoon would
have ensured increased precipitation throughout the entire Alashan Gobi region (Winkler
and Wang, 1993). Higher levels of evapotranspiration would likely have maintained a
desert-steppe environment, despite an increase in the availability of surface and
groundwater.
Studies of Holocene palaeoenvironments on the western Alashan Plateau have
focused on three small lakes, once a part of the Juyanze palaeolake system. Today the
palaeolake system is divided into two sections – the Western and Eastern Juyanze – and
three lakes – Gashun-nuur (Xi/Western Juyanhai), Sogho-nuur (Dong/Eastern Juyanhai),
and the Eastern Juyanze Lake Basin (East Lake Etsina). The latter now exists as two
ephemeral lakes – Jingshoutou and Tian’e Hu (Mischke, 2001: 10). The areal extent of
these lakes has undergone dramatic changes since their final separation by 5.5k cal yr BP
(4.8k yr BP) (Lu et al., 1997). Western and Eastern Juyanze were separated from the
Gashun-nuur system through tectonic movement, either during the Holocene (Becken et
al., 2007; Hölz et al., 2007; Hartmann and Wünnemann, 2009) or possibly much earlier
around 40.0k cal yr BP (35.0k yr BP) (Wang et al., 2004). Western Juyanze did not
become separated into the modern Gashun-nuur and Sogo-nuur until the 17th century CE.
Since the earliest historical records (ca. 300 BCE) until that time, the basin was described
as continuously filled by a large lake (Mischke, 2001: 16).
Changes in the flow of the Ruoshui/Heihe drainage system, including intermittent
variability in the position and flow of its northern branches (Xihe [Moren Gol] and
Donghe [Edsen Gol]), have affected the levels of these terminal lakes. Over the past
322
2000 years, the Ruoshui/Heihe has been known to periodically change its course,
sometimes flowing into the Eastern Juyanze Basin (East Juyanze/Juyanze/Juyan Lake,
including Lake Jingshoutu) and sometimes into Gashun-nuur and Sogo-nuur (Mischke,
2001: 16; Wang et al., 2004).
Palaeoenvironmental records for each minor lake basin
should mirror shifts not only in temperature and precipitation, but in the relative
dominance of minor river branches.
The early to mid-Holocene landscape would have been much different from
today, particularly in terms of hydrology. Human activities in the region have
dramatically altered the hydrological landscape. Changes in lake levels related to human
activities are noted in historical records since 317-534 CE. Intensification of farming
around the town of Ganzhou (established by the Han in 111 BCE) during the Sui and
Tang Dynasties (6th-10th centuries CE) resulted in the inital shrinking of the terminal
lakes, while subsequent rejuventation occurred during the Mongol occupation beginning
in the 13th century CE (Lu et al., 1997). However, even in 1927 when Sven Hedin visited
the region, Sogho-nuur was 2.9 m at its deepest point and it took Hedin almost four hours
to cross the river by canoe (Hedin, 1943: 166-168). At this time the area was home to the
largest poplar forest in China (Lu et al., 1997). More recently, periods of war followed
by industrial and agricultural development have led to desiccation and the destruction of
vegetation (Lu et al., 1997). These examples attest to the vulnerability of both lake levels
and forest vegetation to human activities along the Ruoshui/Heihe, and indicate that
current climatic conditions are a poor indicator of early to mid-Holocene environments.
323
Lake evolution in the Juyanze palaeolake basin illustrates the relationship
between climatic change in the Alashan Gobi and neighbouring territories, and the
stabilization of lakes and water tables throughout the terminal Pleistocene and Holocene.
At 19 ka, Western Juyanze (Gashun/Sogo-nuur) was completely dry until a freshwater
lake was established by 11.3 ka (Lehmkuhl and Hasselein, 2000). Similarly, Eastern
Juyanze was not fully developed until after 11.0k cal yr BP, before which the basin was
characterised by high energy fluvial transport and gravel deposition (Hartmann and
Wünnemann, 2009). After Eastern Juyanze was separated from the larger system, run-off
from the Tibetan Plateau may have ceased to contribute to its formation (Hartmann and
Wünnemann, 2009), but desiccated drainage channels to the north indicate strong fluvial
input during the early to mid-Holocene Gobi-Altai wet period between about 10.5-7.4k
cal yr BP (9.3-6.5k yr BP) (see Hartmann, 2003). Evidence for periods of increased runoff into Eastern Juyanze indicates that high lake levels did occur during highly variable
periods of amplified precipitation in the catchment area between 10.7-8.9k cal yr BP
(Hartmann and Wünnemann, 2009). This correlation indicates that palaeoclimatic studies
at Juyanze may be useful in inferring short-term climatic shifts to the north.
Lake levels at Eastern Juyanze were still unstable in the early Holocene and
fluctuated rapidly in accordance with changes in precipitation. Early Holocene lake
levels reached their height at about 8.9k cal yr BP (Wünneman et al., 1998b), after which
dry conditions prevailed between 8.9-8.1k cal yr BP. There was an abrupt change to
moister conditions with enhanced run-off between 8.1-7.6k cal yr BP (Hartmann and
Wünnemann, 2009). Evidence of this increase in moisture is mirrored in the pollen
324
record by a spike in the relative frequency of steppe pollen between 8.0-7.3k cal yr BP
(Herzschuh et al., 2004). Biome reconstruction based on a pollen record from the Eastern
Juyanze core indicates that the period between 10.7-5.4k cal yr BP was still probably
typified by relatively dry desert-steppe vegetation (Herzschuh et al., 2004), but erosional
factors in run-off might have resulted in an overestimation of desert taxa (Hartmann and
Wünnemann, 2009). Overlapping with greater development of steppe vegetation, lake
levels declined from their early Holocene height between 7.7-5.4k cal yr BP, interrupted
by a period of high run-off (6.2-6.0k cal yr BP) and the expansion of steppe species (6.55.9k cal yr BP) but no distinct lake formation (Herzschuh et al., 2004; Hartmann and
Wünnemann, 2009). This period of high run-off and steppe expansion corresponds with
the northernmost extent of summer monsoon migration (see Winkler and Wang, 1993).
Refilling of the aquifer and lake formation began again between 5.4-5.0k cal yr
BP (Hartmann and Wünnemann, 2009). Deposition of lake mud reduced groundwater
infiltration leading to both increased stability of lake levels and salinization during
periods of low precipitation and increased evaporation (Hartmann and Wünnemann,
2009). Despite increased salinity, complete desiccation during dry spells was buffered by
a newly stabilized water table (Mischke et al., 2003; Hartmann and Wünnemann, 2009).
Lake levels were at their highest between 5.4-4.0k cal yr BP17 (Hartmann and
Wünnemann, 2009), especially 5.1-4.6k cal yr BP (Mischke et al., 2005), when a desertsteppe environment likely predominated (Herzschuh et al., 2004). Stable isotope
analysis, studies of pollen, ostracods and other microfauna indicate significant inflow of
17
Although Wünneman and colleagues (1998b) assert that high lake levels occurred at 8.9 (8.0k yr BP) and
5.8k cal yr BP (5.0k yr BP).
325
river water from the Ruoshui/Heihe drainage system (Mischke et al., 2003; Mischke et
al., 2005). Increased moisture lasting until about 4.0k cal yr BP is consistent with records
from small lakes in the Badain Jaran Desert (Yang et al., 2003; Yang and Williams, 2003;
Yang, 2006). Significant groundwater input between 4.0-3.5k cal yr BP suggests an end
to this period of high inflow and the maintenance of lake levels through local
groundwater (Mischke, 2001: 87).
Aridification and desert expansion occurred again between 3.9-1.7k cal yr BP
(Herzscuh et al., 2004; Mischke et al., 2005), although conditions were probably not as
arid as during the early Holocene (Herzschuh et al., 2004). Pollen levels remained high
until about 3.2k cal yr BP, suggesting that lake level decline beginning since 4.1k cal yr
BP may not be related to any increase in regional aridity (Mischke et al., 2003). Stable
isotope analysis suggests that groundwater may have contributed more significantly to
the water budget around this time (Mischke et al., 2005), perhaps suggesting a decline in
both direct monsoonal precipitation and the influx of water from the Ruoshui/Heihe
drainage system. Palynological studies of mid- to late Holocene lacustrine sediments
from Gashun-nuur indicate an abundance of arboreal and shrub riparian vegetation
around the lake between 3.5-3.2k cal yr BP, including Hippophaë (seabuckthorn18) and
Populus (poplar/aspen), indicating a peak in available moisture when lake levels were
decreasing at Eastern Juyanze (Demske and Mischke, 2003). The divergence between
lake levels and vegetation is probably due to the reliance of lakes on drainage from
18
Seabuckthorn is a traditional food and medicinal plant which grows in sandy soil and is native in East
Asia to cold climates and altitudes above 1200 m a.s.l. The species has recently been widely planted not
only for economic reasons, but because it prevents erosion, controls water loss, increases vegetation cover,
and provides habitat for foxes, hares, and pheasants.
326
higher elevations, with lake levels more closely mirroring changes in precipitation at
river origins than local changes in effective moisture. The spread of montane forests at
high elevations, as indicated by Picea and Betula pollen, supports a shift toward cooler
conditions around the same time (Demske and Mischke, 2003).
Between about 2.9-2.7k cal yr BP a severely arid regional climate is evidenced in
pollen concentrations at Gashun-nur. As lake levels remained higher in Eastern Juyanze
than Western Juyanze, it is expected that the lake was positively influenced by a shift in
the river drainage system (Mischke et al., 2003). Following this period of extreme aridity
lake levels rose again around 2.7k cal yr BP (Demske and Mischke, 2003).
By 3.2k cal yr BP, lake levels in both lakes had became less stable. Eastern
Juyanze dried up three times between 3.2-2.9k cal yr BP and disappeared after ca. 1.7k
cal yr BP (3rd century CE) (Demske and Mischke, 2003; Mischke et al., 2003; Herzschuh
et al., 2004; Mischke et al., 2005). Despite the late Holocene return to a climatic regime
more similar to that of the early Holocene, steppe vegetation may have been more
plentiful in the region than in the early Holocene due to the establishment of more stable
hydrological systems and plant communities during the preceding climatic optimum.
These results are broadly comparable to climatic records from the Tengger Desert,
but the period of climatic optimum is much later in the Juyanze palaeolake region than
that attested to in the south. Lithology and fossil pollen data from Qingtu palaeolake (39o
04’ 15” N, 103o 36’ 43” E, 1302 a.s.l.), a terminal lake in the Shiyang River Basin
illustrates this point. By 7.2k cal yr BP, under moister conditions following a period of
high aridity, a stable, shallow lake surrounded by a steppe desert environment and denser
327
vegetation had formed in the basin of a large Pleistocene lake (Zhao et al., 2008).
Climate was highly variable from 5.2-3.0k cal yr BP, after which a dry climate persisted
and the lake dried up (Zhao et al., 2008). Extreme aridity led to the termination of
fluvial-lacustrine depositional processes beginning at about 3.2k cal yr BP (3.0k yr BP)
(Zhang et al., 2000). The latter date correlates well with decreasing lake levels in both
Gashun-nuur and Eastern Juyanze, but an earlier period of climatic optimum and more
severe aridification by 3.0k cal yr BP are represented. The former may be a signal of
earlier and stronger penetration by the East Asian summer monsoon in the more southerly
location. The latter implies a greater availability of moisture in the Badain Jaran Desert
following the onset of widespread aridification, perhaps as a result of direct positive
moisture influence from the Westerlies or drainage from the north during a humid phase
represented in the Gobi-Altai record after 3.2k cal yr BP (Kniaziev, 1986; Starkel, 1998).
Holocene climates in the Alashan Gobi are characterized by increased
precipitation beginning after 9.0 kya and continuing until at least 4.0 kya. According to
pollen assemblages from the southern Badain Jaran Desert, high elevation and riparian
woodlands may have been established by 7.8 kya, along with wetlands and desert-steppe
to steppe grasslands (Yang and Williams, 2003). Farther north, in the Juyanze region,
two periods of high effective moisture occurred: once in the early Holocene, prior to
8.9kya; and the other in the middle Holocene between 5.1-4.6 kya (Mischke et al., 2003;
Mischke et al., 2005; Hartmann and Wünneman, 2009). There is no evidence of arboreal
development at such an early date; however, pine and birch are represented in pollen
profiles until after 3.2 kya and suggest high elevation forestation in the late middle
328
Holocene (Demske and Mischke, 2003). Riparian arboreal/shrub vegetation and arid to
desert-steppe grasslands remained stable until about 3.2 kya despite lower lake levels
beginning almost 1000 years earlier (Demske and Mischke, 2003; Mischke et al., 2003).
Some of the archaeological sites in this region may date to the late Pleistocene
and early Holocene, but those belonging to the late middle Holocene are most numerous
(Table 3.7c). The earliest dates are from Mantissar 12, a dune-field/wetland site in the
Gurnai Depression (6460 + 700 [5.76-7.16] ka), a major erosion basin east of the
Ruoshui/Heihe located south the Juyanze palaeolake system and north of the Yabulai
Mountains (Figure 5.1). Overlappping, but probably slightly younger dates come from
Yingen-khuduk (5690 + 350 [5.34-6.04] ka), along the Mongolia-China border. This
date can probably be considered approximately contemporaneous with those from the
Ulan Nor Plain site in the Gobi-Altai (see Table 3.1). The high range of error in these
dates makes it difficult to assign the habitations to a period of relative humidity or aridity,
but they appear more closely related to the dry phase between about 7.5-5.5 kya.
The association of early dune-field/wetland sites with a relatively dry phase is
notable since the majority of dates indicate an age of younger than 4.0k cal yr BP (Table
3.1). A similar trend has been noted for the Gobi-Altai region. In addition to the
abundance of studied Oasis 3 sites situated around lakes, there is possible evidence from
Gashun-nuur of anthropogenic burning at about 3.7-3.2k cal yr BP and human or fluvial
transport of coniferous wood at 3.3-3.0k cal yr BP (Demske and Mischke, 2003; Demske,
personal communication, May 2010). Most archaeological sites in the Alashan Gobi can
be attributed to between 4.0-3.2 kya. Residential sites continued to be centred on lakes.
329
This pattern suggests that lakes attracted hunter-gatherers during a drier phase of early to
late middle Holocene humidity.
As discussed in Chapter 3, Oasis 3 belongs to the last phase of the hunter-gatherer
oasis adaptation that predates the florescence of a cultural and economic milieu
associated with early Bronze Age nomadic pastoralists. Additional continuity between
periods is further indicated by such archaeological evidence as fragments of a bronze
vessel or helmet with pottery and typical microliths at Sogho-nuur (Maringer, 1950;
personal observation, October 2008). This late period of hunter-gatherer subsistence in
the Alashan Gobi took place in an environment characterized by an abundance of
arboreal and shrub riparian vegetation, arid steppe to desert-steppe grasslands, and many
stable lakes primarily sustained through groundwater input. Forest growth at higher
elevations would probably have begun in the early to middle Holocene and continued
into the Bronze Age.
330
DATE
(k cal yr
BP)
REGION
12.0
10.0
9.0
8.0
7.0
6.0
5.0
4.0
3.0
2.0
MOISTURE
(relative to
previous)
VEGETATION
FORESTS
East Gobi
TEMPER
ATURE
(relative to
previous)
warmer
wetter
mosaic steppe
mosaic
Gobi-Altai
warmer
wetter
N/A
not establ.
Alashan
warmer
dry
desert
not establ.
East Gobi
cooler
drier
desert-steppe
N/A
Gobi-Altai
warmer
drier
arid to desert-steppe
expansion
Alashan
N/A
wetter
desert
N/A
East Gobi
warmer
wetter
arid steppe
expansion
Gobi-Altai
cooler
high lake level
arid steppe to steppe
expansion
Alashan
N/A
wetter
desert
N/A
East Gobi
cooler
wetter
arid and forest-steppe
mixed
Gobi-Altai
cooler
high lake level
steppe to arid-steppe
expansion
Alashan
N/A
wetter
desert-steppe
mixed
East Gobi
as above
as above
arid steppe
as above
Gobi-Altai
warmer
drier
arid steppe
expansion
Alashan
N/A
drier
desert-steppe
mixed
East Gobi
cooler
drier
arid steppe
coniferous
Gobi-Altai
warmer
drier
arid steppe
developed
Alashan
N/A
high lake level
arid or desert-steppe
mixed
East Gobi
warmer
drier
steppe
coniferous
Gobi-Altai
as above
drier
arid steppe
coniferous
Alashan
N/A
high lake level
arid or desert-steppe
N/A
East Gobi
N/A
drier
arid steppe
decline
Gobi-Altai
warmer
drier
desert-steppe
coniferous
Alashan
N/A
drier
arid or desert-steppe
N/A
East Gobi
N/A
drier
desert-steppe
decline
Gobi-Altai
cooler?
wetter
desert-steppe
decline
Alashan
cooler
wetter
riparian
East Gobi
modern
drier
desert-steppe to
desert
desert-steppe
Gobi-Altai
modern
modern
desert
decline
Alashan
N/A
modern
desert
riparian
decline
Table 5.1 Summary of regional climate change from 12.0k cal yr BP to 2.0k cal yr BP.
331
5.3. Desert forests
Recent research on relic forests of the Gobi-Altai region draws attention to the possible
importance of wooded environments in determining Gobi Desert hunter-gatherer
settlement patterns. Reconstructing vegetation distributions is particularly important to
understanding human land-use throughout the Holocene, as it determines the composition
of available plant and animal species. Compared to steppe or desert-steppe
environments, forests and forest-steppe offer a highly divergent range of useable flora
and fauna. In the East Gobi, where dated Oasis 2 type sites are almost two millennia
older than in the west, forest and forest-steppe were established by at least 8.0k cal yr BP.
Forest development probably began on the southern fringes of the Alashan Gobi by 7.8k
cal yr BP, but the establishment of riparian and higher elevation woodlands is expected to
have been later in the more arid north. The timing of forest development in relation to
the establishment of Oasis 2-type habitation may suggest a correlation between forest
development and the new patterns of land-use.
Currently, there are no detailed studies of the development of post-LGM
vegetation in the western Gobi Desert; however, reconstructed sequences of post-glacial
vegetative colonization on the Ulagan high-mountain plateau (48-59o N, 82-90o E, 19852150 m a.s.l.) in the Altai Mountains of southern Siberia provide a clear example of the
forestation process farther north. After 16.0k cal yr BP glaciers and bare ground gave
way to pioneer herbaceus vegetation (Artemisia, Gramineae, Cyperaceae, Salix, and
Potentilla). By 15.0k cal yr BP, steppe vegetation dominated by Artemisia and
Chenopodiaceae and had encroached into areas of formerly bare ground. Primarily
332
deciduous forests began to fill in areas covered by steppe as glaciers receded and climate
continued to warm; by 9.5-7.5k cal yr BP fully developed forests were present
(Blyakharchuk et al., 2004). Forest-steppe environments were more common across
Central and East Asia at this time, especially at higher altitudes.
Pollen records and wood macrofossils from the Mongolian Altai Mountains
indicate slightly later forest development following deglaciation, though a similar series
of developments probably occurred. Prior to 9.5 ka, the Khangai region of the Altai
Mountains, just north of the Gobi Desert, would probably have been mostly treeless with
cold steppe and dwarf-shrub vegetation dominating the basins (Tarasov et al., 2000).
Larix (larch) appeared later, along with Pinus (pine), both of which probably had a much
greater distribution during the middle Holocene than they do today. Forest vegetation
would have been fully developed by about 6.0 ka. Distribution of vegetation might have
been similar to modern times, only with more widespread arboreal representation due to
the heightened availability of moisture (Tarasov et al., 2000).
Various wood fragments from Bayan Sair (45o 34’ N, 96o 54’ E, 2600 m a.s.l.) in
the Gobi-Altai region, were radiocarbon dated to between 5.2 and 3.8k cal yr BP (4.53.5k yr BP) and included samples of Abies (fir), Picea (spruce), Pinus sibirica (Siberian
pine), and Larix (larch) (Gunin et al., 1999; Tarasov et al., 2000). Fir is particularly
sensitive to moisture, winter temperatures and soil richness. Today, it grows only in a
few of the northernmost boreal forests of Mongolia, along with spruce and Siberian pine.
Spruce is another moisture sensitive species, whose modern habitat is restricted mostly to
333
river valleys north of 48o. Larch is more tolerant of varying conditions and macrofossils
of this species date much later than the others at 2.6-2.0k cal yr BP (2.5-2.0k yr BP).
The late appearance of larch and the disappearance of other species probably
represent high elevation forests under conditions of decreasing precipitation. Although
larch does not grow at Bayan Sair today, it still is found in individual stands throughout
the region (Gunin et al., 1999). According to proxy data from other regions of Mongolia,
we can assume that this macrofossil record captures a period of transition in the decline
of arboreal species after 3.5k cal yr BP, following the end of the Holocene climatic
optimum. This change may be related to the dramatic increase in aridity in the GobiAltai region after 4.0k cal yr BP.
Forests are rare in the Gobi Desert today, but two relict populations of BetulaSalix (birch-willow) forests in the Gobi-Altai region have been studied: the first grows on
the north-exposed slopes of Ikh Bogd (Ih Bogd) Mountain; and the second in the isolated
Gurvan Saikhan (Gurvan Saykhan) mountain ranges. Notably, they contain a variety of
plant species similar to other boreal flora found in conifer forests to the north and as far
away as northeastern Tibet (Miehe et al., 2007). While many of these species are subject
to long-distance dispersal or by birds or humans, others, such as Viola dissecta (violet)
and Adoxa moschatellina (moschatel or muskroot), can only migrate in a step-wise
manner and require fairly continuous habitats (Miehe et al., 2007). The existence of a
relatively continuous Holocene forest belt stretching across the mountain bands of
Mongolia and northwestern China is further supported by the more general lack of
distinct floral and faunal species within the fragmented and distant forests. The lack of
334
connected mountain chains indicates a former continuity between high elevation and
riparian forests, as they stretched across dry, low elevation basins such as those in the
Alashan Desert.
Therefore, the most notable difference between palaeo- and modern
environmental conditions is widespread forest expansion between 10.8-5.2k cal yr BP
(9.5-4.5k yr BP), after which more moisture sensitive arboreal species gradually
vanished. As at Bayan Sair, stands of more resilient species like larch also eventually
retreated. Pollen from dated Eastern Juyanze and Gashun-nuur profiles indicate that
spruce and birch forests receded sometime after 3.2k cal yr BP (3.0k yr BP) in the
Alashan Gobi region (Demske and Mischke, 2003; Mischke et al., 2005).
The widespread decline of forests, though clearly linked to climate, was related to
a series of factors that would have acted on vulnerabilities within the complex system of
sustaining mechanisms. Water retention systems, heat regulation and multiple
reproduction methods have allowed the few remaining Gobi Desert forests to maintain
themselves even in a region which in recent times would otherwise be much too hot and
dry. High-elevation forests in the Gobi-Altai region are distributed on northern
exposures, where there are lower levels of evapotranspiration (Starkel, 1998; Cermak et
al., 2005). The dense forest canopy protects active permafrost layers and rainwater,
prolongs the availability of snow meltwater, and provides shade and protection for young
trees (Cermak et al., 2005; Miehe et al., 2007). When complete deforestation of high
elevation forests occurs, re-establishment is extremely difficult or impossible under
conditions of heightened aridity (see Miehe et al., 2007).
335
Similar positive feedback mechanisms to those recognized in birch-willow forests
would have existed along the stable watercourses and lakes that were once more common
across the Gobi-Altai and the Alashan Gobi. The existence of individual Ulmus, Larix,
or Pinus trees among meadow steppes, rare extended open forests in semi-desert
vegetation like Stipa glareosa (bunchgrass) and Anabasis brevifolia (a type of
Chenopodiaceae), and the dispersal of Ulmus pumila (Siberian elm) in zonal desert
steppes, indicates that sufficient access to groundwater and protection from humans and
animals allows long-lived species to survive in drier environments (Cermak et al., 2005;
Miehe et al., 2007; Wesche et al., 2011). Such evidence has important implications for
the distribution of arboreal species during the mid-Holocene, with increased groundwater
availability and warmer temperatures.
The impact of anthropogenic factors, particularly herding practices on forest
environments, is an especially important consideration in palaeoenvironmental
reconstructions of a region populated primarily by pastoralists. Recent studies of relic
desert forests indicate that climate change can not entirely account for extensive
deforestation in the late Holocene. Studies of reproduction and genetic structure of
Siberian elm stands (single trees and woodlands) in the Mongolian Gobi Desert suggest
that the trees are actually well-adapted to modern conditions of high aridity and extreme
seasonal variation in cold and heat. Trees can sustain themselves for hundreds or even
thousands of years through clonal reproduction (suckering) when conditions are not
suitable for pollination and germination, but genetic diversity among Gobi Desert stands
indicates that elms have managed to persist through sexual reproduction. Considering
336
their ability to reproduce normally, they should be more commonplace along potentially
suitable habitats like drainage lines, riverbeds, and ravines (Wesche et al., 2011). The
authors suggest that extensive grazing is responsible for their absence. The potential for
forestation under modern climatic regimes is exemplified by recent forest expansion Ikh
Bogd. Since a major earthquake and landslide in 1957 covered the main entrance route to
a valley once seasonally occupied by pastoralists, birch-willow forests have recolonized
the cleared land (Cermak et al., 2005; see also Starkel, 1998). Hedin also comments on
the effects of grazing in respect to one frequented oasis, where old popular trees were
common and young ones rare, as wandering camels “would not leave a single new shoot
uneaten” (Hedin, 1943: 143-144). Additionally, he noted that rapid deforestation was
occurring along caravan trails due to the indiscriminate use of trees for campfires.
Overgrazing has also been cited as the primary factor leading to processes of
desertification and deflation in oasis environments of the southernmost Mongolian Gobi
Desert (Pankova, 2008). Just as the destruction of vegetation cover in oasis environments
leads to a loss of surface soils and groundwater retention capacity (Pan and Chao, 2003),
gradual reduction of forest cover through both grazing of the forest fringes, and directed
anthropogenic clearance leads to soil aridification and increases vulnerability. The
intensification of herding practices during the Bronze and Iron Ages should be considered
as contributing factors to widespread deforestation after 2.6k cal yr BP (2.5k yr BP) (see
Gunin et al., 1999). Likewise, the attested decline of riparian shrub and woodland in the
Juyanze region after 3.2k cal yr BP might be connected to the intensified use of local
337
environments by hunter-gatherers, or by the introduction of herd animals into oasis-based
habitation sites.
5.4. Discussion
The interplay between climate change and human land-use is most closely related to the
ecological effects of temperature and precipitation. The story of post-LGM climatic
amelioration summarized above is one of rising groundwater tables, infilling of basins,
and soil formation, all of which contributed to the establishment of stable lakes and
rivers, increasing grassland productivity, and forest habitats. Climate change following
the LGM is broadly characterized by gradual increases in temperature and humidity that
culminated in the Holocene Climatic Optimum and the development of new Holocene
ecozones like stable dune-fields, rich wetlands, and riparian woodlands. As early as the
terminal Pleistocene and early Holocene, dune-field/wetland habitats were beginning to
form around more stable water systems, and woodlands were slowly developing in river
valleys and at higher elevations.
Palaeoenvironmental signatures from the East Gobi indicate that the earliest Oasis
2 sites occur during a period of increased moisture availability and sand stabilization. By
8.0 kya, temperatures and the annual growing season had declined slightly since the
initial Holocene, and mixed coniferous and broad-leaved forests would have been present
at high elevations and around lakes. A high in effective moisture is noted by 7.7 kya.
Together, these data suggest that the wetland-centric focus of Oasis 2 in the East Gobi
338
coincides with high lake levels, increased forest cover, and the expansion of steppe or
arid steppe.
Further west, the earliest dates for Oasis 2 sites suggest a later establishment of
dune-field/wetland specialization, by about 6.0 kya. Additional sampling of western sites
could produce earlier dates, but the current lack of early Holocene dates might simply be
symptomatic of a later shift in organizational strategies. Conditions in the Gobi-Altai,
such as the widespread survival of permafrost until 13.0-10.0 ka (Owen et al., 1998) and
evidence of low effective moisture and flash flooding until the early Holocene (Owen et
al., 1997; Hülle et al., 2009), indicate a lack of stable ecological development within
lowland environments. Lake level pollen data summarized above suggest that lowland
environments would have been well-developed by 8.5-8.1 kya. Forest expansion
probably reached its height throughout the Gobi-Altai and Alashan Gobi by 6.0 kya (see
Tarasov et al., 2000; Miehe et al., 2007). In contrast to the East Gobi, western sites postdate initial peaks in humidity, high lake levels, and cooler temperatures. While the
earliest East Gobi sites are associated with a high in Holocene humidity, western sites
seem more closely associated with relatively arid phases of overall Holocene humidity.
Despite an apparent reliance on dune-field/wetlands, dates for Holocene
archaeological sites indicate that those environments were not typically used for longterm habitation until long after they were fully established, and that there was a closer
relationship between the expansion of forests and dune-field/wetland intensification. In
the East Gobi, the earliest longer-term residential sites date to 7.6 and 6.8k cal yr BP.
This corresponds well to evidence for the spread of high elevation forests and mosaic
339
forest-steppe environments around lakes. Forest development was later in the western
Gobi Desert, with full development by at least 6.0 kya. Here, the earliest such habitation
sites also date to about 6.0 kya, and possibly slightly earlier in the Gurnai Depression of
the southern Alashan. The common characteristic between eastern and western sites is
the relationship between the timing of shifts in land-use and the establishment of high
elevation and riparian forests.
The establishment of these new Holocene habitats has been summarized in this
chapter and shows that their development was a gradual process, periodically interrupted
by phases of aridification and enhanced seasonality. Moreover, the establishment of the
stable dune-fields and rich wetlands favoured by Holocene hunter-gatherers was itself
only temporary and the disappearance of those contributed to the much different
landscape of modern times. According to published palaeoenvironmental data
summarized here, widespread remobilization of dune-fields and the retreat of high
elevation and gallery forests were probably well underway by the end of Oasis 3, when
nomadic pastoralism was established across Mongolia and northern China. Although
many factors contribute to changes in economy, land-use is still inextricably tied to the
availability of local resources. Vegetative and hydrological changes outlined in this
chapter can be compared to patterns of land-use attested in the archaeological record in
order to assess interpretations of hunter-gatherer organization in light of existing
knowledge about hunter-gatherer ecology.
340
10.0k cal yr BP
8.0k cal yr BP
Figure 5.3a Approximation of hydrology and vegetation in three Gobi Desert regions for
10.0k and 8.0k cal yr BP. Reconstructions are based on data cited in text. Key: v
indicates forest, riparian woodland, or forest-steppe mosaic; o indicates desert; t
indicates desert-steppe; x indicates arid steppe; -- indicates steppe.
341
6.0k cal yr BP
3.0k cal yr BP
Figure 5.3b Approximation of hydrology and vegetation in three Gobi Desert regions for
6.0k and 3.0k cal yr BP. Reconstructions are based on data cited in text. Key: v
indicates forest, riparian woodland, or forest-steppe mosaic; o indicates desert; t
indicates desert-steppe; x indicates arid steppe; -- indicates steppe.
342
CHAPTER 6 – DISCUSSION AND CONCLUSIONS
Hunter-gatherer archaeology in the Gobi Desert of Mongolia and China has
received little attention from scholars, despite the importance of this region as a frontier
between two culturally and environmentally divergent zones of East Asia. Exhaustive
archaeological collections exist in Western and Asian museums and universities, but are
largely unstudied. While scholars have suggested various relationships between huntergatherers of this region and contemporary groups in China and Siberia (see Chapter 2),
this thesis constitutes the first attempt to define material culture and establish a
chronology of artefacts and land-use from which further study can be directed.
Results indicate that major reallocations in land-use began by at least 7.6 kya in
the east and 6.0 kya in the west, and were characterized by the dramatically intensified
use of lowland dune-field/wetland habitats, and the provisioning of longer-term
residential sites. Hunter-gatherers retained high residential mobility, despite the
specialized use of dune-field/wetlands. Contemporary communities neighbouring parts
of Northeast Asia show largely divergent economic strategies. By 7.6 kya developments
in North China represent the early stages of agriculture and a shift towards chipped and
polished adze/axes, bone tools, large grinding tools, and other agricultural implements
(Lu, 1999). Archaeological evidence from sites in Northeast China represent both typical
hunter-gatherer assemblages with microblade core, flake, and bifacial technology, and the
occurrence of sedentary villages whose inhabitants used a range of technologies
including microblades, large formal grinding tools, and chipped/polished adze/axes, hoes,
and spades (Guo, 1995a; Tan et al., 1995a; Lu, 1998; Jia, 2007). By 6.0 kya,
343
archaeological sites in northwestern China, just south of the Alashan Gobi, show the
establishment of agricultural economies that incorporated a range of domesticated plant
and animal species (Barton et al., 2010; Bettinger et al., 2010a).
Developmental trajectories in the Gobi Desert differ from reconstructed for
nearby regions dominated by agriculture. The dispersal of chipped and edge-ground
adze/axe technology across the Gobi Desert and the increased importance of large formal
grinding tools in the East Gobi paralleled technological developments in parts of northern
China (particularly Northeast China), but were much more limited in frequency and never
associated with sedentary communities. Pottery appears to have been used in the East
Gobi by at least 9.6 kya. Specialized tools that required time- and labour-intensive
manufacture, such as knobbed rollers or pestles and edge-ground or polished adzes/axes,
were found in the earliest Oasis 2 sites. The use of these new technologies has
traditionally been related to the emergence of small-scale agricultural production
(Derevianko and Dorj, 1992; Cybiktarov, 2002), but the present study suggests that they
were more likely the products of specialized foraging strategies.
6.1. Palaeoenvironment and local ecology
Gobi Desert “dune-dweller” sites are related to early and middle Holocene climatic
amelioration, as recognized by increased precipitation, lake formation and stabilization,
and heightened vegetative biomass. Recent research on relict forests in the Gobi-Altai
suggests widespread forestation across the western Gobi Desert by the middle Holocene
(Miehe et al., 2007). Archaeological sites in the Gobi Desert indicate that the region was
344
once well-populated. New dates on these archaeological sites indicate that such
habitations are associated with a period of increased humidity, when the Gobi Desert
would have been characterized by high elevation forests, scattered lowland riparian
forests, desert-steppes, and extensive developed wetlands around lakes. The Holocene
development of forests at higher elevations and mature wetlands in lowland habitats
would have created distinct upland and lowland foraging patches that are were not typical
of Pleistocene environments.
As such, the stabilization of lowland environments would have played an integral
role in modification of land-use. The formation of alluvial fans between 23.0-9.0 kya in
the Gobi-Altai region suggests that although lake levels had begun to recover from Last
Glacial Maximum lows, aridity and flash flooding were prolonged into the early
Holocene (Owen et al., 1997; Hülle et al., 2009). The Eastern Juyanze in the Alashan
Gobi similarly indicates a period of high run-off with little stable lake development
during the terminal Pleistocene and early Holocene (Hartmann, 2003; Hartmann and
Wünneman, 2009). Localized data on lake development is less available in the East
Gobi, but increased humidity after 8.9 kya and the spread of a woodland-steppe mosaic
just south of the study region (Wang et al., 2001; Peng et al., 2005; Jiang et al., 2006)
indicates an earlier Holocene climatic amelioration.
Unlike in the East Gobi, western areas of the study region were not intensively
occupied during the height of effective moisture. By 6.0 kya, when intensive habitation
of western lowlands began, lake levels had declined from early Holocene highs
(Herzschuh et al., 2004; An et al., 2008; Hartmann and Wünneman, 2009). Increased
345
aridity in the context of a humid early to mid-Holocene climate would not have been
detrimental to human habitation as effective moisture and vegetative biomass would still
have far exceeded modern conditions. While Oasis 1 archaeological sites are usually
distributed around rivers and streams, Oasis 2 and 3 sites are found around both
river/stream and lake/wetland environments (Figures 4.1). This distribution is
particularly notable in the East Gobi (Figure 4.2). Newly developed wetlands around
lakes or slow-moving, shallow rivers would have provided a range of resources attractive
to hunter-gatherers, including reeds, tubers, waterfowl, eggs, and various small aquatic or
semi-aquatic animals. Nutrient-rich wetland soils supported rich vegetation. Wetlands
probably formed following initial infilling and periodic retreat of water, developing first
around shallow lakes, playas, or large pools in interdunal hollows. Here water levels
were more sensitive to annual changes in precipitation and temperature, quickly reaching
a maximal height and retreating to leave rich soils for the benefit of new vegetation and
associated fauna.
During the terminal Pleistocene, in the early stages of post-LGM climatic
amelioration, the Gobi Desert would have been typified by desert and desert-steppe
environments with lake levels supported by infilling from post-glacial melt (see Chapter
5). Where forest development did occur, we can expect a more open woodland mosaic to
have been present (Wang et al., 2001; Wang et al., 2010). Upland environments (>1200
m a.s.l.) are expected to have hosted a range of flora and fauna typical of montane
environments in Mongolia, including medium-sized ungulates like ibex (Capra sibirica),
and argali sheep (Ovis ammon). Lowland environments (< 1200 m a.s.l.) would have
346
offered diverse hunting opportunities in the form of camel (Camelus bactrianus), horse
(Equus ferus przewalskii), khulan (Equus hemionus), marmot (Marmota sibirica), and
various species of antelope, fox, and hare. Ostrich (Struthio sp.) eggs were collected, but
the large birds might also have been hunted (see Janz et al., 2009). An array of
seasonally available plant foods such as Allium, grass and legume seeds, and berries
would have probably been found across the desert-steppes, with a few zones of higher
diversity and productivity.
Distribution of resources during the middle Holocene was much different than
during the terminal Pleistocene and initial Holocene, and the landscape was clearly much
differently used. Forest development in upland zones during the Holocene (8.0 kya in the
East Gobi and 6.0 kya in the western regions) would have altered foraging opportunities,
creating more distinct upland and lowland foraging patches. Heightened moisture
availability and vegetative cover resulting from Holocene climatic amelioration would
have converted sand-covered lowland expanses – intersected with drainage channels, and
dotted with small lakes and interdunal hollows – into large heterogeneous oases of
heightened productivity surrounded by arid steppes. Some of these oases were
expansive. The dune-field/wetland habitat in the Gurnai Depression of the Alashan Gobi
stretched over 100 km (Maringer, 1950: 151-152). While this is not typical of all Gobi
Desert dune-field/wetland locales, dune-fields often covered several kilometres (see
descriptions in Nelson, 1925; Pond, n.d.).
347
Season
Year round
Spring
Summer
(end of
June to
September)
Fall
Winter
Wetlands
Roe deer (riparian)
Père David deer
Fish
Water
Reeds
Saplings (riparian)
Clay
Fuel (dry reeds)
Horses
(pregnant/newborn)
Waterfowl (April)
Eggs (May)
Young greens
Bear
Shellfish
Waterfowl
Fat rodents
Reptiles/
amphibians
Eggs
Green legumes
Allium greens
Chenopodium seed
(flour and jam)
Grass seeds19
(August-September)
Fat bear
Reptiles
Waterfowl
Legume seeds
Caraway seeds
Allium bulbs
Tubers/roots
Dune-fields
Windbreak
Fuel (saxual,
other shrubs)
Fuel (dry rushes)
Wild cat
Rodents
(hibernating)
Birds
Eggs
Herbs
Fat rodents
Wild cat
Birds
Reptiles
Eggs
Succulents
Niter fruit
(Nitraria sibirica,
AugustSeptember)
Berries
Herbs
Wild cat
Bulbs and tubers
Herbs
Plains
Camel
Khulan
Horses
Gazelle
Saiga
Raptors
Stone
Fuel (dung)
Khulan herds
Horses
(pregnant/newborn)
Fox/hare/marmot
Allium greens
Khulan pairs
Gazelle herds (P.
gutturosa aggregation)
Saiga (newborns)
Argali sheep
Fox/hare
Fat marmots/rodents
Reptiles
Green legumes
Allium greens/bulbs20
C. ammannii
Berries
Foothills/Mountains
Boar (forest)
Red deer (forest)
Argali sheep
Windbreak
Saplings
Fuel (dead trees,
leaves, etc.)
