Content uploaded by Kathleen Kuman
Author content
All content in this area was uploaded by Kathleen Kuman on Nov 01, 2017
Content may be subject to copyright.
The oldest Hoabinhian technocomplex in Asia (43.5 ka) at Xiaodong
rockshelter, Yunnan Province, southwest China
Xueping Ji
a
,
b
,
**
, Kathleen Kuman
c
,
d
,
*
, R.J. Clarke
d
, Hubert Forestier
e
, Yinghua Li
f
,
Juan Ma
g
, Kaiwei Qiu
g
, Hao Li
b
,
c
,YunWu
a
a
Yunnan Institute of Cultural Relics and Archaeology, Kunming, Yunnan Province, China
b
Institute for Vertebrate Paleontology and Paleoanthropology, Chinese Academy of Sciences, Beijing, China
c
School of Geography, Archaeology and Environmental Studies, University of the Witwatersrand, Johannesburg, South Africa
d
Evolutionary Studies Institute, University of the Witwatersrand, Johannesburg, South Africa
e
Mus
eum National d'Histoire Naturelle, UMR 7194 CNRS-MNHN-UPVD, Institut de Pal
eontologie Humaine, Paris, France
f
Department Archaeology, College of Humanities, Wuhan University, Wuhan, China
g
Lincang Institute Cultural Relics, Linxiang, Yunnan Province, China
article info
Article history:
Available online xxx
Keywords:
Hoabinhian
Pebble tools
Uniface
Southwest China
Yunnan Province
Southeast Asia
abstract
The Hoabinhian is the most representative technocomplex in Southeast Asian prehistory for the later
hunteregatherer period. As a mainland technology based exclusively on seasonal tropical environments,
this core-tool culture was previously defined in northern Vietnam in 1932 and characterized originally by
its large, flat and long, largely unifacial cobble tools associated with tropical forest fauna. The recent
discoveries and dates obtained at Xiaodong rockshelter in Yunnan Province (southwest China) allow us
to discuss the origin and the homeland of this singular Asian technocomplex which spread to Southeast
Asia during the end of the Late Upper Pleistocene. Here we present the first Chinese Hoabinhian lithic
implements in their stratigraphic and chronological context within a rockshelter site, and we address the
question of the dispersal of modern humans from South China to Southeast Asia.
©2015 Elsevier Ltd and INQUA.
1. Introduction
Xiaodong is a very large rockshelter in Yunnan Province, south-
west China, 40 km from the border with Burma (Fig. 1). It is situated
near the village of Nongke, Mengsheng town (Cangyuan County,
Lincang Prefecture, Yunnan Province) at an elevation of 1195 m
above sea level and lies within a small valley 15 m above the Hemeng
River. The rockshelter consists of a large overhang within Permian
limestone and is approximately 60 45 m large (Figs. 2 and 3).
The name Xiaodong translates as “saltpetre cave”. This mineral
was dug from the site by the local villagers from the 1950s and used
for explosive materials and in the iron and steel manufacturing
process. Bat guano is also present in the deposits and was dug for
fertiliser. The site was discovered during a survey for cultural relics
in 1981 by a joint team from the Yunnan Provincial Museum and
the Lincang Institute of Cultural Relics when artefacts from the
disturbed deposits were first collected. In 2004, Xueping Ji visited
the site again, and with a larger sample of artefacts, he recognized
the importance of the typological and technological information in
the collection, which had not previously been reported (Zhang,
1991; Ji et al., 2006). In January 2007, Ji visited the site again with
Kathleen Kuman and Ronald John Clarke and collected more arte-
facts and some fossils from the surface and disturbed deposits. A
trench was dug for this visit by the Lincang Institute of Cultural
Relics to determine if in situ deposits remained. A considerable
depth of undisturbed deposits was confirmed in this trench, which
had not reached the base of the sequence. In order to confirm the
base of the cultural layers and collect more dating materials, Ji dug
again in the test trench in February 2015. Finds included two core-
axes from the bottom layer (Fig. 4). Due to a layer of large cobbles,
the base of the deposits was still not reached, but more charcoal for
AMS dating was collected. Although the surface of the site is heavily
obscured by a mat of creepers and other vegetation and plant
debris, the sheer size of the shelter indicates that substantial de-
posits remain to be excavated. Fragments of charcoal exposed in the
*Corresponding author. School of Geography, Archaeology and Environmental
Studies, University of the Witwatersrand, Johannesburg, South Africa.
** Corresponding author. Yunnan Institute of Cultural Relics and Archaeology,
Kunming, Yunnan Province, China.
E-mail addresses: jxpchina@foxmail.com (X. Ji), kathleen.kuman@wits.ac.za
(K. Kuman).
Contents lists available at ScienceDirect
Quaternary International
journal homepage: www.elsevier.com/locate/quaint
http://dx.doi.org/10.1016/j.quaint.2015.09.080
1040-6182/©2015 Elsevier Ltd and INQUA.