Stone
Ibex herds
Young greens
Horse herds
Gazelle herds (G.
subgutturosa
aggregation)
Argali sheep
Legume seeds
Allium bulbs
C. ammannii
Berries
Khulan herds
Horse herds
Gazelle herds (G.
subgutturosa
aggregation)
Fur mammals
C. ammannii
Female ibex
Nuts (acorn/pine)
Legume seeds
Allium bulbs
C. ammannii
Herbs
Female ibex
Wild cat
Hare
Green legumes
Allium greens
Flour tubers
(Rheum nanum)
Berries (Ribes
altissimum)
Convoluvus
ammannii
Herbs
Hibernating bear
Ibex herds
Argali herds
Fur mammals
C. ammannii
Water (snow)
Table 6.1 Possible seasonal distribution of raw material and edible resources.
19
Various types of grass seeds unique to the Mongolian Gobi Desert would have been potential sources of
carbohydrates for hunter-gatherers, including Eragrostis pilosa (same genus as teff), and two types of
indigenous barley - Hordeum bogdanii and Hordeum brevisubulatum (see Jigjidsuren and Johnson, 2003).
Seed maturation occurs in late August and September.
20
Allium greens are still an important plant food for herders, as described by Khasbagan et al. 2000.
348
Our understanding of species composition within each ecosystem is limited by a
scarcity of zooarchaeological remains, a lack of local palaeobotanical studies, and the
absence of modern analogs. However, based on modern studies of plant and animal
distributions, there are likely patterns of seasonal resource availability according to
different ecozones. Table 6.1 summarizes hypothesized dispersal of key raw material,
plant and animal resources based on our knowledge of early to middle Holocene
palaeoenvironments, current home ranges, and environmental preferences (see Allen,
1934; Jigjidsuren and Johnson, 2003; Batsaikhan et al., 2010).
Woodlands would have offered a distinct foraging patch along rivers and in high
elevations. Forest environments are notable for the availability of certain raw materials
such as bark and wood, and offer much different foraging opportunities than steppe and
dune-field/wetland environments. Species that are now confined to the forests of
northern Mongolia may have been present in Gobi Desert forests during the middle
Holocene, including wild boar (Sus scrofa nigripes), and red deer (Cervus elaphus). Roe
deer (Capreolus pygargus or C. c. tianschanicus) could have inhabited riparian and
transitional woodlands, including in lowland settings. Edible and fur-bearing species
such as bear (Ursus arctos baikalensis, Ursus arctos isabellinus [or U. a. gobiensis]),
raccoon dog (Nyctereutes procyonoides), marten (Martes spp.), weasel (Mustela spp.),
and lynx (Lynx lynx) are all expected to have extended their ranges beyond modern
boundaries and were probably much more common in Gobi Desert riparian woodlands or
upland forests than modern zoogeography suggests (see Batsaikhan et al., 2010 for
description of habitats). Territories of ungulates adapted to upland grassland
349
environments, such as ibex and argali sheep, would have been more confined. Edible
vegetation would have included berries and tree nuts. Based on the current distribution
of Pinus pumila21, a shrubby nut-bearing species of pine tree, pine-nuts might have been
an important resource for Holocene hunter-gatherers in both upland and lowland forests.
Abundant Quercus pollen, identified in mid-Holocene palaeosols of the Hulun Buir sandy
land (Winkler and Wang, 1993), also suggests that acorns were available in parts of the
East Gobi.
Wetland habitats would have offered a very different range of foods. As uniquely
rich transitional zones between dry land and water, boasting a high diversity of plant and
animal species, they could play a key role in foraging strategies (Nicholas, 1998). Within
arid environments, they are especially notable for offering a stable water source.
Waterfowl, many small rodents, fish, and shellfish are confined to oases of swamplands,
and the marshy margins of lakes and rivers. Carnivorous and omnivorous species,
including humans, are drawn to wetlands in search of such prey. Ungulates tend to be
present around wetlands. In the Gobi Desert, they might have included roe deer in more
wooded settings, and Père David deer (Elaphurus davidianus)22. Edible plant foods like
tubers, cattail pollen, and grass seeds are also more abundant around wetlands. Clay,
21
Pinus pumila is currently distributed across northern Mongolia, Inner Mongolia, and PRC provinces of
Heilongjiang and Jilin. They grow up to 6 m in height, but have creeping branches and (Wu and Raven,
1999).
22
Père David deer (milu/sibuxiang) are large, marshland-dwelling ungulates. They are no longer found in
the wild, but their natural habitat is reed-covered marshes. Although the species are now confined to the
Dafang Nature Reserve in southern China, faunal remains were recovered from the North China Neolithic
site, Nanzhuangtou, along with wolf, pig/boar, dog, red deer, roe deer and fowl (Lu, 1999). Notably, this
suite of faunal remains suggests both wetland and forest exploitation.
350
reeds, and saplings can also be important raw materials for constructing firepits,
containers, hunting equipment, and clothing.
On the adjoining plains, an array of large- and medium-bodied ungulates would
have complemented dune-field/wetland resources. At present, the desert-steppes are
populated by herd ungulates prone to seasonal aggregations of tens (Camelus bactrianus
ferus, Equus przewalskii), hundreds (Equus hemionus – winter/spring; Gazella
subgutturosa – autumn/winter), or even thousands (Procapra [Prodorcas] gutturosa –
especially mid-June/July) of individuals (Allen, 1934; Batsaikhan et al., 2010). Due to
high annual variation in the distribution of resources, herd size and migration are
currently unpredictable. We can not know if they were more predictable in prehistory.
Hunting animals from large herds could have provided high returns and would have been
less risky than the pursuit of solitary animals, but it is not known if hunter-gatherers
attempted to predict and target herds or simply hunted during opportunistic encounters.
In either case, the presence of equid bones in dune-field/wetland sites suggests that largebodied ungulates were available to hunters stationed in those habitats.
Positioning residential bases around lowland dune-field/wetlands would have
allowed hunter-gatherers shelter within the stable dunes, a good view of the surrounding
landscape, and easy access to several key ecozones. Longer-term habitation was possible
due to the range of available resources in the nearby plains, riparian woodlands, wetlands,
and lakes. The availability of plant foods, such as grass seeds, tubers, fruits, and berries,
would have been spatially and temporally predictable. Environmentally restricted small
animal species like waterfowl (and their eggs) and fish would have been present on a
351
reliable schedule. Species like carnivores and reptiles are behaviourally predictable
according to season, but are more dispersed and likely to have been hunted in the course
of foraging for more predictable foods (see Bird et al., 2009). Early to mid-Holocene
lowland dune-field/wetlands in the Gobi Desert would have been characterized by a
range of reliably available resources. Although forests contain an array of important
edible plant and animal species, the reduced accessibility, lack of mobility, and hunting
inefficiency in forests can make them less ideal environments for residential bases (see
Winterhalder, 1981).
6.2. “Dune-dweller” foraging strategies
Despite evidence for exploitation of low-ranked foods (i.e., foods with low caloric returns
compared to energy expended in capture and processing) prior to Oasis 2 and Oasis 3,
site distributions for those periods indicate a substantial change in the way that dunefields and wetlands were used. The discovery of terminal Pleistocene sites in arid
Northeast Asia containing grass seeds (Derevianko et al., 2008) or technologies
associated with processing low-ranked foods (Elston et al., 1997; Bettinger et al., 2007;
Derevianko et al., 2008; Elston et al., 2011; Chapter 3), makes it is clear that such
resources were exploited in the Gobi Desert during Oasis 1; however, the preferential
distribution of sites in upland environments suggests that their use was not a decisive
factor in the positioning of residential bases. In contrast, Oasis 2 and Oasis 3 residential
sites were clearly concentrated in lowland dune-field-wetlands.
352
The presence of Oasis 2 and Oasis 3 task sites in a wider range of habitats (Figure
4.7 and Table 4.9) indicates that task groups were used to procure resources from dunefield/wetland resources, nearby steppes, and more distant upland forests (e.g., ungulate
meat, furs, tool stone, medicines and other highly valued plants). Residential habitation
of uplands and open steppes appears to have been heavily restricted. Evidence of steppe
species in dune-field/wetland assemblages along with the lack of habitation sites in such
environments underscores the likelihood that they were exploited primarily by short-term
task groups.
Unlike wetlands, montane environments are often typified by highly seasonal,
homogeneous and dispersed resources, which favour high residential mobility (Morgan,
2009). Depending on the exact composition of woodland species and the density of
forest growth, forestation might have increased primary productivity of upland zones but
reduced accessibility, ease of mobility, and hunting efficiency (see, for example, Cree in
Winterhalder, 1981). Dune-field/wetland bases offered access to and ease of mobility
across multiple ecozones that would have been highly appealing. Furthermore, while
upland forests would have offered a range of hunting and foraging opportunities,
including solitary ungulate species, berries, and carbohydrate rich plant foods, such
resources would be nutritionally redundant and less diverse when compared to the
composition of mixed lowland (dune-field/wetland combined with arid/desert-steppe)
resources (Table 6.1). The range of medium- to large-bodied steppe-dwelling ungulates
in proximity to reliably available plant and small animal species in lowland dunefield/wetlands may have limited the exploitation of upland zones. Longer term
353
exploitation of forest resources would have necessitated smaller group sizes, and more
frequent moves. Residential B and task sites are more characteristic of a site structure
resulting from such a strategy.
The occurrence of longer-term habitation sites in lowland dune-field/wetlands is
consistant with foraging theory, which predicts that the use of such rich foraging patches
results in less frequent moves due to high overall return rates and the avoidance of
productivity loss during travel time between patches (MacArthur and Pianka, 1966;
Charnov, 1976; Kelly, 1995: 90-97). At the same time, hunter-gatherers still relied on
resources that could not be obtained in these habitats, such as tool stone. Task groups
could have effectively extracted key resources from more distant environments while
allowing other group members to focus on the procurement of plant and animal foods
closer to the residential base. The wide array of resources characterizing early to midHolocene environments and the diverse range of skills required to exploit them would
have favoured a division of labour (Lupo and Schmitt, 2002; Kuhn and Stiner, 2006),
while the use of diverse plant and animal species rather than a reliance on large game
might have allowed for the support of larger group sizes (O’Connell, 2006).
Since the majority of Oasis 2 and Oasis 3 residential sites are situated in dunefield/wetlands, which boasted a wealth of low-ranked resources, it is reasonable to assert
that they would have made active use of small animals and plant foods.23 Use of both
low-ranked and high-ranked foods in the post-LGM diets of Gobi Desert hunter-gatherers
23
Bird (1999: 65) in her article on the sexual division of labor provides a list of ethnographic foraging
strategies that exemplifies the importance of both low- and high-ranked species in a range of settings,
illustrating the ubiquity of low-ranked resources in the diet of hunter-gatherers with access to them.
354
is further supported by the faunal assemblage from Chilian Hotoga, which includes equid,
fox, frog, and bird remains. The use of grinding stones could be related to processing
seeds and/or tubers, while knobbed rollers or pestles from the East Gobi may have been
used for grinding nuts (sensu Wright 1994). Pottery might also be related to exploitation
of low-ranked resources (see Brown, 1989; Hoopes, 1995). Finally, cordage and textile
impressions on pottery suggest that hunter-gatherers had the appropriate technology to
construct nets, which are associated with hunting of small prey (see Soffer, 2000; Lupo
and Schmitt, 2002).
The incorporation of “low-ranked” or “high cost, low return” resources into postLGM foraging strategies defines Epipalaeolithic hunter-gatherers (see Kuhn and Stiner,
2001). Flannery’s (1969) Broad Spectrum Revolution hypothesis proposed that the trend
towards increasing dependence on a widening range of plants and animals was related to
higher population density and increased climatic variability following the LGM. The
Broad Spectrum Revolution was seen as key step in the trajectory towards the agriculture.
The work of Stiner and colleagues supported this theory, showing that animal species
with low caloric returns in relation to energy expended were progressively incorporated
into human diets throughout the Upper Palaeolithic (Stiner et al., 1999; Stiner et al.,
2000; Stiner and Munro, 2002; Munro 2004; Stiner, 2005). Resource depression, as
recognized by the decline of high-ranked prey (i.e., prey that has high caloric return rates
relative to handling costs), was thought to have stimulated increased diet breadth,
suggesting a correlation with demographic packing (Binford, 1968; Flannery, 1969;
Cohen, 1977; Keeley, 1988; Winterhalder and Smith, 2000; Stiner et al., 2000). This
355
relationship has been attested in the archaeological record (Kozłowski and Kozłowski,
1986; Stiner, 2001; Munro, 2004).
More specialized technology and intensive processing are often associated with
diminishing resources, particularly when growing populations or other pressures make it
necessary to acquire more food energy out of the same unit of land (Binford, 1968;
Flannery, 1969; Cohen, 1977; Keeley, 1988; Stiner et al., 2000; Stiner, 2001; Stiner and
Munro, 2002). However, the temporal correlation between Gobi Desert dunefield/wetland habitation and a presumed increase in productivity related to climatic
amelioration indicates that a decline in the regional abundance of plant and animal
species was unlikely. Locally, population density may have increased, but this is difficult
to ascertain since Oasis 1 sites may be either more ephemeral in terms of their material
culture or simply located in less well-surveyed regions. A stable balance between
population density and carrying capacity is suggested by continued high residential
mobility and the lack of territorial behaviour (e.g., the construction of visible graves or
other monuments). Decreased residential mobility probably resulted initially from a
change in resource distribution at the transition to from Oasis 1 to Oasis 2, but constricted
foraging territories would certainly increase local population density, effectively resulting
in resource depression, even if region-wide demography was unaffected.
At the same time, the use of low-ranked resources may not be as reliant on
resource stress as previously modeled. New data has begun to suggest that low-ranked
foods were often targeted alongside high-ranked foods, even in the absence of resource
depression (Elston and Zeanah, 2002; Bird and Bleige Bird, 2005; Bird et al., 2009;
356
Starkovich and Stiner, 2009; also see Revedin et al., 2010). The idea that reliably
obtainable, lower-ranked plant and animal species might have been targeted despite the
availability of high-ranked species is supported by ethnographic data from Australia’s
Western Desert. Researchers observed that when Martu hunters were engaged in the
pursuit of one species, they would forgo higher-ranked species if capture was riskier
(Bird et al., 2009). Therefore, the opportunity to reliably procure lower ranked species
can potentially outweigh the more precarious potential involved in pursuing more mobile
high-return prey types (see also Hawkes et al., 1982: 392). Additionally, the pursuit of
certain low-ranked resources like grass seeds is probably even more underrepresented in
the ethnographic record due to the historical ubiquity of flour rations and other
commercially available starches (e.g., O’Connell and Hawkes, 1981; see also Hawkes et
al., 1982: 384). These data suggest that resource depression may not have been required
to stimulate a focus on dune-field/wetland foods.
A lowland centred pattern of land-use, incorporating exploitation of low-ranked
wetland species, is not unique to the Gobi Desert. Like the Gobi Desert, the Great Basin
in the western United States is an arid, internally-drained basin-range environment. PreArchaic land-use strategies proposed by Elston nd Zeanah (2002) provide an intruiging
comparison. During the Pleistocene-Holocene transition, Pre-Archaic hunter-gatherers
exploited a broad spectrum of food resources that included mountain sheep, elk, bison,
antelope, small mammals, birds, fish, and shellfish (Beck and Jones, 1997; Elston and
Zeanah, 2002). Although small milling stones are occasionally present, the emphasis on
formal, hafted tools such as points, bifaces and scrapers suggests a higher investment in
357
hunting, and minimal investments in the seed processing and storage that defines Archaic
groups. High residential mobility is indicated by the relative rarity and low density of
Pre-Archaic sites, low variability among lithic assemblages, and a lack of residential
structures, middens, or storage facilities. The most extensive Pre-Archaic sites are often
found along lowland beach bars or lunettes associated with pluvial lakes or marshes,
elevated surfaces on valley margins, or Pleistocene stream terraces, suggesting targeted
exploitation of lake/marsh resources and mid- to low-elevation steppe (Elston and
Zeanah, 2002).
Various models from behavioural ecology have been applied to the study of PreArchaic Great Basin hunter-gatherers, including the diet breadth model, patch choice
model, and the Z-score model (O’Connell et al., 1982; Simms, 1987; Elston et al., 1995;
Pinson, 1999); however, all models failed to predict the incorporation of high-cost/lowreturn resources like seeds into Pre-Archaic, Archaic, or ethnohistoric Great Basin diets
(Elston and Zeanah, 2002). Elston and Zeanah (2002) propose that frequent moves by
Pre-Archaic hunter-gatherers from basin to basin represent a complementary set of
strategies that maximize large-game encounters while also focusing on reliable lower
return species like small game and plants. Early Holocene lowland habitats are
envisioned as highly productive environments which allowed low-density populations to
move easily from patch to patch in order to increase frequency of encounters with large
game. Hunters are expected to have targeted large game in low- to mid- elevation brushy
steppe from fall to spring, when they could be easily hunted. Group members not
actively involved in large game hunting could focus on seeds, small game, waterfowl,
358
and fish (Elston and Zeanah, 2002). Extensive aridification during the middle Holocene
desiccated many Great Basin wetlands, encouraging the establishment of Archaic period
residential sites and seed storage facilities around the few remaining wetlands and
perennial springs. Material culture from the Archaic period reflects decreased mobility in
the decline of formal chipped stone industries and the proliferation of milling stones
(Elston and Zeanah, 2002).
There are many notable similarities between the early Holocene archaeological
records of the Great Basin and the Gobi Desert. The use of lowland environments along
with indicators of high residential mobility and largely formal chipped stone technology
are key characteristics in both. A broadly similar pattern of complementary steppic
hunting and wetland foraging has also been recognized. Two important divergences from
the Great Basin record illustrate unique aspects of Gobi Desert foraging: the use of
formal milling technology and pottery during Oasis 2 in the East Gobi; and distinct
artefact assemblage variability between sites in all target regions.
Both of these differences in technological organization probably relate to the
relative importance of wetland resources within the broader subsistence strategy. Based
on the time consuming manufacture of technologies associated with processing dunefield/wetland resources, low-ranked foods would have been more central to Gobi Desert
foraging strategies. Clear differences between residential and task sites indicate that
mobility was still organized more logistically than has been proposed for the Pre-Archaic
Great Basin, but a lack of middens, storage facilities and other architectural features at
Oasis 2 and Oasis 3 sites indicate that hunter-gatherers were still highly mobile.
359
Likewise, although formal and highly portable hunting technologies were maintained
during Oasis 2 and Oasis 3, the diversified lithic assemblages (a combination of
microblade core technology, informal flake/core technology, bifaces, polished stone
adze/axes, and grinding stones) suggest that a wider range of tasks were prioritized and
carried out by group members.
Lower residential mobility in connection with dune-field/wetlands is reflected not
only in the less conservative use of tool stone in some residential sites (see Chapter 4),
but also in the manufacture of large formal grinding tools (e.g., saddle querns, polished
adze/axes, knobbed rollers or pestles) in the East Gobi. Polishing can be accomplished
over time, as smaller tools are transported from site to site, but it is very time consuming.
Large grinding stones are difficult to transport at all without domesticated beasts of
burden. As such, sufficient “down-time” is required for their manufacture. Pottery
manufacture also requires some amount of down-time since mixing clays, building pots,
drying, and firing are all time-consuming tasks that need to be undertaken before pottery
is transported (Arnold, 1985; Brown, 1989).
Although pottery manufacture does imply occasional periods of at least short-term
sedentism (a week or two might be sufficient), transport costs probably do not reflect on
mobility. Studies of pottery-use amongst late Holocene hunter-gatherers in the western
Great Basin indicate that the pottery was used primarily for boiling seeds and levels of
pottery production were not related to residential mobility (Eerkens, 2003). Eerkens
(2003) suggests that pots were cached in low elevation wetlands, where they were
returned to and regularly used. As such, an increase in the importance of pottery during
360
Oasis 3 does not necessarily imply a progressive decrease in residential mobility. During
Oasis 3, the decline of formal grinding technology in the East Gobi, and the use of highly
portable “rubbing stones” across the Gobi Desert, suggests an increase in residential
mobility (i.e., a trend towards more portable technology).
It is not known if Gobi Desert groups transported pottery from site to site, but the
possibility that they could have cached them for return visits, as East Gobi groups almost
certainly would have done with large grinding stones, is intruiging. Investment in
processing technology and caching of site equipment indicates that hunter-gatherers were
at least seasonally tied to specific predictable resources rather than foraging on an
encounter basis (Binford, 1979, 1982). Cached grinding stones and pottery at dunefield/wetland locales implies that hunter-gatherers were heavily invested in the lowranked foods that they were processing.
At the same time, the persistence of microblades and projectile points is
consistent with a focus on high residential mobility and the exploitation of unpredictable
high-ranked resources (e.g., large-game). Microblade cores best exemplify highly
portable, flexible core-tools, which produce standardized components for easilymaintained composite tools (see Elston and Brantingham, 2002). Maintainable and
flexible hunting equipments are associated with “search and encounter” procurement –
where tools can be maintained and employed on a daily basis and in whatever capacity
required – rather than with specialized and predictable use (Bleed, 1986; Ellis, 2008).
Reduction strategies represented in Gobi Desert assemblages make conservative use of
raw materials, suggesting limited access to tool stone associated with persistently high
361
residential mobility (see Chapter 4). Taken together, the use of grinding stones, pottery,
microblade cores, and biface technologies indicate that hunter-gatherers strategically
targeted both seasonally reliable (typically low-ranked), and more unpredictable
(typically high-ranked) prey.
The organization of foraging strategies based on reduced seasonal mobility is
reflected in lithic assemblages. Informal cores are common in Gobi Desert assemblages
and are often associated with reduced mobility since they represent a less conservative
approach to flake production. While higher relative frequencies of microblade cores
should be more commonly associated with high residential mobility, informal cores are
expected to occur in relatively higher frequencies with increasing sedentism. Table 6.2
demonstrates the relationship between the frequency of informal core types and site type.
Residential B sites have significantly fewer informal cores types and notably higher
frequencies of microblade cores than Residential A sites. Task sites are similar in core
type frequencies to Residential A sites. When dune-field/wetland sites are removed from
the sample, there is no significant variation in the relative frequencies of core types
between any types of sites. All sites outside of dune-field/wetland habitats have lower
frequencies of informal cores and higher frequencies of microblade cores.
362
Site type
Mean % informal cores
(all zones)
Mean % microblade cores
(all zones)
Mean % informal cores
(outside wetlands)
Mean % microblade cores
(outside wetlands)
Residential A
33.5
Residential B
16.6
All task sites
31.8
58.1
74.7
60.6
29.5
29.0
45.8
60.3
49.0
45.5
Table 6.2a Mean percentage of core types according to site type for Oasis 2 and Oasis 3,
all Gobi Desert sites.
All site types
(all zones)
Residential site
types
(all zones)
All site types
(outside wetlands)
Residential site
types
(outside wetlands)
Mean % informal
cores
0.0929
Mean %
microblade cores
0.1761
0.0453
0.0812
0.894
0.437
0.4249
0.6518
Table 6.3b P-values associated with Table 6.2a.
Non-wetland Residential B sites have higher frequencies of informal cores and
lower frequencies of microblade cores than is typical of wetland-based Residential B
sites. Decreased access to raw materials may suggest Residential B sites within dunefield/wetland environments are part of a higher mobility strategy, perhaps more focused
on search and encounter foraging than was typical for Residential A occupations or nonwetland task sites. Such sites, indicative of more intensive raw material conservation,
might be related to unprovisioned field camps radiating from Residential A type
habitations. Task sites and Residential A sites show comparable access to raw materials,
363
suggesting that they are related to either raw material procurement or that the groups
were well-provisioned by associated residential sites, being so short term that raw
material conservation was of little consequence.
Lack of variation in the frequency of core types outside of wetland environments
is telling, particularly in the case of Residential A and B sites. Within dune-field/wetland
environments the larger multipurpose Residential A sites indicate better access to raw
materials and more variety in on-site activities. The few residential sites outside lowland
dune-field/wetland zones appear to have more limited access to raw materials. They
were probably not provisioned for reoccupation and may have been related to shorterterm seasonal occupation of habitats that supported smaller populations (i.e., fewer
people for a shorter length of time).
Table 6.3 further illustrates that food processing technologies like pottery and
grinding stones were more common in wetland environments, where vegetative biomass
was most concentrated. The exception to this is that task sites with grinding stones were
always located outside wetlands. Milling activities within wetlands were likely carried
out at residential sites. It is also clear that pottery and grinding stones were not always
associated with residential sites in wetlands: 5% of Residential A wetland assemblages
and 12% of Residential B wetland assemblages contained neither pottery nor grinding
stones (Table 6.4). The use of grinding stones and pottery was restricted to wetlands in
Oasis 2, but use expanded beyond areas of primary plant productivity during Oasis 3
(Table 6.5).
364
Site type
Residential A (all)
Total = 25
Residential A (in wetlands)
Total = 19
Residential A (outside wetlands)
Total = 6
Residential B (all)
Total = 21
Residential B (in wetlands)
Total = 17
Residential B (outside wetlands)
Total = 4
Task site (all)
Total = 31
Task site (in wetlands)
Total = 11
Task site (outside wetlands)
Total = 20
Sites with pottery
N=
%
18
72%
15
79%
3
50%
17
81%
15
88%
2
50%
7
23%
4
36%
3
15%
Sites with grinding stones
N=
%
14
56%
12
63%
2
33%
6
29%
5
29%
1
25%
3
10%
0
0%
3
15%
Table 6.3 Number of Gobi Desert Oasis 2 and Oasis 3 sites with pottery and grinding
stones according to site type.
365
Site type
Residential A (all)
Total = 25
Residential A (in wetlands)
Total = 19
Residential A (outside wetlands)
Total = 6
Residential B (all)
Total = 21
Residential B (in wetlands)
Total = 17
Residential B (outside wetlands)
Total = 4
Task site (all)
Total = 31
Task site (in wetlands)
Total = 11
Task site (outside wetlands)
Total = 20
Sites without pottery or grinding stones
N=
%
3
12%
1
5%
2
33%
3
14%
2
12%
1
25%
22
71%
7
64%
15
75%
Table 6.4 Number of Gobi Desert Oasis 2 and Oasis 3 sites with neither pottery nor
grinding stones according to site type.
Period/
Tool type
Oasis 2
Pottery
Grinding
Oasis 3
Pottery
Grinding
Lowland dunefield/wetland
N=
Lowland river
Lowland dry
Upland
Upland dry
N=
N=
N=
N=
11
11
1
1
0
0
0
0
0
0
29
8
2
3
1
0
3
2
3
1
Table 6.5 Number of Gobi Desert sites with pottery and grinding stones according to
ecozone and period.
366
Core reduction strategies and the distribution of pottery and grinding stones both
support a pattern of land-use characterized by provisioning of dune-field/wetland bases,
which were occupied for longer periods of time and related to the intensified use of lowranked resources. These data indicate significant differences between Residential A and
Residential B type sites and strongly support, as do the data presented in Chapter 4, the
probability that a pattern of radiating mobility was practised at least on a seasonal basis.
According to Binford’s (1980) original definition, this type of land-use would be
characterized by both short-term task groups (task sites) and longer term field camps
(Residential B sites), where select members procured resources for a larger group
associated with a central base camp (Residential A sites). Such a pattern of resource
exploitation is based on the movement of goods to consumers (Binford, 1980). The
scarcity of residential sites outside dune-field wetland environments suggests that that
primary habitation occurred in wetlands year round with varying degrees of mobility
according to the season. In seasons of lowered productivity, the larger group may have
dispersed and followed a pattern of circulating high residential mobility still based in
dune-field/wetlands and characterized by task sites and short-term residential bases
(Residental B sites).
Inferred distribution of various plant and animal resources across environmental
zones give us some indication of possible variability in seasonal land-use. Based on the
ubiquity of grinding stones and pottery, Residential A sites in the East Gobi appear to
have been focused on intensive plant processing, suggesting a possible late summer/early
fall aggregation coinciding with seed and tuber harvesting. Berries and fruit would also
367
have been available around dune-field/wetlands in the late summer. Nearby lowland
steppes would have been ideal locales for the procurement of mammals such as fattened
marmots and herd ungulates. According to modern herd behaviour, the aggregation of
human groups could have corresponded with enormous increases in Procapra gutturosa
herd sizes24, though it is not known how similar herd behaviour was in comparison to
modern times. A distinct decline in the diversity of plant and animal resources occurs in
the winter and spring months, and hunter-gatherers may have needed to increase mobility
to take advantage of more dispersed resources.
Conversely, a decline in residential mobility might also have occurred throughout
the winter months with a return to high residential mobility during the spring and early
summer. Dune-field and wetland foods would have been less readily available in winter
months (see Table 6.1), while some large ungulates like khulan, horses, ibex, and argali
tend to aggregate in the steppes and foothills during the fall and winter (Batsaikhan et al.,
2010). Extremely cold winter conditions might have limited mobility, although
precipitation is greatly reduced in this season and the cold would have been more of a
hindrance to travel than snow. In this case, longer-term seasonal occupation could have
commenced in the late summer, with abundant summer resources being harvested,
processed, and stored for consumption during cold winter months. Logistical habitation
of wetland sites may also have been favoured with return of fatty waterfowl to wetlands
24
Herd sizes of up to 250,000 individuals have been reported (Olson et al., 2009).
368
in April25. Advantages to high mobility during the spring and summer would include
greater ease of mobility, predictable availability of low-ranked resources around dunefields and wetlands, and the ability to procure a wide array of resources across
environmental zones.
Archaeological sites typified by a lack of middens and residential structures do
not necessarily preclude prolonged seasonal occupations. Modern nomadic pastoralists
in the region live in organic portable dwellings (gers or yurts) and leave little refuse upon
their departure. High density sites would probably have accumulated only after many
successive occupations. Neolithic/Eneolithic lithic reduction strategies do indicate the
importance of technologies suitable for high residential mobility and unpredictable
foraging situations, which supports the probability that even seasonal sedentism was
limited. Although periods of reduced mobility are suggested, high residential mobility
should be considered characteristic of post-LGM Gobi Desert hunter-gatherers.
6.3. Oasis 3 and the rise of nomadic pastoralism
Land-use and subsistence in the Gobi Desert during Oasis 3 should inform our
understanding of the processes leading to the rise of nomadic pastoralism, including
developments like the construction of burial monuments and the widespread adoption of
domesticated herd animals. The end of Oasis 3 is contemporaneous with the earliest
25
Decreased body fat in ungulate species during spring can be a severe problem in high latitude
environments. Fats and carbohydrates needed to metabolize meat proteins are especially difficult to
procure this time of year and wetland resources like fattened waterfowl, eggs, tubers or stored
carbohydrates may play an important role in reducing seasonal stress. See Speth and Spielmann, 1983; and
Malainey et al., 2001.
369
Bronze Age burial structures and monuments in Mongolia, a rise in the symbolic
importance of horses in China and Mongolia, and the spread of herd animals throughout
the agricultural regions of China (see Chapter 2). By 3.0 kya, nomadic pastoralism was
widespread in Northeast Asia. However, only Alashan Gobi lithic assemblages reveal
any statistically significant divergence between Oasis 2 and Oasis 3 that might be related
to shifts in residential mobility or raw material access (Table 4.13). Alashan Gobi lithic
assemblages imply increased residential mobility during Oasis 3. A lack of similar
evidence in the East Gobi and Gobi-Altai does not preclude shifts in land-use or
subsistence. High residential mobility and the centralized use of dune-field wetlands
appear to have continued in all three target regions until the end of Oasis 3. Despite some
intriguing finds such as clay spindle whorls, painted pottery (with very similar motifs to
those used by agropastoralist groups farther east), and copper slag, a lack of conclusive
evidence for early pastoralism limits our ability to model the transition from Oasis 3
hunter-gatherers to Bronze Age pastoralists.
Nevertheless, there were a number of widespread shifts in material culture during
Oasis 3. Residential sites in the East Gobi less often contain the large formal grinding
tools that characterize Oasis 2 settlement. Surface treatments on pottery were more
varied and pottery was more widespread during Oasis 3. These two developments might
be related to a decline in the importance of milling or a preference for seed boiling.
Grinding stones and pottery were also dispersed across a wider range of environments in
Oasis 3 than in earlier periods, suggesting that extensive food processing was less
spatially constrained. Other aspects of material culture exhibit variation from earlier
370
styles, including: fully and finely polished adzes/axes; the introduction of high-fired
pottery26, along with more extensive variation in decorative finishes and vessel forms; the
increased visibility of bead-making on both ostrich eggshell and stone27; the more
frequent use of specialized hafted microblade tools like shouldered drills and endscrapers
on microblades; and the introduction of end-hafted bifacial curved knives on bladelets or
thin, elongated chalcedony nodules that may have replaced side-hafted bifacial inset
knives.
One of the most striking aspects of technological change in Oasis 3 is the
emphasis on decorative elements. Manufacture of finely polished adze/axes and eggshell
and stone beads are especially time-consuming tasks. Burnishing, painting, and creating
raised and moulded rims on high-fired Oasis 3 pottery require additional effort, but
produce a more striking appearance than that of earlier vessels. Finer tools like
shouldered drills and endscrapers on microblades would have been used in more detailed
tasks, which may have included some type of decorative work (e.g., drilling holes in
beads, engraving, and carving). Though not directly related to mobility and land-use,
increased emphasis on decorative arts underscores a shift in time allocation. High-fired
pottery and finely polished tools suggest that durability was valued and longer-term
curation intended.
A peak in the production of traditional goods has been frequently witnessed in
situations of new contacts between hunter-gatherers and food producing groups (see
26
Painted pottery and hard red-wares were probably fired at temperatures between 900-1000oC, based on
estimates by Palmgren (1934) of Chinese Neolithic Majiayao Banshan-type pottery.
27
As exemplified by Alashan Gobi assemblage K: 13230 (Maringer, 1950: 109-110).
371
examples in Sadr, 2005), and the shift in production of material culture might be related
to contact with neighbouring herding groups as their influence in the region increased.
Though there is currently insufficient evidence to support a claim for trade between Gobi
Desert hunter-gatherers and nearby groups, it is not unusual for hunter-gatherers to trade
local products with food producing neighbours (Lukacs, 1990; Spielmann and Eder,
1994; Junker, 1996, 2000; Zvelebil, 1996; Sadr, 2005). Furs, feathers, skins, tool stone,
clay, wild meat, turquoise, raw copper, chalcedony beads (such as those being
manufactured at cave site K: 13230 on the Ukh-tokhoi Plateau), and seasonal labour (see
Paterson, 2005) are all products that might have been appealing to contemporary
pastoralist, agro-pastoralist, and even agriculturalist neighbours. Products like
grain/flour, pottery, bronze tools or ornaments, milk products, spun hemp, wool, and even
domesticated animals might have been valued items among hunter-gatherer groups. Such
interactions could have facilitated initial introduction of domestic herd animals into the
Gobi Desert (although some might argue that camels may have already been
domesticated during Oasis 3 as outlined in Chapter 2).
Few habitation sites can be reliably attributed to the Bronze Age. Most
assemblages from Gobi Desert collections are consistent with Oasis 2 and Oasis 3. The
scarcity of recognizable Bronze Age habitation sites suggests a decline in population
density after Oasis 3, a substantial shift in material culture and/or settlement that made
Bronze Age sites more ephemeral, or simply the absence of clear differences between
Oasis 3 and early Bronze Age sites. Increased mobility and a decline in the use of lithic
372
technology, both of which are expected to have occurred with the rise of pastoralism,
could have contributed to reduced visibility in the archaeological record.
Dottore-namak (K: 13248; Maringer 1950: 127) is the only dated Bronze Age site
in this sample and is no later than 2.5 ka (see Table 3.1). The pottery is a high-fired redware tempered with coarse sand and decorated with a moulded band on the shoulder
(Figure 6.1). Two fragments of pottery with copper slag melted into the exterior surface
were recovered from the same site around a spring in the Goitso valley (the oasis-lined
southern edge of major depression, along the northern edge of which Yingen-khuduk and
many other such sites were discovered). Dottore-namak exhibits the same microblade
core reduction technology used during the Neolithic/Eneolithic. The presence of slag and
a distinct pottery type are defining features. Core rejuvenation spalls were the only
evidence of microblade cores in the assemblage so core morphology was not clear. The
assemblage is significant in that it represents a very low density (98 artefacts recovered
[Maringer, 1950: 127]) pottery-bearing site from a valley near a spring, as opposed to a
dune-field/wetland environment. Dottore-namak exemplifies the trend towards more
even dispersal of specialized food processing equipment across ecozones that appears to
have begun in Oasis 3.
373
Figure 6.1 High-fired red-ware from the Bronze Age site, Dottore-namak.
New transportation aids could be responsible for this pattern. Beasts of burden
such as horses or camels could have allowed groups to travel more easily between distant
foraging patches, resulting in the higher residential mobility characteristic of Alashan
Gobi. The Alashan Gobi is notable in that Oasis 3 is characterized by increased mobility,
and dune-field/wetland use intensified during a relatively more arid phase than had been
experienced in the preceding millennia. Lake levels were still high and wetland
vegetation still abundant, however, desertification was probably occurring across the
lowland plains. This may have led to an intensification of dune-field/wetlands that was
less sustainable for longer-term habitation. Contacts with neighbouring pastoralist or
agropastoralist groups might have contributed to local solutions such as the adoption or
374
local domestication of horses or camels. A lack of distinct non-mortuary material culture
associated with the local Bronze Age may support such a model of indigenous
development. Since the use of domestic herd animals for transport would allow all
belongings to be transported during moves in a manner similar to modern pastoralists,
increased ubiquity of pottery and a decline in caching behaviour would be possible
signatures of early Bronze Age sites. Changes in the distribution of pottery and grinding
stones during Oasis 3 may represent the early presence of such elements.
If dune-field/wetland sites were still regularly exploited during the early Bronze
Age, the Oasis 3 signature could easily overlap with and obscure evidence for early
nomadic pastoralism. Sites like Dottore-namak could be confused with
Neolithic/Eneolithic assemblages, particularly in the absence of pottery. Residue analysis
of pottery vessels (Evershed et al., 2008) and sediment micromorphology from habitation
sites (Shahack-Gross et al., 2004) are two methods of recognizing herding signatures in
the archaeological record.
6.4. Conclusion
Despite the last century of heightened archaeological inquiry in East Asia, little work has
been done with the vast collections of archaeological remains from the Gobi Desert. The
dominance of surface assemblages and the inaccessibility of contextual data have
inhibited research. The primary goal of this study has been to mine the extensive existing
collections for a broad comparative sample of cross-regional data in order to provide a
strong interpretative foundation for post-LGM prehistory of the Gobi Desert. Museum
375
collections are shown to hold a diverse wealth of information that could not be replicated
under the contemporary logistical, political, and financial constraints of fieldwork. Most
importantly, the resulting observations of chronology and land-use among
Neolithic/Eneolithic Gobi Desert hunter-gatherers form a basis from which interpretive
investigation of field work and collections analysis can proceed.
Currently, we can assert the following:
1. Beginning around 8.0 kya, Gobi Desert hunter-gatherers began practicing a
mode of subsistence and land-use that diverged greatly from earlier periods.
2. The shift in land-use was related to widespread forestation by 8.0 kya in the
East Gobi and 6.0 kya in the west, along with the stabilization and increased
productivity of lowland habitats.
3. After about 8.0 kya, organizational strategies were centred on the logistical
exploitation of lowland dune-field/wetland and neighbouring ecozones.
Habitation was centred within dune-field/wetlands, while task groups
procured resources from a range of environments. Some locales were
provisioned for longer-term occupation. This pattern of land-use continued
until 3.0 kya.
4. The use of dune-field/wetlands between 8.0-3.0 kya is related to the
complementary exploitation of both unpredictable high-ranked species like
large ungulates, and reliable low-ranked dune-field/wetland foods.
5. Post-LGM hunter-gatherers maintained high residential mobility, but
residential mobility did decline slightly, at least seasonally, after 8.0 kya.
376
6. Due to increased effective moisture and shifts in vegetational distribution,
periodic seasonal reductions in residential mobility almost certainly occurred
by 8.0 kya in the East Gobi, and by 6.0 kya in the Gobi-Altai and Alashan
Gobi.
7. East Gobi sites represent an early intensive focus on milling during Oasis 2
(8.0-5.0 kya) that appears not been replicated in the Gobi-Altai or Alashan
Gobi, and did not continue into Oasis 3.
8. Alashan Gobi sites reflect decreased access to tool stone (either due to a
differential pattern in land-use or more uneven natural dispersal of raw
materials) that seems to have been accentuated during Oasis 3.
9. Oasis 3 represents a period of increased emphasis on decorative elements and
perhaps on the functional durability of certain artefact types such as pottery
and adze/axes.
Several avenues of future research would be highly desirable for testing this new
outline of chronology and land-use systems. First, collections research is necessary to
improve sample size and enable a more refined reconstruction of regional chronology and
organizational systems. Such research should include the integration of additional
chronometric dating. Second, residue analysis on pottery and grinding tools can allow
for a more direct understanding of subsistence and artefact functions, while
compositional analysis of clays (and complementary in-field sourcing studies) can
suggest patterns of transport or trade. Conscientious and conservative use of museum
collections should continue to be an integral part of such research. Third, more holistic
377
approaches to excavation are necessary and should include elements largely absent in
Mongolian archaeology, including flotation for palaeobotanical analysis, systematic
collection of sediments for soil micromorphology and luminescence dating, landform
geomorphology, and coring of nearby lake/playa sediments for complementary
palaeoenvironmental data. Finally, a tool stone sourcing studies would contribute greatly
to our understanding of regional procurement strategies, as would a generally more
detailed understanding of regional and local landscapes.
Due to both a unique geographic and cultural setting, the Gobi Desert region has
great potential for illuminating our understanding of human adaptational and behavioural
processes. This study is intended to lay a foundation for future research on post-LGM
Gobi Desert hunter-gatherers, and to contribute ideas and knowledge to a budding
interest in the transition to nomadic pastoralism in Mongolia (Wright, 2006; Houle,
2010). I have attempted to address the findings and hypotheses of past research;
however, limitations on language and the availability of certain publications might have,
on occasion, inadvertently thwarted my ability to be inclusive. It is my hope that the
models outlined here for chronology, land-use, and subsistence will be refined –
corrected and enriched – as more data becomes available and multiple voices emerge.
378
APPENDIX A – RESULTS AND CONTEXT OF DATED SITES
A.1. Dated Samples
REGION
SITE NAME/#
MATERIAL
METHOD
Jabochin-khure
ORIGINAL
CATALOGUE #
MFEA K.13203: 5
Alashan Gobi
ceramic
Luminescence
Gashun
MFEA K.13207: 1
ceramic
AMS
Yingen-khuduk
MFEA K.13212: 6
ceramic
Luminescence
MFEA K.13212: 123
ceramic
Luminescence
MFEA K.13212: 128
ceramic
Luminescence
MFEA K.13212: 184
eggshell
AMS
MFEA K.13248: 5
ceramic
Luminescence
MFEA K.13248: 6
ceramic
Luminescence
Mantissar 4
MFEA K.13290: 44
eggshell
AMS
Mantissar 7
MFEA K.13293: 29
eggshell
AMS
Mantissar 12
MFEA K.13298: 15
ceramic
Luminescence
MFEA K.13298: 25
ceramic
Luminescence
MFEA K.13298: 55
eggshell
AMS
MFEA K.13298: 60-01
eggshell
AMS
MFEA K.13298: 60-02
eggshell
AMS
MFEA K.13298: 60:03
eggshell
AMS
AMNH 73/648 A
eggshell
AMS
AMNH 73/648 B
eggshell
AMS
AMNH 73/655 A
ceramic
AMS
AMNH 73/655 A
ceramic
AMS
AMNH 73/763-01
eggshell
AMS
AMNH 73/763-02
eggshell
AMS
AMNH 73/764-01
eggshell
AMS
AMNH 73/790-01
eggshell
AMS
AMNH 73/790-02
eggshell
AMS
AMNH 73/790-03
eggshell
AMS
AMNH 73/887 A
ceramic
AMS
AMNH 73/894 A
eggshell
AMS
AMNH 73/984 A
eggshell
AMS
Dottore-namak
Gobi-Altai
Shabarakh-usu 1
Shabarakh-usu 2
Shabarakh-usu 2 in situ
Shabarakh-usu 4
379
Gobi-Altai
(continued)
Shabarakh-usu 4
(continued)
AMNH 73/998 A
eggshell
AMS
Shabarakh-usu 7
AMNH 73/1034-01
eggshell
AMS
AMNH 73/1034-02
eggshell
AMS
AMNH 73/1034-03
eggshell
AMS
AMNH 73/1035-01
eggshell
AMS
AMNH 73/1189 A
ceramic
AMS
AMNH 73/1194 A
ceramic
AMS
AMNH 73/1609 A
ceramic
AMS
AMNH 73/1609 C
ceramic
AMS
Barun Daban
AMNH 73/1702 A
ceramic
AMS
Orok Nor
AMNH 73/1790 A
eggshell
AMS
AMNH 73/1790 B
eggshell
AMS
AMNH 73/1792 A
ceramic
AMS
Shara Kata Well
AMNH 73/466A
ceramic
AMS
Baron Shabaka Well
(Site 19)
AMNH 73/2229 A
ceramic
AMS
AMNH 73/2231 A
ceramic
AMS
AMNH 73/2231 C
ceramic
AMS
AMNH 73/2236 A
ceramic
AMS
AMNH 73/2237 B
ceramic
AMS
AMNH 73/2225-01
eggshell
AMS
AMNH 73/2225-02
eggshell
AMS
Shara Murun Crossing
(Site 3)
AMNH 73/2303 A
eggshell
AMS
Ta Sur Heigh (Site 7)
AMNH 73/2403 A
eggshell
AMS
Spring Camp (Site 16)
AMNH 73/2526 A
ceramic
AMS
Alkali Wells (Site 26)
AMNH 73/2646 A
eggshell
AMS
Chilian Hotoga
(Site 35)
AMNH 73/2796 B
ceramic
AMS
AMNH 73/2796 C
ceramic
AMS
AMNH 73/2797 A
ceramic
AMS
AMNH 73/2797 A
ceramic
AMS
AMNH 73/2800 A
Eggshell
AMS
AMNH 73/2800 C
Eggshell
AMS
Shabarakh-usu 10
Ulan Nor Plain
East Gobi
380
A.2. Context of Dated Sites
REGION
SITE NAME
Alashan Gobi
Jabochinkhure
Gobi-Altai
# OF
SAMPLES
1
MATERIALS
METHODS
NOTES
ceramics
L
Gashun
1
ceramics
AMS
Yingenkhuduk
4
ceramics,
eggshell
L, AMS
Dottorenamak
2
ceramics
L
Mantissar 4
1
eggshell
AMS
Mantissar 7
1
eggshell
AMS
Mantissar 12
6
ceramics,
eggshell
L, AMS
Shabarakh
usu 1
4
ceramics,
eggshell
AMS
Shabarakhusu 2
6
ceramics,
eggshell
AMS
Shabarakhusu 4
5
ceramics,
eggshell
AMS
Shabarakh
usu 7
4
eggshell
AMS
Plains/basin, near square
ruin, diagnostic pottery,
evidence of microblade
cores, probably cohesive
Dunes/basin, paddled bowl,
microblade tools, cohesive
Dunes/basin, many small
sites, diagnostic shards,
grinding stone, polished
stone, mixed
Dunes/basin, associated
with copper slag, evidence
of microblade cores,
cohesive
Painted pottery, textile
pottery, small microblades,
probably cohesive
Painted pottery, incised
pottery, bifaces, small
microblades, probably
cohesive
Painted pottery, diagnostic
pottery (textile, engraved,
handles, etc.), microblade
cores, mixed
Dunes, partially excavated,
higher than S-u 2, few
microblade/cores, mostly
white chalcedony, bird bone
artefact, small/large bifaces,
cohesive
Dunes, partially excavated,
below high water lines,
adze/axes, bifaces, stamped
pottery, grinding stones,
cohesive?
Dunes, hearths, several
small sites, diagnostic
pottery (paddled, comb, net,
engraved, moulded,
channeled), barrel-shaped
microblade cores, small
bifaces, cohesive?
Dunes, below high water
line, one small shard, rough
small bifaces, small cobbles,
eggshell beads, probably
cohesive
381
REGION
SITE NAME
Gobi-Altai
(continued)
Shabarakh
usu 10
East Gobi
# OF
SAMPLES
3
MATERIALS
METHODS
NOTES
ceramics
AMS
Dunes, promontory, above
high water line, partially
excavated, few lithics,
diagnostic pottery (paddled,
stamped, channeled), few
microblades/cores, cohesive
Ulan Nor
Plain
4
ceramics
AMS
Barun Daban
1
ceramics
AMS
Orok Nor
4
ceramics,
eggshell
AMS
Shara Kata
Well
1
ceramics
AMS
ceramics,
eggshell
AMS
Several hearths, sand, near
raw material source, many
test pieces, small bifaces,
diagnostic pottery, mixed
pottery, probably cohesive
Dunes, many hearth sites
around lakes, above lake
deposits with possible
exception of two, one from
later period, small bifaces,
polished stone, some mixing
Dunes near lake, hearths,
several small sites,
diagnostic pottery (paddled,
incised), small bifaces,
copper ore, some mixing
Mountains near river,
excavated, fibre-tempered
shards with light cording,
wedge-shaped microblade
cores on chalcedony,
cohesive
Dunes, hearths, many small
sites, many diagnostics,
formal grinding stones,
small bifaces, mixed
Dunes near river, mostly
debitage, no pottery,
microblades, possibly mixed
Mesa/hill near river,
stamped pottery, cowry
shell, small biface,
microblades/cores, probably
cohesive
Mesa, grinding stone
fragment,
microblades/cores, stamped
pottery, small bifaces,
mixed
Hills with dunes, grinding
stones, polished stone,
microblades/cores, small
bifaces, bronze arrowhead
Baron
Shabaka
Shara Murun
Crossing
1
eggshell
AMS
Ta Sur Heigh
1
eggshell
AMS
Spring Camp
1
ceramics
AMS
Alkali Wells
1
eggshell
AMS
382
REGION
SITE NAME
East Gobi
(continued)
Chilian
Hotoga
# OF
SAMPLES
6
MATERIALS
METHODS
NOTES
ceramics,
eggshell
AMS
Dunes, hearth, diagnostic
pottery (textile, string,
toothed), ochre, formal
grinding stones, faunal
remains, fox canine
ornaments, bone tools,
small unifacial points,
microblades/cores, probably
cohesive
383
A.3. Results of Chronometric Dating
14
C AGE BP +
1σ/ ka + 1 σ
KYA (CAL.
TO 68%
RANGE)
3500 + 300
3.20-3.80
3385 + 40
3.59-3.68
C
3910 + 300
3.61-4.21
212: 123
C
5690 + 350
5.34- 6.04
UW2360
212: 128
C
3910 + 230
3.68-4.14
AA87198
212: 184
E
41,900 + 1500
44.04-47.25
UW2856
248: 5
D-n
C
3540 + 1060
2.48-4.60
UW2355
248: 6
D-n
C
2740 + 200
2.54-2.94
AA87197
290: 44
M4
E
-11.8
14,857 + 85
17.92-18.44
-9.1
>49,900
discarded
LAB. NO.
CAT. NO.
(K.13 OR
73/)
SITE
MAT.
UW2361
203: 5
J-k
C
AA91693
207: 1
G
C
UW2358
212: 6
Y-k
UW2357
δ13C
-32.4
-2.4
+
AA87200
293: 29
M7
E
UW2362
298: 15
M 12
C
6460 + 700
5.76-7.16
UW2359
298: 25
M 12
C
3840 + 340
3.50-4.18
AA87201
AA87202
AA87199
298: 55
298: 60-01
298: 60-02
E
+
-8.1
>49,900
discarded
E
+
-8.1
>49,900
discarded
E
+
-8.8
>48,500
discarded
E
+
-9.2
>48,800
discarded
E
-10.4
7483 + 47
8.23-8.36
AA87203
298: 60:03
AA89869
648 A
AA89870
648 B
E
-8.4
8522 + 50
9.49-9.54
AA89872
655 A
C
-20.9
4308 + 40
4.85-4.95
AA89872
655 A
C*
-7.3
10,039 + 57
discarded
AA76420
763-01
E
-10.3
8159 + 43
9.04-9.20
AA76421
763-02
E
-9.6
8184 + 44
9.06-9.23
AA76419
764-01
E
~ -9.3
7969 + 37
8.75-8.95
AA76416
790-01
E
-9.0
8396 + 52
9.34-9.48
AA76417
790-02
E
-11.1
8268 + 44
9.17-9.37
AA76418
790-03
E
-10.7
30,490 + 780
34.05-35.50
C
-21.9
3680 + 76
3.92-4.13
S-u 1
S-u 2
S-u 2
(in
situ)
S-u 4
+
AA89873
887 A
AA89874
894 A
E
-10.0
7589 + 47
8.37-8.42
AA89875
984 A
E
-10.0
8473 + 64
9.44-9.52
AA89876
998 A
E
-10.0
8254 + 47
9.15-9.34
AA76422
1034-01
E
-11.3
8054 + 43
8.82-9.02
S-u 7
384
LAB. NO.
CAT. NO.
(K.13 OR
73/)
SITE
MAT.
δ13C
14
C AGE BP +
1σ/ ka + 1 σ
KYA (CAL.
TO 68%
RANGE)
AA76423
1034-02
S-u 7
E
-11.6
38,600 + 1000
42.20-43.85
AA76424
1034-03
E
-10.7
8439 + 60
9.41-9.51
AA76427
1035-01
E
-11.0
8081 + 49
8.89-9.08
AA89877
1189 A
C
-24.6
3595 + 41
3.86-3.96
AA89878
1194 A
C
-23.4
3246 + 39
3.42-3.54
AA89879
1609 A
C
-23.1
5116 + 41
5.78-5.91
AA89880
1609 C
C
-23.3
5061 + 49
5.75-5.88
AA89881
1702 A
BD
C
-27.5
1661 + 42
1.53-1.62
AA89882
1790 A
ON
E
-9.5
8307 + 56
9.22-9.41
AA89883
1790 B
-9.5
8307 + 56
9.29-9.43
C
-26.8
10,030 + 140
discarded
S-u 10
E
+
AA89884
1792 A
AA89868
466A
SKW
C
-24.4
8604 + 51
9.54-9.63
AA89885
2229 A
BS 19
C
-25.7
5609 + 47
6.34-6.44
AA89886
2231 A
C
-24.3
5954 + 52
6.73-6.86
AA89887
2231 C
C*
-3.8
1173 + 58
discarded
2231 C
+
-22.5
5825 + 85
discarded
+
AA89887
C
AA89888
2236 A
C
-23.2
1445 + 86
discarded
AA89889
2237 B
C
-24.0
3115 + 47
3.28-3.38
AA76426
2225-01
E
-12.0
12,509 + 59
14.53-15.13
AA76427
2225-02
E
-10.7
12,450 + 74
14.38-15.05
AA89890
2303 A
SMC 3
E
-12.3
12,497 + 70
14.56-15.12
AA89891
2403 A
TSH 7
E
-11.4
14,129 + 80
17.13-17.61
AA89892
2526 A
SC 16
C
-20.1
866 + 51
0.74-0.88
AA89893
2646 A
AW 26
E
AA89895
2796 B
CH 35
-10.4
9562 + 51
10.79-11.05
+
-26.7
1866 + 88
discarded
+
C
AA89896
2796 C
C
-27.6
17,120 + 220
discarded
AA89897
2797 A
C*
+0.6
33,160 + 540
discarded
AA89897
2797 A
C
-25.5
6728 + 45
7.56-7.64
AA89898
2800 A
E
-7.2
10,586 + 56
12.42-12.68
AA89899
2800 C
E
-6.9
10,103 + 55
11.49-11.90
* indicates that sample was taken on carbonate fraction of ceramic without pretreatment,
using selective dissolution; + indicates that the date may be unreliable due to a carbon
yield of under 0.10 mg C or the sample produced an infinite date.
385
APPENDIX B – DETAILED SUMMARY OF DATED SITES
East Gobi
Baron Shabaka Well, Site 19
The Baron Shabaka locality, or Site 19, was located near a well in a narrow valley
covered with weathered sand dunes. The vegetation was considerable relative to the rest
of the region. “Camel sage” and a tough wire-like grass were most plentiful. Pond (nd:
90B) reports that they were led to the site by a Mongol hunter with a flint-lock gun, who
claimed that the place yielded large quantities of material suitable for gun flints. The site
stretched over an area of about 0.4 x 1.2 km. Archaeological remains were recovered 3
km south of a mesa and about 3 km west of a well. The majority of artefacts were found
in blown out wind hollows, but smaller quantities were collected from the top of the
valley sides and the tops of the dunes. The latter are probably related to historic use of
the locale, but were mixed during curation.
The dunes were partially solidified and then weathered into many hollows by
aeolian activity. Many distinct hearth sites were reported but the integrity of these sites
was not maintained during curation as they were for the Shabarakh-usu locality in the
Gobi-Altai (see below). Pond described two of the sites. At Site 1, fire-cracked rocks
suggested a hearth site, forming a loose group near a partially buried adze/axe. They
were partially embedded in a dark grey soil that might have been so coloured due to the
inclusion of ash in the sediment. A broken rectangular metate was also found next to a
knobbed grinding bar or pestle. At Site 2, a broken rectangular metate was also found
next to a knobbed grinding bar or pestle. Several small groups of lithics, all made of the
386
same material, and sometimes “very crude pottery,” were found in a small area of less
than a square meter. Elsewhere fine microlithic and coarser chipped implements were
mixed and found together.
The archaeological assemblage is extremely diverse and includes some historic
emains (including fragments of iron cookware), as is typical of many larger dune field
sites. Ostrich eggshell fragments and unfinished beads from Baron Shabaka Well date to
between about 14.5-15.0k cal yr BP (Janz et al., 2009). Radiocarbon dating produced
three good dates from both Oasis 2 and the end of Oasis 3: 6.8k cal yr BP (5954 + 52 BP
[AA89886, AMNH #73/2231A]), 6.4k cal yr BP (5609 + 47 BP [AA89885, AMNH
#73/2229A]), 3.3k cal yr BP (3115 + 47 BP [AA89889, AMNH #73/2237B]). The
collection of artefacts and associated dates illustrates the reuse of dune-field/wetland sites
over many millennia. Pond’s description of the site indicates that several of the collected
scatters were probably temporally coherent. A similar situation is recognized at the
Shabarakh-usu locality (Nelson, 1925).
Almost 7000 artefacts were recovered and curated from the Baron Shabaka
locality, including: 402 ceramic shards, a stone spindlewhorl or circular disk, fragments
of an iron cooking vessel, fragments of a mollusc shell, beads and fragments of ostrich
eggshell, a partially drilled piece of talc, red paint stone, a fragment of a stone ring, 116
grinding stones (e.g., grinding slabs, hand stones, saddle querns, pestles, stone
vessels/mortars), chipped and/or partially polished adze/axe/gouges, hammerstones,
unifacial and bifacial points, bifacial knives, perforators, drills, burins, a diverse array of
endscrapers, microblade cores, microblades, rough flake cores, and flakes. Cores and
387
endscrapers are especially numerous which is probably partially due to collection bias,
but still illustrates the nature of the site as a center of production. Pottery from Baron
Shabaka was dated to both Oasis 2 and Oasis 3, and the range of artefact types indicates
that the locality was used intermittently during both periods. Due to the temporal range
of habitation, clear chronological distinctions are difficult to assess within the
assemblage.
Microblade tools were often steeply retouched and several examples of unifacially
retouched projectile points and perforators were recovered. Blade tools (2 to 2.5 cm
wide) with lightly retouched distal ends are unusual and not typical of the Neolithic
period. Likewise, several examples of unifacial tools made on large flakes (3 to 9 cm
long and up to 4 cm wide) or flat-backed cobbles may be related to earlier Palaeolithic or
Epipalaeolithic occupations, as they have no clear parallel in the Neolithic/Eneolithic.
They are oval shaped and steeply retouched on all edges (Figure B.1). Large flakes and
blades struck from prepared cores were also found, though some may have been
retouched more recently. Artefacts that are most closely aligned with Oasis 2
assemblages include semi-lunar knives or thick bifacial preforms, short bifacial blades
with parallel edges that were probably used as insets, and low-fired corded brown ware.
The wide range of pottery types suggests intermittent use of the site throughout the
Neolithic.
388
a.
b.
Figure B.1 Macrotools from Baron Shabaka Well.
389
Certain other artefact types are likely to have been associated with Oasis 3
habitations. One pottery shard produced a date of 3.3k cal yr BP (Table 3.1). This shard
is very hard and was probably fired at a higher temperature than is typical of Oasis 2
pottery. The shard is uniformly coloured – grey on the exterior surface and light
brown/buff on the interior surface. Some coarse sand or gravel temper was used, but
organics were also included, as indicated by the porous and spongy texture of the fabric
as seen in a cross-section of the interior paste (Figure 3.9). Manufacture by coiling and
slow turning on a wheel are suggested by the undulating interior surface with fine,
parallel striations. The type of temper used is not known, but is common in later East
Gobi pottery. Although the surface was eroded, traces of the decorative finish show the
use of a roller stamp, resulting in widely (~1.5 cm) and evenly spaced rows of square
punctates – “toothed” impressions (Figure 3.8e). Similar shards recovered from other
East Gobi sites indicate a tall, cylindrical vessel with a flat bottom.
Other examples of pottery from Baron Shabaka show the frequent use of coarse
sand temper, large quantities of very fine sand temper, unidentified organics, mica or
nacre, and possibly shell. Red, grey and brown wares were all found in high numbers at
Baron Shabaka. Undecorated shards made of untempered homogeneous paste, or very
lightly sand-tempered red-ware, are reminiscent of those from the Alashan Gobi. Coarser
red-ware is usually sand-tempered and fired at a lower temperature. One shard is a sandtempered red-ware with faint traces of string-paddled markings. The exterior surface was
darkened, suggesting use over fire. Another fragment of sand-tempered red-ware shows
possible traces typical of geometric-incised pottery recovered from western sites (Figure
390
3.8c, d). One shard of brown-ware resembles a shard from Chilian Hotoga, dated to
20.5k cal yr BP (AA89896, AMNH #73/2796C). The shard appears to be unfired and is
porous, with a high content of extremely fine sand particles. The radiocarbon date for the
Chilian Hotoga sample is almost certainly much too old and probably resulted from
contamination of the porous unfired shard.
Surface treatments show parallels with other Gobi Desert collections, as well as
some unique styles. The majority of shards are undecorated. Common finishes include
stamped, string-paddled, and textile impressions. A smeared net impression made on
buff clay was noted on several shards having heavily blackened interior surfaces. Textile
impressions make up about 9% of this collection. Many of these shards are of thick grey
ware with the spongy textured paste (Figure 3.9). Parallel and intersecting rows of cord
impressions, like those found at Mantissar 12, were also recovered from Baron Shabaka.
Several examples of incised designs included linear and checked patterns. Slightly
diagonal vertical scraped incisions were identified on several shards and are reminiscent
of paddle markings. As in the western Gobi Desert sites, moulded rims are also present.
High-fired grey ware with angular punctate impressions is included in the collection and
was typical of Mongolian pottery during the Khitan Period/Lao Dynasty.
As at other Gobi Desert sites, shards are very small and vessel form difficult to
ascertain. One fragment of a round, thick drilled clay disk is probably a spindle whorl.
Similar artefacts were recovered from the Alashan Gobi collections. As at other Gobi
Desert sites, drilled fragments may indicate mending and curation. Discernable examples
include one brown shard with a finely finished outward projecting rim which was thick
391
and rounded. The vessel was decorated with vertical scraped incisions. Handle
fragments and miniature lugs were found, supporting evidence from the western sites of a
new type of vessel design and perhaps function. The latter were attached to low-fired
spongy ware with incised patterns or smeared paddling. Flat bottomed vessels are
evidenced by some fragments. One fragment indicates the use of a pedestal bottom with
blackened patches near the bottom on the exterior surface.
Polished and chipped macrotools and grinding stones occurred in high frequency.
Grinding slabs of many different shapes and sizes are numerous (at least 116 examples
were retained in the AMNH collections for this site alone), and include pestles or
knobbed/ball-headed rollers. The labour invested in their manufacture underscores the
importance of milling tools. Such tools are associated with Oasis 2 assemblages such as
Chilian Hotoga and Jira Galuntu. Two examples of what might be construed as hoes
were also found. Such artefacts were also found in the west at Yingen-khuduk and
Gashun Well. They were crudely chipped into semi-circular shapes. A large pick (~30
cm x ~5 cm wide) was partially flaked from an oblong stone is another unique artefact
that does not have any parallel in more western Gobi Desert sites. Fragments of thick
polished stone rings are also included in the assemblage and might be similar to artefacts
in other regions described as “counter-weights.”
Macrotools were mostly made on basalt. Some are roughly chipped adze/axe
preforms, occasionally exhibiting localized patches of light polishing. Polished versions
of similarly shaped tools were also recovered, including a broken axe with a bulbous
proximal end that was roughly polished, with numerous deep striations. There are also
392
large bifacially flaked tools with one purposely flattened end for hafting. One such
biface is a leaf-shaped point on basalt. The others are on flint and somewhat spatulate in
shape, but with an angled distal end (Figure B.1). Some unifacial specimens may be
blanks for such tools. Small bifacially flaked adze/axes (N = 8) on fine quality flints are
notable for this collection and are probably associated with Oasis 2 occupations.
Other bifacially flaked tools include rough thick bifaces, microadze/axes,
projectile points, and blade knives or inset knife blades. Fine bifacially retouched awls or
drills with thin points were made on chalcedony microblades and have parallels with
examples from other Oasis 3 Gobi Desert sites. Knife blades at Baron Shabaka are short
and have parallel edges rather than the curved blades of western sites. Such artefacts are
common in Oasis 2 assemblages and were probably used as inset knife blades. Fine
bifacially flaked points or arrowheads exhibit concave, convex, and straight bases.
The wide variety of core types is typical of both Oasis 2 and Oasis 3. Many different
sorts of raw materials were used and there is little evidence of the red jasper that is
common in the Gobi-Altai and Alashan Gobi. Wedge-shaped cores are more common
here than is typical of Oasis 3 sites. Some wedge-shaped cores might have been made on
prepared bifaces. This group is similar to those from the Oasis 1 site of Shara KataWell
in that some of the cores were prepared by initial unifacial retouch on a flat cobble. A
long spall was then removed to form the platform, resulting in a D-shaped perform.
Microblades were struck from the platform without further preparation.
Scrapers are typical of Neolithic/Eneolithic assemblages from the Gobi Desert.
Many are heavy-duty, thick specimens, and there are several large rectangular scrapers
393
with a squared distal end. Most scrapers are on amorphous microlithic flakes (35%), but
a high frequency (20%) were made on amorphous blocks, chips, or cobbles. Scrapers on
the ends of microblades or thick elongated flakes are characteristic of Oasis 3
assemblages. One large macroflake is reminiscent of Epipalaeolithic assemblages.
Many archaic forms are recognized within assemblage and may support an
undated Early Epipalaeolithic or Oasis1 occupation. At the same time, long term
stockpiling of raw materials could simply have resulted in less emphasis on microlithic
tool types. Closer comparison of the Baron Shabaka collection with dated
Epipalaeolithic assemblages from Northeast and North China should be undertaken in
order to offer a better assessment of this situation.
Baron Shabaka South, Site 21
Site 21, or Baron Shabaka South, was found in weathered dunes on a hillside almost 5 km
south of Baron Shabaka Well (Pond, n.d.: 91) and appears to be a temporally coherent
occupation site with less than 500 recovered artefacts. Artefacts from the site were not
dated, but the presence of diagnostic artefacts indicates an early Oasis 2 occupation. No
pottery was found at this site. A unifacial point made on a microblade and a very finely
retouched bifacial flake point with light fluting on one face were collected. Formal
grinding stones are limited to one handstone or mano. Two roughly chipped bifaces,
small and thick, may have been used as adzes. The core assemblage comprises 57%
microblade cores. All are of the wedge-shaped variety and made on small nodules of
chert and chalcedony. Other artefacts include one finished ostrich eggshell bead,
394
endscrapers, flakes, microblades, and a perforator on a microblade. Raw materials
included chert, chalcedony, and various other metavolcanic rocks. Based on the
transitional form of the unifacial and bifacial points, an age of about 8.0-7.5 kya is
proposed.
Chilian Hotoga, Site 35
Textile-impressed pottery from Chilian Hotoga, or Site 35 was dated to 7.6k cal yr BP
(Figure 3.4c). The site was found near a well among dunes in a large wind hollow. On
south side of hollow, worked bone, pierced teeth, and shells were found at foot of an
escarpment. On surface, loose sand was mixed with small artefacts. The site was
excavated to reveal a hearth site about 2.5 m in diameter, containing burned stone, bone
fragments, charcoal, and one roller for grinding. Artefacts were found all over the
eastern half of the large wind hollow around the well, but no other artefacts were
embedded.
Aside from the bifacial and unifacial points, a bifacially-flaked bladelet knife was
recovered. This specimen is typical of bifaces found in many Neolithic sites and
resembles a large microblade or bladelet that was completely retouched in order to create
a parallel-sided tool suggestive of a knife blade. Such artefacts are typical of Oasis 2 and
may have been used as inset blades for composite knifes. One edge usually shows
usewear. The remainder of the lithic assemblage included chipped and partially polished
adzes, a lightly flaked bifacial semi-lunar knife, rough cores, flakes, sidescrapers,
numerous and varied endscrapers, perforators, microblade cores (subprismatic, conical,
395
cylindrical), microblades, endscrapers on microblades, and a stone pendant. A fragment
of “red paintstone”, ostrich eggshell fragments, drilled bivalve shells, drilled fox canines,
fragments of avian long bones (one of which was incised with transverse grooves), and
fragments of a bone awl and two bone needles are also included. Faunal remains are
from rabbit (MNI = 6), equid (including teeth, long bone fragments, phalanges, an
astragalus, and a sesamoid), frog (Ranidae), and some type of small Galliformes
(identified by spur). Possible Oasis 3 elements include several shards of reddish-brown,
low-fired pottery decorated with toothed or roller-stamp punctate impression and
endscrapers on microblades.
Jira Galuntu, Site 18
Jira Galuntu was a large site yielding 6,340 artefacts. Artefacts from this site were not
dated, but the site is assigned to the beginning of Oasis 2 based on the presence of
diagnostic artefacts. There is evidence of intrusive pottery from later occupations. Pond
(n.d.: 90A) wrote that the site was discovered in a wind hollow on a hillside with some
sand (perhaps a former dune), near a wash that drained into the nearby lake bottom.
Some of the associated pottery was clearly intrusive – high-fired grey ware with a
punctate design typical of the Khitan period/Liao Dynasty (AD 911-1125). Pond
suggested that the site was composed of remains from two distinct periods – one
representing a microlithic component and the other characterized by coarsely chipped
material.
396
The majority of pottery at this site is coarse sand-tempered red-ware with organic
inclusions. Formal grinding stones, including a pestle or ball-headed/knobbed roller, are
typical of Oasis 2. Microblade cores include wedge-shaped, conical, and sub-prismatic
core, made on a variety of poor quality raw materials that were probably local. The
remainder of the lithic assemblage comprises a massive chipped and finely polished
adze/axe fragment, hammerstones, various types of endscrapers, fragmentary biface
blanks, microblades, burins, perforators, drills on microblades, and unused amorphous
flakes. Decorated pottery includes fragments of a straight-walled, wide-mouthed vessel
with an undulating moulded rim and vertical cording. These shards are not typical of
Oasis 2 sites and could be intrusive elements from Oasis 3. Jira Galuntu is assigned to
the earlier period of the unifacial point phase at about 8.0-7.6 kya.
Gobi-Altai
Shabarakh-usu 1
Shabarakh-usu 1 was found largely deflated from an eroded dune. Artefacts were
scattered along the surface of the valley floor, continuing to a spot in the base of a low
escarpment, where some artefacts were still in place (Nelson, 1925: 33). Neither hearths
nor the firecracked rocks typical of other Shabarakh-usu sites were found in the vicinity.
The site is presumed to have been a workshop for lithic production. Over 7000 artefacts
were recovered including a some pottery, a small metate or “rubbing stone”, a crudely
chipped adze/axe, an incised fragment of shale, many lithics (about 85% of which were
of white chalcedony), a worn and polished bird bone, weathered mammal phalanges,
397
angular fragments of ostrich eggshell, pendants made on bivalve shells, and a perforated
shell. Only a small portion (N = < 1000) of that total assemblage was removed from the
site and included in the museum collections. According to Nelson’s (1925) description
and the abundance of white chalcedony used as a raw material, the assemblage should be
considered temporally coherent.
One potshard from Shabarakh-usu 1 (AA89872, AMNH #73/655A) was dated
using radiocarbon analysis to the Oasis 2-Oasis 3 transition (4.9k cal yr BP [4308 + 40
BP]). It is a piece of roughly paddled light reddish-brown ware. The interior portion of
the paste is blackened and shows darkened pits where organic inclusions were
carbonized. Sand grains are present in the paste, though not in a quantity suggestive of
intentional tempering. The surfaces of the shard are cracked and appear friable, perhaps
due to erosion. Ostrich eggshell from this site was dated to 8.3 and 9.5k cal yr BP. The
former date is so far the latest date yet recorded for ostrich in East Asia, suggesting that
the species might still been present in the Gobi-Altai region at the beginning of Oasis 2.
The dates do not reflect the age of the archaeological assemblage (see Chapter 3).
Several distinctly decorated pottery shards are associated with the site, but some
of the pieces may be intrusive, judging from the style of manufacture. One of these
fragments is from a finely made and relatively high-fired sandy pottery. It refits with a
larger repaired section of the same vessel recovered by other researchers in the general
vicinity of Shabarakh-usu 1 and 2. The shard features a dispersed diagonally slanted stab
and drag pattern, each based on two sets of four lines (see Figure 26.b’. in Fairservis,
1993: 51). Another conspicuous shard features moulded curvilinear designs and is from
398
an unusually large, thick-walled, high-fired vessel that was made on a sand-tempered
paste of very homogeneous character (see Figure 64.i, Fairservis, 1993:154). Another
thick, high-fired shard is stamped with a curvilinear design similarly inconsistent with the
mid-Holocene date. As such, we must consider the possibility that these artefacts were
from the dune surface and may have been mixed during deflation. Despite evidence of
intrusive elements, the consistency of lithic raw material and collection context support a
strong temporal coherency.
Most of the pottery from Shabarakh-usu 1 is reddish-brown and string-paddled
(some may have been cord-marked). The blackened interior paste characteristic indicates
low firing and the use of organic temper (or highly organic clay). Only some of the
shards appear to have been sand-tempered. One partially reconstructed sand-tempered
vessel with a raised band of moulded clay placed in an undulating pattern along the rim,
and string-paddled on the lower section is reminiscent of decorated pottery from Jira
Galuntu, though the design is less developed. The shape suggests straight walls and there
is a thick, flat handle. Other shards have raised clay bands below the rim, moulded and
pressed, with some evidence of vertical string-paddling on the exterior surface below.
Two additional Oasis 3 shards appear to have been fired at higher temperatures. One is
decorated with incised slanting lines reminiscent of a pattern recovered at Site 10 that
will be referred to as “geometric incised” (see Figure 8.d). The other is a tiny dark red
shard with a raised band incised with vertical troughs at regular intervals (see Figure 8.f).
In general, this pottery is similar to that from Oasis 2 sites, but with distinct additions like
raised moulded bands and handles.
399
The lithics from this site also represent a continuation of Oasis 2 reduction
strategies. A series of finely made projectile points are included in the sample of
bifacially flaked points (N = 26), as are some large rough bifaces which might be
unfinished blanks. All were made on either jasper or chalcedony. New biface forms
include stemmed, leaf-shaped with a deeply concave or fishtail base, and a tear-drop
shaped point (see Figure 3.11). A variety of sizes are represented and small projectile
points may have been reworked from larger damaged specimens. Bifacial straight or
slightly curved blades with rounded ends were also found (N = 6), and are similar to
specimens from Ulan Nor Plain (see Figure 3.10). One large adze/axe of brown and
black mottled silicified sandstone was recovered. One end had been modified for hafting.
Thick microblades or elongated flakes often possess steep retouch along parallel
vertical edges. Many were clearly being formed into minimally retouched perforators or
completely retouched drills, including forms with expanded bases, which are common in
many Oasis 3 sites but absent from Oasis 2 assemblages. Several of the finely finished
awls on microblades are probably more typical of Oasis 2 tool kits. Some large elongated
flakes show steep retouches and may have been used as scrapers or knives. Many other
elongated flakes and microblades were used without retouch and discarded. Some have
possible evidence of hafting at the distal or proximal end. A few very slender, pointed –
almost needle-like – microblades were recovered. Some have signs of light use along the
lateral edges. Uniformity and consistency of edge damage distinguishes usewear from
post-depositional chipping.
400
Only 12 cores are included in the Shabarakh-usu 1 assemblage. This number may
not accurately reflect the actual number of cores at the site since almost 90% of the
assemblage was discarded (Nelson, 1925: 51c-51d). Still, judging from the high number
of unused debitage flakes (N = 469) retained in the collection, it is probable that few
cores were present in the original site. The relatively higher number of bifacial points (N
= 26) is significant and probably related to site function, since microblade cores and
bifaces are expected to be equally favoured during collection.
The assemblage includes two examples of each of the following core forms:
amorphous cores, informal bladelet or elongated flake cores, unsuccessful microblade
cores of unknown type, unknown microblade cores, and flat-backed conical cores with
blunt angled distal ends. One wedge-shaped and one flat-backed funnel shaped core are
also included. There are twenty-two scrapers: ten on microlithic flakes; five on the ends
of microblades or elongated flakes; four on blocks, chips, or cobbles; two reworked from
thick elongated flakes; and one formal thumbnail scraper. From a chronological
perspective the most important characteristics to consider are the use of endscrapers on
elongated flakes or microblades, the high number of finely worked projectile points, the
occurrence of curved bifacial blades, and the importance of bifacial retouch in the
preparation of formal tools.
401
Shabarakh-usu Site 4
The Shabarakh-usu 4 assemblage (see also Janz, 2006: Appendix A) is comprised of
several distinct concentrations that were deflated or in the process of deflating from the
dune matrix. Some of the spots were partially excavated. According to Nelson (1925:
37), the lithics from subsite 4E were separately grouped, but the museum assemblage
does not include the one potshard reported by Nelson. The majority of pottery was found
in a separate area about 9 m in diameter. Faunal remains were not analyzed, but included
an ungulate metapodial, a possible tarsal, and a rib fragment. Additional bone fragments
were reported but not retained for the museum collections. The distribution of these sites,
spread out over about 920 m2, suggest clusters of task sites related to the primary hearth
group of 4A and 4B. Nelson believed that what remained of the in situ artefact groups
lay at about the same stratigraphic level.
Despite low carbon yields (0.04%), one shard from Shabarakh-usu 4 (AA89873,
AMNH #73/887A) produced a radiocarbon date of 4.0k cal yr BP (3680 + 76 BP), which
is contemporaneous with the earlier date from Shabarakh-usu 10 (see below). As low
carbon yields regularly produce erroneously old dates, the shard could actually be more
similar in age to the slightly later date from Shabarakh-usu 10. The sample had a visibly
darkened interior paste suggestive of carbonaceous remnants from organic inclusions.