Quaternary International xxx (2015) 1e9
Please cite this article in press as: Ji, X., et al., The oldest Hoabinhian technocomplex in Asia (43.5 ka) at Xiaodong rockshelter, Yunnan Province,
southwest China, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.080
Fig. 1. The geographic location of the Xiaodong rockshelter site in Yunnan Province, south China.
Fig. 2. a) the large rockshelter in Permian limestone; yellow circle shows sitting excavators for scale; b) a view of the hilly karst landscape typical of the region. (For interpretation of
the references to colour in this figure legend, the reader is referred to the web version of this article.)
X. Ji et al. / Quaternary International xxx (2015) 1e92
Please cite this article in press as: Ji, X., et al., The oldest Hoabinhian technocomplex in Asia (43.5 ka) at Xiaodong rockshelter, Yunnan Province,
southwest China, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.080
Fig. 3. Plan and profile of the Xiaodong rockshelter.
Fig. 4. The stratigraphic sequence exposed by the 2015 test trench and matching
14
C dates.
X. Ji et al. / Quaternary International xxx (2015) 1e93
Please cite this article in press as: Ji, X., et al., The oldest Hoabinhian technocomplex in Asia (43.5 ka) at Xiaodong rockshelter, Yunnan Province,
southwest China, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.080
2008 and 2015 test excavations taken for
14
C dating have produced
good results (Table 1). They show that Xiaodong preserves a
sequence of deposits from >24,000 years BP in the uppermost layer
to 43.5 ka BP in the base of the excavated trench (see also Fig. 4). In
this paper we assess the affinities of the heavy duty tools and other
artefacts retrieved from the disturbed deposits and test trench
within the context of the
14
C dates.
Although the site has not yet been systematically excavated, the
extensive digging by local people has brought a large number of
artefacts to the surface that reveals clear Hoabinhian affinities and
indicates the value of the site for future excavations. The term
Hoabinhian derives from the northern Vietnamese province of Hoa
Binh where numerous rock shelters and caves were excavated
during the 1920s by French archaeologists (Colani, 1927, 1929).
During the First Congress of Prehistorians of the Far East in 1932, M.
Colani gave the first definition of the Hoabinhian as, “a culture
composed of implements that are in general flaked with some
varied types of primitive workmanship. It is characterized by tools
often worked on one face, by hammerstones, by implements of sub-
triangular cross section, by discs, short axes and almond shaped
artefacts, with an appreciable number of bone tools”(Collectif,
1932:11, translated in English by; Matthews, 1966). Almost one
century after its discovery, the Hoabinhian remains the most
important technocomplex for late hunteregatherer subsistence in
the Southeast Asian mainland, including Sumatra between the end
of the Upper Pleistocene to the Early Holocene (Matthews, 1966;
Heekeren and Knuth, 1967; Boriskovsky, 1969; Mourer and
Mourer, 1970; Gorman, 1972; Van Tan, 1980, 1997; McKinnon,
1991; Forestier, 2000, 2005; Taha, 2000; Viet, 2000; Moser, 2001;
Kamaruzaman, 2002; Majid, 2003; Schoocongdej, 2006; Marwick,
2007; Yi et al., 2008; Bacon, 2012; Rabett, 2012; Higham, 2013;
Forestier et al., 2015).
Overall, the Hoabinhian is known as a late Pleistocene to Ho-
locene technocomplex based in its earlier stages on hunting and
gathering, and it is well known from many sites in mainland
Southeast Asia (Gorman, 1969, 1971; Moser, 2001; Zeitoun et al.,
2008; Rabett, 2012; Higham, 2014). However, it has long been
thought that the Hoabinhian technocomplex may have existed in
southwest China in rockshelters in mountainous regions, and from
this “Hoabinhian Homeland”, populations dispersed to Southeast
Asia (Dai, 1988). Xiaodong now confirms not only the existence of
this archaeological tradition of large cobble tools in dated deposits
in China for the first time, but it also dates its occurrence as the
earliest site in Asia for this technocomplex. Xiaodong both extends
the northern borders and the chronology of this Southeast Asian
lithic phenomenon.
2. The Hoabinhian technocomplex: time and space
Most sites belonging to the Hoabinhian technocomplex are
located in mainland Southeast Asia, associated with a humid,
tropical hunteregatherer subsistence ecology. It lasted until the
transition to the Neolithic (around 6th millennia B.C.), when evi-
dence of plant domestication or “proto-agriculture”appears
(Gorman, 1969,1971, 1972; Glover, 1977; Reynolds, 1989). The
earliest Hoabinhian levels at the important site of Spirit Cave,
Thailand are estimated to be ca 13e14,000 years old, with the in-
dustry persisting until ca 7000 years ago when a more settled
Neolithic culture with pottery and edge-ground tools appears
(Gorman, 1971, 1972). InThailand, the oldest Hoabinhian site, Tham
Lod rockshelter in northern Thailand, dates back to 26,580 ±250 BP
(Beta-17222, MHSTLAR2-918; layer 8 in Area2; Schoocongdej,
2006). However the great majority of Hoabinhian sites in main-
land Southeast Asia and Sumatra date to the Last Glacial Maximum
(<c. 20,000 BP) to the Late Pleistocene to Early/Holocene transition,
corresponding with the large-scale modern human settlement of
the tropical Southern Provinces (Moser, 2001; Rabett et al., 2009;
Bacon, 2012; Higham, 2013, 2014; Forestier et al., 2015).