The piece is thick-walled, heavily sand-tempered, and light reddish-brown. Curvature is
suggestive of the lower portion of a globular vessel. From the same site a lightly netimpressed shard of similar material, though lacking a darkened interior paste, was dated
using luminescence (AMNH #73/890A) and should provide a comparative date range
402
when dates are available. Three dates on ostrich eggshell from this site yielded ages of
8.4, 9.5, and 9.2k cal yr BP, suggesting prehistoric scavenging from multiple sources.
These dates do not reflect the age of the site. Dated pottery shows contemporaneity with
Shabarakh-usu 10 – about a thousand years younger than Shabarakh-usu 1.
Many pottery types from the site are reminiscent of those from Shabarakh-usu 1.
One dark grey, thick-walled, high-fired shard with a curvilinear stamp decoration is
similar to a shard found at Shabarakh-usu 1 and may be intrusive. Some undecorated
brown, red and grey shards, heavily tempered with sand, show evidence of carbonized
organic residues in the interior paste. Reddish-brown string-paddled wares with raised
clay bands are common, making up almost 75% of the total pottery sample. They are
sand-tempered and many show evidence of possible organic temper. Two groups of such
ceramics can be distinguished – those with narrow raised bands and those with thicker
raised bands that were moulded or incised. The latter often have wider ridges on the
body suggestive of cord paddled marks rather than string-paddling. Again, they are
comparable to the majority of pottery found at Shabarakh-usu 1. String-paddled shards
are reminiscent of the ceramic bowl found in the Alashan Gobi at Gashun (MFEA
#K.13207:1), dated to about 3.6k cal yr BP.
Other shards are similar to those recovered from Shabarakh-usu 10. A larger
reconstructed rim fragment of sand-tempered ware is decorated with “stamped bands of
parallel corrugations” (AMNH catalogue) that is similarly made to other examples of
channelled ware from Shabarakh-usu 10 (Figure 30.d. in Fairservis, 1993:59; this
volume, Figure 3.8a). One small bead-like lug is attached to the fragment near the rim.
403
Another tiny shard is marked with a series of short rows made up of diagonal circular
punctuates, which is a decorative technique also represented at Shabarakh-usu 10 (see
Figure 30.i in Fairservis, 1993: 59). Finally, a very interesting reconstructed fragment of
a large vessel with distinctive decorations was recovered. The shard is sand-tempered
and the paste is reminiscent of that used in the stab-and-drag decorated vessel from
Shabarakh-usu 1. Two fragments, probably from the same vessel, show a pattern of
geometric incision (linear and triangular designs) common in late Gobi Desert sites.
Along the rim are diagonally angled almond-shaped indentations placed in a band just
below rim (see Figure 30.a, 30.b in Fairservis 1993: 59, Figure 3.6.). In view of the date
on pottery from this site and from Shabarakh-usu 10 (see below), these shards should all
be considered representative of the later phase of Oasis 3.
Chalcedony was the most common raw material used in tool manufacture. Core
types include amorphous cores, a variety of informal microblade cores, wedge-shaped,
wedge or flat-backed funnel-shaped, conical, and massive barrel-shaped cylindrical cores.
The latter are more common in Alashan Gobi Desert sites and are sometimes
characterized by opposed platforms. At Shabarakh-usu 4, barrel-shaped cores were made
on coarser-grained cryptocrystallines. These microblade cores have only one platform,
and probably represent nuclei discarded at an early stage of use. Endscrapers on
microblades and thick elongated flakes, drill points, larger sidescrapers, and a bifacially
flaked adze made on jasper were also contained in the site assemblage. Bifacial tools
include points and knives. Points include the fragmentary base of the concave/fish-tailed
type and the teardrop shape with a convex base. Two bifacial blades, one finer bladelet-
404
sized version, and a larger reconstructed one (2.5 cm at the base and about 7.5 cm long)
were collected. Both were curved with straight bases. The range of lithic types is similar
to Shabarakh-usu 1, which suggests the continuation of earlier reduction strategies.
Shabarakh-usu 10
Shabarakh-usu 10 is the last dated site for this locality and produced two dates: 3.9k cal
yr BP (3595 + 41 BP [AA89877, AMNH #73/1189A]); and 3.5k cal yr BP (3246 + 39 BP
[AA89878, AMNH #73/1194 A]). The distinction of surface and subsurface components
at the site is probably arbitrary. Nelson (1925: 47) reported that the pottery at this site
was found in a streak of charcoal under the point of a small finger-like bluff. His
diagram shows that surface artefacts were found scattered around site. As tends to be the
case at the Shabarakh-usu locality, the site was probably temporally coherent and
deflated from the same layer of the original matrix.
The difference in dates between the two radiocarbon samples is about 400 years.
This divergence may be related to differences between the date of manufacture as
recorded by the date on the interior paste and the last burning episode as recorded in the
subsurface sample. The difference in dates might also be accounted for by the use of clay
heavily impregnated with decomposed organics (represented by samples with darkened
interior paste), which could have been much older than the date of manufacture. The lack
of decomposed organics in the later sample might reflect a more accurate date for the site.
In any case, Shabarakh-usu 10 dates to between about 3.9 and 3.5k cal yr BP.
405
The shard dated to 3.9k cal yr BP (AMNH #73/1189A) is light greyish-brown
with a thick corded impression, perhaps created by a cord-wrapped paddle. The interior
paste is heavily blackened. More sparse sand inclusions suggest either incidental mixing
or a light sand temper. Judging from the interior surface, the fabric of the shard is
smoother and more homogeneous than the sample from Shabarakh-usu 1 and was
probably fired at a higher temperature. Another shard, dated to 3.4k cal yr BP, is from
the excavated component of the site (AMNH #73/1194A) and produced a slightly
younger date than the surface sample of about 3.5k cal yr BP (3246 + 39 BP). The
interior paste suggests an original colour of light reddish-brown, but carbonization on the
exterior and interior surfaces gives the artefact a grey hue. Incidental traces of darkened
hollows associated with combusted organic remains are visible, but the interior paste
does not show darkening associated with organic temper. The clay on the interior surface
is heavily infiltrated with carbonaceous residues. The exterior surface of the vessel
shows a smoothed thick corded impression overlain with striations that suggest light
scraping prior to firing. Aside from this smoothing or scraping, the cord pattern is similar
to that from the sample dated to 3.9k cal yr BP.
Another brown or buff shard from the same surface component was dated using
luminescence (AMNH #73/1190A), but dates are not yet available. The exterior surface
is somewhat haphazardly impressed with a distinct pattern of raised ridges surrounding
indented squares (referred to by Nelson in the original catalogue as “checker-stamped”)
(Figure 3.8b). The interior paste is slightly darkened and small holes throughout the
fabric are indicative of combusted organic particles or perhaps weathered minerals.
406
Indentations from some type of fibre (e.g., grass, thread, or hair) are plainly visible on the
exterior surface. Fibres resembling clumps of hair were found in the temper of several
other Gobi-Altai shards. Sand inclusions are visible, but probably incidental.
Additional diagnostic pottery types are identified from Shabarakh-usu 10. The rim
fragment of heavily blackened channelled ware is similar to shards from Shabarakh-usu
4, further suggesting broadly contemporaneous habitation episodes (compare Figure 37 in
Fairservis, 1993: 68; Figure 3.8a). Other pottery displays decorations of incised raised
bands, both string and cord paddling, linear designs made with circular punctuates and
thin diagonal troughs. Clay fabrics are highly variable and included sand-tempered redware, high-fired dark grey or black ware, light reddish-brown wares, and dark red-brown
wares that were heavily tempered with very fine sand grains resulting in a surface texture
reminiscent of coarse sand-paper. Some pieces exhibited the remnants of combustion of
long, narrow organic elements that might have been small clumps of fibres or hairs.
Blackened interior pastes are common, though high-fired wares are more homogeneous
and sometimes exhibit hollow impressions of fibres or other possible combusted organics
which left no traces of carbon residue.
The few lithics from Shabarakh-usu 10 clearly exemplify the continuation of postLGM microblade reduction sequences complemented by the use of amorphous flakes.
One small (2001-5000 cm3) sandstone metate or “rubbing stone” was recovered. Two
chalcedony cylindrical microblade cores and one chert (probably silicified volcanic ash)
conical microblade cores were recovered, along with a chalcedony test piece (AMNH
#73/1172, catalogued as a rough core). Microlithic scrapers were made on the end of
407
elongated flakes and microblades (N = 2), or other amorphous microlithic flakes (N = 3),
including one on a spall derived from a microblade core. One formal thumbnail scraper
is also included. As in other Gobi Desert assemblages, there is a variety of debitage and
used microblade and amorphous flakes. The lithic assemblage is typical of other Oasis 3
sites, although the cores are smaller and more heavily reduced than is typical of GobiAltai assemblages.
Ulan Nor Plain
Pottery from the Ulan Nor Plain site was dated to about 5.8k cal yr BP (5116 + 41 BP
[AA89879, AMNH #73/1609A], 5061 + 49 BP [AA89880, AMNH #73/1609C]). The
assemblage appears to be relatively temporally coherent. Decorated potsherds were
limited to the upper level and distinctly located at the eastern extent of the site (Nelson,
1925). These decorated pieces, along with high-fired scraped red ware, can be associated
with the Turkic Period (AD 552-630). The assemblage was recovered from stabilized
reddish sand dunes (characteristic of Gobi-Altai Oasis 2 and 3 finds and referred to by
Nelson as a “Shabarakh-usu deposit”) in a hollow of the Ulan Nor-Artsa Bogdo plain.
The hollow was just over 1 km in diameter, bow shaped, and drained by an east flowing
stream. The main deposit was central to the hollow and most densely concentrated on the
western front, stretching over 275 m north to south and 185 m east to west. Several
fireplaces were found with ashes already exposed. The Ulan Nor-Artsa Bogdo plain was
a major raw material procurement locality during the Palaeolithic and Neolithic, and the
408
abundance of raw materials in the vicinity is reflected in the numerous test pieces
(Nelson, 1925; AMNH catalogue).
The assemblage is consistent with other Oasis 2 sites, but includes slightly curved
bifacial blade knives and one endscraper on a microblade. Both are more typical of Oasis
3 tool kits and probably represent the early use of such tools. The site was rich in high
quality materials and discarded test pieces. Finely made and heavily reduced microblade
cores as well as more expedient forms were collected. The high frequency of unknown
core fragments (48%) and informal elongated flake and microblade cores (15%) is
indicative of raw material abundance. Of the seventeen formal microblade cores seven
are wedge-shaped cores, six cylindrical, and four conical. Of the seventeen scrapers,
eight are on amorphous microlithic flakes, seven on blocks, chips or cobbles, and one on
the end of a microblade. Bifaces include a large semi-lunar knife, a poorly formed and
rather large bifacial point, and a curved bladelet knife.
The pottery is also representative of Oasis 2 types and is a coarse brownish-grey
ware. Many of the shards show extensive exfoliation on the interior and exterior
surfaces. The clay contains sand grains and organic temper. In some pieces, individual
fibres are evident, including what appear to be coarse black hairs. The two radiocarbon
dates are essentially identical and dated shards are probably from two pieces of the same
vessel.
409
Alashan Gobi
Gashun Well (K. 13207)
The Gashun collection is listed in the MFEA catalogue as K.13207. Artefacts were
picked up from the surface near a well 10 km northeast of Hoyar-amatu, an
archaeological locality and stopping point on the caravan route. The location is near the
border between Mongolia and the Inner Mongolia Autonomous Region. It appears to
have been a workshop site with many large lithic specimens, probably unfinished
adze/axes, all made on the same brown jasper (Maringer, 1950: 129). Similar specimens
are included in collections from Abderungtei (K.13209: 128) and Mongol (K.13210:
132), slightly farther west.
The dated shard, which places the site assemblage at 3.6k cal yr BP (3385 + 40
BP; AA91693, MFEA #K.13207: 1), is from a partially reconstructed sand-tempered
bowl with very thin walls and string-paddled markings on the exterior surface. The bowl
is uniformly blackened on the interior surface and has several darkened patches on the
exterior. What appears to be a short (0.5 cm long) raised band of clay may have been
applied near the rim as a sort of miniature or decorative lug, but might also be incidental.
Four pieces of higher-fired, thicker, and heavily sand-tempered shards (MFEA #K.13207:
2) are similar to the Shabarakh-usu 4 shard radiocarbon dated to 4.0k cal yr BP. They are
reddish- and greyish-brown and darkened on the interior surface.
The lithic assemblage is consistent with Gobi-Altai assemblages in both the range of raw
materials and tool types. Microblade cores include the same type of barrel-shaped cores
found at Shabarakh-usu 4, as well as wedge-shaped, conical, and cylindrical forms. Core
410
reduction strategies appear biased towards the cylindrical form and many of the wedgeshaped specimens are more cylindrical with round bases. Raw material used in the
manufacture of microblade cores and microblades is highly consistent, suggesting that
microblades were detached and discarded at the same site with parent cores. Brown and
yellow jasper are most common, followed by red jasper similar to the kind found in GobiAltai sites. Perforators include fine unifacially retouched specimens and one
perforator/point on a thick, steeply retouched microblade. One scraper is on the end of a
microblade or elongated flake, while the other four are on microlithic flakes, a large
amorphous flake, and a chip or cobble.
Bifacial macrotools are notable. There are five unfinished implements made on
the same brown jasper and retouched primarily along the edges. The other two forms are
one plane-like tool and one large, thin unfinished biface that may have been an axe
preform. The similarity to macrotools from Abderungtei (K. 13209) and Mongol (K.
13210) is notable. Curved bifacial blade knives were found at all three sites. This
association further supports the chronological association of curved blade knives with
Oasis 3 tool kits. Steeply retouched microblade tools and endscrapers on microblades are
also consistent with Oasis 3 assemblages.
Jabochin-khure (K. 13203)
Jabochin-khure was dated using luminescence on pottery to about 3.5 ka (3490 + 285 ka),
making it broadly contemporaneous with Shabarakh-usu 4 and 10, as well as Gashun
Well (K: 13207). A wide range of error is unavoidable because we were unable to
411
control for external dose rates, but it does not preclude a solid Oasis 3 date. This group
of artefacts was found near an unidentified square ruin just over 40 km south of Hoyaramatu. The dated pottery (MFEA #K: 13203: 5) is decorated with a geometric incised
design that is diagnostic of Oasis 3 sites (Figure 3.8c, d). The paste is sand-tempered and
low-fired. While the exterior surface was light brown with darkened patches, the interior
paste is dark grey and somewhat porous, indicating the organic content of the clay. Sandtempered plain-ware manufactured from a similar paste may be from an undecorated
portion of the same vessel, or from another undecorated pot. Debitage flakes include
microblade platform core rejuvenation and preparation spalls made of yellowish-brown
and red jasper. Cores include one wedge-shaped microblade core and one test piece.
Mantissar 12 (K. 13298)
Mantissar 12 was one of many small sites located in the Gurnai (Gurrunai) Depression, a
large erosion basin bounded by the Ruoshui/Heihe (Etsin-gol) drainage system to the
west and the Badain Jaran Desert in the east. The depression extends about 50 km from
west to east. The most deeply eroded regions reach ground-water level and are
overgrown with reeds. Aeolian deposition has resulted in sand from the depression being
redeposited on the southern and eastern border of the basin in the form of a dune belt
almost 40 m high, which intermittently juts out in a spur-like formation. This stretch was
reported to be about 100 km from north to southwest, along which a “nearly
uninterrupted series of small prehistoric sites” was discovered within the scrub and reeds
or along the reed-saxual transitional zone (K. 13277-13319) (Maringer, 1950: 151-152).
412
Many of the collected sites were found within about 10 km of a small lake or playa
feature called Ulan Nor. Current conditions suggest that the area was once occupied by
extensive dune-fields and marshland, complemented by a small freshwater lake. Gurnai
Depression sites are notable in the high frequency of ostrich eggshell fragments
associated with bead manufacture. Radiocarbon dates on ostrich eggshell artefacts verify
that inhabitants were collecting and modifying Pleistocene ostrich eggshell, most of
which provided infinite-age radiocarbon dates (Appendix A.3).
Two shards were selected from Mantissar 12 for luminescence dating, including
one cord-marked (K. 13298: 15, described as having “vigorous textile impressions” –
Maringer, 1950: 160; see Figure 3.8d) and one plain brown burnished shard (K. 13298:
25). The first produced a date of about 6.5 ka (6460 + 700; UW2362, MFEA #K. 13298:
15). The second shard was dated to 3.9 ka (3870 + 340; UW2359, MFEA #K. 13298:
25). About 22% of the 169 shards from this site are high-fired and untempered or lightly
sand-tempered red-ware with traces of painted black lines, including a checkered or
lattice design. Amongst the untempered shards, the striking homogeneity of the fabric,
including uniform coloration revealed in cross-sections of the interior paste, indicate that
before use the clay was probably cleaned of impurities and then fired in a kiln under
controlled conditions (compare to descriptions of Banshan pottery manufacture in
Palmgren, 1934: 1, 3-4 and contra 5-6).
There is a high range of error in the two dates, but they securely date occupation
of the Gurnai Depression to both Oasis 2 and Oasis 3. The first date is broadly
contemporaneous with Yangshao sites of the middle Neolithic in China (Chang, 1987;
413
Liu, 2004). The later Oasis 3 date spans both the Qijia (4.2-3.8 ka – after DebaineFrancfort, 1995) and Siba (3.7-3.6 ka – after An, 1992b) archaeological cultures. Qijia
pottery is characterized by fine red-ware, coarse reddish-brown ware, and some greyware. Surface treatments include rare occurrences of burnishing, smoothing, white slip,
cord impressions, and basket-impressions. Painted pottery is rare, but designs include
lines and checks (An, 1992b). Siba pottery is less well known, but noted as
characteristically poorer in quality and less diverse than that of contemporaneous groups
in the region, or preceding periods. Sand- and gravel-tempered plain-ware is most
typical. Occasional finds of painted pottery show very simple and repetitive patterns.
Red and black paints were applied in thick layers. Impressed “N” and “Z” patterns and
wedge shapes are most the most frequent type of surface design (Yang, 1998).
The majority of painted pottery from Mantissar 12 is consistent with either Qijia
or Majiayao pottery. Majiayao (4.7-4.3 kya [2700-2300 BC] – after An, 1992a) pottery is
usually red, and painted in black with zoomorphic or geometric designs in delicate lines,
curves, dots, triangles, and impressed checks. Cord-impressed and moulded surface
decorations were also used (An, 1992a). Since the shards recovered from the Gurnai
Depression are small, it is difficult to compare them with the wealth of complete Chinese
Neolithic vessels. The pottery from Mantissar 12 is most consistent with either Majiayao
or Qijia ceramic traditions. Luminescence dates coincide best with Qijia, but indicate
that the Gurnai Depression was probably inhabited throughout both periods.
In general, the Mantissar 12 assemblage is characterized by red and reddishbrown pottery. Fabric types include an untempered ware with only a few incidental
414
inclusions, light sand temper, and some heavily sand-tempered shards. Evidence of
organic inclusions is extremely rare. Even sand-tempered shards show a paste that is
very uniform. Most pottery fragments are consistent with a higher firing temperature
than is common at other Gobi Desert sites, though a few shards appear to have been fired
at lower temperatures. One shard exhibits a heavily blackened core, despite appearing to
have been fired at a high temperature. One rim fragment of a fine, high-fired black-ware
was also recovered from this site, but is atypical. Some indications of vessel design are
evident in this group. Handle fragments found at Mantissar 12 are typical of Oasis 3 and
were also recovered from Shabarakh-usu. Some handles are from painted red-ware and
others from a heavily sand-tempered dark red-ware. Another heavily sand-tempered
fragment of greyish-brown plain-ware is from a flat-bottomed vessel. A third fragment
of sand-tempered dark red-ware is from a globular vessel with constricted neck and flared
rim. Well-rounded thicker rims are characteristic and, based on the appearance of fine
uniform striations, may have been smoothed using a slow wheel (as described for
Banshan [Panshan] pottery by Palmgren, 1934: 3). Drill holes are rather common in sites
from this region and suggest curation by extensive pot repair.
Surface treatments are highly variable for such a small site. Plain-ware is most
common, making up 36% of all shards. Vessel surfaces are often smoothed. There are a
few examples of burnished brown-ware, including one of the dated shards. Decorative
treatments include intersecting and parallel incised lines, hand-moulded undulating rims
(beneath which several shards are ringed by distinct vertical clusters of small circular
punctates), intersecting rows of parallel rolled cord impressions, faint textile impressions
415
(including on low-fired, drilled, brown-ware), textile or basket impressions, and cordpaddled. One shard is decorated with parallel rows of rectangular grooves subtly
reminiscent of channelled ware from Shabarakh-usu. Pottery types are indicative of both
Oasis 2 and Oasis 3 occupations.
The lithic assemblage from Mantissar 12 is representative of Gurnai Depression
sites, showing both similarities and notable divergences from other Gobi Desert Oasis 3
sites. Chalcedony (either yellow and white, or translucent) is the main raw material,
along with some poor quality jasper or siliceous sandstone and quartzite. The lithic
assemblage includes cores and core fragments, unused and retouched flakes, perforators
on microblades, drills on microblades, scrapers, and two small splinters of polished stone
implements. Drills were very extensively retouched and there are distinctive forms like
the double-ended drill. Only six scrapers are included in this collection, four from
amorphous microlithic flakes, one is on the end of a microblade and the other on a thick
elongated flake. No bifaces were recovered from Mantissar 12, but fragments of blade
knives and bifacial projectile points were collected in other Gurnai Depression sites,
including one stemmed point and one with a straight base.
The six cores from Mantissar 12 are particularly notable in form and manufacture.
They are unusually small, with a mean approximate volume of 33.7 cm3, in comparison
to the overall Alashan Gobi mean of 179.1 cm3 (see Chapter 4). The smallest of these is
0.4 x 1.4 x 1.1 cm, while the largest measures 3.2 x 2.1 x 1.4 cm. Keel flakes from initial
core preparation range in length from between about 2.5 cm to under 1.0 cm, further
suggesting high variability in the size of the original prepared cores. The range of raw
416
materials used at Gurnai Depression sites suggests a similar source – perhaps the use of
small locally available pebbles. The diminutive size may be related to small nodule size
and/or intensive conservation of raw material due to limited access. Only three of the 25
cores analyzed from the Gurnai Depression retain any cortical remnant. Cores are
indicative of typical Neolithic/Eneolithic Gobi Desert reduction sequences that result in
flat-backed conical and cylindrical specimens. Less formal types were also made on
small cobbles. They were reduced in a similar sequence to standard wedge-shaped
microblade cores, but lacked a bifacial wedge opposite the striking face.
Yingen-khuduk (K. 13212, 48)
The Yingen-khuduk locality includes both MFEA site collections K. 13212: 1-186 and
48: 1-100. The locality was particularly rich and comprised numerous site groups
collected from the desert surface. Yingen-khuduk sits on the Mongolian border and was
an oasis of scrub with a few isolated trees, where drift sand had come to form dunes. The
locality was situated in a vast dune-filled basin, just north of an ancient lake bottom and
lined on the north with red cliffs (Bergman, 1945: 158; Montell, 1945: 367; Maringer,
1950: 127, 130). Folke Bergman, the Sino-Swedish Expedition archaeologist, reported
encountering the carcass of a large web-footed bird along the road and a flock of swans
flying south on the day before arriving at Yingen-khuduk (April 7, 1931) (Bergman,
1945: 158). During wetter periods, the area was rich in avian fauna. The locality is south
of a transitional zone of scattered foothills belonging to the Gobi-Altai range, alluvial
plains, and open desert plains intersected with drainage channels and small lakes.
417
Numerous sites scattered across the open territory along the Mongolian border suggest
extensive prehistoric occupation. Judging by dates for the Yingen-khuduk assemblage,
the region was exploited throughout the Neolithic.
Occupation of the Yingen-khuduk site in the Alashan Gobi, as with Baron
Shabaka and Mantissar 12, appears to span Oasis 2 and Oasis 3 (5690 + 350 ka
[UW2357, MFEA #K. 13212: 123], 3910 + 300 ka [UW2358, MFEA #K. 13212: 6],
3910 + 230 ka [UW2360, MFEA #K. 13212: 128]). As with the Ulan Nor Plain site, the
Yingen-khuduk artefact assemblage is different from early Oasis 2 sites, as represented
by Jira Galuntu, Chilian Hotoga, and Baron Shabaka South. It is not clear which artefact
types are associated with which date; therefore, it is difficult to identify temporally
diagnostic artefacts. Despite this, directly dated “net-impressed” pottery (Figure 3.8a) is
considered diagnostic of Oasis 2. There is an abundance of diagnostic Oasis 3 artefacts,
but as we have seen with the Ulan Nor Plain site it is possible that many such specimens
are transitional technologies, belonging to a late phase of Oasis 2.
Two essentially identical Oasis 3 luminescence dates of about 3.6 ka were derived
from Yingen-khuduk ceramic shards. These are consistent with Oasis 3 dates from
Shabarakh-usu and other Alashan Gobi sites. The first is from a fragment of red highfired,string-paddled pottery, the fabric of which was porous and lightly tempered with
coarse sand or gravel. The second is from a fragment of high-fired plain red-ware with a
homongeneous, untempered paste. Darker patches of red are suggestive of a red slip or
paint on the exterior surface. A similar fragment is derived from the same collection and
is lightly sand-tempered. Such shards are found in other Alashan Gobi sites as well,
418
including Hoyar-nor (Khoburin-nor, K. 13176). Examples of a similar fabric are found in
the Gurnai Depression sites, but often painted in black lines or swirls (Figure 3.7).
Much of the pottery is high-fired. A more porous texture suggests organic temper
or highly organic clays, but shard cross-sections show consistent colour and texture
throughout. Organic temper appears to have been less favoured in the Alashan Gobi than
in the Gobi-Altai or East Gobi. Heavy sand or coarser gravel is a typical temper. Surface
finishes at Yingen-khuduk include, in order of descending frequency, textile or basket
impressions (often smeared), paddled, plain or slipped, and net-impressed. Some highfired red-wares have traces of red or black paint. Surface treatments in this group are
representative of Oasis 3 Alashan Gobi ceramics. One specimen is from the rim
fragment of a large narrow-mouthed jar with a wavy moulded rim and an exterior surface
covered in a smeared basket or net impressions (Maringer, 1950: Pl. XXII, 1). The walls
are rather thin and the paste untempered. Another partially reconstructed fragment is
from the belly of a more globular vessel. The exterior surface shows a smeared cordmarked finish. The shard is thick-walled, brown with reddish patches, and appears to
have been more highly fired. The paste is untempered or only lightly sand-tempered.
Geometric-incised pottery was also recovered at Yingen-khuduk. Various pieces
resemble those from Shabarakh-usu 4 and Jabochin-khure. Other plain-ware shards have
darkened interior pastes and are tempered with coarse sand or gravel and rather porous.
Much of the pottery, particularly the high-fired types, probably belongs to the
later occupation and can be compared to Oasis 3 shards from other sites. While it is
tempting to associate the coarser pieces with the Oasis 2 component, dates from Gashun
419
Well and Shabarakh-usu clearly indicate that both types of pottery were used in later
periods. Several examples are of a coarse, sand-tempered red-ware with miniature luglike protrusions similar to those on the dated shard from Gashun Well. Thin-walled
greyish-brown wares from the site are similarly string-paddled and sand-tempered (K:
13212: 5). Only some of the shards are blackened on the interior surface, but all have
darkened patches on the exterior surface. One possible intrusive element is a high-fired,
sand-tempered red-ware with a scraped exterior surface, possibly dating to the Turkic
Period.
According to dates on pottery from Yingen-khuduk, the lithic assemblage should
be considered representative of both Oasis 2 and Oasis 3 occupations. Formal macrotools
from Yingen-khuduk likely belong to Oasis 3 and include two fully polished and one
chipped specimen. The first of the polished pieces is a shattered and partially
reconstructed adze/axe, and another is a thin (0.8 cm), finely polished axe of green,
mottled translucent stone similar to jade. The latter shows heavy use on the distal end
and was broken about 2.5 cm from the working end. Striations are clearly visible across
the surface. Two chips of polished stone implements were also recovered. The third
macrotool is expediently and roughly chipped. It is a rounded hoe-like tool narrowing at
the neck. A comparable artefact was recovered at Shine-usu (K. 13259), a large surface
site lying between Gashun-nor and Sogho-nor in the Juyanze region. One red sandstone
hand-stone or runner was found at Yingen-khuduk and another is associated with the
collection made by Bergman and Hedin in 1933 (MFEA #48: 1-100). The former shows
an elongated oval profile, with one end mostly unfinished and the other heavily used and
420
slightly rounded. Partially finished ostrich eggshell beads and fragments of worked fossil
eggshell were also found with the primary assemblage, one of which was dated to 41.9k
cal yr BP (41,900 + 1500 BP [AA87198, MFEA #K. 13212:184]).
Jasper and chalcedony are the most common raw materials at Yingen-khuduk,
with red jasper predominating. The majority of cores are of jasper (41%), while bifaces
and scrapers are mostly on chalcedony (67% and 40%, respectively). Core types include
conical (24%) and cylindrical (19%), as well as wedge-shaped (14%) forms. Both
informal amorphous flake cores and formal biface cores are common. Tools are made on
amorphous flakes, microblades, and thick bladelets or elongated flakes. Many such
flakes were used without retouch or only lightly retouched and used. More formal tool
types include scrapers, awls, and drills – including the expanding base variety. Scrapers
on amorphous microlithic flakes are most common (46%), followed by those reworked
from thick elongated flakes (16%), and scrapers on the end of a microblade or elongated
flake (13%). Other scraper types include amorphous scrapers on cobbles or chips (9%),
formal thumbnail (6%), elongated thumbnail or tongue-shaped scrapers (6%), one scraper
on a macrolithic flake, and one heavy-duty scraper or small plane (2.8 x 2.9 x 1.6 cm).
Bifacially retouched flakes or thin, reduced chalcedony cobbles appear to be unfinished
preforms or expedient knives. Two blade knives constitute the sample of more formal
bifaces.
421
APPENDIX C – ARTEFACT TYPOLOGIES
C. 1. Artefact summary for each site
SITE
EAST
GOBI
SKW
PERIOD
CORES
SCRAPERS
O1
N=2
123 = 2
N=1
320 = 1
3
O1
N=1
131 = 1
N=1
330 = 1
6D
O1
N = 22
101 = 8
120 = 5
122 = 2
123 = 4
124 = 1
126 = 1
140 = 1
N = 11
301 = 3
320 = 5
321 = 1
324 = 1
325 = 1
7
O3
N = 12
120 = 1
122 = 4
126 = 7
N = 38
300 = 1
301 = 10
320 = 24
325 = 3
9
N/A
N=1
122 = 1
9B
P?
N=1
301 = 1
9C
O2/3
N=8
320 = 4
321 = 1
325 = 3
9D
O2/3
10/10A/10B
O2/3
N=5
101 = 1
122 = 1
123 = 1
126 = 1
128 = 1
N=9
301 = 1
320 = 7
325 = 1
BIFACES
POTTERY
OTHER
N = 11
1=2
12 = 9
Dates
Formal wedgeshaped core
Chalcedony
Large flakes
Microblades
Chalcedony/lava
Ostrich eggshell
Chalcedony/lava
N=5
1=1
4=1
6=1
10 = 1
12 = 1
Rough wedgeshaped cores
Backed
microblades
Chaldecony
Ostrich eggshell
Cowry
Core preform
Lava
Rough macrotool
Lava
N=1
220 = 1
N=3
201 = 2
214 = 1
Lava
N=1
200 = 1
N = 10
200 = 4
211 = 6
Silicified lava or
sandstone
Lava and
chalcedony
422
SITE
EAST
GOBI
11/11A
PERIOD
CORES
SCRAPERS
BIFACES
POTTERY
OTHER
O3
N = 32
301 = 14
320 = 14
321 = 1
323 = 1
325 = 1
330 = 1
N = 12
200 = 1
20 1= 3
212 = 1
214 = 1
216 = 1
220 = 2
222 = 1
260 = 1
261 = 1
N=1
1=1
Chalcedony and
lava
12
O2
12/12A/12B
P, E, O2
13
O2?
N = 51
101 = 18
120 = 6
122 = 9
123 = 3
124 = 2
125 = 1
126 = 3
127 = 1
129 = 2
130 = 1
140 = 5
N=5
101 = 1
120 = 1
123 = 1
126 = 2
N = 19
101 = 9
120 = 2
122 = 2
123 = 3
125 = 1
131 = 1
140 = 1
N=5
101 = 3
131 = 2
13A
O2/3
14
O1
15
O2
N=8
101 = 4
120 = 1
123 = 2
124 = 1
N=4
123 = 1
126 = 1
127 = 2
N = 11
101 = 5
123 = 1
130 = 1
131 = 3
140 = 1
N=4
301 = 3
320 = 1
Chalcedony and
lava
N = 10
301 = 3
320 = 5
323 = 2
Lava and
chalcedony
N=8
301 = 1
320 = 3
321 = 2
325 = 1
330 = 1
N=8
320 = 6
321 = 1
324 = 1
Chalcedony, lava,
jasper
N=1
211 = 1
N=2
325 = 2
N=8
301 = 4
320 = 3
321 = 1
Jasper
N=2
1=1
9=1
N=1
214 = 1
Chalcedony and
lava
Chalcedony and
lava
423
SITE
EAST
GOBI
19
PERIOD
CORES
SCRAPERS
BIFACES
POTTERY
OTHER
P/E?,
O2, O3
N = 509
100 = 1
101 = 17
120 = 94
121 = 1
122 = 52
123 = 99
124 = 33
125 = 21
126 = 85
127 = 31
128 = 9
129 = 9
130 = 2
131 = 27
132 = 10
140 = 18
N = 487
301 = 125
310 = 18
320 = 170
321 = 31
322 = 2
323 = 51
324 = 13
325 = 18
330 = 59
N = 61
200 = 12
211 = 3
214 = 6
215 = 2
216 = 3
222 = 4
241 = 7
260 = 19
261 = 5
N = 323
0=4
1 = 166
3 = 31
4 = 36
5=4
5/8/10 = 1
8=2
8/11 = 1
9 = 29
9/10 = 1
9/11 = 1
10 = 15
11 = 3
12 = 22
14 = 7
20
O2/3
N = 10
101 = 5
122 = 2
123 = 2
126 = 1
N = 11
301 = 6
320 = 5
N=5
4=2
5/8 = 1
12 = 2
20A
Various
N=2
123 = 2
N=1
320 = 1
21
Early
Oasis 2
N = 15
101 = 4
120 = 2
123 = 8
130 = 1
N = 17
301 = 7
320 = 8
323 = 1
325 = 1
N=9
200 = 2
201 = 2
215 = 3
216 = 1
222 = 1
N=3
201 = 1
214 = 1
224 = 1
N=5
201 = 2
214 = 2
241 = 1
Dates
Unifacial knives
Unifacial points
Adze/axes
Small adze
Pick
Gouge
Chisel
“Hoe”
“Whetstones”
Ground ring
Manos
Rollers
Pestles
Metates
Spindle whorls
Chalcedony
dominant
Iron (Historic)
Ostrich eggshell
Pestle
Metate
Drill
23/23A
O3
N = 46
100 = 2
101 = 15
102 = 1
120 = 3
122 = 10
123 = 1
124 = 5
126 = 7
127 = 1
140 = 1
N = 10
301 = 3
320 = 3
325 = 4
N=2
201 = 1
216 = 1
N = 26
1=7
9 = 18
12 = 1
Mano
Ostrich eggshell
Unifacial points
Formal wedgeshaped core
Metate
Perforator
Ostrich eggshell
Mano
Metate
424
SITE
EAST
GOBI
28
PERIOD
CORES
SCRAPERS
BIFACES
POTTERY
OTHER
O3
N = 24
301 = 12
320 = 12
N = 19
200 = 1
201 = 9
211 = 6
261 = 2
263 = 1
N=3
3=1
Unifacial knife
Adze/axe
Small adze
Large axe
“Whetstone”
Manos
Metates
Beads
29
O2/O3
N = 25
301 = 12
320 = 12
324 = 1
N=3
201 = 1
222 = 2
Polished adze/axe
Mano
Roller
Handheld mortar?
Drills
30/30A
Late O3
N = 13
301 = 3
320 = 5
321 = 2
325 = 3
N=2
200 = 1
201 = 1
Adze/axe
Metate
31
O2/O3
N = 44
301 = 18
320 = 21
323 = 1
325 = 2
330 = 2
N = 21
200 = 3
211 = 8
214 = 1
215 = 1
217 = 1
220 = 3
222 = 2
241 = 2
“Whetstones”
Manos
Pestles
Metate
Small adze
Drills
Shell bead
34
O3?
N=6
201 = 6
36
O1
N = 42
101 = 28
120 = 3
122 = 2
126 = 3
130 = 2
131 = 1
132 = 1
140 = 2
N = 24
101 = 9
120 = 2
122 = 5
123 = 1
126 = 2
127 = 1
128 = 1
132 = 1
140 = 1
N = 22
101 = 7
120 = 5
122 = 2
123 = 5
125 = 1
126 = 1
127 = 1
N = 75
101 = 18
120 = 14
122 = 13
123 = 17
124 = 2
125 = 1
126 = 2
127 = 5
129 = 1
140 = 2
N = 19
101 = 14
120 = 5
N=5
101 = 3
123 = 1
140 = 1
Chalcedony
biface blanks
Ostrich eggshell
2.5 km north of
Chilian Hotoga
N=1
320 = 1
425
SITE
GOBIALTAI
Cemetery
Mesa (CM)
PERIOD
CORES
SCRAPERS
BIFACES
E, O3
N=2
101 = 2
N=5
320 = 4
323 = 1
Gashuin
Bologai
Well
(GBW)
Early O2
N=2
123 = 2
Barongi Usu
Valley
(BUV)
O1?
N=4
301 = 1
320 = 1
330 = 2
Dubshi Hills
(DH)
P
N = 14
101 = 6
120 = 4
123 = 1
129 = 1
130 = 1
131 = 1
N=1
120 = 1
Ulan Nor
Plain (UNP)
O2
N = 19
301 = 4
310 = 4
320 = 4
321 = 1
322 = 1
324 = 1
325 = 2
326 = 1
330 = 1
N = 37
200 = 4
201 = 4
216 = 3
222 = 2
224 = 1
260 = 23
Jichirun
Wells (JW)
Early O2
N=8
301 = 4
310 = 1
320 = 1
325 = 1
330 = 1
N=5
220 = 3
260 = 1
261 = 1
Jichirun
Wells, in
situ
Early O2
N = 139
101 = 67
120 = 3
122 = 18
123 = 7
124 = 1
125 = 1
126 = 3
127 = 3
128 = 2
129 = 1
130 = 31
140 = 2
N = 25
101 = 9
120 = 10
122 = 1
127 = 1
128 = 1
130 = 1
131 = 1
140 = 1
N=6
101 = 3
120 = 2
130 = 1
N=2
301 = 1
320 = 1
N=3
214 = 1
226 = 1
260 = 1
POTTERY
OTHER
N=2
1=2
“Whetstone”
Camel effigy
Vicinity of
burials and stone
monuments
Petroglyphs
Lava stone
mortar
N=1
200 = 1
Jasper
Intrusive
elements,
including Chinese
glazed shard
N=1
301 = 1
N = 42
1=4
2 = 23
4=1
5=1
6=3
10 = 10
Macroflakes
Yellow
chalcedony
Dates
Mano
Curved bifacial
knife
Jasper
Bones and equid
teeth
426
SITE
GOBIALTAI
Sairim
Gashoto
(SG)
PERIOD
CORES
SCRAPERS
E
N=2
300 = 1
301 = 1
Arts Bogd
(AB)
O2
Khunkhur
Ola (KhO)
Metal
Ages?