In terms of geographic distribution, the Hoabinhian tech-
nocomplex and Hoabinhian-like tools have also been suggested to
occur as far west as the southern fringes of the Himalayas, where
large adze/axe-like tools with a transverse cutting edge are noted,
in association with choppers and flake tools (Gaillard, 2010; Soni
and Soni, 2010; Gaillard et al., 2011). In Burma, the industry is
present from the late Pleistocene to ca 6500 years when edge-
ground tools occur (Thaw, 1971). The eastern extent of the com-
plex is even suggested to extend to Australia (Bowdler, 1994). As the
earliest dates currently available for Xiaodong are at least 43,500
years, this site is thus the earliest recorded presence for the in-
dustry in Asia.
Hoabinhian lithics are poorly standardised in comparison with
some others lithic production systems (e.g., the production of
flakes, blades, etc.). Nevertheless, the larger Hoabinhian tools are
consistently shaped on cobbles and are distinguished by a consis-
tent plano-convex cross-section. The common theme for formal
tools centres around the unifacial flaking of water-rounded cobbles,
with flaking usually around the circumference of a unifacial tool,
which has been called the “sumatralith”(Gorman, 1972; Forestier,
2000; White and Gorman, 2004; Marwick, 2008). The earlier
term of sumatralith (or unifaced pebble tool) similarly refers to
unifacial core tools made on elongated rolled cobbles
(Shoocongdej, 2000; Forestier, 2000, 2013). Bifacial flaking is also
present but less well represented in the assemblages, and choppers,
chopping-tools and small flake tools are common. There is much
discussion of the variability of the Hoabinhian core tools and debate
on its widespread distribution in Southeast Asia in association with
small tools (Shoocongdej, 2000). At the earliest dated sites in
Thailand ca 26,000 years BP (Tham Lod level 8, Area2,
26,580 ±250 BP) and Lang Kamnan Cave, with an uncalibrated date
of 27,110 years, the diagnostic tools are simply called utilised cores
(ibid.). Although there are some rare cases of an earlier appearance
probably around 29,000 years (for the basal level at Hang Cho cave–
Yi et al., 2008), the Hoabinhian was only well established at sites in
North Vietnam by 20,000 years BP. The majority of Hoabinhian
deposits in Southeast Asia thus date between ca 20,000 and 5000
years BP and are interpreted as “Later HuntereGatherer”groups.
3. Dating and stratigraphy of Xiaodong
As a large rock shelter formed within Permian limestone,
Xiaodong has good preservation of mammalian fauna. Stalactites
and stalagmites cover the walls and roof of the cavern (see Fig. 2)
and are also found in areas of cemented deposits. Fossils can be
seen embedded in the hard calcareous matrix of some deposits, and
others have been collected from softer sediments. Fauna collected
between 1981 and 2007 include Rhinoceros sinensis,Selenarctos
thibetanus,Macaca sp., Cervus unicolor,Muntiacus sp., Bos gaurus,
Table 1
Radiocarbon dates for five of the six layers in the 2015 test trench at Xiaodong.
Sample Age Layer
XDCH-1 24,450 to 24,910 BP 1
XDCH-2 23,840 to 24,250 BP 2
XDCH-4 29,420 to 29,590 BP 4
XDCH-5 31,240 to 31,490 BP 5
XDCH-6 31,380 to 31,640 BP 6
XDCH-7 42,695 to 44,085 BP 6
XDCH-10 43,500 BP 6
X. Ji et al. / Quaternary International xxx (2015) 1e94
Please cite this article in press as: Ji, X., et al., The oldest Hoabinhian technocomplex in Asia (43.5 ka) at Xiaodong rockshelter, Yunnan Province,
southwest China, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.080
and Hystrix brachyuran yunnanensis. These species suggest a late
Pleistocene age for some of the deposits. As deeper, intact deposits
at depth are preserved, we anticipate significant potential for older
cultural layers at greater depth. Today the habitat around Xiaodong
consists of tropical and sub-tropical semi-moist rainforest and
evergreen broad leaf forest. The modern fauna includes more than
60 species of mammals, more than 340 species birds, and more
than 50 species of reptiles. The fossil fauna suggests that ancient
habitats were similar to those of today, but ca 20e30 ka tempera-
tures were about three degrees lower.
The profile shown in Fig. 4 indicates that all six layers preserve
artefacts, although relative quantities need to be determined
through systematic excavations. All layers are dominated by silt-
sized sediments, but layers 1 and 5 contain more limestone
rubble, and layer 5 is also notable for boulder-sized clasts. Colour
differences between the strata are easily distinguished, correlating
with changes in vegetation due to temperature, although precipi-
tation was usually similar. The dates indicate that occupation took
place at intervals for at least 20,000 years. Layers younger than
24,000 years unfortunately were missing from our test trench
probably due to disturbance of the site by the local people, but large
scale excavations may hopefully find them preserved in other parts
of the site.