N=3
101 = 1
120 = 1
130 = 1
N=4
101 = 1
120 = 2
124 = 1
N=3
101 = 3
Barun
Daban (BD)
O2, O3
N = 30
301 = 3
310 = 1
320 = 20
322 = 1
323 = 5
N = 38
1 = 12
3=8
4 = 17
9=1
Orok Nor
(ON)
O2, O3
N = 13
101 = 1
120 = 1
122 = 1
123 = 5
126 = 3
127 = 2
N = 31
120 = 2
122 = 13
123 = 8
126 = 4
128 = 3
140 = 1
N = 35
301 = 5
320 = 19
322 = 2
323 = 5
325 = 4
N = 62
1 = 25
3 = 28
5=3
8=3
10 = 1
12 = 2
Salt Creek
(SC)
O3
N=2
320 = 1
323 = 1
Shabarakhusu 1
O3
N=7
101 = 3
122 = 1
123 = 1
126 = 1
129 = 1
N = 12
101 = 2
120 = 2
122 = 4
123 = 1
126 = 2
129 = 1
Perforators
Drills
Pendant
Shell bead
Ostrich eggshell
Vicinity of stone
monuments and
burials
“Whetstone”
Bronze
arrowpoint
Stone structures
N = 72
1 = 12
2=2
3 = 48
3/5 = 5
4=1
5=1
8=2
11 = 1
Dates
Axe
Small metate
Perforators
Drills
White chalcedony
Ostrich eggshell
N=3
320 = 2
321 = 1
BIFACES
POTTERY
Jasper
N=2
211 = 2
Blade tools
N=1
301 = 1
N = 22
301 = 4
320 = 10
321 = 1
323 = 5
324 = 2
OTHER
N = 43
200 = 7
201 = 1
211 = 2
212 = 3
214 = 9
215 = 6
216 = 6
220 = 2
221 = 1
222 = 6
Jasper and
chalcedony
Vicinity of stone
monuments
Polished adze/axe
427
SITE
GOBIALTAI
Shabarakhusu 1A
PERIOD
CORES
SCRAPERS
O2
N = 25
101 = 12
120 = 3
122 = 6
123 = 1
126 = 1
130 = 1
140 = 1
N = 58
301 = 4
310 = 1
320 = 39
321 = 4
323 = 6
324 = 1
325 = 2
330 = 1
Shabarakhusu 2a
E/O1
N=2
101 = 2
Shabarakhusu 2b
O1
N=3
120 = 1
122 = 1
131 = 1
N = 40
101 = 9
120 = 14
122 = 3
123 = 6
129 = 1
130 = 2
131 = 3
132 = 2
N=3
125 = 2
126 = 1
Shabarakhusu 7
Shabarakhusu 10
ALASHAN
GOBI
176
O3
N = 15
101 = 1
120 = 1
122 = 5
123 = 1
125 = 1
126 = 1
127 = 4
128 = 1
BIFACES
POTTERY
OTHER
Grooved slabs
Small metate
Jasper
“Mossy chert”
From valley floor
- base of dunes
Perforator
Ostrich eggshell
N = 52
301 = 3
320 = 41
321 = 1
323 = 3
324 = 3
325 = 1
N = 75
200 = 36
201 = 6
211 = 21
216 = 2
220 = 1
221 = 5
222 = 2
260 = 2
N=1
1=1
Shouldered drills
Perforators
Jasper
Tools/cores on
very small
nodules
Ostrich eggshell
N=7
320 = 3
321 = 1
323 = 2
325 = 1
N = 97
1 = 15
2 = 19
3 = 50
4=1
5=6
5/8 = 1
8=1
8/13 = 1
13 = 3
Dates
Small metate
Jasper
N = 12
301 = 1
320 = 3
321 = 2
324 = 6
N = 109
0 = 47
1 = 48
2=3
5=1
7=4
10 = 2
11 = 1
14 = 3
Partially polished
axe
Within 600 m of
scatter with
“Ordos-style”
bronzes
428
SITE
ALASHAN
GOBI
179
PERIOD
CORES
SCRAPERS
O3
N=3
122 = 2
131 = 1
N=1
301 = 1
183
O3
N=2
122 = 1
127 = 1
186
O2?
N=4
122 = 3
128 = 1
188
unknown
202
E?
N=2
101 = 2
N=5
120 = 1
131 = 4
203
O3
204
E
207
O3
208
O3
N=4
101 = 1
122 = 2
123 = 1
N=2
101 = 1
131 = 1
N = 17
120 = 1
123 = 3
125 = 1
126 = 5
127 = 3
128 = 1
131 = 3
N = 11
123 = 1
124 = 2
125 = 4
127 = 3
129 = 1
N = 22
300 = 1
310 = 1
320 = 18
321 = 1
323 = 1
N=1
320 = 1
N = 12
301 = 4
320 = 5
321 = 2
322 = 1
BIFACES
POTTERY
OTHER
N = 13
0=5
1=7
11/12/13 = 1
N=1
6=1
Within 600 m of
scatter with
“Ordos-style”
bronzes
Axe
Drill
Painted pottery
Large
microblades
N=2
200 = 1
260 = 1
N=1
10 = 1
N = 15
1=9
11/14 = 6
N = 32
300 = 1
301 = 9
310 = 4
320 = 9
330 = 9
N=5
301 = 1
310 = 1
320 = 2
323 = 1
N=1
200 = 1
N=5
260 = 5
N=3
260 = 1
261 = 2
Large flakes,
cobbles
Pottery intrusive?
Dates
No microblades
N=5
1=4
3=1
Dates
Adze/axes
Perforators
Adze/axe
429
SITE
ALASHAN
GOBI
212
PERIOD
CORES
SCRAPERS
BIFACES
POTTERY
OTHER
O2,O3
N = 45
301 = 2
310 = 1
320 = 21
321 = 3
322 = 4
323 = 4
324 = 9
330 = 1
N=9
200 = 2
222 = 2
260 = 5
N = 52
0=3
1=6
2=5
3 = 14
4=1
5=2
5/10 = 1
6=3
9 = 10
10 = 2
11 = 2
14 = 1
15 = 1
20 = 1
Dates
Adze/axes
Thin polished axe
Mano
Drills
Ostrich eggshell
213
O2
216
O3
N = 88
100 = 1
101 = 3
120 = 2
122 = 12
123 = 17
124 = 7
125 = 3
126 = 14
127 = 5
128 = 9
129 = 1
130 = 1
131 = 1
132 = 5
133 = 5
140 = 2
N=4
120 = 1
124 = 1
125 = 1
126 = 1
N=2
126 = 1
127 = 1
218
O3
219
P
220
O2
222
O3
N = 24
100 = 1
101 = 2
122 = 17
124 = 1
125 = 1
126 = 2
N=2
320 = 2
N=8
320 = 6
321 = 1
324 = 1
N = 18
300 = 1
320 = 7
321 = 3
323 = 6
324 = 1
Partially polished
adze/axe
N=1
200 = 1
Perforators
Awl
N=1
310 = 1
N=4
101 = 1
127 = 1
131 = 2
N=2
123 = 1
129 = 1
N=1
222 = 1
N=6
301 = 1
320 = 3
323 = 2
N=1
200 = 1
N=1
1=1
430
SITE
ALASHAN
GOBI
223
PERIOD
CORES
SCRAPERS
BIFACES
POTTERY
OTHER
O2/O3
N = 102
300 = 12
301 = 19
310 = 14
320 = 43
321 = 6
324 = 7
325 = 1
N=5
200 = 1
230 = 1
240 = 3
N=6
0=1
1=2
3=3
Spearpoint
226
O1
229
unknown
N = 23
100 = 1
101 = 5
122 = 7
123 = 3
125 = 1
126 = 2
127 = 2
128 = 2
N=6
122 = 1
123 = 5
N=1
140 = 1
230
O3
(Metal
Ages?)
231
O2
237
P?
247
O2
248
Metal
Age
N=2
301 = 1
320 = 1
N = 87
101 = 2
120 = 12
121 = 1
122 = 10
123 = 7
125 = 2
126 = 9
127 = 9
130 = 9
131 = 11
132 = 13
140 = 2
N=4
130 = 2
132 = 1
140 = 1
N=3
122 = 1
128 = 1
132 = 1
N=8
122 = 8
N = 69
0=5
1 = 39
9=2
10 = 16
11 = 7
Macro or core
tool
Grooved slab
Chalcedony beadmaking
N = 34
301 = 5
310 = 9
320 = 6
323 = 4
324 = 2
330 = 8
N = 13
221 = 1
260 = 9
261 = 3
Axe
N = 13
301 = 3
310 = 10
N=3
260 = 2
261 = 1
Handaxe
N=2
320 = 2
N=2
226 = 1
260 = 1
Axe
Drills
Awls
N=3
301 = 1
320 = 2
N = 38
0 = 16
1 = 17
10 = 2
3=2
5=1
Dates
Slag
431
SITE
ALASHAN
GOBI
251
PERIOD
CORES
258
259
O2 or
O3
O2?
277
O3?
287
O3?
N=2
122 = 2
290
O3
N=6
122 = 1
123 = 2
124 = 1
126 = 1
127 = 1
N = 18
301 = 3
320 = 11
323 = 1
324 = 3
293
O3
N=5
122 = 3
126 = 1
140 = 1
N = 14
301 = 2
320 = 9
323 = 3
294
O3
N=1
125 = 1
N=2
301 = 1
320 = 1
unknown
SCRAPERS
BIFACES
POTTERY
OTHER
N=1
320 = 1
N=1
125 = 1
N = 197
100 = 1
101 = 21
120 = 33
122 = 45
123 = 16
124 = 4
125 = 8
126 = 23
127 = 24
128 = 3
130 = 1
131 = 10
132 = 4
140 = 4
N = 93
301 = 10
310 = 4
320 = 54
321 = 10
323 = 4
324 = 8
325 = 2
330 = 1
N=8
301 = 2
320 = 4
323 = 2
N=5
301 = 2
320 = 2
323 = 1
N=9
225 = 1
260 = 5
261 = 3
Unifacial knives
Adze/axes
Chisel
“Hoe”
Awl
Unifacial point
(perforator?)
N=7
0=3
1=3
9=1
N=1
5=1
Slag
Ostrich eggshell
N=1
222 = 1
N = 13
1=5
6=2
12 = 5
20 = 1
N=2
212 = 1
221 = 1
N = 11
0=3
1=4
4/13 = 1
6=1
12 = 2
N=1
6=1
Painted pottery
Polished stone
frag.
Turquoise frag.
Drills
Perforators
Ostrich eggshell
Painted pottery
Drills
Perforators
Ostrich eggshell
N=1
222 = 1
Unifacial knife
Ostrich eggshell
Painted pottery
Ostrich eggshell
432
SITE
ALASHAN
GOBI
298
PERIOD
CORES
SCRAPERS
O2, O3
N=6
101 = 1
122 = 3
126 = 1
127 = 1
N=7
301 = 1
320 = 4
323 = 1
324 = 1
303
O3
307
O3?
311
O3
316
O3
321
Early O2
322
P
323
E?
N=3
123 = 1
126 = 1
127 = 1
N=1
126 = 1
N=1
126 = 1
N = 25
101 = 5
120 = 4
121 = 1
122 = 1
123 = 8
126 = 4
127 = 2
N=2
101 = 1
131 = 1
N=2
127 = 1
131 = 1
BIFACES
N=2
323 = 1
324 = 1
N=6
320 = 5
323 = 1
N=4
320 = 2
323 = 1
324 = 1
N=9
301 = 1
320 = 5
322 = 1
323 = 2
N=8
301 = 4
310 = 1
321 = 1
322 = 1
330 = 1
N=5
300 = 1
310 = 1
320 = 1
321 = 1
330 = 1
N=1
222 = 1
POTTERY
OTHER
N = 169
0 = 10
1 = 61
4=1
4/5 = 4
5 = 15
6 = 39
9 = 11
10 = 1
11 = 11
12 = 4
13 = 12
Dates
Painted pottery
Drills
Perforators
Ostrich eggshell
N=3
1=2
12 = 1
Slag
Ostrich eggshell
Mixed stray finds
N=1
0=1
Ostrich eggshell
Mixed stray finds
Mixed stray
finds?
Axe
Perforators
Macroflake tools
Unifacial knife
Blade
Drill
433
SITE
ALASHAN
GOBI
324
PERIOD
CORES
unknown
N=1
123 = 1
SCRAPERS
BIFACES
POTTERY
OTHER
* Counts are only of studied artefacts: a percentage of each tool type was selected for
analysis in some of the largest assemblages (Baron Shabaka Well, Ulan Nor Plain).
Numbers refer to artefact codes listed below.
434
C.2. Codes for artefact types
ARTEFACT
CODE
ARTEFACT
CODE
Core type
100 unknown fragment
101 amorphous/expedient/
test piece
102 reworked tool
110 Levallois-style
120 informal blade/bladelet/
elongated flake
121 blade
122 unknown/unsuccessful
microblade
123 wedge-shaped
microblade
124 conical microblade
125 cylindrical microblade
126 flat or biface-backed,
pointed microblade
127 flat or biface-backed,
round bottomed microblade
128 cylindrical microblade
with wedge
129 boat-shaped microblade
130 biface
131 scraper core tool (usu.
bifacial)
132 formal biface core tool
140 core tool, informal
Scraper type
300 unknown scraper
301 amorphous scraper on block,
chip, cobble, etc.
310 on macrolithic flake
320 on microlithic flake
321 thumbnail scraper
322 tongue-shaped (extended
thumbnail)
323 end of bladelet/elongated flake
324 reworked from thick blade or
elongated flake
325 reworked from broken core or
tool
330 heavy-duty scraper
Macrotool
type
400 unknown macrotool
410 unifacial knife
420 adze/axe/chisel
421 small knife/adze
422 heavily reduced adze/axe
430 small wedge
440 pick
200 unknown fragment
211 point fragment
212 stemmed point
213 shouldered point
214 concave point
215 convex point
216 straight base point
217 shouldered point
220 unknown knife
221 knife fragment
222 knife – blade
223 knife – leaf-shaped
224 knife – lunate
225 knife – pillow
226 knife - spatulate
230 spear point/knife
(large, triangular point)
240 adze/axe
241 small adze/axe
260 large biface
261 semi-lunar
Pottery
finish
1 none/slipped
2 net
3 paddle
4 stamped
5 moulded
6 painted
7 glazed
8 incised
9 textile
10 brushed/smoothed
11 combed
12 corded
13 punctate
20 other
Biface type
435
APPENDIX D – ASSEMBLAGE CHARACTERISTICS
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
Oasis 1
11-100
4
Cooking
Manufacturing
Lithic
Reduction
Residential
B
3
Shara
Murun
River
Oasis 1
1011000
1
Manufacturing
Lithic reduction
6D
Shara
Murun
River
Oasis 1
11-100
4
Manufacturing
Lithic reduction
7
Shara
Murun
River
Oasis 3
1011000
4
Cooking
Manufacturing
Weaponry
Ornaments
Lithic reduction
9
Shara
Murun
River
Unknown
< 10
4
Lithic reduction
9B
Shara
Murun
River
Palaeo.
< 10
5
Manufacturing
9C
Shara
Murun
River
Oasis 2 or
Oasis 3
11-100
2
Manufacturing
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, W
% microblade: 100
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, K,
CT
% microblade: 0
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, W,
CT
% microblade: 36
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: A
Generalized: S
% microblade: 92
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: no
% microblade:100
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S
No cores
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S
No cores
EAST
GOBI
Shara
KataWell
Shara
Murun
River
Task site
Task site
Residential
B
Task site
Task site
Task site
436
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
EAST
GOBI
9D
Shara
Murun
River
Oasis 2 or
Oasis 3
< 10
5
Weaponry
Lithic reduction
Task site
10A
Southwest
Oasis 2 or
Oasis 3
11-100
4
Manufacturing
Weaponry
Lithic reduction
11/11A
Southwest
Oasis 3
1011000
1
Cooking
Manufacturing
Weaponry
Ornaments?
Lithic reduction
12
Southwest
Oasis 2
11-100
2
Manufacturing
Lithic reduction
12/12A/12B
Southwest
Palaeo.,
Epipalaeo.,
Oasis 2
11-100
4
Manufacturing
Lithic reduction
13
Southwest
Late Oasis
2?
11-100
4
Manufacturing
Lithic reduction
13A
Southwest
Oasis 2 or
Oasis 3
1011000
4
Manufacturing
Weaponry
Lithic reduction
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: A
Generalized: S
No cores
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: A
Generalized: S, W
% microblade: 80
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: A
Generalized: S, K,
B, CT, W
% microblade: 42
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, W
% microblade: 60
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, W,
CT
% microblade: 32
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, CT
% microblade: 0
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: A
Generalized: S, W
% microblade: 38
Task site
Residential
A
Task site
Task site
Task site
Task site
437
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
Oasis 1
1011000
4
Manufacturing
Lithic reduction
Bone
Task site
15
Southwest
Oasis 2
1011000
4
Manufacturing
Weaponry
Lithic reduction
19
Shara
Murun
River
Oasis 2,
Oasis 3
10015000
1
Cooking
Manufacturing
Woodworking
Weaponry
Ornaments
Lithic reduction
20
Shara
Murun
River
Oasis 2
and/or 3
11-100
1
Cooking
Manufacturing
Weaponry
Lithic reduction
20A
Shara
Murun
River
Various
11-100
1
Cooking
Manufacturing
Weaponry
Lithic reduction
21
Shara
Murun
River
Early Oasis
2
1011000
1
Cooking
Manufacturing
Woodworking
Weaponry
Ornaments
Lithic reduction
23/23A
Shara
Murun
River
Oasis 3
1011000
4
Cooking
Manufacturing
Weaponry
Lithic reduction
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, W
% microblade: 100
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: A
Generalized: K, B,
W, CT
% microblade: 9
Pottery: yes
Grinding: F/I, Sm-L
Adze/Axe: both
Specialized: A, P, D
Generalized: S, K,
B, W, CT
% microblade: 67
Pottery: yes
Grinding: I, Sm
Adze/Axe: no
Specialized: A, D
Generalized: S, K,
W
% microblade: 50
Pottery: no
Grinding: I, Small
Adze/Axe: no
Specialized: A
Generalized: S, K
% microblade: 100
Pottery: no
Grinding: Small
Adze/Axe: chipped
Specialized: A
Generalized: S, K,
B, W
% microblade: 53
Pottery: yes
Grinding: I, S-M
Adze/Axe: no
Specialized: A
Generalized: S, W,
CT
% microblade: 52
EAST
GOBI
14
Southwest
Residential
B
Residential
A
Residential
B
Residential
B
Residential
A
Residential
A
438
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
EAST
GOBI
28
Great Lake
Basin
Oasis 3
1011000
1
Cooking
Manufacturing
Woodworking
Weaponry
Ornaments
Lithic reduction
Residential
A
29
Great Lake
Basin
Late Oasis
2 or early
Oasis 3
1011000
5
Cooking
Manufacturing
Woodworking
Lithic reduction
30/30A
Great Lake
Basin
Late Oasis
3
1011000
2
Cooking
Manufacturing
Woodworking
Lithic reduction
31
Great Lake
Basin
Late Oasis
2 or early
Oasis 3
10015000
1
Cooking
Manufacturing
Weaponry
Ornaments
Lithic reduction
Bone
34
Great Lake
Basin
Oasis 3?
1011000
4
Lithic reduction
(includes 1
biface blank)
36
Great Lake
Basin
Oasis 1
11-100
1
Manufacturing
Lithic reduction
Pottery: yes
Grinding: F/I, S-M
Adze/Axe: both
Specialized: A
Generalized: S, B,
CT
% microblade: 12
Pottery: no
Grinding: F, S-M
Adze/Axe: ground
Specialized: D
Generalized: S, K,
B, W, CT
% microblade: 46
Pottery: no
Grinding: I, L
Adze/Axe: chipped
Specialized: N/A
Generalized: S, W
% microblade: 45
Pottery: no
Grinding: F/I, S-M
Adze/Axe: both
Specialized: A, D
Generalized: S, K,
B, W, CT
% microblade: 55
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: N/A
Generalized: no
% microblade: 0
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, W,
CT
% microblade: 20
Residential
A
Residential
B
Residential
A
Task site
Task site
439
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
GOBIALTAI
Cemetery
Mesa (CM)
Shabarakhusu Region
Epipalaeo.,
Oasis 3
11-100
4
Manufacturing
Lithic reduction
Camel effigy
Burial
monuments
Task site
Gashuin
Bologai
Well
(GBW)
Arts Bogd –
Ulan Nor
Barongi Usu
Valley
(BUV)
Arts Bogd –
Ulan Nor
Region
Early
Oasis 2
< 10
2
Cooking
Lithic reduction
Oasis 1?
11-100
1
Manufacture
Weaponry (or
knife?)
Lithic reduction
Dubshi Hills
(DH) Arts
Bogd – Ulan
Nor Region
Palaeo.
11-100
2
Manufacturing
Lithic reduction
Ulan Nor
Plain (UNP)
Arts Bogd –
Ulan Nor
Region
Oasis 2
1011000
1
Cooking
Manufacture
Weaponry
Lithic reduction
Jichirun
Wells (JW)
Arts Bogd –
Ulan Nor
Region
Early
Oasis 2
1011000
1
Manufacture
Lithic reduction
Jichirun
Wells, in
situ
Arts Bogd –
Ulan Nor
Region
Late Oasis
2
1011000
1
Manufacture
Weaponry
Lithic reduction
Bone
Pottery: no
Grinding:
whetstone
Adze/Axe: no
Specialized: no
Generalized: S
% microblade: 0
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: W
% microblade: 100
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: N/A
Generalized: S, B,
W, CT
% microblade: 14
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S
% microblade: 0
Pottery: yes
Grinding: F, Med.
Adze/Axe: no
Specialized: A
Generalized: S, K,
B, W, CT
% microblade: 26
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, K,
B, CT
% microblade: 12
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: A
Generalized: S, B
% microblade: 0
Task site
Residential
B
Task site
Residential
A
Residential
B
Task site
440
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
Epipalaeo.
11-100
5
Manufacturing
Lithic reduction
Task site
Oasis 2
11-100
4
Manufacturing
Weaponry
Lithic reduction
Khunkhur
Ola (KhO)
Arts Bogd –
Ulan Nor
Region
Metal
Ages?
11-100
4
Lithic reduction
Barun
Daban (BD)
Arts Bogd –
Ulan Nor
Region
Oasis 2
and Oasis
3
1011000
1
Orok Nor
(ON)
Valley of
the Lakes
Region
Oasis 2
and Oasis
3
1011000
1
Cooking
Manufacturing
Woodworking
Weaponry
Lithic reduction
Bone
Cooking
Manufacturing
Weaponry
Ornamentation
Lithic reduction
Bone
Salt Creek
(SC)
Shabarakhusu Region
Oasis 3
11-100
2
Manufacturing
Lithic reduction
Shabarakhusu, Oasis 2
Shabarakhusu Region
Oasis 2
5000+
1
Cooking
Manufacture
Weaponry
Ornamentation
Lithic reduction
Bone
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, B
% microblade: 0
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: A
Generalized: S
% microblade: 25
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S
% microblade: 0
Pottery: yes
Grinding: no
Adze/Axe: ground
Specialized: A
Generalized: S, W
% microblade: 85
Pottery: yes
Grinding: F, Med.
Adze/Axe: no
Specialized: A, P,
D?
Generalized: S, W,
CT
% microblade: 90
Pottery: no
Grinding: I?
Adze/Axe: no
Specialized: no
Generalized: S, W
% microblade: 57
Pottery: yes
Grinding: I/F, Small
Adze/Axe: N/A
Specialized: A, P, G
Generalized: S, K?,
B, W, CT
% microblade: 1a =
32, 8 = 71, 11 = 18
GOBIALTAI
Sairim
Gashoto
(SG)
Arts Bogd –
Ulan Nor
Region
Arts Bogd
(AB)
Arts Bogd –
Ulan Nor
Region
Task site
Task site
Residential
B
Residential
A
Task site
Residential
A
441
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
Oasis 3
5000+
1
Cooking
Manufacturing
Woodworking
Weaponry
Ornaments
Lithic reduction
Bone
Residential
A
S-u 2a
Shabarakhusu Region
Epipalaeo.
or Oasis 1
11-100
1
Lithic reduction
S-u 2b
Shabarakhusu Region
Oasis 1
11-100
1
Manufacturing
Lithic reduction
Pottery: yes
Grinding: I, Med.
Adze/Axe: both
Specialized: A, P,
D, G
Generalized: S, K,
B, W, CT
% microblade: 1 =
67, 2 = 34, 4 = 89, 7
= 25, 10 = 100
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: no
% microblade: 0
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: P?
Generalized: CT
% microblade: 33
Oasis 3
1011000
1
Cooking
Manufacturing
Woodworking
Lithic reduction
Residential
B
179
Eastern
Alashan
Oasis 3
11-100
1
Cooking
Lithic reduction
183
Eastern
Alashan
Oasis 3
11-100
1
Cooking
Manufacturing
Woodworking
Lithic reduction
186
Eastern
Alashan
Oasis 2?
11-100
1
Manufacturing
Lithic reduction
Pottery: yes
Grinding: no
Adze/Axe: ground
Specialized: no
Generalized: S, W
% microblade: 87
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, CT
% microblade: 66
Pottery: yes
Grinding: no
Adze/Axe: ground
Specialized: no
Generalized: no
% microblade: 100
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S
% microblade: 100
GOBIALTAI
Shabarakhusu, Oasis 3
Shabarakhusu Region
ALASHAN
GOBI
176
Eastern
Alashan
Task site
Task site
Task site
Residential
B
Task site
442
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
ALASHAN
GOBI
188
Eastern
Alashan
Unknown
< 10
3
Manufacturing
Lithic reduction
Task site
202
Galbain
Gobi
Epipalaeol.
?
11-100
3
Manufacturing
Lithic reduction
203
Galbain
Gobi
Oasis 3
11-100
3
Cooking
Lithic reduction
204
Galbain
Gobi
Epipalaeo.
< 10
3
Manufacturing
Lithic reduction
207
Galbain
Gobi
Oasis 3
1011000
1
Cooking
Manufacturing
Woodworking
Lithic reduction
208
Galbain
Gobi
Oasis 3
11-100
1
Cooking
Manufacturing
Woodworking
Lithic reduction
212
Galbain
Gobi
Oasis 2
and Oasis
3
1011000
1
Cooking
Manufacturing
Woodworking
Ornaments
Lithic reduction
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S
% microblade: 0
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, K,
B, CT
% microblade: 0
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: W
% microblade: 75
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, CT
% microblade: 0
Pottery: yes
Grinding: no
Adze/Axe: chipped
Specialized: P
Generalized: S, B,
W, CT
% microblade: 76
Pottery: yes
Grinding: no
Adze/Axe: ground
Specialized: no
Generalized: B, W
% microblade: 100
Pottery: yes
Grinding: F, Small
Adze/Axe: both
Specialized: D,
Aw?
Generalized: S, K,
B, W, CT
% microblade: 78
Residential
B
Task site
Task site
Residential
A
Residential
B
Residential
A
443
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
ALASHAN
GOBI
213
Galbain
Gobi
Oasis 2
11-100
1
Manufacturing
Woodworking
Lithic reduction
Task site
216
Ukh-tokhoi/
Khara Dzag
Oasis 3
11-100
3
Manufacturing
Weaponry
Lithic reduction
218
Ukh-tokhoi/
Khara Dzag
Oasis 3
1011000
4
Cooking
Manufacturing
Lithic reduction
219
Ukh-tokhoi/
Khara Dzag
Palaeo.
< 10
5
Lithic reduction
220
Ukh-tokhoi/
Khara Dzag
Oasis 2
11-100
4
Manufacturing
Woodworking
Lithic reduction
222
Ukh-tokhoi/
Khara Dzag
Oasis 3
11-100
5
Cooking
Manufacturing
Weaponry? (or
knife)
Lithic reduction
223
Ukh-tokhoi/
Khara Dzag
Late Oasis
2 or early
Oasis 3
1011000
5
Cooking
Manufacturing
Woodworking
Weaponry
Lithic reduction
Pottery: no
Grinding: no
Adze/Axe: ground
Specialized: no
Generalized: S
% microblade: 75
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: A
Generalized: S
% microblade: 100
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: P, Aw?
Generalized: S, K?
% microblade: 88
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S
No cores
Pottery: no
Grinding: no
Adze/Axe: chipped
Specialized: no
Generalized: K, B
% microblade: 25
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: N/A
Generalized: S, K?,
B, W
% microblade: 50
Pottery: yes
Grinding: no
Adze/Axe: chipped
Specialized: D, Sp
Generalized: S, K,
W
% microblade: 74
Task site
Residential
A
Task site
Task site
(workshop)
Residential
B
Residential
A
444
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
ALASHAN
GOBI
226
Ukh-tokhoi/
Khara Dzag
Oasis 1
11-100
2
Manufacturing
Lithic reduction
Task site
229
Ukh-tokhoi/
Khara Dzag
Unknown
< 10
1
Manufacturing
Lithic reduction
230
Ukh-tokhoi/
Khara Dzag
Oasis 3
11-100
5
Cooking
Manufacturing
Ornaments
Lithic reduction
231
Ukh-tokhoi/
Khara Dzag
Oasis 2
1011000
4
Manufacturing
Woodworking
Lithic reduction
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: W
% microblade: 100
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: CT
% microblade: 0
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: G
Generalized: S
No cores
Pottery: no
Grinding: no
Adze/Axe: chipped
Specialized: D
Generalized: S, K,
B, W, CT
% microblade: 42
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, K,
B, CT
% microblade: 0
Pottery: no
Grinding: no
Adze/Axe: chipped
Specialized: D, Aw
Generalized: S, K,
B
% microblade: 67
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S
% microblade: 100
237
Ukh-tokhoi/
Khara Dzag
11-100
Manufacturing
Lithic reduction
247
Goitso
Valley
Oasis 2
1011000
1
Manufacturing
Woodworking
Lithic reduction
248
Goitso
Valley
Metal
Ages
11-100
1
Cooking
Smelting
Lithic reduction
Task site
Task site
(beadmaking)
Residential
A
Residential
B
Residential
B
Residential
B
445
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
Unknown
< 10
3
Manufacturing
( 1 scraper)
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S
No cores
Task site
258
Juyanze
Oasis 2 or
3
< 10
1
Lithic reduction
Task site
259
Juyanze
Oasis 2?
1001 5000
1
Manufacturing
Woodworking
Hoes?
Lithic reduction
277
Gurnai
Depression
Oasis 3?
11-100
1
Cooking
Manufacturing
Smelting
Ornaments
Lithic reduction
287
Gurnai
Depression
Oasis 3?
1011000
1
Cooking
Manufacturing
Lithic reduction
290
Gurnai
Depression
Oasis 3
1011000
1
Cooking
Manufacturing
Ornaments
Lithic reduction
293
Gurnai
Depression
Oasis 3
1011000
1
Cooking
Manufacturing
Weaponry
Ornaments
Lithic reduction
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: no
% microblade:100
Pottery: no
Grinding: no
Adze/Axe: both
Specialized: P, Aw
Generalized: S, K,
B, W, CT
% microblade: 62
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S
No cores
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, K
% microblade: 100
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: P
Generalized: S, K,
W
% microblade: 100
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: A, P, D
Generalized: S, K,
CT
% microblade: 80
ALASHAN
GOBI
251
Juyanze
Residential
A
Residential
B
Residential
B
Residential
B
Residential
B
446
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
ALASHAN
GOBI
294
Gurnai
Depression
Oasis 3
11-100
1
Cooking
Manufacturing
Lithic reduction
Task site
(related to
293?)
298
Gurnai
Depression
Oasis 3
1011000
1
Cooking
Manufacturing
Ornaments
Lithic reduction
303
Gurnai
Depression
Oasis 3
11-100
1
Manufacturing
Ornaments (?)
Lithic reduction
307
Gurnai
Depression
Oasis 3?
11-100
1
Cooking
Manufacturing
Lithic reduction
311
Gurnai
Depression
Oasis 3
11-100
1
Manufacturing
Ornaments (?)
Lithic reduction
316
Gurnai
Depression
Oasis 3
< 10
1
Lithic reduction
(1 core)
321
Black Gobi
Early
Oasis 2
1011000
3
Manufacturing
Woodworking
Lithic reduction
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S
% microblade: 100
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: P, D
Generalized: S
% microblade: 83
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: P
Generalized: S
No cores
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S, W
% microblade: 100
Pottery: yes
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: S
% microblade: 100
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: no
% microblade: 100
Pottery: no
Grinding: no
Adze/Axe: chipped
Specialized: P
Generalized: S, K,
W
% microblade: 60
Residential
B
Task site
Task site
Task site
Task site
Residential
A
447
SITE
PERIOD
SITE
SIZE
ECOZONE
ARTEFACT
CATEGORIES
ARTEFACT
TYPES
SITE TYPE
Palaeo.
< 10
3
Manufacturing
Lithic reduction
Task site
323
Black Gobi
Epipalaeo.
11-100
5
Manufacturing
Lithic reduction
324
Black Gobi
Unknown
< 10
5
Lithic reduction
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: CT
% microblade: 0
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: D
Generalized: S, Bl,
CT
% microblade: 50
Pottery: no
Grinding: no
Adze/Axe: no
Specialized: no
Generalized: W
% microblade: 100
ALASHAN
GOBI
322
Black Gobi
Residential
B
Task site
*Grinding: I = informal, F = formal, S = small, M = medium, L = large. Specialized and
generalized tools: A = arrowpoint, D = drill, Aw = awl, P = perforator, G = grooved slab,
S = scraper, K = knife, B = large biface, Bl = blade tool, CT = core tool, W = wedgeshaped microcore.
448
APPENDIX E - SUMMARY OF MEANS AND FREQUENCIES
SITE
East Gobi*
means
SKW
3
6D
7
9
9B
9C
9D
10/10A/10B
11/11A
12
12A
12B
12/12A/12B
13
13A
14
15
19
20
20A
CORE
VOLUME
(CM3)
AND CV
188.2
1.32
117.5
0.27
1,360.8
N=1
103.5
0.80
62.6
0.74
77.3
N=1
N/A
N/A
N/A
CORE
PLATFORM
(CM2)
REMNANT
CORE
CORTEX
REMNANT
SCRAPER
CORTEX
MEAN t/T
CORE
REDUCTION
TYPE
43.4
29% none
36% > 25%
low
58% none
12% > 25%
low
0.50
0.50
46% F
54% I
100% F
0.00
100% F
low
0.52
low
0.48
N/A
N/A
36% F
64% I
91% F
8% I
100% F
0.64
0.42
0.73
N/A
N/A
N/A
165.3
1.25
186.2
0.92
104.5
0.74
159.0
0.60
290.6
0.79
127.3
1.32
326.6
1.08
242.8
2.06
79.8
0.37
93.3
0.74
155.3
1.27
82.0
0.93
42.6
0.14
40.7
high
low
50%
50%
low
0.62
49.6
low
0.46
45.0
low
0.59
48.4
N/A
N/A
80% F
20% I
43% F
57% I
71% F
28% I
100% I
84.4
N/A
N/A
100% I
34.1
low
0.16
low
0.40
low
0.59
low
0.21
41% F
58% I
40% F
60% I
38% F
62% I
100% F
47.5
216.0
34.1
22.8
low
27.6
N/A
N/A
N/A
67.3
N/A
N/A
N/A
low
47.3
26.1
low
29.6
0.37
41.6
low
0.61
30.3
low
0.55
21.0
0.00
45% F
54% I
74% F
25% I
70% F
30% I
100% F
449
SITE
East Gobi*
means
21
23/23A
28
29
30/30A
31
34
36
Gobi-Altai
means
C. M.
G. B. W.
B.U.V.
D. H.
Ulan Nor
Plain
J. W. surface
J. W. in situ
Sairim
Gashato
Arts Bogd
Kh. O.
Barun Daban
Orok Nor
S. C.
CORE
VOLUME
(CM3)
AND CV
188.2
1.32
109.5
0.97
96.4
0.78
CORE
PLATFORM
(CM2)
REMNANT
CORE
CORTEX
REMNANT
SCRAPER
CORTEX
MEAN t/T
CORE
REDUCTION
TYPE
43.4
58% none
12% > 25%
low
0.50
low
0.46
494.2
0.86
141.2
1.30
119.3
0.55
92.6
29% none
36% > 25%
low
47% none
high
44%
> 25%
high
46% F
54% I
60% F
40% I
52% F
43% I
4% other
22% F
78% I
50% F
50% I
45% F
54% I
152.9
1.08
262.8
0.82
511.1
0.98
327.3
0.96
952.8
1.23
88.2
N/A
266.4
0.67
338.4
N/A
408.5
0.76
442.3
0.66
270.0
0.69
385.1
0.42
205.3
0.96
155.2
0.36
101.1
0.70
122.6
1.62
382.4
1.15
38.2
33.3
30.7
0.53
36.8
37.8
63.0
0.55
high
49%
> 25%
low
0.58
low
0.50
high
N/A
75.6
68.0
113.6
0.61
0.75
21% none
46% > 25%
high
24.8
51%none
16% > 25%
low
0.54
0.48
20% F
80% I
50% F
50% I
100% I
N/A
N/A
100% F
29% F
71% I
100% I
61.6
high
high
0.10
72.0
high
high
0.0
81.5
high
0.36
86.5
high
0.47
58.5
high
0.0
85.3
0.25
47.1
low
0.80
40.5
high
high
0.55
30.1
low
low
0.63
34.0
low (48%
none)
low
low
0.62
low
0.37
79.6
54% F
45% I
100% I
48% F
52% I
17% F
83% I
20% F
80% I
33% F
66% I
25% F
74% I
100% I
85% F
15% I
90% F
10% I
57% F
43% I
450
SITE
Gobi-Altai
means
S. U. 1
S. U. 1A
S. U. 2a in
situ
S. U. 2b in
situ
S. U. 7
S. U. 10
Alashan
Gobi means
176
179
183
186
188
202
203
204
207
208
212
213
216
218
219
CORE
VOLUME
(CM3)
AND CV
327.3
0.96
165.5
0.65
442.7
0.88
498.7
0.03
717.7
0.50
141.0
0.76
34.0
0.55
179.1
1.57
85.3
1.42
5.1
0.70
CORE
PLATFORM
(CM2)
REMNANT
CORE
CORTEX
REMNANT
SCRAPER
CORTEX
MEAN t/T
CORE
REDUCTION
TYPE
68.0
21% none
46% >25%
high
51%none
16% > 25%
low
0.54
50% F
50% I
67% F
33% I
36% F
64% I
100% I
43.8
99.4
95.9
0.45
0.60
low
N/A
N/A
133.8
N/A
N/A
40.4
low
0.53
(average)
0.54
(average)
0.60
11.5
low
low
44.3
25.2
53% none
21% > 25%
low
63% none
20% > 25%
low
4.1
low
high
35.9
0.65
23.6
0.62
352.3
0.46
15.4
low
N/A
7.6
low
low
91.5
low
290.1
0.40
66.7
1.01
683.6
0.07
218.8
0.95
130.0
1.20
91.2
0.98
187.0
1.02
154.8
0.46
67.9
1.86
N/A
71.4
50% none
50%
51-90%
high
45.0
low
N/A
112.2
high
high
42.3
high
low
28.8
low
N/A
28.6
low
low
48.6
low
low
30.3
low
low
26.9
low
low
N/A
N/A
low
0.62
average
0.14
low
N=1
N/A
0.54
low
0.25
low
N=1
0.53
low
N/A
0.44
low
0.49
low
N/A
0.62
average
0.84
high
0.48
low
0.56
low
0.18 (N = 1)
low
66% F
33% I
42% F
58% I
100% F
80% F
20% I
93% F
7% I
100% F
100%F
100%F
100% I
80% F
20% I
75% F
25%I
50% F
50% I
94% F
6%I
100% F
90% F
9% I
100% F
100% F
87% F
8% I
N/A
451
SITE
Alashan
Gobi means
220
222
223
226
229
230
231
237
247
248
251
258
259
277
287
290
293
294
298
303
307
311
CORE
VOLUME
(CM3)
AND CV
179.1
1.57
606.1
0.30
443.8
1.05
171.8
0.57
226.4
0.65
387.1
N=1
N/A
CORE
PLATFORM
(CM2)
REMNANT
CORE
CORTEX
REMNANT
SCRAPER
CORTEX
MEAN t/T
CORE
REDUCTION
TYPE
44.3
53% none
21% > 25%
63% none
20% > 25%
N/A
0.60
80% F
20% I
75% F
25% I
100% F
362.3
1.09
1,852.5
0.30
37.2
0.62
14.1
1.22
N/A
79.1
129.4
N/A
79.7
low
low
54.5
low
low
53.5
low
N/A
0.61
average
0.59
average
N/A
N/A
N/A
100% I
low
0.32
low
0.59
average
0.39
low
0.42
low
0.39
low
0.33
low
N=1
N/A
N/A
0.67
high
0.88
high
0.58
average
0.73
high
0.75
high
0.56 average
70% F
29% I
N/A
0.68
high
0.55
average
0.76
high
0.43
Low
83% F
17% I
N/A
75.9
N/A
227.1
N/A
high
14.9
low
low
10.0
low
low
N/A
N/A
high
38.9
N/A
147.1
1.51
N/A
16.9
low
N/A
39.6
low
low
N/A
N/A
low
19.0
0.08
15.7
0.69
19.6
0.71
15.4
N=1
33.7
1.09
N/A
12.7
low
low
10.9
low
low
12.9
low
low
15.4
low
low
23.0
low
low
N/A
N/A
low
8.4
0.44
13.0
N=1
4.7
low
low
5.4
low
low
74% F
22% I
100% F
82% F
18% I
75% F
25% I
100% F
100% F
N/A
100% F
100% F
100% F
80% F
20% I
100% F
100% F
100% F
452
SITE
Alashan
Gobi means
316
321
322
323
CORE
VOLUME
(CM3)
AND CV
179.1
1.57
15.8
N=1
130.5
0.54
161.7
0.24
96.1
1.25
CORE
PLATFORM
(CM2)
REMNANT
CORE
CORTEX
REMNANT
SCRAPER
CORTEX
MEAN t/T
CORE
REDUCTION
TYPE
44.3
53% none
21% > 25%
low
63% none
20% > 25%
N/A
0.60
80% F
20% I
100% F
low
0.99
high
0.36
low
0.41
low
8.8
39.4
50.3
high
high
25.5
low
low
N/A
72% F
28% I
50% F
50% I
100% F
*Due to the relatively high number of artefacts from Baron Shabaka Well (over 500 each
of cores and scrapers), the site was excluded from calculation of regional mean.