4. Artefacts from Xiaodong
4.1. Core-axes/adzes
The Xiaodong artefacts are made mainly on igneous rocks that
occur in the river gravels within the vicinity of the shelter. Today
the gravels below the site contain large cobbles, often well rounded
or oval in shape, but even facetted cobbles have well-rounded
edges. The occasional quartzite and quartz piece is noted in the
cave and probably also derives from these gravels. The igneous raw
materials are dominated by coarse-grain sizes and are generally of
quite poor flaking quality, a fact which is confirmed by the domi-
nance of step-terminated flake scars on cores and core tools. Large
cobbles are flaked directly or split where suitable angles are not
available to initiate flaking. Our own experiments with flaking of
the local cobbles showed that the streambed gravels below the site
today are very unsuitable due to weathering. It was difficult to find
hammerstones that did not break and equally difficult to produce
good flakes (whole or without step-terminations) on igneous rocks
that are so heavily weathered.
The current sample of artefacts is dominated by heavy-duty
tools, as these were the most easily recognised pieces during
collection. Table 2 provides details of the core axes/adzes, choppers,
scrapers and other tools in this selected but informative assem-
blage of artefacts. The most prominent tool type is a form of core-
axe or adze (formerly sumatralith). This type is specific to Hoa-
binhian and Hoabinhian-related assemblages. At Xiaodong, the
core-axe/adze (Figs. 5e7) is a high-backed, heavy-duty tool with a
plano-convex profile and cross-section. In five of the 10 complete
examples, cobbles are identified as the blank, one blank is clearly a
flake, and the remainder is indeterminate due to heavy working.
The planar face is dominated by cortex and has no or only limited
removals. In the example made on a large flake, the planar face is
formed by the flake's ventral surface. A plano-convex shape is easy
to achieve with these cobbles if a flat or only slightly convex surface
is selected, which is a common shape for cobbles within the gravels.
Fig. 5. Core-axes/adzes. 1) CYO4010; 2) CYO4012; 3) CYO4013; 4) 003; 5) CYO4001; 6) CYP4007.
X. Ji et al. / Quaternary International xxx (2015) 1e95
Please cite this article in press as: Ji, X., et al., The oldest Hoabinhian technocomplex in Asia (43.5 ka) at Xiaodong rockshelter, Yunnan Province,
southwest China, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.080
Fig. 6. Core-axes/adzes. 1) 001; 2) CYO4022; 3) CYO4044; 4) CYO4023; 5) CYO8001; 6) CYP4021.
Fig. 7. Photographs of selected pieces in the illustrations. 1) CYO4010; 2) CYO4012; 3) CYO4013; 4) 003; 5) CYO4001; 6) CYO4007.
X. Ji et al. / Quaternary International xxx (2015) 1e96
Please cite this article in press as: Ji, X., et al., The oldest Hoabinhian technocomplex in Asia (43.5 ka) at Xiaodong rockshelter, Yunnan Province,
southwest China, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.080
Each core-axe is characterised by an adze-like working end, which
occurs at the distal end of the long axis of the tool. Often this
working end is enhanced by one or more flake removals to the side.
It is also characterised by small tip damage from use, which is
usually bifacial. The shapes of working ends vary from transverse,
to rounded, to convergent. At the proximal end of each piece is a
tapering base, formed by flaking, either for ease of handling or for
mounting of the tool in a handle.
A prominent feature of these tools is that one or both lateral
edges of the piece are blunted, either naturally by cortex or more
commonly by steep flaking and battering. This pattern is recorded
for 12 out of the 13 pieces in Table 2 and is illustrated in the figures
for some of the pieces as the dotted areas. Such blunting could
provide a comfortable hand-hold for the tool but it could also
facilitate hafting. This trait also gives most pieces a slightly waisted
appearance at some or other position behind the working end. The
steep, adze-like working end of many of the core-axes is suggestive
of rejuvenation through further flaking, giving the business end of
the tool a stepped appearance. Marwick (2008) has demonstrated
this stepped and overhanging flaking as a characteristic feature of
reduction intensity.
One tool is somewhat unique in the collection (CYO4001,
Fig. 5.5) because the working end is not stepped back through such
flaking. However, the planar surface is, like the other examples,
cortical, and there is use damage on the tip. The difference may be
due to function but it is more likely that it is due to the lack of
Table 2
Details of the major tool types in the Xiaodong artefact sample. The angle of the working edge is an overall approximation of what is judged to be the most functional area. The
two tools with the smallest lengths give the impression of being highly rejuvenated specimens. Tapering of the bases is created through shaping, but in the case of CYO7023,
also partly by the shape of the cobble, and in the case of CYO4004 by the flake striking platform in conjunction with shaping. As four tools are made on cobbles, it is possible
that others were as well, particularly as planar surfaces are mainly cortical, but split cobbles cannot be ruled out. In the one example made on a large flake, the planar face is the
ventral surface of the flake.