453
REFERENCES
Ackerman, R. E., 2007. The microblade complexes of Alaska and the Yukon: early
interior and coastal adaptations in: Kuzmin, Y. V., Keates, S. G., Shen, C. (Eds.),
Origin and Spread of Microblade Technology in Northern Asia and North
America. Archaeology Press, Simon Fraser University, Burnaby, pp. 147-170.
Aikens, C. M., Akazawa, T., 1996. The Pleistocene-Holocene transition in Japan and
adjacent Northeast Asia: climate and biotic change, broad-spectrum diet, pottery,
and sedentism in: Straus, L. G., Eriksen, B. V., Erlandson, Yesner, D. R. (Eds.),
Humans at the End of the Ice Age: the Archaeology of the Pleistocene-Holocene
Transition. Plenum Press, New York, pp. 215-227.
Aiken, M. J., 1985. Thermoluminescence Dating. Academic Press, New York.
Akishinonomiya, F., Tetsuo, M., Shin-Ichiro, S., Asaru, T., Susumu, O., Norio, K., 1994.
One subspecies of the red junglefowl (Gallus gallus gallus) suffices as the
matriarchic ancestor of all domestic breeds. PNAS 91(26), 12505-12509.
Allen, G. M., 1938. The Mammals of Central Asia, Parts 1 and 2. Natural History of
Central Asia, Volume XI. Central Asiatic Expeditions. American Museum of
Natural History, New York.
Andersson, J .G., 1943. Researches into the Prehistory of the Chinese. Bulletin of the
Museum of Far Eastern Antiquities 15. Museum of Far Eastern Antiquities,
Stockholm.
454
Andrefsky, W., Jr., 1987. Diffusion and innovation from the perspective of wedge
shaped cores in Alaska and Japan in: Johnson, J. K., Morrow, C. A. (Eds.), The
Organization of Core Technology. Westview Press, Boulder, pp. 13-44.
Andrefsky, W., Jr., 1994. Raw material availability and the organization of technology.
American Antiquity 59(1), 21-34.
Andrefsky, W., Jr., 1998. Lithics: Macroscopic Approaches to Analysis. Cambridge
Manuals in Archaeology. Cambridge University Press, Cambridge.
An, C., Chen, F.-H., Barton, L., 2008. Holocene environmental changes in Mongolia: a
review. Global and Planetary Change 63, 283-289.
An, C., Feng, Z., Barton, L., 2006. Dry or humid? Mid-Holocene humidity changes in
arid and semi-arid China. Quaternary Science Reviews 25, 351-361.
An, Z., 1992a. Neolithic communities in eastern parts of Central Asia in: Dani, A. H.,
Masson, V. M. (Eds.), History of Civilization of Central Asia, Volume 1: The
Dawn of Civilization, Earliest Times to 700 BC. UNESCO Publishing, Paris, pp.
153-168.
An, Z., 1992b. The Bronze Age in eastern parts of Central Asia in: Dani, A. H., Masson,
V. M. (Eds.), History of Civilization of Central Asia, Volume 1: The Dawn of
Civilization, Earliest Times to 700 BC. UNESCO Publishing, Paris, pp. 319-336.
Anthony, D. W., Brown, D. R., 1991. The origins of horseback riding. Antiquity 65, 2238.
Anthony, D. W., Brown, D. R., 2000. Eneolithic horse exploitation in the Eurasian
steppes: diet, ritual and riding. Antiquity 74, 75-86.
455
Army Map Services, Corps of Engineers, U.S. Army, Washington, D.C., 1949. “KueiSui” [map]. 1:1,000,000. In: Norin, E., Montell, G. (Eds.), 1969, Sven Hedin
Central Asia Atlas. Reports from the Scientific Expedition to the North-western
Provinces of China under the Leadership of Sven Hedin, Sino-Swedish
Expedition Publication, Publication 47. Statens Etnografiska Museum,
Stockholm, NK 49.
Army Map Services, Corps of Engineers, U.S. Army, Washington, D.C., 1950. “Dalan
Dzadagad” [map]. 1:1,000,000. In: Norin, E., Montell, G. (Eds.), 1969, Sven
Hedin Central Asia Atlas. Reports from the Scientific Expedition to the Northwestern Provinces of China under the Leadership of Sven Hedin, Sino-Swedish
Expedition Publication, Publication 47. Statens Etnografiska Museum,
Stockholm, NK 48.
Army Map Services, Corps of Engineers, U.S. Army, Washington, D.C., 1954. “O-ChiNa Ho” [map]. 1:1,000,000. In: Norin, E., Montell, G. (Eds.), 1969, Sven Hedin
Central Asia Atlas. Reports from the Scientific Expedition to the North-western
Provinces of China under the Leadership of Sven Hedin, Sino-Swedish
Expedition Publication, Publication 47. Statens Etnografiska Museum,
Stockholm, NK 47.
Arnold, D. E., 1985. Ceramic Theory and Cultural Process. Cambridge University Press,
Cambridge.
Ascough, P., Cook, G., 2005. Methodological approaches to determining the marine
reservoir affect. Progress in Physical Geography 29(4), 532-547.
456
Aseyev, I. V., 2008. Horseman image on an ostrich eggshell fragment. Archaeology,
Ethnology & Anthropology of Eurasia 34, 96-99.
Bae K., Kim, J. C., 2003. Radiocarbon chronology of the Palaeolithic complexes and the
transition to the Neolithic in Korea. The Review of Archaeology 24(2), 46-49.
Bamforth, D. B., 1986. Technological efficiency and tool curation. American Antiquity
51, 38-50.
Bamforth, D. B., Bleed, P., 1997. Technology, flaked stone technology, and risk in:
Clark, G. A., Barton, C. M. (Eds.), Rediscovering Darwin: Evolutionary Theory
and Archaeological Explanation. Archaeological Papers of the American
Anthropological Association 7. American Anthropological Association,
Arlington, Virginia, pp. 109-140.
Barnes, G. L., 1999. The Rise of Civilization in East Asia: the Archaeology of China,
Korea and Japan. Thames and Hudson, London.
Barton, L., Brantingham, P. J., Ji, D., 2007. Late Pleistocene climate change and
Paleolithic cultural evolution in northern China: implications from the Last
Glacial Maximum in: Madsen, D. B., Chen, F. H., Gao, X. (Eds.), Late
Quaternary Climate Change and Human Adaptation in Arid China, Developments
in Quaternary Science, Vol. 9. Elsevier, Amsterdam, pp. 105-128.
Barton, L., Newsome, S. D., Chen, F. –H., Wang, H., Guilderson, T. P., Bettinger, R. L.,
2009. Agricultural origins and isotopic identity of domestication in northern
China. PNAS 106(14), 5523-5528.
457
Bar-Yosef, O., 1996. The impact of Late Pleistocene-Early Holocene climatic changes
on humans in Southwest Asia in: Straus, L. G., Erikson, B. V., Erlandson, J. M.,
Yesner, D. R. (Eds.), Humans at the End of the Ice Age: the Archaeology of the
Pleistocene-Holocene Transition. Plenum Press, London, pp. 61-78.
Bar-Yosef, O., 2002. The role of the Younger Dryas in the origin of agriculture in West
Asia in: Yasuda, Y. (Ed.), The Origins of Pottery and Agriculture. Roli Books,
New Delhi, pp. 39-54. Bazaliiskiy, V. I., Savelyev, N. A., 2003. The wolf of
Baikal: the “Lokomotiv” Early Neolithic cemetery in Siberia. Antiquity 77(295),
20-30.
Batsaikhan, N., Samiya, R., Shar, S., King, S. R. B., 2010. A Field Guide to Mammals of
Mongolia. Zoological Society of London, London.
Bazaliiskiy, V. I., Savelyev, N. A., 2003. The Wolf of Baikal: the “Lokomotiv” early
Neolithic cemetery in Siberia. Antiquity 77(295), 20-30.
Becken, M., Hölz, S., Fiedler-Volmer, R., Hartmann, K., Wünnemann, B., Burkhardt, H.,
2007. Electrical resistivity image of the Jingsutu Graben at the NE margin of the
Ejina Basin (NW China) and its implications for the basin development.
Geophysical Research Letters 34, L09315.
Beck, C., Jones, G. T., 1994. The terminal Pleistocene/early Holocene archaeology of
the Great Basin. Journal of World Prehistory 11, 161-236.
458
Benecke, N., von den Dreisch, A., 2003. Horse exploitation in the Kazakh steppes during
the Eneolithic and Bronze Age in: Levine, M., Renfrew, C., Boyle, K., Prehistoric
Steppe Adaptations and the Horse. McDonald Institute for Archaeological
Research, Cambridge, pp. 69-82.
Bennett, D., Hoffmann, R. S., 1999. Equus caballus. Mammalian Species 628, 1-14.
Bergman, F., 1945. Travels and archaeological field-work in Mongolia and Sinkiang – a
diary of the years 1927-1934 in: Hedin, S. (Ed.), History of the Expedition in Asia
1927-1935. Reports from the Scientific Expedition to the North-western
Provinces of China under the Leadership of Sven Hedin, Sino-Swedish
Expedition Publication, Publication 26, Part IV. Statens Etnografiska Museum,
Stockholm, pp. 1-192.
Berkey, C. P., Nelson, N. C., 1926. Geology and prehistoric archaeology of the Gobi
Desert. American Museum Novitates 222, 9-16.
Bettinger, R. L., Barton, L., Morgan, C., 2010a. The origins of food production in North
China: a different kind of agricultural revolution. Evolutionary Anthropology 19
(1), 9-21.
Bettinger, R. L., Barton, L., Morgan, C., Chen, F., Wang, H., Guilderson, T. P., Ji, D.,
Zhang, D., 2010b. The transition to agriculture at Dadiwan, People’s Republic of
China. Current Anthropology 51(5), 703-714.
459
Bettinger, R. L., Barton, L., Richerson, P. J., Boyd, R., Wang, H., Choi, W., 2007. The
transition to agriculture in northwestern China in: Madsen, D. B., Chen, F. H.,
Gao, X. (Eds.), Late Quaternary Climate Change and Human Adaptation in Arid
China, Developments in Quaternary Science, Vol. 9. Elsevier, Amsterdam, pp.
83-101.
Bettinger, R. L., Madsen, D. B., Elston, R. G., 1994. Prehistoric settlement categories
and settlement systems in the Alashan Desert of Inner Mongolia, PRC. Journal of
Anthropological Archaeology 13, 74-101.
Bettinger, R. L., Winterhalder, B., McElreath, R., 2006. A simple model of technological
intensification. Journal of Archaeological Science 33(4), 538-545.
Binford, L. R., 1968. Post-Pleistocene adaptations in: Binford, S. R., Binford, L. R.
(Eds.), New Perspectives in Archaeology. Aldine Press, Chicago, pp. 313-341.
Binford, L. R., 1979. Organization and formation processes: looking at curated
technologies. Journal of Anthropological Research 35(3), 255-273.
Binford, L. R., 1980. Willow smoke and dogs’ tails: hunter-gatherer settlement systems
and archaeological site formation. American Antiquity 45(1), 4-20.
Binford, L. R., 1982. The archaeology of place. Journal of Anthropological
Archaeology 1(1), 5-31.
Bird, D. W., Bliege Bird, R., 2005. Evolutionary and ecological understandings of the
economics of desert societies: comparing the Great Basin USA and the Australian
deserts in: Veth, P., Smith, M., Hiscock, P. (Eds.), Desert Peoples: Archaeological
Perspectives. Blackwell Publishing, Oxford, pp. 81-99.
460
Bird, D. W., Bliege Bird, R., Codding, B. F., 2009. In pursuit of mobile prey: Martu
hunting strategies and archaeofaunal interpretation. American Antiquity 74(1), 329.
Bird, M. I., Ayliffe, L. K., Fifield, J., Cresswell, R., Turney, C., 2003. Radiocarbon
dating of organic- and carbonate-carbon in Genyornis and Dromaius eggshell
using stepped combustion and stepped acidification. Quaternary Science Reviews
22, 1805-1812.
Bird, R., 1999. Cooperation and conflict: the behavioral ecology of the sexual division of
labor. Ecological Anthropology 8, 65-75.
Blades, B. S., 1999. Aurignacian lithic economy and early modern human mobility: new
perspectives from classic sites in the Vézère valley of France. Journal of Human
Evolution 37, 91-120.
Blades, B. S., 2003. End scraper reduction and hunter-gatherer mobility. American
Antiquity 68(1), 141–156.
Bleed, P., 1986. The optimal design of hunting weapons: maintainability or reliability.
American Antiquity 51(4), 737-747.
Bleed, P., 2002. Obviously sequential, but continuous or staged? Refits and cognition in
three late Paleolithic assemblages from Japan. Journal of Anthropological
Archaeology 21, 329-343.
461
Blyakharchuk, T. A., Wright, H. E., Borodavko, P. S., van der Knaap, W. O., Ammann,
B., 2004. Late glacial and Holocene vegetational changes on the Ulagan highmountain plateau, Altai Mountains, southern Siberia. Palaeogeography,
Palaeoclimatology, Palaeoecology 209, 259-279.
Böhner, J., 2006. General climatic controls and topoclimatic variations in Central and
High Asia. Boreas 35, 279-295.
Bollong, C. A., Sampson, C. G., Smith, A. B., 1997. Khoikhoi and Bushman pottery in
the Cape Colony: ethnohistory and later Stone Age ceramics of the South African
interior. Journal of Anthropological Archaeology 16, 269-299.
Bourguignon, L., Faivre, J.-P., Turq, A., 2004. Ramifaction des chaîne opératoires: une
spécificité du Moustérien. Paleo 16, 37-48. (in French)
Bousman, C. B., 1993. Hunter-gatherer adaptations, economic risk, and tool design.
Lithic Technology 18(1/2), 59-86.
Brantingham, P. J., 2003. A neutral model of stone raw material procurement. American
Antiquity 68, 487-509.
Bright, J. R., Ugan, A., 1999. Ceramics and mobility: assessing the role of foraging
behaviour and its implications for culture-history. Utah Archaeology 12, 17-29.
Broccoli, A. J., Manabe, S., 1992. The effects of orography on midlatitude northern
hemisphere dry climates. Journal of Climate 5, 1181-1201.
Brown, J. A., 1989. The beginning of pottery as an economic process in: van der Leeuw,
S. E., Torrence, R. (Eds.), What’s New? A Closer Look at the Process of
Innovation. Unwin Hyman, London, pp. 203-224.
462
Bureau of Geological Investigation, Geological Survey of Mongolia, 2003a. “Orog nuur”
[map]. 1:500,000. Mongol Uls, L-47-Г. Ulaanbaatar.
Bureau of Geological Investigation, Geological Survey of Mongolia, 2003b. “Ongi”
[map]. 1:500,000. Mongol Uls, L-48-B. Ulaanbaatar.
Bureau of Geological Investigation, Geological Survey of Mongolia, 2003c. “Urt”
[map]. 1:500,000. Mongol Uls, K-47-Б. Ulaanbaatar.
Bureau of Geological Investigation, Geological Survey of Mongolia, 2003d.
“Dalanzadgad” [map]. 1:500,000. Mongol Uls, K-48-A, B. Ulaanbaatar.
Bureau of Geological Investigation, Geological Survey of Mongolia, 2003e.
“Baruuncuu” [map]. 1:500,000. Mongol Uls, K-48-Б, Г. Ulaanbaatar.
Burr, G. S., Edwards, R. L., Donahue, D. J., Druffell, E. R. M., Taylor, F. W., 1992.
Mass spectrometric 14C and U-Th measurements in coral. Radiocarbon 34, 611618.
Buvit, I., Waters, M. R., Konstantinov, M. V., Konstantinov, A. V., 2003.
Geoarchaeological investigations at Studenoe, an Upper Palaeolithic sites in the
Transbaikal Region, Russia. Geoarchaeology 18(6), 649-673.
Cermak, J., Opgenoorth, L., Miehe, G., 2005. Isolated mountain forests in Central Asian
Deserts: a case study from the Govi Altay, Mongolia in: Broll, G., Keplin, B.
(Eds.), Mountain Ecosystems: Studies in Treeline Ecology. Earth and
Environmental Science, Part 4. Springer, New York, pp. 253-273.
Chang, K. –C., 1987. The Archaeology of Ancient China. Fourth Edition. Yale
University Press, New Haven and London.
463
Chard, C. S., 1974. Northeast Asia in Prehistory. University of Wisconsin Press,
Madison.
Charnov, E. L., 1976. Optimal foraging, the Marginal Value Theorem. Theoretical
Population Biology 9(2), 129-136.
Chen, C., An, J., Chen, H., 2010. Analysis of the Xiaonanhai lithic assemblage,
excavated in 1978. Quaternary International 211, 75-85.
Chen, C., 2007. Techno-typological comparison of microblade cores from East Asia and
North America in: Kuzmin, Y. V., Keates, S. G., Shen, C. (Eds.), Origin and
Spread of Microblade Technology in Northern Asia and North America.
Archaeology Press, Simon Fraser University, Burnaby, pp.7-38.
Chen, C., Wang, X. –Q., 1989. Upper Palaeolithic microblade industries in North China
and their relationships with Northeast Asia and North America. Arctic
Anthropology 26(2), 127-156.
Chen, F.-H., Cheng, B., Zhao, H., Fan, Y.-X., Madsen, D. B., Jin, M., 2007. Post-glacial
climate variability and drought events in the monsoon transition zone of western
China in: Madsen, D. B., Chen, F.-H., Xing, G. (Eds.), Late Quaternary Climate
Change and Human Adaptation in Arid China. Elsevier, Amsterdam, pp. 25-39.
Chen, F., Wu, W., Holmes, J. A., Madsen, D. B., Zhu, Y., Jin, M., Oviatt, C. G., 2003. A
mid-Holocene drought interval as evidenced by lake desiccation in the Alashan
Plateau, Inner Mongolia, China. Chinese Science Bulletin 48(14), 1401-1410.
464
Chen, F., Yu, Z., Yang, M., Ito, E., Wang, S., Madsen, D., Huang, X., Zhao, Y., Sato, T.,
Birks, H. J. B., Boomer, I., Chen, J., An, C., Wünnemann, B., 2008. Holocene
moisture evolution in arid central Asia and its out-of-phase relationship with
Asian monsoon history. Quaternary Science Reviews 27, 351-364.
Chen, S., Lin, B. –Z., Baig, M., Mitra, B., Lopes, R. J., Santos, A. M., Magee, D. A.,
Azevedo, M., Tarroso, P., Sasazaki, S., Ostrowski, S., Mahgoub, O., Chaudhuri,
T. K., Zhang, Y. –P., Costa, V., Royo, L. J., Goyache, F., Luikart, G., Boivin, N.,
Fuller, D. Q., Mannen, H., Bradley, D. G., Beja-Pereira, A., 2010. Zebu cattle are
an exclusive legacy of the South Asia Neolithic. Molecular Biology and
Evolution 27(1), 1-6.
Chifeng International Collaborative Archaeological Research Project (CICARP), 2003.
Regional Archeology in Eastern Inner Mongolia: A Methodological Exploration.
Science Press, Beijing.
Childe, V. G., 1953. New Light on the Most Ancient Near East. Praeger, New York.
Christian, D., 1998. A History of Russia, Central Asia, and Mongolia, Vol. 1: Inner
Eurasia from Prehistory to the Mongol Empire. Blackwell Publishing, Oxford.
Clark, J., 1987. Politics, prismatic blades, and Mesoamerican civilization in: Johnson, J.
K., Morrow, C. A. (Eds.), The Organization of Core Technology. Westview
Press, Boulder, pp. 259-284.
465
Close, A. E., 1995. Few and far between: early ceramics in North Africa in: Barnett, W.
K., Hoopes, J. W. (Eds.), The Emergence of Pottery, Technology and Innovation
in Ancient Societies. Smithsonian Institution Press, Washington, D. C., pp. 2337.
Cohen, D. J., 2003. Microblades, pottery, and the nature and chronology of the
Palaeolithic-Neolithic transition in China. The Review of Archaeology 24(2), 2136.
Cohen, M. N., 1977. The Food Crisis in Prehistory: Overpopulation and the Origins of
Agriculture. Yale University Press, New Haven.
Crawford, G. W., 1997. Anthropogenesis in prehistoric Northeastern Japan in:
Gremillion, K. (Ed.), People, Plants, and Landscapes: Studies in
Paleoethnobotany. University of Alabama Press, Tuscaloosa, pp. 86-103.
Crawford, G. W., 2006. East Asian plant domestication in: Stark, M. (Ed.), Archaeology
of Asia. Blackwell Publishing, Oxford, pp. 77-95.
Crawford, G. W., Lee, G. -A., 2003. Agricultural origins in the Korean peninsula.
Antiquity 77, 87-95.
Crawford, G., Underhill, A., Zhao, Z., Lee, G.-A., Feinman, G., Nicholas, L., Luan, F.,
Yu, H., Cai, F., 2005. Late Neolithic plant remains from northern China:
preliminary results from Liangchengzhen, Shandong. Current Anthropology
46(2), 309-317.
466
Cybiktarov, A. D., 2002. Eastern Central Asia at the dawn of the Bronze Age: issues in
ethno-cultural history of Mongolia and the southern Trans-Baikal region in the
late third-early second millennium BC. Archaeology, Ethnology & Anthropology
of Eurasia 3(11), 107-123.
Cybiktarov, A. D., 2003. Central Asia in the Bronze and early Iron Ages (Problems of
ethno-cultural history of Mongolia and the southern Trans-Baikal region in the
middle 2nd – early 1st millennia BC). Archaeology, Ethnology & Anthropology of
Eurasia 1(13), 80-97.
Danzeglocke, U., Jöris, O., Weninger, B., 2010. CalPal-2007online. http://www.calpalonline.de/, accessed 2011-12-07.
Debaine-Francfort, C., 1995. Du Néolithique à l’Âge du Bronze en Chine du NordOuest: la culture de Qijia et ses connexions. Mémoires de la mission
archéologique Française en Asie Centrale, Vol. VI. Éditions recherche sur les
civilizations, Paris. (in French)
Demske, D., Mischke, S., 2003. Palynological investigation of a Holocene profile
section from the Palaeo-Gaxun-Nur-Basin. Chinese Science Bulletin 48(14),
1418-1422.
Deng, T., 2005. The fossils of the Przwalski’s horse and the climatic variation of the
Late Pleistocene of China in: Mashkour, M. (Ed.), Equids in Time and Space.
Oxbow Books, Oxford, pp. 12-19.
467
Derevianko, A. P. (Ed.), 2000. Paleolithic and Neolithic of the Northern Face of the
Valley of the Lakes. Institute of Archaeology and Ethnography SB RAS Press,
Novosibirsk. (in Russian, French, and English)
Derevianko, A. P., Dorj, D., 1992. Neolithic tribes in northern parts of Central Asia in:
Dani, A. H., Masson, V. M. (Eds.), History of Civilization of Central Asia,
Volume 1: The Dawn of Civilization, Earliest Times to 700 BC. UNESCO
Publishing, Paris, pp. 169-189.
Derevianko, A. P., Markin, S., 1998. Palaeolithic sites along the Yenisei in: Derevianko,
A. P., Shimkin, D. B., Powers, W. R. (Eds.), The Palaeolithic of Siberia: New
Discoveries and Interpretations. University of Illinois Press, Urbana and Chicago,
pp. 116-118.
Derevianko, A. P., Olsen, J. W., Tseveendorj, D. (Eds.), 1996. A Preliminary Report on
Archaeological Studies carried out by the Joint Russian-Mongolian-American
Expedition in Mongolia in 1995. Institute of Archaeology and Ethnography SB
RAS Press, Novosibirsk. (in Russian, Mongolian, and English)
Derevianko, A. P., Olsen, J. W., Tseveendorj, D. (Eds.), 1998. Archaeological Studies
carried out by the Joint Russian-Mongolian-American Expedition in Mongolia in
1996. Institute of Archaeology and Ethnography SB RAS Press, Novosibirsk. (in
Russian, Mongolian, and English)
468
Derevianko, A. P., Olsen, J. W., Tseveendorj, D. (Ed.), 2000. Archaeological Studies
carried out by the Joint Russian-Mongolian-American Expedition in Mongolia in
1997-1998. Institute of Archaeology and Ethnography SB RAS Press,
Novosibirsk. (in Russian, Mongolian, and English)
Derevianko, A. P., Gladyshev, S. A., Nohrina, T. I., Olsen, J. W., 2003. The Mongolian
Early Holocene excavations at Chikhen Agui Rockshelter in the Gobi Altai. The
Review of Archaeology 24(2), 50-56.
Derevianko, A. P., Olsen, J. W., Tseveendorj, D., Gladyshev, S. A., Nokhrina, T. I.,
Tabarev, A. V., 2008. New insights into the archaeological record at Chikhen
Agui Rockshelter (Mongolia). Archaeology, Ethnology & Anthropology 34(2),
2-12.
Dibble, H. L., 1984. Interpreting typological variation of Middle Palaeolithic scrapers:
function, style, or sequence of reduction? Journal of Field Archaeology 11, 431436.
Dibble, H. L., Roth, B. J., Lenoir, M., 1995. The use of raw material at Combe-Capelle
Bas in: Dibble, H. L., Lenoir, M. (Eds.), The Middle Palaeolithic Site of CombeCapelle Bas (France). The University Museum Press, Philadelphia, pp. 259-287.
Dibble, H. L., Schurmans, U. A., Iovita, R. P., McLaughlin, M. V., 2005. The
measurement and interpretation of cortex in lithic assemblages. American
Antiquity 70(3), 545-560.
Di Cosmo, N., 1994. Ancient Inner Asian nomads: their economic basis and its
significance in Chinese history. The Journal of Asian Studies 53(4), 1092-1126.
469
Dorj, D., Derevianko, A. P., 1970. Noviye materiali dlya izucheniya neolita vostochnoy
Mongolii [New material for the study of the Neolithic of eastern Mongolia].
Izvestiya Akademii Nauk Mongol’skoy Narodnoy Respubliki Institut Istorii,
Ulaanbaatar. (in Russian)
Dorj, D., 1971. Neolit vostochnoy Mongolii [Neolithic of eastern Mongolia]. Izvestiya
Akademii Nauk Mongol’skoy Napodnoy Respubliki Institut Istorii, Ulaanbaatar.
(in Russian)
Douglass, M. J., Holdaway, S. J., Fanning, P. C., Shiner, J. I., 2008. An assessment and
archaeological application of cortex measurement in lithic assemblages.
American Antiquity 73(3), 513-526.
Dunnell, R. C., Feathers, J. K., 1994. Thermoluminescence dating of surficial
archaeological material in: Beck, C. (Ed.), Dating in Exposed and Surface
Contexts. University of New Mexico Press, Albuquerque, pp. 115-137.
Eerkens, J. W., 2003. Residential mobility and pottery use in the western Great Basin.
Current Anthropology 44(5), 728-738.
Eerkens, J. W., 2004. Privatization, small-seed intensification, and the origins of pottery
in the Western Great Basin. American Antiquity 69(4), 653-670.
Eerkens, J. W., Neff, J., Glascock, M. D., 2002. Ceramic production among small-scale
and mobile hunters and gatherers: a case study from the southwestern Great
Basin. Journal of Anthropological Archaeology 21, 200-229.
470
Egami, N., Mizuno, S., 1935. Nai-Moko chojo chitai (Innermost Mongolia and the
region of the Great Wall). Archaeologia Orientalis, Series B, Vol. 1. (in
Japanese)
Elisseeff, V., 1950. La Mongolie dans l’Antiquitié. Kokogaku Zasshi 36(4). (in French)
Ellis, C., 2008. The Fluted Point tradition and the Arctic Small Tool tradition: what’s the
connection? Journal of Anthropological Archaeology 27, 298-314.
Elston, R. G., Brantingham, P. J., 2002. Microlithic technology in northern Asia: a riskminimizing strategy of the Late Palaeolithic and Early Holocene in: Elston, R. G.,
Kuhn, S. L. (Eds.), Thinking Small: Global Perspectives on Microlithization.
Archaeological Papers of the American Anthropological Association, 12.
American Anthropological Association, Arlington, Virginia, pp. 103-116.
Elston, R. G., Chen, X., Madsen, D. B., Kan, Z., Bettinger, R. L., Li, J., Brantingham, P.
J., Wang, H., Yu, J., 1997. New dates for the North China Mesolithic. Antiquity
71, 985-993.
Elston, R. G., Dong, G., Zhang, D., 2011. Late Pleistocene intensification technologies
in northern China. Quaternary International 242(2), 401-415.
Elston, R. G., Zeanah, D. W., Carter, J. A., Dugas, D. P., 1995. Paleoenvironmental
variation and human land use in the Carson Desert in: Zeanah, D. W. (Ed.), An
Optimal Foraging Model of Hunter-Gatherer Land Use in the Carson Desert.
Intermountain Research, Silver City, pp. 293-326.
471
Elston, R. G., Zeanah, D. W., 2002. Thinking outside the box: a new perspective on diet
breadth and sexual division of labor in the Prearchaic Great Basin. World
Archaeology 34(1), 103-130.
Erdenebaatar, D., 2004. Burial materials related to the history of the Bronze Age in the
territory of Mongolia in: Linduff, K. M. (Ed.), Metallurgy in Ancient Eastern
Eurasia from the Urals to the Yellow River. Edwin Mellen Press, Lewiston, NY,
pp. 189-223.
Eren, M. I., Dominguez-Rodrigo, M., Kuhn, S. L., Adler, D. S., Le, I., Bar-Yosef, O.,
2005. Defining and measuring reduction in unifacial stone tools. Journal of
Archaeological Science 32, 1190-1201.
Evershed, R. P., Payne, S., Sherratt, A., G., Copley, M. S., Coolidge, J., Urem-Kotsu, D.,
Kotsakis, K., Özdoğan, M., Özdoğan, A., E., Nieuwenhuyse, O., Akkermans, P.
M. M. G., Bailey, D., Andeescu, R. –R., Campbell, S., Farid, S., Hodder, I.,
Yalman, N., Özbașaran, M., Biçakcı, E., Garfinkel, Y., Levy, T., Burton, M. M.,
2008. Earliest date for milk use in the Near East and southeastern Europe linked
to cattle herding. Nature 455, 528-531.
Fairservis, W. A., 1993. The Archaeology of the Southern Gobi, Mongolia. Carolina
Academic Press, Durham.
Feathers, J. K., 2003. Use of luminescence dating in archaeology. Measurement Science
and Technology 14, 1493-1509.
472
Feng, Z.-D., Zhai, X. W., Ma, Y. Z., Huang, C. Q., Wang, W. G., Zhang, H. C.,
Khosbayar, P., Narantsetseg, T., Liu, K.-B., Rutter, N. W., 2007. Eolian
environmental changes in the Northern Mongolian Plateau during the past
~35,000 yr. Palaeogeography, Palaeoclimatology, Palaeoecology 245, 505-517.
Fernandez‐Gimenez, M. E., 1999. Sustaining the steppes: a geographical history of
pastoral land use in Mongolia. Geographical Review 89(3), 315‐342.
Fiedel, S. J., 2005. Man’s best friend – mammoth’s worst enemy? A speculative essay
on the role of dogs in Paleoindian colonization and megafaunal extinction. World
Archaeology 37, 11-25.
Fiskesjö, M., Chen, X., 2004. China before China. Museum of Far Eastern Antiquities,
Stockholm. (in English and Chinese)
Fitzhugh, W. W., 2009. Stone shamans and flying deer of northern Mongolia: Deer
Goddess of Siberia or Chimera of the Steppe? Arctic Anthropology 46(1-2), 7288.
Flad, R. K., Yuan, J., Li, S., 2007. Zooarchaeological evidence for animal domestication
in northwest China in: Madsen, D. B., Chen, F. H., Gao, X. (Eds.), Late
Quaternary Climate Change and Human Adaptation in Arid China, Developments
in Quaternary Science, Vol. 9. Elsevier, Amsterdam, pp. 167-203.
Flannery, K. V., 1969. Origins and ecological effects of early domestication in Iran and
the Near East in: Ucko, P. J., Dimbleby, G. W. (Eds.), The Domestication and
Exploitation of Plants and Animals. Aldine Press, Chicago, pp. 73-100.
473
Flenniken, J. J., 1987. The Paleolithic Dyuktai pressure blade technique of Siberia.
Arctic Anthropology 24, 117-132.
Formozov, A. A., 1961. Microlithic sites in the Asiatic USSR. American Antiquity
27(1), 82-92.
Frachetti, M., 2002. Bronze Age exploitation and political dynamics of the eastern
Eurasian steppe zone in: Boyle, K., Renfrew, C., Levine, M. (Eds.), Ancient
Interactions: East and West in Eurasia. McDonald Institute of Archaeological
Research, Cambridge, pp. 161-170.
Frachetti, M. D., 2008. Pastoralist Landscapes and Social Interaction in Bronze Age
Eurasia. University of California Press, Berkeley and Los Angeles.
Freundlich, J. C., Kuper, R., Breunig, P., Bertram, H. –G., 1989. Radiocarbon dating of
ostrich eggshells. Radiocarbon 31, 1030-1034.
Fuller, D. Q., Harvey, E., Qin, L., 2007. Evidence for wild rice cultivation and
domestication in the fifth millennium BC of the Lower Yangtze region. Antiquity
81, 316-331.
Gábori, M., 1963. Gisements et industries de l’âge de la pièrre en Mongolie. Acta
Archaeologia XV(1/4), 11-32. (in French)
Gai, P., 1984. Dynamic typology of Yangyuan core and analysis of its technological
ideas. Acta Anthropologica Sinica 3(3), 244-251. (in Chinese with English
summary)
Gai, P., Wei, Q., 1977. Hutouliang jiushiqi shidai wanqi yizhi (The discovery of the
Upper Palaeolithic site at Hutouliang). G. D. X. 15(4), 286-300. (in Chinese)
474
Ganopolski, A., Kubatzki, C., Claussen, A., Brovkin, V., Petoukhov, V., 1998. The
influence of vegetation-atmosphere-ocean interaction on climate during the midHolocene. Science 280, 1916-1919.
Garrod, A., Colledge, S., Martin, L., 1996. The emergence of crop cultivation and
caprine herding in the “Marginal Zone” of the southern Levant in: Harris, D.
(Ed.), The Origins and Spread of Agriculture and Pastoralism in Eurasia.
Smithsonian Institution Press, Washington, D.C., pp. 204-226.
Germonpré, M., Sablin, M. V., Stevens, R. E., Hedges, R. E. M., Hofreiter, M., Stiller,
M., Després, V. R., 2009. Fossil dogs and wolves from Palaeolithic sites in
Belgium, the Ukraine and Russia: osteometry, ancient DNA and stable isotopes.
Journal of Archaeological Science 36, 473-490.
Giuffra, E., Kijas, J. M. H., Amarger, V., Carlborg, Ö., Jeon, J. –T., Andersson, L., 2000.
The origin of the domestic pig: independent domestication and subsequent
introgression. Genetics 154, 1785-1791.
Gladyshev, S. A., 1987. Tekhniko-tipologicheskaya kharakteristika kamennovo
inventarya mestonakhozhdehiya Bygat-2 (Techno-typological character of lithic
assemblage of Mesolithic Bygat-2) in: Derevianko, A. P. (Ed.), Arkheologia,
Ethnografia i Antropologia Mongolii. Institute of Archaeology and Ethnography
SB RAS Press, Novosibirsk, pp. 97-103. (in Russian)
Gladyshev, S. A., Olsen, J. W., Tabarev, A. V., Kuzmin, Y. V., 2010. Chronology and
periodization of Upper Palaeolithic sites in Mongolia. Archaeology, Ethnology,
and Archaeology of Eurasia 38(3), 33-40.
475
Goebel, T., 2002. The “microblade adaptation” and recolonization of Siberia during the
late Upper Pleistocene in: Elston, R. G., Kuhn, S. L. (Eds.), Thinking Small:
Global Perspectives on Microlithization. Archaeological Papers of the American
Anthropological Association, 12. American Anthropological Association,
Arlington, Virginia, pp. 115-131.
Godfrey-Smith, D. I., Deal, M., Kunelius, I., 1997. Thermoluminscence dating of St.
Croix ceramics: chronology building in southwestern Nova Scotia.
Geoarchaeology 12(3), 251-273.
Goriunova, O. I., Khlobystin, L. P., 1991. Datirovka kompleksov poselenii i nogrebenii
bykhtyi Ulan-Khada [Dating of habitation and burial complexes at Ulan-Khada
Cove] in: Masson, V. M. (Ed.), Drevnosti Baikala. Irkutsk National University,
Irkutsk, pp. 41-56. (in Russian)
Gramly, R. M., 1980. Raw material source areas and “curated” tool assemblages.
American Antiquity 45(4), 823-833.
Grunert, J., Lehmkuhl, F., 2004. Aeolian sedimentation in arid and semi-arid
environments of western Mongolia in: Smykatz-kloss, W., Zöller, L., FelixHenningsen, P. (Eds.), Paleoecology of Quaternary Drylands. Lecture Notes in
Earth Sciences, Vol. 102. Springer, Heidelberg, pp. 195-218.
Grunert, J., Lehmkuhl, F., Walther, M., 2000. Paleoclimatic evolution of the Uvs Nuur
Basin and adjacent areas (Western Mongolia). Quaternary International 65/66,
171-192.
476
Gryaznov, M. P., 1969. The Ancient Civilization of Southern Siberia. Cowles Book
Company, New York.
Gunin, P. D., Vostokova, E. A., Dorofeyuk, N. I., Tarasov, P. E., Black, C. C., 1999.
Vegetation Dynamics of Mongolia. Geobotany 26. Kluwer Academic Publisher,
Dordrecht.
Guo, D., 1995a. Hongshan and related cultures in: Nelson, S. M. (Ed.), The Archaeology
of Northeast China: Beyond the Great Wall. Routledge, London, pp. 21-64.