Core-axes/adzes
Number Length Width Thickness Cortex on planar
face
Edge battering or
blunting
Waisting Tapering
base
Working end
damage
Working end
angle
Blank
001 126 85 55 90% Both sides Yes Yes Bifacial 80
Unknown
003 130 64 48 70% Both sides Yes Yes Bifacial 80
Unknown
CYO4004 111 79 41 Flake ventral Right side Yes Yes Bifacial 50 to 90
Flake
CYO4009 104 73 55 50% Left side, cortical right Yes Yes Bifacial 83
Cobble
CYO4010 114 82 51 75% Both sides Yes/on
left
Yes Bifacial 75
Cobble
CYO4012 126 81 48 60% Both sides Yes Yes Bifacial 70
Cobble
CYO0413 97 66 31 90% Left side Yes Yes Bifacial 52
Unknown
CYO4021 105 67 41 95% Both sides Yes Yes Bifacial 79
Unknown
CYO4031 87 67 44 80% Both sides Yes Yes Bifacial 78
Cobble
CYO7023 82 74 44 85% Left side, cortex on right Yes Yes Bifacial 70 to 90
Unknown
CYO4014 Distal end/broken
tool
Yes Left side Bifacial 80
Unknown
CYO4007 125 89 39 100% No No No Indet Indet Cobble
CYO4001 127 73 46 95% Left side No Yes Bifacial Unknown
Choppers
Number Length Width Thickness Flaking Flaking position Working edge damage Blank
CYO4018 97 89 33 Bifacial Two sides and end Bifacial Flat cobble
CYO4023 140 90 63 Unifacial One side and end Bifacial Split cobble
CYO4035 122 96 56 Bifacial Side and end Unclear Cobble
CYO7001 98 58 69 Unifacial End Battering Cobble
CYO7007 110 76 57 Unifacial One side and end Bifacial Cobble
CYO7009 194 158 70 Bifacial Two sides and end Bifacial Cobble
CYO7010 170 147 72 Unifacial Side Unclear Split cobble
CYO7012 108 91 54 Unifacial Side Bifacial Split cobble
CYO7013 119 102 35 Bifacial Two sides Bifacial Cobble
CYO7018 93 81 47 Unifacial Side Unclear Cobble
CYO7020 84 82 26 Unifacial End Bifacial Cobble
Denticulated and notched scrapers
Number Max length Backing Blank
CYO4003 94 Natural Cortical slab
CYO4019 144 Unknown/broken piece Flat cobble
CYO4020 113 Steep flaking Unknown
CYO4022 87 Probably hafted Unknown
CYO4030 103 Steep flaking Unknown
CYO4127 99 Steep flaking Unknown
CYO7002 129 Cortical Flat cobble
CYO7004 95 Cortical Unknown
CYO7005 92 Cortical Cobble
CYO7019 87 Natural Natural flake
CYO7011 97 Cortical Flat cobble
CYO7025 94 Steep removals&cortex Cobble
CYO7026 88 Steep flaking Flat cobble fragment
X. Ji et al. / Quaternary International xxx (2015) 1e97
Please cite this article in press as: Ji, X., et al., The oldest Hoabinhian technocomplex in Asia (43.5 ka) at Xiaodong rockshelter, Yunnan Province,
southwest China, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.080
rejuvenation flaking. Both lateral edges of the tool are steeply
flaked and extensively blunted, and there is some waisting at the
proximal end.
Another unusual piece (CYO4007, Fig. 5.6) is so convergent that
it somewhat resembles a unifacial handaxe, demonstrating the
variability in these Hoabinhian heavy duty tools. It is made on a flat
cobble and is totally cortical on the planar face. The worked face
shows fine retouch on three-quarters of its perimeter, creating an
acute edge. The left side of the specimen provides an effective
cutting edge, which continues around the tip for about one-third of
the right lateral, while the remainder of the right side is blunted by
edge-grinding. This uniface may thus be a form of hand-held cut-
ting tool. The collection also contains two distal tool portions that
appear to be broken unifaces of a similar nature (CYO7022 and
CYO7027).
4.2. Choppers
The choppers are characterised by contiguous removals, either
unifacial or bifacial, along the side or end of the piece. In African
assemblages, many researchers make a distinction between
chopper-cores (which lack evidence of utilisation) and choppers (as
tools), which shows signs of use damage. The chopper is a very
common tool type in Asian core and flake assemblages, where little
attention has been given to this important distinction. At Xiaodong,
it appears that most of the eight examples in this collection may
have indeed been used as tools, as some have bifacial damage on
the working edge. However, the coarse nature of the igneous raw
material makes it difficult to be certain about use damage on some
of the pieces. As a group, these examples appear to be heavy duty
tools, possibly used in woodworking, a feature which is suggested
by damage patterns on various Hoabinhian artefacts (Gorman,
1971 ).
4.3. Pick
The one pick in the collection is roughly made on a cobble,
minimally worked to a simple point, and has some bifacial tip
damage. One side of the piece is blunted through battering and the
other side is cortical. The base also has a cortical rounded butt
(Dimensions in mm are 119 long, 73 wide, 53 thick).
4.4. Core scrapers
There are seven heavy duty scrapers made on cobbles. Each
possesses one or more steeply flaked working edges, usually at a
right-angle in relation to a flat surface of the cobble. The cobbles
appear to have been selected for one or more such flat surfaces. In
this respect they resemble single platform cores, but bifacial
damage on many specimens suggests tool use. This steep, right-
angled working edge gives the appearance of a push-plane, and
hence they may have been used in wood-working activities.
Working edges can range from fairly smooth to irregular and
denticulated. The latter are thus similar to large, heavy examples of
denticulated and notched scrapers (Maximum lengths:
CYO4015 ¼111; CYO4016 ¼109; CYO4017 ¼100; CYO7008 ¼108;
CYO7015 ¼136; CYO7016 ¼118; CYO7017 ¼112).
4.5. Denticulated and notched scrapers
This type of scraper is not as heavy as a core scraper, and the
working edge is much more irregular in form, either with dentic-
ulation or notching, or a combination of the two. The working edge
is steep enough to suggest a scraping, rather than cutting function.
4.6. Denticulates
Flakes with more acutely angled working edges and denticu-
lated retouch or scars are termed denticulates. They are presumed
to be used for cutting, because of their more acute working edge.
They often show some bifacial use damage, which is characteristic
of cutting functions, and sometimes there is minor flaking on the
opposite side.
4.7. Other
A few irregular cores, not flaked to any pattern, occur in the
collection, as well as some core fragments and several flakes.
5. Conclusions
Although most of the tools belong to a selected collection from
disturbed and surface deposits and just a few pieces are from the
test excavations, the type most diagnostic of the Hoabinhian
technocomplexdthe core-axe/adze–is present, including in the
bottom layer of our test trench. The high-backed, plano-convex
shape of these heavy duty tools is distinctive, with cortex often
forming the flat surface. The function of such tools is widely
considered to be for woodworking, with the working of bamboo in
particular a most likely possibility (Gorman, 1971; Thaw, 1971). In
this respect, the type (although unrelated culturally and techno-
logically) has some similarities with the Sangoan industry core-
axes produced during the late Middle Pleistocene in the more
wooded habitats of central Africa (Clark, 1970). The best description
of such tools, also argued to be used in woodworking, was pub-
lished by Clark (2000) for the Kalambo Falls site in Zambia. For the
Hoabinhian, the adze-like nature of the working ends of these high-
backed tools is particularly diagnostic of woodworking. The
blunted edges and tapering bases of the typical core-axe/adze could
assist use as either a hand-held tool or a tool fixed into some form
of haft. The wood-working function of the site is also suggested by
the associated heavy-duty core scrapers with steep working edges,
as well as the denticulated and notched scrapers and the choppers.
The calibrated
14
C dates from the 2008 and 2015 trench indicate
the preservation of layers that date from >24,000 BP at the youngest
to 43,500 BP at the oldest.While these levels are pre-Neolithic in age
and an expression of the earliest Hoabinhian, some artefacts from
our surface collection indicate that younger deposits belonging to
the Neolithic are also present but were not sampled in the test
trench. Excavations in future will hopefully be able to document an
even lengthier sequence at Xiaodong for the Hoabinhian and will
place these tools in securely stratified contexts. For now, the Xiao-
dong artefacts clearly link this site with the karst upland subsistence
ecology that was a very widespread and successful adaptation to a
particular environment and a huntingegathering lifestyle that led to
the early appearance of domesticated plants. Thus the key site of
Xiaodong securely confirms the existence of the Hoabinhian tech-
nocomplex in southwestern China and places it chronologically as
the oldest site in Asia. In this sense, the Yunnan Province takes its
place as the most likely strategic location of the “Hoabinhian
Homeland”leading to gene flow and cultural dispersion for the
peopling of Southeast Asia.
Acknowledgments
This research was financially supported by the State Administra-
tion of Cultural Heritage of China (Chinese-Burma &Laos Frontier
Archaeology Survey project), Yunnan Cultural Relics and Archae-
ology (A-201301), Lincang Cultural Instituteand the University of the
Witwatersrand. Thanks go to Eric Bo€
eda and Val
ery Zeitoun for
X. Ji et al. / Quaternary International xxx (2015) 1e98
Please cite this article in press as: Ji, X., et al., The oldest Hoabinhian technocomplex in Asia (43.5 ka) at Xiaodong rockshelter, Yunnan Province,
southwest China, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.080
inviting to Xueping Ji, Yun Wuand Yinghua Li to a Thailand workshop
in 2011 for beneficial discussions.Thanks to Dequan Hu, Hongmei Xie
and Qiang Xiao for assisting field work. Kathleen Kuman, Ronald
Clarke and Hao Li thank the ChinaeSouth Africa Bilateral Programme
in Palaeosciences for funding provided by the Chinese Ministry of
Science and Technology to Liu Wu (IVPP–Institute of Vertebrate
Paleontology and Paleoanthropology, Beijing) and to R.J. Clarke
(University of the Witwatersrand), South African National Research
Foundation (NRF) grant number 68625. Since 2014 this research has
been funded by grants in the same programme to Gao Xing (IVPP,
ChinaeSouth Africa Joint Research Program-7) and K. Kuman (NRF
grant number 88480). Artefact illustrations for this paper have been
done by Wendy Voorvelt from original drawings by R.J Clarke. We
thank Prof Gao Xing for his invitation to contribute to this volume.
References
Bacon, A.M., 2012. Les sites de Tham hang, Nam Lot et Tam P
a Ling au nord du Laos.
Des gisements
a vert
ebr
es du Pl
eistoc
ene aux origines des Hommes modernes.
CNRS
editions, Paris.
Boriskovsky, P., 1969. Vietnam in primeval times. Soviet Anthropology and
Archaeology 8, 214e257.
Bowdler, S., 1994. The Hoabinhian in Australia: a retrospective. Vietnam Social
Sciences 5, 87e94.
Clark, J.D., 1970. The Prehistory of Africa. Thames and Hudson, London.
Clark, J.D., 2000. Kalambo Falls Prehistoric Site, vol. III. Cambridge University Press.
Colani, M., 1927. L'^
age de pierre dans la province de Hoa Binh (Tonkin). In: M
emoire
du Service G
eologique de l'Indochine 13. Hanoi.
Colani, M., 1929. Quelques pal
eolithes hoabinhiens typiques de l’abri-sous-roche de
Lang Kay. Bulletin de l'Ecole Française d'Extr^
eme-Orient 26, 353e384.
Collectif, 1932. Praehistorica Asiae Orientalis: 1, Premier Congr
es des pr
ehistoriens
d'Extr^
eme-Orient, Hanoi. Imprimerie d’Extr^
eme-Orient, Hanoi.
Dai, G.H., 1988. On the relationship between Southeast Asian ‘Hoabinhian Culture’
and the culture of Southern China. Southeast Asian 1, 52e61.
Forestier, H., 2000. De quelques chaînes op
eratoires lithiques en Asie du Sud-Est au
Pl
eistoc
ene sup
erieur final et au d
ebut de l'Holoc
ene. L'Anthropologie 531e548.
Forestier, H., 2013. The open-air site of Huai Hin (Northwestern Thailand): chro-
nological perspectives for the Hoabinhian. Comptes Rendu Palevol 12, 45e55.
Forestier, H., Simanjuntak, H.T., Guillaud, D., Driwantoro, D., Wiradnyana, K.,
Siregar, D., Due Awe, R., Budiman, 2005. Le site de T€
ogi Ndrana, île de Nias,
Sumatra Nord: les premi
eres traces d'une occupation hoabinhienne en grotte
en Indon
esie. C. R. Palevol 4, 727e733.
Forestier, H., Sophady, H., Puaud, S., Celiberti, V., Fr
ere, S., Zeitoun, V., Mourer-
Chauvir
e, C., Mourer, R., Than, H., Billault, L., 2015. The Hoabinhian from Laang
Spean Cave in its stratigraphic, chronological, typo-technological and envi-
ronmental context (Cambodia, Battambang province). Journal of Archaeological
Science Reports 3, 194e206.
Gaillard, C., 2010. Reply: do not confuse large cutting tool types. Quaternary In-
ternational 223e224, 245e247.
Gaillard, C., Singh, M., Dambricourt Malass
e, A., 2011. Late Pleistocene to Early
Holocene lithic industries in the southern fringes of the Himalaya. Quaternary
International 229, 112e122.
Glover, I.C., 1977. The Hoabinhian: hunteregatherers or early agriculturalists in
South-east Asia? In: Megaw, J.V.S. (Ed.), Hunters, Pastoralists and Early Agri-
culturalists in South Asia. Leicester University Press, Leicester, pp. 145e166.
Gorman, C., 1969. Hoabinhian: a pebble tool complex with early plant associations
in Southeast Asia. Science 163, 671e673.
Gorman, C., 1971. The Hoabinhian and after: subsistence patterns in Southeast Asia
during the Late Pleistocene and early recent periods. World Archaeology 2,
300e320.
Gorman, C., 1972. Excavations at spirit cave, North Thailand: some interim in-
terpretations. Asian Perspectives 13, 79e107.
Heekeren, H., Knuth, E., 1967. Sai Yok Archaeological Excavations in Thailand, vol. I.
Munksgaard, Copenhagen.
Higham, C., 2013. Hunteregatherers in Southeast Asia: from prehistory to the
present. Human Biology 85 (1e3), 21e44.
Higham, C., 2014. Early Mainland in Southeast Asia. From First Humans to Angkor.
River Books, Bangkok.
Ji, X.P., Zhang, P., Ma, J., Dong, H., Xie, H.M., 2006. A preliminary study on the
technology and typology of the artefacts from Xiaodong site of Cangyuan in
Yunnan province. In: Collected Works for the 40th Anniversary of Yuanmou
Man Discovery and the International Conference on Paleoanthropological
Studies. Yunnan Science and Technology Press, Kunming, pp. 122e132 .
Kamaruzaman, A.R., 2002. The Hoabinhian site at Taat Hill cave, upper Terengganu,
Malaysia. Jabat 29, 69e78.
Majid, Z., 2003. Archaeology in Malaysia. Centre for Archaeological Research
Malaysia, University Sains, Malaysia.
Marwick, B., 2007. Approaches to flaked stone artefact archaeology in Thailand: a
historical review. Silpakorn University International Journal 7, 49e88.
Marwick, B., 2008. What attributes are important for the measurement of assem-
blage reduction intensity? Results from an experimental stone artefact
assemblage with relevance to the Hoabinhian of mainland Southeast Asia.
Journal of Archaeological Science 35, 1189e
1200.
Matthews, J.M., 1966. A review of the Hoabinhian in Indo-China. Asian Perspectives
9, 86e95.
McKinnon, E., 1991. The Hoabinhian in the Wampu/Lau Biang Valley of North-
eastern Sumatra: an update. Indo-Pacific Prehistory Association Bulletin 10,
132 e142.
Moser, J., 2001. Hoabinhian: Geographie und Chronologie eines steinzeitlichen
Technokomplexes in Südostasien. Linden Soft, Koln.
Mourer, C., Mourer, R., 1970. The prehistoric industry of Laang Spean, province of
Battambang, Cambodia. Archaeology and Physical Anthropology in Oceania 5
(2), 128e146.
Rabett, R.J., 2012. Human Adaptation in the Asian Palaeolithic. Hominid Dispersal
and Behaviour during the Late Quaternary. Cambridge University Press,
Cambridge.
Rabett, R.J., Barker, G., Hung, G.O., Naruse, T., Piper, P., Raddatz, E., Reynolds, T.,
Nguyen, Van Son., Strimpson, C., Szabo, K., Nguyen, Cao Tan., Wilson, J., 2009.
The Trang an project: late to post Pleistocene settlement of the lower Song
Hong Valley, North Vietnam. Journal of the Royal Asiatic Society 19, 73e109.
Reynolds, T.E.G., 1989. Techno-typology in Thailand: a case study of Tham Khao Khi
Chan. Bulletin of the Indo-Pacific Prehistory Association 9, 33e43.
Schoocongdej, R., 2006. Late Pleistocene activities at the Tham Lod Rockshelter in
Highland Pang Mapha, Mae Hong Son Province, Northwestern Thailand. In:
Bacus, E.A., Glover, I.C., Piggott, V. (Eds.), Uncovering Southeast Asia's Past:
Selected Papers from the 10th Annual Conference of the European Association
of Southeast Asian Archaeologists. NUS Press, Singapore, pp. 22e37.
Shoocongdej, R., 2000. Forager mobility organization in seasonal tropical environ-
ments of western Thailand. World Archaeology 32 (1), 14e40.
Soni, V.S., Soni, A.S., 2010. Large size cleaver-like flake and Hoabinhoidal elements
from terminal Pleistocene to mid-Holocene epoch sites. Quaternary Interna-
tional 223e224, 242e244.
Taha, A.T., 2000. Archaeological Investigation in Ulu Kelantan, Peninsular Malaysia
(Unpublished PhD thesis). Australian National University, Canberra.
Thaw, A., 1971. The ‘Neolithic’culture of the Padah-lin Caves. Asian Perspectives
XIV, 123e133.
Van Tan, Ha, 1980. Nouvelles recherches pr
ehistoriques et protohistoriques au
Vietnam. Bulletin de l'Ecole Française d’Extr^
eme-Orient 113e154.
Van Tan, Ha, 1997. The Hoabinhian and before. Bulletin of Indo-Pacific Prehistory
Association 16, 35e41.
Viet, Nguyen, 2000. Homeland of the Hoabinhian in Vietnam. In: Unpublished
Communication Presented at the European Southeast Asian Archaeologists
Association, Sarteano, Italy.
White, J., Gorman, C., 2004. Patterns in “amorphous”industries: the Hoabinhian
viewed through a lithic reduction sequence. In: Paz, V. (Ed.), Southeast Asian
Archaeology. University of the Philippines Press, Quezon City, pp. 411e441.
Yi, S., Lee, J.J., Kim, S., Yoo, Y., Kim, D., 2008. New data on the Hoabinhian: in-
vestigations at Hang Cho cave, northern Vietnam. Bulletin of Indo-Pacific As-
sociation 28, 73e79.
Zeitoun, V., Forestier, H., Nakbunlung, S., 2008. Pr
ehistoires au sud du Triangle d’Or.
Editions IRD, Paris.
Zhang, X.Y., 1991. Stone artefacts and Pleistocene fauna of Xiaodong, Cangyuan. In:
Yunnan Human Origin and Prehistory. Yunnan People's Press, Kunming,
pp. 198e199.
X. Ji et al. / Quaternary International xxx (2015) 1e99
Please cite this article in press as: Ji, X., et al., The oldest Hoabinhian technocomplex in Asia (43.5 ka) at Xiaodong rockshelter, Yunnan Province,
southwest China, Quaternary International (2015), http://dx.doi.org/10.1016/j.quaint.2015.09.080