Guo, D., 1995b. Lower Xiajiadian Culture in: Nelson, S. M. (Ed.), The Archaeology of
Northeast China: Beyond the Great Wall. Routledge, London, pp. 147-181.
Habu, J., 1996. Jomon sedentism and intersite variability: collectors of the Early Jomon
Moroiso phase in Japan. Arctic Anthropology 33(2), 38-49.
Habu, J., 2004. Ancient Jomon of Japan. Cambridge University Press, Cambridge.
Halkett, D., Hart, T., Yate, R., Volman, T. P., Parkington, J. E., Orton, J., Klein, R. G.,
Cruz-Uriba, K., Avery, G., 2003. First excavation of intact Middle Stone Age
layers at Ysterfontein, Western Cape Province, South Africa: implications for
Middle Stone Age ecology. Journal of Archaeological Science 30, 955-971.
Hämäläinen, P., 2003. The rise and fall of Plains Indian horse culture. The Journal of
American History 90(3), 833–862.
Handwerker, W. P., 1983. The first demographic transition: an analysis of subsistence
choices and reproductive consequences. American Anthropologist 85(1), 5-27.
477
Han, D., Xu, C., 1985. Pleistocene mammalian faunas of China in: Wu, R., Olsen, J. W.
(Eds.), Palaeoanthropology and Palaeolithic Archaeology in the People’s
Republic of China. Academic Press, Orlando, pp. 267-289.
Han, Y., Fang, X., Kang, S., Wang, H., Kang, F., 2008. Shifts in dust source over central
Asia and the Tibetan Plateau: connections with the Arctic oscillation and the
westerly jet. Atmospheric Environment 42, 2358-2368.
Hartmann, K., 2003. Spätpleistozäne und Holozäne Morphodynamik im nördlichen
Gaxun Nur Becken, Innere Mongolei, NW China, unter besonderer
Berücksichtigung des Juanze-Paläosees. Unpublished Ph.D. Thesis. Freie
Universität Berlin. (in German)
Hartmann, K., Wünnemann, B., 2009. Hydrological changes and Holocene climate
variations in NW China, inferred from lake sediments of Juyanze palaeolake by
factor analyses. Quaternary International 194, 28-44.
Hassan, F., 1973. On mechanisms of population growth during the Neolithic. Current
Anthropology 14, 535-540.
Hassan, F., 1981. Demographic Archaeology. Academic Press, New York.
Hawkes, K., Hill, K., O’Connell, J. F., 1982. Why hunters gather: optimal foraging and
the Aché of eastern Paraguay. American Ethnologist 9: 379-398.
Hayden, B., 1989. From chopper to celt: the evolution of resharpening techniques in:
Torrence, R. (Ed.), Time, Energy and Stone Tools. New Directions in
Archaeology. Cambridge University Press, Cambridge, pp. 7-16.
478
Hemphill, B. E., Mallory, J. P., 2004. Horse-mounted invaders from the Russo-Kazakh
steppe or agricultural colonists from western Central Asia? A craniometric
investigation of the Bronze Age settlement in Xinjiang. American Journal of
Physical Anthropology 124, 199-222.
Hedin, S., 1943. History of the Expedition in Asia 1927-1935, Part I. Reports from the
Scientific Expedition to the North-Western Provinces of China under the
Leadership of Dr. Sven Hedin, Publication 23. Statens Etnografiska Museum,
Stockholm.
Herbert, J. M., Feathers, J. K., Cordell, A. S., 2002. Building ceramic chronologies with
thermoluminescence dating: a case study from the Carolina Sandhills.
Southeastern Archaeology 21(1), 92-109.
Herzschuh, U., 2006. Palaeo-moisture evolution in monsoonal Central Asia during the
last 50,000 years. Quaternary Science Reviews 25, 163-178.
Herzschuh, U., Kürschner, H., Ma, Y., 2003. The surface pollen and relative pollen
production of the desert vegetation of the Alashan Plateau, western Inner
Mongolia. Chinese Science Bulletin 48(14), 1488-1493.
Herzschuh, U., Liu, X., 2007. Vegetation evolution in arid China during Marine Isotope
Stages 3 and 2 (~65-11 ka) in: Madsen, D.B., Chen, F.-H., Xing, G. (Eds.), Late
Quaternary Climate Change and Human Adaptation in Arid China. Elsevier,
Amsterdam, pp. 41–49.
479
Herzschuh, U., Tarasov, P., Wünnemann, B., Hartmann, K., 2004. Holocene vegetation
and climate of the Alashan Plateau, NW China, reconstructed from pollen data.
Palaeogeography, Palaeoclimatology, Palaeoecology 211, 1-17.
Hiendleder, S., Kaupe, B., Wassmuth, R., Janke, A., 2002. Molecular analysis of wild
and domestic sheep questions current nomenclature and provides evidence for
domestication from two different subspecies. Proceedings: Biological Sciences
269(1494), 893-904.
Hiendleder, S., Mainz, K., Plante, Y., Lewalski, H., 1998. Analysis of mitochondrial
DNA indicates that domestic sheep are derived from two different ancestral
maternal sources: no evidence for contributions from Urial and Argali sheep.
Journal of Heredity 89, 113-120.
Hill, W. P. T., N.d. “Inner Mongolia” [map]. 1:200,000. American Museum of Natural
History, Central Asiatic Expeditions Route Maps, Sheet Nos. 20, 21, 22, 28, 29.
Hill, W. P. T., Roberts, L. B., N.d. “Inner Mongolia” [map]. 1:200,000. American
Museum of Natural History, Central Asiatic Expeditions Route Maps, Sheet Nos.
24, 26.
Hill, W. P. T., Roberts, L. B., Butler, F. B., Robinson, H. G., N.d. “Inner Mongolia”
[map]. 1:200,000. American Museum of Natural History, Central Asiatic
Expeditions Route Maps, Sheet No. 27.
480
Hillman, G. C., 1989. Late Palaeolithic plant foods from Wadi Kubbaniya in Upper
Egypt: dietary diversity, infant weaning, and seasonality in a riverine environment
in: Harris, D. R., Hillman, G. C. (Eds.), The Evolution of Plant Exploitation.
Unwin Hyman, London, pp. 207-239.
Hillman, G. C., 1996. Late Pleistocene changes to wild plant-foods available to huntergatherers of the northern Fertile Crescent: possible preludes to cereal cultivation
in: Harris, D. (Ed.), The Origins and Spread of Agriculture and Pastoralism in
Eurasia. Smithsonian Institution Press, Washington, D.C., pp. 159-204.
Hodder, I., 1990. The Domestication of Europe. Blackwell Publishers, Oxford.
Hofman, J., 1992. Recognition and interpretation of Folsom technological variability on
the southern plains in: Stanford, D., Day, J. (Eds.), Ice Age Hunters of the
Rockies. Denver Museum of Natural History and University Press of Colorado,
Niwot, pp. 193-224.
Hole, F., 1996. The context of caprine domestication in the Zagros region in: Harris, D.
(Ed.), The Origins and Spread of Agriculture and Pastoralism in Eurasia.
Smithsonian Institution Press, Washington, D.C., pp. 263-281.
Hölz, S., Polag, D., Becken, M., Fiedler-Volmer, R., Zhang, H., Hartmann, K.,
Burkhardt, H., 2007. Electromagnetic and Geoelectric Investigation of the
Gurinai Structure, Inner Mongolia, NW China. Tectonophysics 445, 26-48.
Honeychurch, W., Amartuvshin, C., 2006. States on horseback: the rise of Inner Asian
confederations and empires in: Stark, M. (Ed.), Archaeology of Asia. Blackwell
Publishing, Oxford, pp. 255-278.
481
Hoopes, J. W., 1995. Interaction in hunting and gathering societies as a context for the
emergence of pottery in the Central American Isthmus in: Barnett, W. K.,
Hoopes, J. W. (Eds.), The Emergence of Pottery: Technology and Innovation in
Ancient Societies. Smithsonian Series in Archaeological Inquiry. Smithsonian
Institution Press, Washington, D. C., pp. 185-198.
Houle, J. –L., 2010. Emergent Complexity on the Mongolian Steppe: Mobility,
Territoriality, and the Development of Early Nomadic Polities. Unpublished
Ph.D. Dissertation. University of Pittsburgh, Pittsburgh.
Hülle, D., Hilgers, A., Ratke, U., Stolz, C., Hempelmann, N., Grunert, J., Felauer, T.,
Lehmkuhl, F., 2009. OSL dating of sediments from the Gobi Desert, Southern
Mongolia. Quaternary Geochronology 5(2-3), 107-113.
Ikawa-Smith, F., 1976. On ceramic technology in East Asia. Current Anthropology 17,
513-515.
Ikawa-Smith, F., 1986. Late Pleistocene and early Holocene technologies in: Pearson, R.
J., Barnes, G. L., Hutterer, K. L. (Eds.), Windows on the Japanese Past: Studies in
Archaeology and Prehistory. Center for Japanese Studies, The University of
Michigan, Ann Arbor, pp. 199-216.
Ikawa-Smith, F., 2008. Japanese Archipelago, Paleolithic cultures in: Pearsall, D. M.
(Ed.), Encyclopedia of Archaeology, vol. 1. Academic Press, New York, pp. 632637.
482
Ingbar, E., 1992. The Hanson site and Folsom on the northwestern plains in: Stanford,
D., Day, J. (Eds.), Ice Age Hunters of the Rockies. Denver Museum of Natural
History and University Press of Colorado, Niwot, pp. 169-192.
Jacobsen, W., 1940. Mongoliets arkaeologi en kort oversight [Mongolian archaeology, a
short oversight] in: Fra Danmarks Ungtid, Arkaeologiske Studier til Johannes
Brøndsted PAA 50 – Aarsdagen. Ejnar Munksgaard, Copenhagan, pp. 206-221.
(in Danish)
Jacobson, E., 1988. Beyond the frontier: a reconsideration of cultural interchange
between China and the early nomads. Early China 13, 201-240.
Jacobson, E., 2002. Petroglyphs and the qualification of Bronze Age mortuary
archaeology. Archaeology, Ethnology and Anthropology of Eurasia 3(11), 32-47.
Jansen, T., Forster, P., Levine, M. A., Oelke, H., Hurles, M., Renfrew, C., Weber, J.,
Olek, K., 2002. Mitochondrial DNA and the origins of the domestic horse.
PNAS 99(16), 10905-10910.
Janz, L., 2006. Shabarakh-usu and the Dune Dwellers of the Gobi: Explanations for
Lithic Assemblage Variability in the Gobi Desert, Mongolia. Unpublished
Master’s Thesis, Department of Anthropology, University of Arizona.
Janz, L., 2007. Pastoralism and ideological resistance to agriculture: Mongolia from
prehistory to modern times. Arizona Anthropologist 18, 28-52.
Janz, L., Elston, R. G., Burr, G. S., 2009. Dating Northeast Asian surface assemblages
with ostrich eggshell: implications for palaeoecology and extirpation. Journal of
Archaeological Science 36, 1982-1989.
483
Jaubert, J., Bertran, P., Fontugne, M., Jarry, M., Lacombe, S., Leroyer, C., Marmet, E.,
Taborin, Y., Tsogtbaatar, Brugal, J. P., Desclaux, F., Poplin, F., Rodière, J.,
Servelle, C., 2004. Le Paléolithique supérieur ancient de Mongolie: Dörölj 1
(Egïïn Gol). Analogies avec les données de l’Altaï et de Sibériein in: Le
Secrétariat du Congrès (Ed.), The Upper Palaeolithic – General Sessions and
Posters. Acts of the XIVth UISPP Congress, University of Liège, Belgium, 2-8
September 2001. Archaeopress, Oxford, pp. 245-251. (in French)
Jia, W. M., 2007. Transition from Foraging to Farming in Northeast China. BAR
International Series 1692. Archaeopress, Oxford.
Jiang, L., Liu, L., 2006. New evidence for the origins of sedentism and rice
domestication in the Lower Yangzi River, China. Antiquity 80, 355-361.
Jiang, W., Guo, Z., Sun, X., Wu, H., Chu, G., Yuan, B., Hatté, C., Guiot, J., 2006.
Reconstruction of climate and vegetation changes of Lake Bayanchagan (Inner
Mongolia): Holocene variability of the East Asian monsoon. Quaternary
Research 65, 411-420.
Jigjidsuren, S., Johnson, D. A., 2003. Forage Plants in Mongolia (Mongol oroni malin
tejeeliin urgamal). Admon, Ulaanbaatar. (in English and Mongolian)
Ji, R., Cui, P., Ding, F., Geng, J., Gao, H., Zhang, H., Yu, J., Hu, S., Meng, H., 2009.
Monophyletic origin of domestic bactrian camel (Camelus bactrianus) and its
evolutionary relationship with the extant wild camel (Camelus bactrianus ferus).
Animal Genetics 40, 377-382.
484
Junker, L. L., 1996. Hunter-gatherer landscapes and lowland trade in the Prehispanic
Philippines. World Archaeology 27(3), 389-410.
Junker, L. L., 2000. Raiding, Trading, and Feasting: the Political Economy of Philippine
Chiefdoms. Ateneo de Manila University Press, Manila.
Keally, C. T., Taniguchi, Y., Kuzmin, Y. V., 2003. Understanding the beginnings of
pottery technology in Japan and neighbouring East Asia. The Review of
Archaeology 24(2), 3-14.
Keates, S. G., 2007. Microblade technology in Siberia and neighbouring regions: an
overview in: Kuzmin, Y. V., Keates, S. G., Shen, C. (Eds.), Origin and Spread of
Microblade Technology in Northern Asia and North America. Archaeology
Press, Simon Fraser University, Burnaby, pp. 125-146.
Keeley, L. H., 1988. Hunter-gatherer economic complexity and “population pressure”: a
cross-cultural analysis. Journal of Anthropological Archaeology 7, 373-411.
Kelly, R., 1988. The three sides of a biface. American Antiquity 51: 737-747.
Kelly, R., 1995. The Foraging Spectrum. Smithsonian Institution Press, Washington.
Khasbagan, Huai, H.- Y., Pei, S.- J., 2000. Wild plants in the diet of Arhorchin Mongol
herdsmen in Inner Mongolia. Economic Botany 54(4), 528-536.
Khlobystin, L. P., 1969. The stratified settlement of Ulan-Khada on Lake Baikal (Based
on materials excavated by B. E. Petri). Arctic Anthropology 6(1), 88-94.
Kim, K. S., Choi, C. B., 2002. Genetic structure of Korean native pig using microsatellite
markers. Korean Journal of Genetics 24, 1–7.
485
Kirillov, I., Derevianko, A. P., 1998. The Paleolithic of the Trans-Baikal area in:
Derevianko, A. P., Shimkin, D. B., Powers, W. R. (Eds.), The Palaeolithic of
Siberia: New Discoveries and Interpretations. University of Illinois Press, Urbana
and Chicago, pp. 137-150.
Kislenko, A., Tatarintseva, N., 1999. The eastern Ural steppe at the end of the Stone Age
in: Levine, M., Rassamakin, Y., Kislenko, A., Tatarinteseva, N. (Ed.), Late
Prehistoric Exploitation of the Eurasian Steppe. McDonald Institute for
Archaeological Research, Cambridge, pp. 183-216.
Kniaziev, A. W., 1986. Changes in the extent of the carnivore (mammals) in the uplands
during the Holocene in: Deserts of the Trans-Altai Gobi, Natural Resources,
Ecosystems and their Regionalization. Nauka, Moscow, pp. 131-135. (in
Russian)
Köhler-Rollefson, I., 1996. The one-humped camel in Asia: origin, utilization and
mechanisms of dispersal in: Harris, D. (Ed.), The Origins and Spread of
Agriculture and Pastoralism in Eurasia. Smithsonian Institution Press,
Washington, D.C., pp. 282-294.
Kohl, P. L., 2007. The Making of Bronze Age Eurasia. Cambridge University Press,
Cambridge.
Komatsu, G., Brantingham, P. J., Olsen, J. W., Baker, V. R., 2001. Paleoshoreline
geomorphology of Böön Tsagaan Nuur, Tsagaan Nuur and Orog Nuur: the Valley
of Lakes, Mongolia. Geomorphology 39, 83-98.
486
Kozłowski, J. K., 1972. Research on the Stone Age in South Mongolia in 1968.
Archaeologia Polona 13, 231-261.
Kozłowski, J. K., Kozłowski, S. K., 1986. Foragers of central Europe and their
acculturation in: Zvelebil, M. (Ed.), Hunters in Transition: Mesolithic Societies of
Temperate Eurasia and their Transition to Farming. Cambridge University Press,
Cambridge, pp. 95-108.
Krader, L., 1959. The ecology of nomadic pastoralism. International Social Science
Journal 11, 499-510.
Kuhn, S. L., 1990. A geometric index of reduction for unifacial stone tools. Journal of
Archaeological Science 17, 583-593.
Kuhn, S. L., 1991. “Unpacking” reduction: lithic raw material economy in the
Mousterian of West-Central Italy. Journal of Anthropological Archaeology 10,
76-106.
Kuhn, S. L., 1992. On planning and curated technologies in the Middle Palaeolithic.
Journal of Anthropological Research 48(3), 185-214.
Kuhn, S. L., 1994. A formal approach to the design and assembly of mobile toolkits.
American Antiquity 59(3), 426-442.
Kuhn, S. L., 1995. Chapter 2. Technology, foraging, and land use: a strategic approach
in: Kuhn, S. L., Mousterian Lithic Technology: An Ecological Approach.
Princeton University Press, Princeton, New Jersey, pp. 18-37.
Kuhn, S. L., 2004. Upper Paleolithic raw material economies at Üçağızlı cave, Turkey.
Journal of Anthropological Archaeology 23, 431-448.
487
Kuhn. S. L., 2007. Cores, tools, and the priorities of lithic analysis in: McPherron, S. P.,
Tools versus Cores: Alternative Approaches to Stone Tool Analysis. Cambridge
Scholars Publications, Newcastle, pp. 267-276.
Kuhn, S. L., Stiner, M. C., 2001. The antiquity of hunter-gatherers in: Panter-Brick, C.,
Layton, R. H., Rowley-Conwy, P. A. (Eds.), Hunter-Gatherers: An
Interdisciplinary Perspective. Cambridge University Press, Cambridge, pp. 99142.
Kuhn, S. L., Stiner, M. C., 2006. What’s a mother to do? The division of labor among
Neandertals and modern humans in Eurasia. Current Anthropology 47(6), 953980.
Kulik, N. A., Nokhrina, T. I., Milyotin, K. I., 2006. Petrografiya artefaktov
geoarkheologicheskovo ob”ekta Shabarak-usu (yozhnaya-Mongoliya)
(Petrography of artefacts and a geoarchaeological object at Shabarakh-usu
[southern Mongolia]). Problemyi arkheologii, etnografii, antropologii Sibiri i
sopredel’nyikh territoriy XII(I), 167-172. (in Russian)
Kurochkin, E. N., Kuzmin, Y. V., Antoshchenko-Olenev, I. V., Zabelin, V. I.,
Krivonogov, S. K., Nohrina, T. I., Lbova, L. V., Burr, G. S., Cruz, R. J., 2009.
The timing of ostrich existence in Central Asia: AMS 14C age of eggshell from
Mongolia and southern Siberia (a pilot study). Nuclear Instruments and Methods
in Physics Research Section B: Beam Interactions with Materials and Atoms
268(7-8), 1091-1093.
488
Kutzbach, J. E., 1981. Monsoon climate of the early Holocene: climate experiment using
the earth’s orbital parameters for 9000 years ago. Science 214, 59-61.
Kuzmina, E. E., 1998. Cultural connections of the Tarim Basin people and pastoralists of
the Asian steppes in the Bronze Age in: Mair, V. (Ed.), The Bronze Age Eastern
Central Asia. University of Pennsylvania Museum Publications, Philadelphia, pp.
63-98.
Kuzmina, E. E., Mallory, J. P. (Ed.), 2007. The Origin of the Indo-Iranians. Brill,
Leiden.
Kuzmin, Y. V., 1995. People and environment in the Russian Far East from Paleolithic
to Middle Ages: chronology, palaeogeography, interaction. Geojournal 35(1), 7983.
Kuzmin, Y. V., 1997. Vertebrate animal remains from prehistoric and medieval
settlements in Primorye (Russian Far East). International Journal of Osteology 7,
172-180.
Kuzmin, Y. V., 1998. Early agriculture in Primorye, Russian Far East: new radiocarbon
and pollen data from late Neolithic sites. Journal of Archaeological Science 25,
813-816.
Kuzmin, Y. V., 2003. Introduction: changing the paradigm. The Review of Archaeology
24(2), 1-3.
Kuzmin, Y. V., 2006. Chronology of the earliest pottery in East Asia: progress and
pitfalls. Antiquity 80, 362-371.
489
Kuzmin, Y. V., 2007. Geoarchaeological aspects of the origin and spread of microblade
technology in northern and central Asia in: Kuzmin, Y. V., Keates, S. G., Shen, C.
(Eds.), Origin and Spread of Microblade Technology in Northern Asia and North
America. Archaeology Press, Simon Fraser University, Burnaby, pp. 115-124.
Kuzmin, Y. V., 2010. Extinction of the woolly mammoth (Mammuthus primigenius) and
woolly rhinoceros (Coelodonta antiquitatis) in Eurasia: review of chronological
and environmental issues. Boreas 39(2), 247-261.
Kuzmin, Y. V., Orlova, L. A., 2000. The Neolithization of Siberia and the Russian Far
East: radiocarbon evidence. Antiquity 74, 356-364.
Kuzmin, Y. V., Keates, S. G., Shen, C. (Eds.), 2007. Origin and Spread of Microblade
Technology in Northern Asia and North America. Archaeology Press, Simon
Fraser University, Burnaby.
Kuzmin, Y. V., Shewkomud, I. Y., 2003. The Palaeolithic-Neolithic transition in the
Russian Far East. The Review of Archaeology 24(2), 37-45.
Kuzmin, Y. V., Speakman, R. J., Glascock, M. D., Popov, V. K., Grebennikov, A. V.,
Dikova, M .A., Ptashinsky, A. V., 2008. Obsidian use at the Ushki Lake
complex, Kamchatka Penninsula (Northeastern Siberia): implications for terminal
Pleistocene and early Holocene human migrations in Beringia. Journal of
Archaeological Science 35: 2179-2187.
Larichev, V. E., 1962. The microlithic character of Neolithic cultures in Central Asia,
Trans-Baikal, and Manchuria. American Antiquity 27(3), 315-322.
490
Larson, G., Dobney, K., Albarella, U., Fang, M., Matisoo-Smith, E., Robins, J., Lowden,
S., Finlayson, H., Brand, T., Willerslev, E., Rowley-Conway, P., Andersson, L.,
Cooper, A., 2005. Worldwide phylogeny of wild boar reveals multiple centers of
pig domestication. Science 307, 1618-1621.
Lechtman, H., 1977. Style in technology – some early thoughts in: Lechtman, H.,
Merrill, R. (Eds.), Material Culture: Styles, Organization, and Dynamics of
Technology. The 1975 Proceedings of the American Ethnological Society. West
Coast Publishing Co., St. Paul, pp. 3-20.
Lemonnier, P., 1992. Elements for an Anthropology of Technology. Anthropological
Papers No. 88. Museum of Anthropology, University of Michigan, Ann Arbor.
Lee, G. –A., Crawford, G. W., Liu, L., Chen, X., 2007. Plants and people from the early
Neolithic to Shang periods in North China. PNAS 104(3), 1087-1092.
Lee, J. –J., 2001. From Shellfish Gathering to Agriculture in Prehistoric Korea: the
Chulmun to Mumun Transition. Unpublished Ph.D. Dissertation. University of
Wisconsin, Madison.
Legrand, S., 2006. The emergence of the Karasuk culture. Antiquity 80, 843-879.
Lehmkuhl, F., Haselein, F., 2000. Quaternary paleoenvironmental change on the Tibetan
Plateau and adjacent areas (Western China and Western Mongolia). Quaternary
International 65-66, 121-145.
Lehmkhul, F., Lang, A., 2001. Geomorphological investigations and luminescence
dating in the southern part of the Khangay and the Valley of the Gobi Lakes
(Central Mongolia). Journal of Quaternary Science 16(1), 69-87.
491
Levine, M., 1999a. Botai and the origins of horse domestication. Journal of
Anthropological Archaeology 18, 29-78.
Levine, M., 1999b. The origins of horse husbandry on the Eurasian steppe in: Levine,
M., Rassamakin, Y., Kislenko, A., Tatarintseva, N. (Eds.), Late Prehistoric
Exploitation of the Eurasian Steppe. McDonald Institute for Archaeological
Research, Cambridge, pp. 5-58.
Lieberman, D. E., Belfer-Cohen, A., Henry, D. O., Kaufman, D., Mackie, Q., Olszewski
D. I., Rocek, T. R., Sheppard, P. J., Trinkaus, E., Valla, F. R., 1993. The rise and
fall or seasonal mobility among hunter-gatherers: the case of the southern Levant
[and comments and replies]. Current Anthropology 34(5), 599-631.
Li, S., 2002. The interaction between Northwest China and Central Asia during the
second millennium BC: an archaeological perspective in: Boyle, K., Renfrew, C.,
Levine, M. (Eds.), Ancient Interactions: East and West in Eurasia. McDonald
Institute for Archaeological Research, Cambridge, pp.171-182.
Li, S.-H., Sun, J.-M., Zhao, H., 2002. Optical dating of dune sands in the northeastern
deserts of China. Palaeogeography, Palaeoclimatology, Palaeoecology 181, 419429.
Li, X., Dodson, J., Zhou, X., Zhang, H., Masutomoto, R., 2007. Early cultivated wheat
and broadening of agriculture in Neolithic China. The Holocene 17(5), 555-560.
Li, X., Shang, X., Dodson, J., Zhou, X., 2009. Holocene agriculture in the Guanzhong
Basin in NW China indicated by pollen and charcoal evidence. The Holocene
19(8), 1213-1220.
492
Liu, H., Xu, L., Cui, H., 2002a. Holocene history of desertification along the woodlandsteppe border in northern China. Quaternary Research 57, 259-270.
Liu, J., Yu, G., Chen, X., 2002b. Palaeoclimate simulation of 21ka for the Tibetan
Plateau and Eastern Asia. Climate Dynamics 19, 575-583.
Liu, L., Chen, X., 2006. Sociopolitical change from Neolithic to Bronze Age China in:
Stark, M. (Ed.), Archaeology of Asia. Blackwell Publishing, Oxford, pp. 149176.
Liu, L., 2004. The Chinese Neolithic: Trajectories to Early States. Cambridge
University Press, Cambridge.
Liu, Z., 1995. Recent Neolithic discoveries in Jilin Province in: Nelson, S. M. (Ed.), The
Archaeology of Northeast China: Beyond the Great Wall. Routledge, London,
pp. 89-117.
Loftus, R. T., MacHugh, D. E., Bradley, D. G., Sharp, P. M., Cunningham, P., 1994.
Evidence for two independent domestications of cattle. PNAS 91, 2757-2761.
Losey, R. J., Bazaliiskii, V. L., Garvie-Lok, S., Germonpré, M., Leonard, J. A., Allen, A.
L., Katzenberg, M. A., Sablin, M. V., 2011. Canids as persons: early Neolithic
dog and wolf burials, Cis-Baikal, Siberia. Journal of Anthropological
Archaeology 30(2), 174-189.
Luikart, G., Gielly, L., Excoffier, Vigne, J. –D., Bouvet, J., Taberlet, P., 2001. Multiple
maternal origins and weak phylogeographic structure in domestic goats. PNAS
98(10), 5927-5932.
493
Lukacs, J. R., 1990. On hunter-gatherers and their neighbors in prehistoric India: contact
and pathology. Current Anthropology 31(2), 183-186.
Lupo, K. D., Schmitt, D. N., 2002. Upper Palaeolithic net-hunting, small prey
exploitation, and women’s work effort: a view from the ethnographic and
ethnoarchaeological record of the Congo Basin. Journal of Archaeological
Method and Theory 9(2), 147-179.
Lu, H., Zhang, J., Liu, K., Wu, N., Li, Y., Zhou, K., Ye, M., Zhang, T., Zhang, H., Yang,
X., Shen, L., Xu, D., Li, Q., 2009. Earliest domestication of common millet
(Panicum miliacum) in East Asia extended to 10,000 years ago. PNAS 106(18),
7367-7372.
Lu, L. D., 1998. The microblade tradition in China: regional chronologies and
significance in the transition to Neolithic. Asian Perspectives 37(1), 84-112.
Lu, T. L. D., 1999. The Transition from Foraging to Farming and the Origin of
Agriculture in China. BAR International Series 774. Archaeopress, Oxford.
Lu, X., Yuan, B., Guo, Z., Li, K., 1997. The 5000-year environmental change and
associated human activities at Sokho-nor Lake, Inner Mongolia, P.R. China.
Nuclear Instruments and Methods in Physics Research B 123, 460-463.
MacArthur, R., Pianka, E., 1966. On optimal use of a patchy environment. American
Naturalist 100, 603-609.
MacHugh, D. E., Bradley, D. G., 2001. Livestock genetic origins: goats buck the trend.
PNAS 98(10), 5382-5384.
494
Madsen, D. B., Elston, R. G., Bettinger, R. L., Xu, C., Zhong, K., 1996. Settlement
patterns reflected in assemblages from the Pleistocene/Holocene transition of
north central China. Journal of Archaeological Science 23, 217-231.
Madsen, D. B., Elston, R. G., 2007. Variation in Late Quaternary central Asian climates
and the nature of human response in: Madsen, D. B., Chen, F. H., Gao, X. (Eds.),
Late Quaternary Climate Change and Human Adaptation in Arid China,
Developments in Quaternary Science, Vol. 9. Elsevier, Amsterdam, pp. 69-82.
Madsen, D. B., Li, J., Brantingham, P. J., Gao, X., Elston, R. G., Bettinger, R. L., 2001.
Dating Shuidonggou and the Upper Palaeolithic blade industry in North China.
Antiquity 75, 706-716.
Madsen, D. B., Li, J., Elston, R. G., Xu, C., Bettinger, R. L., Geng, K., Brantingham, P.
J., Zhong, K., 1998. The loess/paleosol record and the nature of the Younger
Dryas climate in Central China. Geoarchaeology 13(8), 847-869.
Malainey, M. E., Przybylski, R., Sherriff, B. L., 1999. Identifying the former contents of
Late Precontact period pottery vessels from Western Canada using gas
chromatography. Journal of Archaeological Science 26, 425-438.
Malainey, M. E., Przbylski, R., Sherriff, L., 2001. One person’s food: how and why fish
avoidance may affect the settlement and subsistence patterns of hunter-gatherers.
American Antiquity 66(1), 141-161.
Mallory, J. P., 1989. In Search of the Indo-Europeans: Language, Archaeology, and
Myth. Thames and Hudson, New York.
495
Manne, T., Bicho, N., 2009. Vale Boi: Rendering new understandings of resource
intensification, diversification, and specialization in southwestern Iberia. Before
Farming [online version] 2 (article 1), 1-21.
Manne, T., Cascalheira, J., Évora, M., Marreiros, J. and Bicho, N., 2012. Intensive
subsistence practices at Vale Boi, an Upper Palaeolithic site in southwestern
Portugal. Quaternary International. doi:10.1016/j.quaint.2012.02.026 (in press)
Manne, T., 2012. Vale Boi: 10, 000 years of Upper Paleolithic bone boiling in: Graff, S.
R., Rodríguez-Alegria, E. (Eds.), The Menial Art of Cooking: Archaeological
Studies of Cooking and Food Preparation. University Press of Colorado, Boulder,
pp. 173-200.
Mannen, H., Kohno, M., Nagata, Y., Tsuji, S., Bradley, D. G., Yeo, J. S., Nyamsamba,
D., Zagsuren, Y., Yokohama, M., Nomura, K., Amano, T., 2004. Independent
mitochondrial origin and historical genetic differentiation in north eastern Asian
cattle. Molecular Phylogenetics and Evolution 32, 539-544.
Maringer, J., 1950. Contribution to the Prehistory of Mongolia. Reports from the
Scientific Expedition to the North-western Provinces of China under the
Leadership of Sven Hedin, Sino-Swedish Expedition Publication, Publication 34.
Statens Etnografiska Museum, Stockholm.
Maringer, J., 1963. Mongolia before the Mongols. Arctic Anthropology 1(2), 75-85.
496
Markin, S., 1998. Paleolithic sites in the Western Sayan Mountains in: Derevianko, A.
P., Shimkin, D. B., Powers, W. R. (Eds.), The Palaeolithic of Siberia: New
Discoveries and Interpretations. University of Illinois Press, Urbana and Chicago,
pp. 118-120.
Marshall, F., Hildebrand, E., 2002. Cattle before crops: the beginnings of food
production in Africa. Journal of World Prehistory 16(2), 99-142.
Masson, V. M., Sarianidi, V. I., 1972. Central Asia: Turkmenia before the Achaemenids.
Thames and Hudson, Ltd., London.
Matsui, A., Ishiguro, N., Hongo, H., Minagawa, M., 2005. Wild pig? Or domesticated
boar? An archaeological view on the domestication of Sus scrofa in Japan in:
Vigne, J. –D., Peters, J., Helmer, D. (Eds.), The First Steps of Animal
Domestication. Proceedings of the 9th ICAZ Conference, Durham 2002. Oxbow
Books, Oxford, pp. 148-159.
Matsui, A., Kanehara, M., 2006. The question of prehistoric plant husbandry during the
Jomon period in Japan. World Archaeology 38(2), 259-273.
McBrearty, S., Brooks, A. S., 2000. The revolution that wasn’t: a new interpretation of
the origin of modern human behavior. Journal of Human Evolution 39, 453-563.
McKenzie, H. G., 2009. Review of early hunter-gatherer pottery in eastern Siberia in:
Jordan, P., Zvelebil, M. (Eds.), Ceramics before Farming: the Dispersal of Pottery
among Prehistoric Eurasian Hunter-Gatherers. Left Coast Press, Walnut Creek,
pp., 167-208.
497
Meadow, R. H., 1996. The origins and spread of agriculture and pastoralism in
northwestern South Asia in: Harris, D. R. (Ed.), The Origins and Spread of
Agriculture and Pastoralism in Eurasia. Smithsonian Institution Press,
Washington, D.C., pp. 390-412.
Mercader, J., Garcia-Heras, M., Gonzalez-Alvarez, I., 2000. Ceramic tradition in the
African forest: characterisation analysis of ancient and modern pottery from Ituri,
D. R. Congo. Journal of Archaeological Science 27, 163-182.
Michael, H. N., 1984. Absolute chronologies of late Pleistocene and early Holocene
cultures of northeastern Asia. Arctic Anthropology 21(2), 1-68.
Miehe, G., Schlütz, F., Miehe, S., Opgenoorth, L., Cermak, J., Samiya, R., Jäger, E. J.,
Wesche, K., 2007. Mountain forest islands and Holocene environmental changes
in Central Asia: a case study from the southern Gobi Altay, Mongolia.
Palaeogeography, Palaeoclimatology, Palaeoecology 250, 150-166.
Miller, G. H., Magee, J. W., Johnson, B. J., Fogel, M. L., Spooner, N. A., McCulloch, M.
T., Ayliffe, L. K., 1999. Pleistocene extinction of Genyornis newtoni: human
impact on Australian megafauna. Science 283, 205-208.
Mills, P., 1993. An axe to grind: a functional analysis of Anasazi stone axes from Sand
Canyon Pueblo Ruin (5MT765), southwestern Colorado. Kiva 58(3), 393-413.
Mischke, S., Demske, D., Schudack, M. E., 2003. Hydrologic and climatic implications
of a multi-disciplinary study of the mid to late Holocene Lake Eastern Juyanze.
Chinese Science Bulletin 48(14), 1411-1417.
498
Mischke, S., Demske, D., Wünnemann, B., Schudack, M. E., 2005. Groundwater
discharge to a Gobi desert lake during mid and late Holocene dry periods.
Palaeogeography, Palaeoclimatology, Palaeoecology 225, 157-172.
Mischke, S., 2001. Mid and Late Holocene palaeoenvironment of the lakes Eastern
Juyanze and Sogo Nur in NW China, based on ostracod species assemblages and
shell chemistry. Berliner Geowissenschaftliche Abhandlungen, Reihe E, Band
35. FU-TU-TFH, Berlin.
Montell, G., 1945. As ethnographer in China and Mongolia 1929-1932 in: Hedin, S.
(Ed.), History of the Expedition in Asia 1927-1935. Reports from the Scientific
Expedition to the North-western Provinces of China under the Leadership of Sven
Hedin, Sino-Swedish Expedition Publication, Publication 26, Part IV. Statens
Etnografiska Museum, Stockholm, pp. 327-449.
Moore, A. M. T., 1995. The inception of potting in western Asia and its impact on
economy and society in: Barnett, W. K., Hoopes, J. W. (Eds.), The Emergence of
Pottery, Technology and Innovation in Ancient Societies. Smithsonian Institution
Press, Washington, D. C., pp. 39-53.
Morgan, C., 2009. Climate change, uncertainty and prehistoric hunter-gatherer mobility.
Journal of Anthropological Archaeology 28, 382-396.
Morlan, R. E., 1967. The preceramic period of Hokkaido: an outline. Arctic
Anthropology 4(1), 164-220.
499
Morlan, R. E., 1976. Technological characteristics of some wedge-shaped cores in
northwestern North America and Northeast Asia. Asian Perspectives 19(1), 96106.
Munro, N. D., 2004. Zooarchaeological measures of hunting pressure and occupation
intensity in the Natufian: implications for agricultural origins. Current
Anthropology 45, S5-S33.
Myers, A., 1989. Reliable and maintainable technological strategies in the Mesolithic of
mainland Britain in: Torrence, R. (Ed.), Time, Energy and Stone Tools. New
Directions in Archaeology. Cambridge University Press, Cambridge, pp. 78-91.
Nelson, M., 1991. The study of technological organization. Archaeological Method and
Theory 3, 57-100.
Nelson, N. C., 1925. Diary of the Central Asiatic Expedition into Mongolia, Saturday,
April 18 to Wednesday, September 16. Department of Anthropology Archives,
American Museum of Natural History, New York City.
Nelson, N. C., 1926a. The dune dwellers of the Gobi. Natural History 26, 246-251.
Nelson, N. C., 1926b. Notes on the archaeology of the Gobi Desert. American
Anthropologist (New Series) 28, 305-308.
Nelson, N. C., 1939. Archaeology of Mongolia. Compte Rendue de la Deuxième
Session 1938, Congrès International des Sciences Anthropologiques et
Ethnologiques, Copenhague. Einar Munksgaard, Copenhague, pp. 259-262.
Nelson, S. M., 1993. The Archaeology of Korea. Cambridge World Archaeology Series.
Cambridge University Press, Cambridge.
500
Nelson, S. M. (Ed.), 1995. The Archaeology of Northeast China: Beyond the Great Wall.
Routledge, London.
Nicholas, G. P., 1998. Wetlands and hunter-gatherers: a global perspective. Current
Anthropology 39(5), 720-731.
Nobis, G., 1979. Der älteste Haushund lebte vor 14.000 jahren. UMSHAU 19, 215-225.
Norin, E., Montell, G. (Eds.), 1969. Sven Hedin Central Asia Atlas. Reports from the
Scientific Expedition to the North-western Provinces of China under the
Leadership of Sven Hedin, Sino-Swedish Expedition Publication, Publication 47.
Statens Etnografiska Museum, Stockholm.
Norin, E., 1978. “The Edsengol-Gurnai Basin” [map]. 1:500,000. In: Norin, E.,
Montell, G. (Eds.), 1969, Sven Hedin Central Asia Atlas. Reports from the
Scientific Expedition to the North-western Provinces of China under the
Leadership of Sven Hedin, Sino-Swedish Expedition Publication, Publication 47.
Statens Etnografiska Museum, Stockholm, NK 47: VII-XII, d-h.
Norton, C. J., Bae, K., Lee, H., Harris, J. W. K., 2007. A review of Korean microlithic
industries in: Kuzmin, Y. V., Keates, S. G., Shen, C. (Eds.), Origin and Spread of
Microblade Technology in Northern Asia and North America. Archaeology
Press, Simon Fraser University, Burnaby, pp. 91-102.
Novgorodova, E. A., 1989. Drevnyaya Mongoliya [Of Ancient Mongolia]. Nauka,
Moscow.
501
O’Connell, J. F., Hawkes, K., 1981. Alywara plant use and optimal foraging theory in:
Winterhalder, B., Smith, E. A. (Eds.), Hunter-Gatherer Foraging Strategies:
Ethnographic and Archaeological Analyses. University of Chicago Press,
Chicago, pp. 99-125.
O’Connell, J. F., Jones, K. T., Simms, S. R., 1982. Some thoughts on Prehistoric
archaeology in the Great Basin in: Madsen, D. B., O’Connell, J. F. (Eds.), Man
and Environment in the Great Basin. Society for American Archaeology Paper 2.
Society for American Archaeology, Washington, D. C., pp. 227-241.
O'Connell, J. F., 2006. How did modern humans displace Neanderthals? Insights from
hunter-gatherer ethnography and archaeology in: Conard, N. (Ed.), Neanderthals
and Modern Humans Meet? Kerns Verlag, Tübingen, pp. 43-64.
Oda, S., Keally, C. T., 1992. The origin and early development of axe-like and edgeground stone tools in the Japanese Palaeolithic. IPPA Bulletin 12, 23-31.
Okladnikov, A. P., Derevianko, A. P., 1970. Tamsa-bulagskaya neoliticheskaya kul’tura
vostochnoy Mongolii [Tamsagbulag Neolithic cultures of eastern Mongolia].
Materialy po istorii i filologii Tsentral’noy Azii 5, 3-20. (in Russian)
Okladnikov, A. P., 1951. Novye dannye po drevneishei istorii vnutrennei Mongolii [New
data on the ancient history of Inner Mongolia]. Vestnik Drevnei Istorii 4, 162174. (in Russian)
Okladnikov, A. P., 1962. Novoe v izuchenii drevneyshikh kul’tur Mongolii (po rabotam
1960 g.) [New in studies on the ancient culture of Mongolia – work to 1960].
Sovetskaya etnografiya 1, 83-90. (in Russian)
502
Okladnikov, A. P., 1981. Paleolit Tsentral’noi Azii (Palaeolithic of Central Asia).
Institute of Archaeology and Ethnography SB RAS Press, Novosibirsk. (in
Russian with English summary)
Okladnikov, A. P., 1986. Paleolit Mongolii (Palaeolithic of Mongolia). Institute of
Archaeology and Ethnography SB RAS Press, Novosibirsk. (in Russian)
Olsen, S. J., Olsen, J. W., 1977. The Chinese wolf, ancestor of New World dogs.
Science 197, 533-535.
Olsen, S. J., 1985. Origins of the Domestic Dog. University of Arizona Press, Tucson.
Olsen, S. J., 1988. The camel in ancient China and an osteology of the camel.
Proceedings of the Academy of Natural Sciences of Philadelphia 140(1), 18-58.
Olsen, S. J., 1990. Fossil ancestry of the yak, its cultural significance and domestication
in Tibet. Proceedings of the National Academy of Sciences Philadelphia 142, 73100.
Olsen, S. J., 1991. Confused yak taxonomy and evidence of domestication. Illinois State
Museum Scientific Papers 23, 387-393.
Olsen, S. L., 2003. The exploitation of horses at Botai, Kazakhstan in: Levin, M.,
Renfrew, C., Boyle, K. (Eds.), Prehistoric Steppe Adaptation and the Horse.
McDonald Institute for Archaeological Research, Cambridge, pp.83-103.
Olson K. A., Mueller T., Bolortsetsega S., Leimgruber P., Fagan W. F., Fuller T. K.,
2009. A mega-herd of more than 200,000 Mongolian gazelles Procapra
gutturosa: a consequence of habitat quality. Oryx 43, 149–153.
503
O’Malley, J. M., Kuzmin, Y. V., Burr, G. S., Donahue, D. J., Jull, A. J. T., 1998. Direct
radiocarbon accelerator mass spectrometric dating of the earliest pottery from the
Russian Far East and Transbaikal in: Groupe des Méthodes Pluridisciplinaires
Contribuant à l'Archéologie (Eds.), Actes du colloque “14C et Archéologie”.
Revue d'Archéométrie, Paris, pp.19-24.
Orton, J., 2008. Later Stone Age ostrich eggshell bead manufacture in the Northern
Cape, South Africa. Journal of Archaeological Science 35, 1765-1775.
Outram, A. K., Stear, N. A., Bendry, R., Olsen, S., Kasparov, A., Zaibert, V., Thorpe, N.,
Evershed, R. P., 2009. The earliest horse harnessing and milking. Science 323,
1332-1335.
Owen, D., 2007. An exercise in experimental archaeology on Chinese stone spades.
Indo-Pacific Prehistory Association Bulletin 27, 87-92.
Owen, L. A., Richards, B., Rhodes, E. J., Cunningham, W. D., Windley, B. F.,
Badamgarav, J., Dorjnamjaa, D., 1998. Relic permafrost structures in the Gobi of
Mongolia: age and significance. Journal of Quaternary Science 13(6), 539-547.
Owen, L. A., Windley, B. F., Cunningham, W. D., Badamgarav, J., Dorjnamjaa, D.,
1997. Quaternary alluvial fans in the Gobi of southern Mongolia: evidence for
neotectonics and climate change. Journal of Quaternary Science 12(3), 239-252.
Pachur, H., Wünnemann, B., and Hucai, Z., 1995. Lake evolution in the Tengger Desert,
Northwestern China, during the Last 40,000 Years. Quaternary Research 44, 171–
180.
504
Palmgren, N., 1934. Kansu mortuary urns of the Pan Shan and Ma Chang groups.
Palaeontologica Sinica Series D. 3(1).
Pang, J. –F., Kluetsch, C., Zou, X. –J., Zhang, A. –B., Luo, L. –Y., Angleby, H., Ardalan,
A., Ekström, C., Sköllermo, A., Lundeberg, J., Matsumura, S., Leitner, T., Zhang,
Y. –P., Savolainen, P., 2009. MtDNA data indicate a single origin for dogs south
of Yangtze River, less than 16,300 years ago, from numerous wolves. Molecular
Biology and Evolution 26(12), 2849-2864.
Pan, X., Chao, J., 2003. Theory of stability, and regulation and control of ecological
system in oasis. Global and Planetary Change 37, 287-295.
Pankova, E. I., 2008. Environmental conditions and soils of natural oases in the Alashan
Gobi Desert, Mongolia. Eurasian Soil Science 41(8), 827-836.
Parry, W. J., Kelly, R. L., 1987. Expedient core technology and sedentism in: Johnson, J.
K., Morrow, C. A. (Eds.), The Organization of Core Technology. Westview
Press, Boulder, pp. 285-304.
Paterson, A., 2005. Hunter-gatherer interactions with sheep and cattle pastoralists from
the Australian arid zone in: Veth, P., Smith, M., Hiscock, P. (Eds.), Desert
Peoples: Archaeological Perspectives. Blackwell Publishing, Oxford, pp. 276292.
Pavlů, I., 1997. Pottery Origins. Initial Forms, Cultural Behavior and Decorative Styles.
Vydavatelství University Karlovy, Prague.
Peng, K., Zhu, Y., 1995. New research on the origin of cowries in ancient China. SinoPlatonic Papers 68.
505
Peng, Y., Xiao, J., Nakamura, T., Liu, B., Inouchi, Y., 2005. Holocene East Asian
monsoonal precipitation pattern revealed by grain-size distribution of core
sediments of Daihai Lake in Inner Mongolia of north-central China. Earth and
Planetary Science Letters 233, 467-479.
Peters, J., von den Driesch, A., 1997. The two-humped camel (Camelus bactrianus): new
light on its distribution, management and medical treatment in the past. Journal
of Zoology, London 242, 651-679.
Pinson, A. O., 1999. Foraging in Uncertain Times: The Effect of Risk on Subsistence
Behavior during the Pleistocene-Holocene Transition in the Oregon Great Basin.
Unpublished Ph.D. dissertation. University of New Mexico, Albuquerque.
Pond, A. W., 1928. Gobi Diary. Micro 1178, Reel 2. Wisconsin Historical Society,
Madison.
Pond, A. W., N.d. Unpublished Manuscript summarizing archaeological sites from 1928
Central Asiatic Expedition. 47 pages. Division of Anthropology, American
Museum of Natural History, New York. Note: original copy is missing, only
remaining copy is contained in W. A. Fairservis’ archival material, which is
housed at the Peabody Museum, Harvard University.
Potts, D. T., 2001. Ostrich distribution and exploitation in the Arabian Peninsula.
Antiquity 75, 182-190.
Potts, D. T., 2004. Camel hybridization and the role of Camelus bactrianus in the ancient
Near East. Journal of the Economic and Social History of the Orient 47(2), 143165.
506
Pratt, J. A. F., 1999. Determining the function of one of the New World’s earliest pottery
assemblages: the case of San Jacinto, Colombia. Latin American Antiquity 10(1),
71-85.
Rassamakin, Y., 1999. The Eneolithic of the Black Sea steppe: Dynamics of culture and
economic development 4500-2300 BCE in: Levine, M., Rassamakin, Y.,
Kiselenko, A., Kiselenko, T. N. (Eds.), Late Prehistoric Exploitation of the
Eurasian Steppe. McDonald Institute for Archaeological Research, Cambridge,
pp. 59-182.
Rea, D. K., Leinen, M., 1988. Asian aridity and the zonal westerlies: late Pleistocene and
Holocene record of eolian deposition in the northwest Pacific Ocean.
Palaeogeography, Palaeoclimatology, Palaeoecology 66, 1-8.
Redding, R., 1988. A general explanation of subsistence change: from hunting and
gathering to food production. Journal of Anthropological Archaeology 7, 56-97.
Reher, C. A., 1991. Large scale lithic quarries and regional transport systems on the
High Plains of Eastern Wyoming: Spanish Diggings revisited in: Montet-White,
A., Holen, S. (Eds.), Raw Material Economies among Prehistoric HunterGatherers. University of Kansas Publications in Anthropology 19. University of
Kansas, Lawrence, pp. 375-398.
Renfrew, C., Bahn, P., 1996. Archaeology: Theories, Methods, and Practice. Second
Edition. Thames and Hudson, London.
507
Revedin, A., Aranguren, B., Becattini, R., Longo, L., Marconi, E., Lippi, M. M., Skakun,
N., Sinitsyn, A., Spiridonova, E., Svoboda, J., 2010. Thirty thousand-year-old
evidence of plant food processing. PNAS 107(44), 18815-18819.
Rhode, D., Madsen, D. B., Brantingham, P. J., Dargye, Ts., 2007. Yaks, yak dung, and
prehistoric human habitation of the Tibetan Plateau in: Madsen, D. B., Chen, F.
H., Gao, X. (Eds.), Late Quaternary Climate Change and Human Adaptation in
Arid China, Developments in Quaternary Science, Vol. 9. Elsevier, Amsterdam,
pp. 205-224.
Rice, P. M., 1999. On the origins of pottery. Journal of Archaeological Method and
Theory 6(1), 1-54.
Richerson, P. J., Boyd, R., Bettinger, R. L., 2001. Was agriculture possible during the
Pleistocene but mandatory during the Holocene? A climate change hypothesis.
American Antiquity 66(3), 387-411.
Roberts, L. B., Butler, F. B., Robinson, H. G., N.d. Mongolia Route Maps: Covering
Traverse of the Expedition from Kalgan Westward, 1925. American Museum of
Natural History, New York.
Rowley-Conway, P., 1984. Postglacial foraging and early farming economies in Japan
and Korea: a west European perspective. World Archaeology 16, 28-42.
Rudaya, N. A., Tarasov, P. E., Dorofeyuk, N. I., Kalugin, I. A., Andreev, A. A.,
Diekmann, B., Daryin, A. V., 2008. Environmental changes in the Mongolian
Altai during the Holocene. Archaeology, Ethnology and Anthropology of Eurasia
36(4), 2-14.
508
Sablin, M. V., Khlopachev, G. A., 2002. The earliest Ice Age dogs: evidence from
Eliseevichi I. Current Anthropology 43, 795-799.
Sadr, K., 2005. Hunter-gatherers and herders of the Kalahari during the late Holocene in:
Veth, P., Smith, M., Hiscock, P. (Eds.), Desert Peoples: Archaeological
Perspectives. Blackwell Publishing, Oxford, pp. 206-221.
Sampson, C. G., Bailiff, I., Barnett, S., 1997. Thermoluminescence dates from Later
Stone Age Pottery on Surface Sites in the Upper Karoo. The South African
Archaeological Bulletin, 52(165), 38-42.
Sandelowsky, B. H., 1971. Ostrich egg-shell caches from south-west Africa. South
African Archaeological Bulletin 37, 57-62.
Sano, K., 2007. Emergence and mobility of microblade industries in the Japanese Islands
in: Kuzmin, Y. V., Keates, S. G., Shen, C. (Eds.), Origin and Spread of
Microblade Technology in Northern Asia and North America. Archaeology
Press, Simon Fraser University, Burnaby, pp.79-90.
Sassaman, K. E., 1993. Early Pottery in the Southeast: Tradition and Innovation in
Cooking Technology. University of Alabama Press, Tuscaloosa.
Sato, H., Tsutsumi, T., 2007. The Japanese microblade industries: technology, raw
material procurement, and adaptations in: Kuzmin, Y. V., Keates, S. G., Shen, C.
(Eds.), Origin and Spread of Microblade Technology in Northern Asia and North
America. Archaeology Press, Simon Fraser University, Burnaby, pp. 53-78.
Savolainen, P., Zhang, Y. –P., Luo, J., Lundeberg, J., Leitner, T., 2002. Genetic evidence
for an East Asian origin of domestic dogs. Science 298, 1610-1613.
509
Schiffer, M. B., 1972. Archaeological context and systemic context. American Antiquity
37(2), 156-165.
Schiffer, M. B., 1975. An alternative to Morse’s Dalton settlement pattern hypothesis.
Plains Anthropologist 20(70), 253-266.
Schiffer, M B., 1987. Formation Processes of the Archaeological Record. 1996 edition.
University of Utah Press, Salt Lake City.
Schiffer, M. B., 1992. Technological Perspectives on Behavioral Change. University of
Arizona Press, Tucson.
Schiffer, M. B., 2001. The explanation of long-term technological change in: Schiffer,
M. B. (Ed.), Anthropological Perspectives on Technology. University of New
Mexico Press, Albuquerque, pp. 215-235.
Schiffer, M. B., 2005. The electric lighthouse in the nineteenth century: aid to navigation
and political technology. Technology and Culture 46(2), 275-305.
Sealy, J., Yates, R., 1994. The chronology of the introduction of pastoralism to the Cape,
South Africa. Antiquity 68, 58-67.
Séfériadès, M. L., 2006. Ancient hunter-gatherers, first sedentary farmers and nomad
stock herders of Mongolia (8000-3000 BC): new researches at Tamsagbulag
(Dornod aimak) in: Le Secrétariat du Congrès (Ed.), Préhistoire de l’Asie et de
l’Océanie (Asian and Oceanic Prehistory). Actes du XIVèm Congrès UISPP,
Université de Liège, Belgium, 2-8 September 2001, Section 16. BAR
International Series 1523.
510
Seong, C., 1998. Microblade technology in Korea and adjacent Northeast Asia. Asian
Perspectives 37(2), 245-278.
Seong, C., 2007. Late Pleistocene microlithic assemblages in Korea in: Kuzmin, Y. V.,
Keates, S. G., Shen, C. (Eds.), Origin and Spread of Microblade Technology in
Northern Asia and North America. Archaeology Press, Simon Fraser University,
Burnaby, pp. 103-114.
Seong, C., 2008. Tanged points, microblades, and Late Palaeolithic hunting in Korea.
Antiquity 82, 871-883.
Shahack-Gross, R., Marshall, F., Ryan, K., Weiner, S., 2004. Reconstruction of spatial
organization in abandoned Maasai settlements: implications for site structure in
the Pastoral Neolithic of East Africa. Journal of Archaeological Science 31,
1395-1411.
Shao, W., 2000. The Longshan period and incipient Chinese civilization. Journal of East
Asian Archaeology 2(1-2), 195-226.
Shelach, G., 2000. The earliest Neolithic cultures of Northeast China: recent discoveries
and new perspectives on the beginning of agriculture. Journal of World
Prehistory 14(4), 363-413.
Sherratt, A. G., 1983. The secondary exploitation of animals in the Old World. World
Archaeology 15, 90-104.
Shigehara, N., Hongo, H., 2000. Ancient remains of Jomon dogs from Neolithic sites in
Japan in: Crockford, S. J. (Ed.), Dogs through Time: An Archaeological
Perspective. BAR International Series 889. Archaeopress, Oxford, pp. 61-67.
511
Shigehara, N., S., Qi, G., Komiya, H., Yuan, J., 1998. Morphological study of the
ancient dogs from three Neolithic sites in China. International Journal of
Osteoarchaeology 8(1), 11-22.
Shi, P., Song, C., 2003. Palynological records of environmental changes in the middle
part of Inner Mongolia, China. Chinese Science Bulletin 48(14), 1433-1438.
Shi, Y., Yu, G., Liu, X., Li, B., Yao, T., 2001. Reconstruction of the 30-40 ka BP
enhanced Indian monsoons climate based on geological records from the Tibetan
Plateau. Palaeogeography, Palaeoclimatology, Palaeoecology 169, 69-83.
Shott, M. J., 1986. Technological organization and settlement mobility: an ethnographic
examination. Journal of Anthropological Research 42(1), 15-51.
Shott, M. J., 1996. An exegesis of the curation concept. Journal of Anthropological
Research 52(3), 259-280.
Simms, S. R., 1987. Behavioral Ecology and Hunter-Gatherer Foraging: An Example
from the Great Basin. BAR International Series 381. British Archaeological
Reports, Oxford.
Slobodin, S., 1999. Northeast Asia in the late Pleistocene and early Holocene. World
Archaeology, 30(3), 484-502.
Soffer, O., 2000. Gravettian technologies in social contexts in: Roebroeks, W., Mussi,
M., Svoboda, J., Fennema, K. (Eds.), Hunters of the Golden Age: the Mid Upper
Palaeolithic of Eurasia, 30,000-20,000 BP. University of Leiden, Leiden, pp. 5969.
512
Sommerström, B., 1956. Archaeological Researches in the Edsen-Gol Region Inner
Mongolia, Part I. Reports from the Scientific Expedition to the North-Western
Provinces of China under the Leadership of Dr. Sven Hedin, Publication 39.
Statens Etnografiska Museum, Stockholm.
Speth, J. D., Spielmann, K. A., 1983. Energy source, protein metabolism, and huntergatherer subsistence strategies. Journal of Anthropological Archaeology 2, 1-31.
Spielmann, K. A., Eder, J. F., 1994. Hunters and farmers: then and now. Annual Review
of Anthropology 23, 303-323.
Starkel, L., 1998. Geomorphic response to climatic and environmental changes along a
Central Asian transect during the Holocene. Geomorphology 23, 293-305.
Starkovich, B. M., Stiner, M. C., 2009. Hallan Çemi Tepesi: high-ranked game
exploitation alongside intensive seed processing at the Epipalaeolithic-Neolithic
transition in southeastern Turkey. Antropozoologica 44(1), 41-61.
Stiner, M. C., Munro, N. D., 2002. Approaches to prehistoric diet breadth, demography,
and prey ranking systems in time and space. Journal of Archaeological Method
and Theory 9(2), 181-214.
Stiner, M. C., 2001. Thirty years on the “Broad Spectrum Revolution” and Paleolithic
demography. PNAS 98(13), 6993-6996.
Stiner, M. C., 2005. The Faunas of Hayonim Cave, Israel: A 200,000-Year Record of
Paleolithic Diet, Demography, and Society. American School of Prehistoric
Research Bulletin 48. Peabody Museum of Archaeology and Ethnology, Harvard
University, Cambridge.
513
Stiner, M C., Munro, N. D., Surovell, T. A., 2000. The tortoise and the hare: small game
use, the Broad Spectrum Revolution, and Paleolithic demography. Current
Anthropology 41(1), 39-73.
Stiner, M. C., Munro, N. D., Surovell, T. A., Tchernov, E., Bar-Yosef, O., 1999.
Paleolithic population growth pulses evidenced by small animal exploitation.
Science 283, 190-194.
Surovell, T. A., 2003. The Behavioral Ecology of Folsom Lithic Technology.
Unpublished Ph.D. Dissertation. University of Arizona, Tucson.
Svyatko, S. V., Mallory, J. P., Murphy, E. M., Polyakov, A. V., Reimer, P. J., Schulting,
R. J., 2009. New radiocarbon dates and a review of the chronology of prehistoric
populations from the Minusinsk Basin, southern Siberia, Russia. Radiocarbon
51(1), 243-252.
Tabarev, A. V., 1997. Paleolithic wedge-shaped microcores and experiments with pocket
devices. Lithic Technology 22(2), 139-149.
Tang, C., Gai, P., 1986. Upper Palaeolithic cultural traditions in North China in:
Wendorf, F., Close, A. E. (Eds.), Advances in World Archaeology 5. Academic
Press, New York, pp. 339-364.
Tang, C., 2000. The Upper Palaeolithic of North China: the Xiachuan culture. Journal of
East Asian Archaeology 2 (1-2), 37-49.
Tan, Y. –J., Sun, X. –R., Zhao, H. –G., Gan, Z. –G., 1995a. The Neolithic in
Heilongjiang province in: Nelson, S. M. (Ed.), The Archaeology of Northeast
China. Routledge, London, pp. 118-144.
514
Tan, Y. –J., Sun, X. –R., Zhao, H. –G., Gan, Z. –G., 1995b. The Bronze Age of the
Song-nen Plain in: Nelson, S. M. (Ed.), The Archaeology of Northeast China.
Routledge, London, pp. 225-250.
Tarasov, P. E., Dorofeyuk, I. N., Sevastyanov, D. V., Nakagawa, T., 2000. Holocene
vegetation and climate changes in Mongolia derived from lake status, pollen and
plant macrofossil records. Berliner geowissensschaftliche abhandlungen, Reihe A
205, 94-99.
Taylor, W. W., 1964. Tethered nomadism and water territoriality: an hypothesis in:
Actas y Memorias 35th Congreso Internacional Instituto de Americanistas.
Instituto Nacional de Antropologia e Historia, Mexico City, pp. 197-203.
Teilhard de Chardin, P., Pei, W.-C., 1944. Le Neolithique de la Chine. Peking.
Torrence, R., 1983. Time budgeting and hunter-gatherer technology in: Bailey, G. (Ed.),
Hunter-gatherer Economy in Prehistory. Cambridge University Press,
Cambridge, pp. 11-22.
Torrence, R., 1989. Retooling: towards a behavioral theory of stone tools in: Torrence,
R. (Ed.), Time, Energy and Stone Tools. New Directions in Archaeology.
Cambridge University Press, Cambridge, pp. 57-66.
Trigger, B. G., 1989. A History of Archaeological Thought. Cambridge University
Press, Cambridge.
Tseveendorj, D., Khosbayar, P., 1982. “Dulaani gobiig” arkeologi geologiin talaar
dakhin sudalsan n’. Studia Archaeologica Instituti Historiae Academiae
Scientiarum Republicae Populi Mongolici 10(3), 22-31. (in Mongolian)
515
Underhill, A. P., Habu, J., 2006. Early communities in East Asia: economic and
sociopolitical organization at the local and regional levels in: Stark, M. (Ed.),
Archaeology of Asia. Blackwell Publishing, Oxford, pp. 121-148.
Underhill, A. P., 1997. Current issues in Chinese Neolithic archaeology. Journal of
World Prehistory 11, 103-160.
Vasil’ev, S. A., Kuzmin, Y. V., Orlova, L. A., Dementiev, V. N., 2002. Radiocarbonbased chronology of the Paleolithic of Siberia and its relevance to the peopling of
the New World. Radiocarbon 44 (2), 503–530.
Vasil’ev, S. A., Semenov, V. A., 1993. Prehistory of the Upper Yensei area (southern
Siberia). Journal of World Prehistory 7(2), 213-242.
Vilà, C., Leonard, J. A., Götherström, A., Marklund, S., Sandberg, K., Lidén, K., Wayne,
R. K., Ellegren, H., 2001. Widespread origins of domestic horse lineages.
Science 291, 474-477.
Vilà, C., Savolainen, P., Maldonado, J. E., Amorim, I. R., Rice, J. E., Honeycutt, R. L.,
Crandall, K. A., Lundeberg, J., Wayne, R. K., 1997. Multiple and ancient origins
of the domestic dog. Science 276, 1687-1689.
Vogel, J. C., Visser, E., Fuls, A., 2001. Suitability of ostrich eggshell for radiocarbon
dating. Radiocarbon 43, 133-137.
Volkov, V. V., 1967. Bronzovy i ranniy zhelezny veka Severnoy Mongolii [Bronze and
early Iron Age of Northern Mongolia]. Izvestiya Akademii Nauk Mongol’skoy
Napodnoy Respubliki Institut Istorii, Ulaanbaatar. (in Russian)
516
Volkov, V. V., 1981. Olennye kamni Mongolii [Deer stones of Mongolia]. Izvestiya
Akademii Nauk Mongol’skoy Napodnoy Respubliki Institut Istorii, Ulaanbaatar.
(in Russian)
von Holdt, B. M., Pollinger, J. P., Lohmueller, K. E., Han, E., Parker, H. G., Quignon, P.,
Degenhardt, J. D., Boyko, A. R., Early, D. A., Auton, A., Reynolds. A., Bryc, K.,
Brisbin, A., Knowles, J. C., Mosher, D. S., Spady, T. C., Elkahloun, A., Geffen,
E., Pilot, M., Jedrzejewski, W., Greco, C., Randi, E., Bannasch, D., Wilton, A.,
Shearman, J., Musiani, M., Cargill, M., Jones, P. G., Qian, Z., Huang, W., Ding,
Z. –L., Zhang, Y. –P., Bustamante, C. D., Ostrander, E. A., Novembre, J., and
Wayne, R. K., 2010. Genome-wide SNP and haplotype analyses reveal a rich
history underlying dog domestication. Nature 464, 898-903.
von Wehrden, H., Zimmerman, H., Hanspach, J., Ronnenberg, K., Wesche, K., 2009.
Predictive mapping of plant species and communities using GIS and Landsat data
in a southern Mongolian mountain range. Folia Geobot 44, 211-225.
Vostretsov, Y. E., Sergeisheva, E. A., Masayaki, K., Miyamoto, K., Obata, H., 2003.
Novye dannye o rannem zemledelii v Primorye: Neoliticheskii kompleks
posledeniya Kpounovka-1 [New data on early stage of agriculture in Primorye:
Neolithic complex of the Krounovka 1 settlement]. Problemy arkeologii i
paleoekologii Severnoi, Vostochnoi i Tsentralnoi Azii. Institute of Archaeology
and Ethnography, Novosibirsk, pp. 373-377. (in Russian)
517
Wa, Y., 1992. Neolithic tradition in Northeast China in: Aikens, C. M., Rhee, S. N.
(Eds.), Pacific Northeast Asia in Prehistory: Hunter-Fisher-Gatherers, Farmers
and Sociopolitical Elites. Washington State University Press, Washington, pp.
139-156.
Wallace, I. J., Shea, J. J., 2006. Mobility patterns and core technologies in the Middle
Palaeolithic of the Levant. Journal of Archaeological Science 33, 1293-1309.
Wang, H., Liu, H., Cui, H., Abrahamsen, N., 2001. Terminal Pleistocene/Holocene
palaeoenvironmental changes revealed by mineral-magnetism measurements of
lake sediments for Dali Nor area, southeastern Inner Mongolia Plateau, China.
Palaeogeography, Palaeoclimatology, Palaeoecology 170, 115-132.
Wang, H., Liu, H., Zhu, J., Yin, Y., 2010. Holocene environmental changes as recorded
by mineral magnetism of sediments from Anguli-nuur Lake, southeastern Inner
Mongolia Plateau, China. Palaeogeography, Palaeoclimatology, Palaeoecology
285, 30-49.
Wang, S., Foote, W. C., Bunch, T. D., 1990. Genetic variability in domesticated and
wild sheep based on blood protein characters. Comparative Biochemistry and
Physiology Part B: Comparative Biochemistry 96(1), 201-207.
Wang, X., Guo, H., Chang, Y., Zha, L., 2004. On paleodrainage evolution in mid-late
Epipleistocene based on radar remote sensing in northeastern Ejin Banner, Inner
Mongolia. Journal of Geographic Sciences 14(2), 235-241.
518
Watanobe, T., Okumura, N., Ishiguro, N., Nakano, M., Matsui, A., Sahara, M., Komatsu,
M., 1999. Genetic relationship and distribution of the Japanese wild boar (Sus
scrofa leucomystax) and Ryukyu wild boar (Sus scrofa riukiuanus) analysed by
mitochondrial DNA. Molecular Ecology 8, 1509-1512.
Weber, A.W., 1995. The Neolithic and early Bronze Age of the Lake Baikal region: a
review of recent research. Journal of World Prehistory 9(1), 99-165.
Weber, A. W., Bettinger, R., 2010. Middle Holocene hunter-gatherers of Cis-Baikal,
Siberia: an overview for the new century. Journal of Anthropological
Archaeology 29, 491-506.
Weber, A. W., Link, D. W., Goriunova, O. I., Konopatskii, A. K., 1998. Patterns of
prehistoric procurement of seal at Lake Baikal: a zooarchaeological contribution
to the study of past foraging economies in Siberia. Journal of Archaeological
Science 25, 215-227.
Weber, A.W., Katzenberg, M. A., Schurr, T. G. (Eds.), 2010. Prehistoric HunterGatherers of the Baikal Region, Siberia: Bioarchaeological Studies of Past Ways
of Life. University of Pennsylvania Museum of Archaeology and Anthropology,
Philadelphia.
Wesche, K., Walther, D., von Wehrden, H., Hensen, I., 2011. Trees in the desert:
reproduction and genetic structure of fragmented Ulmus pumila forests in
Mongolian drylands. Flora 206: 91-99.
West, B., Zhou, B. –X., 1988. Did chickens go north? New evidence for domestication.
Journal of Archaeological Science 15, 515-533.
519
Winkler, M. G., Wang, P. K., 1993. The late-Quaternary vegetation and climate change
of China in: Wright, H. E. Jr., Kutzbach, J. E., Webb, T. III, Ruddiman, W. F.,
Street-Perrott, F. A., Bartlein, P. J. (Eds.), Global Climates since the Last Glacial
Maximum. Minnesota Press, Minneapolis, pp. 221-261.
Winterhalder, B., 1981. Foraging strategies in the boreal forest: an analysis of Cree
hunting and gathering in: Winterhalder, B., Smith, E. A. (Eds.), Hunter-Gatherer
Foraging Strategies: Ethnographic and Archaeological Analyses. University of
Chicago Press, Chicago, pp. 66-98.
Winterhalder, B., Goland, C., 1993. On population, foraging efficiency, and plant
domestication. Current Anthropology 34, 710-715.
Winterhalder, B., Smith, E. A., 2000. Analyzing adaptive strategies: human behavioral
ecology at twenty-five. Evolutionary Anthropology 9(2), 51-72.
Wood, N. J., Phua, S. H., 1996. Variation in the control region sequence of the sheep
mitochondrial genome. Animal Genetics 27, 25-33.
Wright, J. S.-C., 2006. The adoption of pastoralism in Northeast Asia: monumental
transformation in the Egiin Gol Valley, Mongolia. Unpublished Ph.D.
Dissertation. Harvard University, Cambridge.
Wright, K. I., 1994. Ground-stone tools and hunter-gatherer subsistence in southwest
Asia: implications for the transition to farming. American Antiquity 59(2), 238263.
520
Wünnemann, B., Hartmann, K., Janssen, M., Hucai, C.Z., 2007. Responses of Chinese
desert lakes to climate instability during the last 45,000 years in: Madsen, D.B.,
Chen, F.-H., Xing, G. (Eds.), Late Quaternary Climate Change and Human
Adaptation in Arid China. Elsevier, Amsterdam, pp. 11–24.
Wünnemann, B., Pachur, H. J., Zang, H., 1998a. Climatic and environmental changes in
the deserts of Inner Mongolia, China, since the Late Pleistocene in: Alsharhan, A.
S., Glennie, K. W., Whittle, G. L., Kendall, C. G.St.C. (Eds.), Quaternary Deserts
and Climate Change. Balkema, Rotterdam, pp. 381-394.
Wünnemann, B., Pachur, H. J., Li, J., Zang, H., 1998b. Chronologie der pleistozänen und
holozänen Seespiegelschwankungen des Gaxun Nur/Sogo Nur und Baijian Hu,
Innere Mongolei, Nordwestchina. Petermanns Geographische Mitteilungen 142:
191-206. (in German)
Wu, Z.-Y., Raven, P., H. (Eds.), 1999. Flora of China. Volume 4. Missouri Botanical
Garden, St. Louis. http://www.efloras.org
Wu, X., Zhao, C., 2003. Chronology of the transition from Palaeolithic to Neolithic in
China. The Review of Archaeology 24(2), 15-20.
Xia, Z., Chen, F., Chen, G., Zheng, G., Xie, F., Mei, H., 2001. Environmental
background of evolution from the Paleolithic to Neolithic culture in Nihewan
Basin, North China. Science in China (Series D) 44(9), 779-788.
Xu, Q., Li, Y., Tian, F., Cao, X., Yang, X., 2009. Pollen assemblages of tauber traps and
surface soil samples in steppe areas of China and their relationships with
vegetation and climate. Review of Palaeobotany and Palynology 153, 86-101.
521
Xu, Y. –L., 1995. The Houwa site and related issues in: Nelson, S. M. (Ed.), The
Archaeology of Northeast China: Beyond the Great Wall. Routledge, London,
pp. 65-88.
Yamazaki, K., Takahashi, O., Sugawara, H., Ishiguro, N., Endo, H., 2005. Wild boar
remains from the Neolithic (Jomon Period) sites on the Izu Islands and in
Hokkaido Island, Japan in: Vigne, J. –D., Peters, J., Helmer, D. (Eds.), The First
Steps of Animal Domestication. Proceedings of the 9th ICAZ Conference,
Durham 2002. Oxbow Books, Oxford, pp. 160-176.
Yang, J., 1998. Siba: Bronze Age culture of the Gansu Corridor. Sino-Platonic Papers
86, 1-18.
Yang, X., Liu, T., Xiao, H., 2003. Evolution of megadunes and lakes in the Badain Jaran
Desert, Inner Mongolia, China during the last 31,000 years. Quaternary
International 104, 99-112.
Yang, X., Rost, K. T., Lehmkuhl, F., Zhenda, Z., Dodson, J., 2004. The evolution of dry
lands in northern China and in the Republic of Mongolia since the Last Glacial
Maximum. Quaternary International 118-119, 69-85.
Yang, X., 2006. Chemistry and late Quaternary evolution of ground and surface waters
in the area of Yabulai Mountains, western Inner Mongolia, China. Catena 66(12), 135-144.
Yang, X., Williams, M. A. J., 2003. The ion chemistry of lakes and late Holocene
desiccation in the Badain Jaran Desert, Inner Mongolia. Catena 51, 45-60.
522
Yang, X., Zhu, B., Wang, X., Li, C., Zhou, Z., Chen, J., Wang, X., Yin, J., Lu, Y., 2008.
Late Quaternary environmental changes and organic carbon density in the
Hunshandake Sandy Land, eastern Inner Mongolia, China. Global and Planetary
Change 61, 70-78.
Yasuda, Y., 2002. Origins of pottery and agriculture in East Asia in: Yasuda, Y. (Ed.),
The Origins of Pottery and Agriculture. Roli Books, New Delhi, pp. 151-156.
Yellen, J., Harpending, H., 1972. Hunter-gatherer populations and archaeological
inference. World Archaeology 4(2), 244–253.
Yellen, J. E., 1977. Long term hunter-gatherer adaptation to desert environments: a
biogeographical perspective. World Archaeology 8(3), 262-274.
Yerkes, R. W., Barkai, R., Gopher, A., Bar Yosef, O., 2003. Microwear analysis of early
Neolithic (PPNA) axes and bifacial tools from Netiv Hagdud in the Jordan
Valley, Israel. Journal of Archaeological Science 30, 1051-1066.
Yoshida, K., Ohmichi, J., Kinose, M., Iijima, H., Oono, A., Abe, N., Miyazaki, Y.,
Matsuzaki, H., 2004. The application of 14C dating to potsherds of the Jomon
period. Nuclear Instruments and Methods in Physics Research B 223-224, 716722.
Yuan, J., Flad, R. K., 2002. Pig domestication in ancient China. Antiquity 76, 724-732.
Yuan, J., Flad, R. K., 2003. Two issues concerning ancient domesticated horses in
China. Bulletin of the Museum of Far Eastern Antiquities 75, 110-126.
523
Yuan, J., Flad, R. K., 2005. Research on early horse domestication in China in:
Mashkour, M. (Ed.), Equids in Time and Space. Oxbow Books, Oxford, pp. 124131.
Yuichiro, K., 2005. The temporal correspondence between archaeological chronology
and environmental changes in the Final Pleistocene in Eastern Honshu Island.
Daiyonki Kenkyu 44, 51-64. (in Japanese with English summary)
Yu, G., Chen, X., Ni, J., Cheddadi, R., Guiot, J., Han, H., Harrison, S.P., Huang, C., Ke,
M., Kong, Z., Li, S., Li,W., Liew, P., Liu, G., Liu, J., Liu, Q., Liu, K.-B., Prentice,
C., Qui,W., Ren, G., Song, C., Sugita, S., Tang, L., Van Campo, E., Xia, Y., Xu,
Q., Yan, S., Yang, X., Zhao, J., Zheng, Z., 2000. Palaeovegetation of China: a
pollen data based synthesis for the mid-Holocene and last glacial maximum.
Journal of Biogeography 27, 635–664.
Zeder, M. A., 2001. A metrical analysis of a collection of modern goats (Capra hircus
aegargus and C. h. hircus) from Iran and Iraq: implications for the study of
caprine domestication. Journal of Archaeological Science 28, 61-79.
Zeder, M. A., 2005. A view from the Zagros: new perspectives on livestock
domestication in the Fertile Crescent in: Vigne, J. –D., Peters, J., Helmer, D.
(Eds.), The First Steps of Animal Domestication. Proceedings of the 9th ICAZ
Conference, Durham 2002. Oxbow Books, Oxford, pp. 125-146.
Zhang, H. C., Ma, Y. Z., Wünnemann, B., Pachur, H.-J., 2000. A Holocene climatic
record from arid northwestern China. Palaeogeography, Palaeoclimatology,
Palaeoecology 162, 389-401.
524
Zhang, S., 2000. The Epipalaeolithic in China. Journal of East Asian Archaeology 2(12), 51-66.
Zhao, Y., Xu, Q., Huang, X., Guo, X., Tao, S., 2009. Differences of modern pollen
assemblages from lake sediments and surface soils in arid and semi-arid China
and their significance for pollen-based quantitative climate reconstruction.
Review of Palaeobotany and Palynology 156, 519-524.
Zhao, Y., Yu, Z., Chen, F., 2009. Spatial and temporal patterns of Holocene vegetation
and climate changes in arid and semi-arid China. Quaternary International 194, 618.
Zhao, Y., Yu, Z., Chen, F., Li, J., 2008. Holocene vegetation and climate change from a
lake sediment record in the Tengger Sandy Desert, northwest China. Journal of
Arid Environments 72, 2054-2064.
Zvelebil, M., 1996. The agricultural frontier and the transition to farming in the circumBaltic region in: Harris, D. R. (Ed.), The Origins and Spread of Agriculture and
Pastoralism in Eurasia. University College London Press, London, pp. 323-345.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement