The Cave of the Monk (Ban Fa Suai, Chiang Dao wildlife sanctuary, northern Thailand)

lable at ScienceDirect

Quaternary International 220 (2010) 160–173

Contents lists avai

Quaternary International

journal homepage: www.elsevier .com/locate/quaint

The Cave of the Monk (Ban Fa Suai, Chiang Dao wildlife sanctuary,northern Thailand)

Valery Zeitoun a,*, Arnaud Lenoble b, Frederic Laudet c, Jeroen Thompson d, William Jack Rink e,Jean-Baptiste Mallye b, Winayalai Chinnawut f

a UMR 9993 du CNRS, Musee Guimet, 19 avenue d’Iena, 75116 Paris, Franceb PACEA, UMR 5199 du CNRS, Universite Bordeaux 1, Avenue des Facultes, 33405 Talence Cedex, Francec UMR 5608 du CNRS, Universite de Toulouse le Mirail CNRS, 5 allees Antonio-Machado, 31058 Toulouse Cedex 1, Franced Medical Physics and Applied Radiation Sciences, McMaster University, 1280 Main St. West, Hamilton, ON L8S 4K1, Canadae School of Geography and Earth Sciences, McMaster University, 1280 Main St. W., Hamilton, Ontario L8S 4J5, Canadaf 11th Regional Office of Fine Arts Department, Thambon Jaeramae, Muang Ubon Ratchathani 34 000, Thailand

a r t i c l e i n f o

Article history:Available online 4 December 2009

* Corresponding author.E-mail address: [email protected] (V. Zeito

1040-6182/$ – see front matter � 2010 Elsevier Ltd adoi:10.1016/j.quaint.2009.11.022

a b s t r a c t

Biostratigraphic and paleoecological analyses of Southeast Asian Pleistocene faunal sites are based on theassumption that paleontological assemblages are homogeneous. This means that the sites formed duringa time range shorter than biological evolution range and, above all, without faunal replacement inducedby environmental successions. Detailed study of the Cave of the Monk paleontological site, in Thailand,has lead to the conclusion that the complex pattern of this Southeast Asian paleontological site cannotsupport such an assumption.

The Cave of the Monk provides an Ailuropoda–Stegodon assemblage typical of Southeast AsiaPleistocene Fauna. The pluridisciplinary study presented here includes (i) site morphology description,(ii) sedimentological analysis of fossiliferous deposits, (iii) taphonomy study of the bone assemblage, and(iv) electron spin resonance (ESR) dating of tooth samples. This integrated approach demonstrates theexistence of a Pleistocene porcupine den inhabited during MIS 3 and 2. The features of this site arecommonly reported for other paleontological sites of the area, indicating that this kind of site formationis a very general one in Southeast Asia.

ESR dating and porcupine ethological reports are used to question the site time averaging. A timerange of a few tens of thousands of years is suggested. The paleoecological value of the site can then bediscussed. On one hand, the assemblage is thought to be a valuable paleoenvironmental recordconsidering that the bone accumulation agent is reputed to represent the fauna present in theenvironment without bias. On the other hand, the time averaging of the site indicates that site formationrange likely covered the environmental fluctuations of the last climatic cycle. This last hypothesis isconfirmed by the alternating association of species indicated by the stratigraphic distribution of thefaunal remains.

The research demonstrates the need for integrated and pluridisciplinary taphonomic study ofSoutheast Asian paleontological sites. It also reveals the utility of a microstratigraphic analysis ofa fossiliferous karst deposit in order to disentangle the succession of Pleistocene mammals in response toenvironmental changes.

� 2010 Elsevier Ltd and INQUA. All rights reserved.

1. Introduction

A specific association of mammals, called the Sino-Malayanfauna by von Koenigswald (1938–1939) but more commonlynamed the Ailuropoda–Stegodon fauna complex, is described as an

un).

nd INQUA. All rights reserved.

indicator of the Upper Middle Pleistocene in Southeast Asia (Bienand Chia, 1938; Granger, 1938; Pei, 1938; Kahlke, 1961). Thisfaunistic group includes Asian taxa endemic to or strongly associ-ated with tropical environments. This complex includes: Stegodon,Asian Elephant, different kinds of Rhinoceros, the large primatesGigantopithecus and Pongo, and also numerous species of Suids,Cervids and Bovids. The most common Carnivora are Hyena, Tiger,Panther, Cuon and Tibetan Bear, together with Giant Panda,Ailuropoda.

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173 161

This particular faunistic complex was initially identified in SouthChina (Matthew and Granger, 1923) in connection with tropicaltaxa such as Hylobates and Tapirus. Later, the same complex wasfound in Vietnam (Patte, 1928), Laos (Fromaget, 1936), Myanmar(De Terra, 1938), Cambodia (Beden and Guerin, 1973) and Thailand(Pope et al., 1981; Ginsburg et al., 1982). It is associated with the so-called Indochinese biogeographic area that spreads from theYangtse River (Pei, 1957) to Isthmus of Kra (Tougard, 1998). Thisassociation is mostly present at sites in karst areas. Some sites wereonce swallow holes, such as the Yenchingkuo site (Granger, 1938);others are sinkholes such as Longuppo (Wanpo et al., 1995). Most ofthe sites, however, are included in caves (Liucheng, Hoschangtung,Maba, etc.). The sites are highly important for an understanding ofthe Pleistocene in Southeast Asia, and all the more so becausehuman remains were found which could provide crucial informa-tion regarding the dynamics of peopling of this part of the world(Brown, 2001). The faunistic series from these sites is being used toestablish a regional biostratigraphy (Kahlke, 1961; Han and Xu,1989; Tougard, 1998) and is a useful tool for reconstructingevolutionary scenarios in these environments (Tougard andMontuire, 2006). All of these studies are based on the workinghypothesis that assemblages are homogeneous in terms ofchronology and environmental origin. This assumption, however, isnot shared by all people working on this question. Thus, De Vos(1984), following Patte (1928), assessed that Indochinese paleon-tological assemblages are a faunal mixture from different periodsand environments.

The processes of site formation are central to the interpretationof this faunistic complex. However, there is to date no generalstudy describing the formation of faunistic sites in a karstic contextin Southeast Asia. The main reason is that paleontological seriesfrom Southeast Asia sites are poorly documented. Generally,excavation reports only include lists of fauna, positions of all theremains in the site, and the description of the stratigraphy of thedeposits, while the documentation of other remains other thandeterminable bones and teeth (coproliths, surficial state of bones,etc.) is missing.

The first reason for this lack of data is historical. Indeed, in thisregion of the world excavations are either old or modern methodsare still not used. One of the first excavations was at Yenchingkou(Granger, 1938). Prior to this, Pleistocene fauna was only knownfrom Chinese drugstores where remains of fossil hominids, Pongoor Gigantopithecus (Schlosser, 1903) were discovered. Despite themomentum generated by the Zhoukoudian site in the 1920s, it wasonly at the end of the 1970s that a project was developed to fullydocument Asian sites and to attempt to determine the origin of thebone assemblages (White, 1975).

The second reason for the lack of complete data is because thisgoal is difficult to reach. As emphasized by Pope et al. (1981),numerous sites were already plundered by Dragon bone collectors.Therefore, preserved sites where a detailed excavation can beundertaken are rare. Furthermore, new sites are hard to find due totheir often hidden nature and difficult access.

A major limitation to the thorough exploitation of known sites isthat there is little existing data on which to base criteria for theidentification of the origin of the accumulated remains. Thus, themechanisms involved in deposition of human remains, biaseslinked to the formation of the sites and the limitations of paleo-environmental information are rarely discussed (Simons and Ettel,1970; White, 1975).

The absence of data on the mechanisms of deposition in existingsites suggests that important advances could be made by exca-vating and extensively studying new sites. A few projects have beenundertaken with this aim in China (Bakken, 1997; Schepartz et al.,2001, 2003). In Vietnam, sedimentological studies of the Tham

Khuyen (Ciochon et al., 1996) and Ma U’Oi (Bacon et al., 2004) caveshave been published. In Ma U’Oi, the authors concluded thata fluvial accumulation of bones occurred in the endokarst. Never-theless, biases in the composition of the assemblage related to thiskind of deposition mechanism (Lyman, 1994) were not noted. Alsonot noted was the exclusive preservation of teeth versus bones,which can hardly be ascribed to fluvial sorting (e.g. Voorhies, 1969;Behrensmeyer, 1988).

Bakken (1997) carried out a compared taphonomic study on sixChinese sites among which two are found in the Indochinese area:Yenchingkou and Longuppo. The first site is a swallow hole and thesecond is a sinkhole among which, ‘‘the representation and conditionof the Longuppo fossils also recall the assemblages from Yanjinggou[Yenchingkou], where carcasses accumulated in vertical passages, theresult of predation and falls’’ (Wanpo et al., 1995, p. 275). Never-theless, the results obtained cannot be extended to the whole ofSouth China, where the majority of sites are caves. Finally, tapho-nomic and geological approaches were not undertaken conjointlyon any site.

Due to these limitations, no synthetic model or general rule canbe derived from research into the formation of the paleontologicalsites in the karstic context of Southeast Asia. The biases whichinfluenced the accumulation of bones are unknown, as are thelimitations of the usefulness of the collected series for doc-umenting the biostratigraphy, reconstructing paleoenvironments,or providing evidence of human activity in the past. For thesereasons, the mechanisms involved in the formation of thesepaleontological sites remain to be elucidated in spite of theirimportance for an understanding of the fauna and environmentpresent in this part of the world during the Pleistocene.

This paper presents the results of a pluridisciplinary approachcarried out in the Cave of the Monk in northern Thailand(19�24.600N; 98�48.980E). The analysis of both geological andtaphonomic data allows identification of bone accumulationprocesses. The results are completed by the dating of the paleon-tological deposits. In this way, the time averaging of the site can beestimated and, consequently, the consistency of the paleoenvir-onmental and biostratigraphic inferences based on such sites.

2. Site context

The Cave of the Monk, also called Ban Fa Suai I, is close to Ban FaSuai village, 80 km north of Chiang Mai (Zeitoun et al., 2005)(Fig. 1). The cave opening is at an altitude of 880 m in a dry tropicalforest. The surrounding countryside is characterised by mountains,dominated by Doi Chiang Dao, which at 2716 m is a major peak onthe regional horizon. This limestone block is highly karstic(Delange, 1997), as are the surrounding limestone outcrops. The sitewas found within one of these karst outcrops.

The cave is in the highest section of three levels of galleriesconnected by separate narrow shafts (Fig. 2). The lower section ofthe network is active, and a tributary of the Huai Mae Pla Ao riverflows through it. The survey showed that the configuration of thecavity was constrained by the fracturing of the retaining rock inregional directions: north–south and south–southwest/north–northeast, as commonly observed in northern Thai karsts(Kiernan, 1991; Dunkley, 1995). The deposits which flow on theslope are trapped at the entrance to the cavity; they forma detrital cone which partially blocks the entrance. This infillingwas largely removed by a monk who inhabited the entrance ofthe cave.

An initial corridor around 5 m wide and 2–5 m high continuesfor some 30 m. Two main corridors extend the length and make upthe higher network. The highest corridor, in the north, is full offallen rock material which obstructs the corridor in places.

Fig. 1. Map of location of the site.

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173162

Leading from the first corridor, a narrow passage situated in theprolongation of a joint pattern permitted exploration into thenetwork towards the south. A descent of 3 m in a ‘‘toboggan’’permitted progression along a gallery. The principal tunnel is

terminated by a passage of 40 m long where it was necessary tocrawl before reaching a low-ceilinged chamber which ended up asa constriction. This low chamber communicates via a narrow holewith a parallel corridor connected with the gallery further uphill.

NorthN 19° 24.603'E 98°48.985'Z = 880 m

10 metres

mainentrance

secondaryentrance

well towards the lower cave

low ceilinged chamber

Upper level of the Cave of the Monk

narrow passage

reported by Seveau, Paggrasa, Chumdeedrawing Zeitoun

30 metres

North

well towards the upper cavelower

entrance

Lower level of the Cave of the Monk

Fig. 2. Map of the upper level of the Cave of the Monk.

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173 163

3. Material

Some faunal remains (N¼ 247) were collected on the soil of thesediment at the entrance of a first chamber and under a fine coat ofclay which covers the rocky substratum elsewhere in the northernnetwork of the cave. Most are dental remains of Artiodactyls(N¼ 156) amongst which principally can be determined Suids(N¼ 47), Cervids (N¼ 38), Bovines (N¼ 29), and Nemorheds(N¼ 22). Rodent teeth are quite numerous (N¼ 56), amongst whichthe porcupine (Hystrix brachura) dominates. A second Hystricid(Atherurus macrorousus) and the mandible of a giant rat (Sundamysmuelleri) are also present. This material came from a shaft whichcommunicates with the outside, as demonstrated by the presenceof bamboo leaves. The few remains of Murids and the very rareremains of Rhinocerotids as well as a unique fragment of Probos-cidian tooth and the tooth of a Macaque were found close to thisshaft. Several anatomical elements of a porcupine covered in calciteand manganese oxide were also found in a side cavity.

Due to the site formation process the age of these remains candiffer from those found in the constriction. The faunal remainscollected in the northern branch are not included in this study.

Most of the remains (N¼ 3709) have indeed been provided bythe excavation located in the low-ceilinged chamber, approxi-mately 100 m from the entrance. The studied series was collectedfrom the surface in the entrance to the chamber, but mainly duringthe excavation. The excavation consisted of two test-pits (Fig. 3).The first was near the entrance to the room and the second deepwithin the chamber. The total surface covered was 5 m2. Thepaleontological remains were found within the first 50 cm ofsediment. To ensure that the recovery of material was complete, allof the articles found during the dig were removed after havingnoted their exact position. Then, each layer of removed sedimentwas sieved with water and a 1.3 mm sieve, and, once dry, allremains larger than this were kept.

The paleontological identifications lead to the determination of24 genera and 38 mammalian species including 3 rodent genera.

North

Area 2

1 aerA

1 meter

profile

profile

Fig. 3. Map of the Low chamber.

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173164

The determination rate at the genus level was high: 31.6%, but only9.1% at the species level. In the faunal listing (Table 1) the term ‘‘cf’’specifies that the identification was made using the closest present-day taxon. The assemblage includes pieces of teeth of Stegodon andtwo taxa which have now vanished from this area: Ailuropoda andPongo (Zeitoun et al., 2005).

The faunal association from the Cave of the Monk belongs to theSino-Malayan fauna of von Koenigswald (1938–1939), morecommonly named the Ailuropoda–Stegodon complex and describedas a marker of Middle or Upper Pleistocene in Southeast Asia(Matthew and Granger, 1923; Pei, 1935; Bien and Chia, 1938;Granger, 1938; Kahlke, 1961). The assemblage described alsocontains species found in other assemblages in Southern China,Vietnam, Laos, and Thailand where the Cervids and the Bovinesdominate. The genus Ursus is systematically found as well asRhinoceros and Proboscidians with the genera Stegodon and/orElephas, a few Carnivores, and almost always remains of Hystrix.

The absence of archaic species at the site suggests that theassemblage could be similar to the fauna found at Yenchingkuo II,which pertains to the Upper Pleistocene of southern Chinaaccording to Han and Xu (1989). Nevertheless it is not possible tofully assert this fact because the modern Hyena Crocuta crocutaultima that typically replaced Hyena sinensis (Kurten, 1956; Gins-burg et al., 1982) during this period was not found in the Cave of theMonk. The precise taxonomic attribution of the single dentalfragment of discovered Hyenidae could not be determined.

4. Methods

4.1. Geological study

The geological study is based on a detailed description of the siteand on the sedimentological study of the fossiliferous deposits. Thesite was described by examining the organisation of the karstnetwork, the morphology of the cave and its infilling. A cavetemperature profile was also determined by measuring theminimum and maximum daily temperatures at different placeswithin the cave.

The stratigraphy of the site has been established by examiningeach layer uncovered during the excavation. The deposits weredescribed directly on site by noting their general organisation, thecharacteristics of the stratification, the structure and texture of thesediment, the occurrence of grain size sorting or grading, and gravelmorphology and generation mechanisms. Sediment colour wasdetermined by reference to the Munsell colour code (Munsell,1954). The macroscopic examination was completed with severallaboratory analyses: particle size distribution in the deposit matrix(i.e. the sediments less than 2 mm) was determined by mechanicalsieving or laser diffractometer analysis for the smallest fractions(<50 mm), X-ray diffraction was used to identify mineral species,and thin sections, prepared from undisturbed blocks of sedimentimpregnated with polyester resin, were observed with a petro-graphic microscope. The genesis of the deposits has been inter-preted by comparison with actual analogues.

4.2. Taphonomic study

Each paleontological remain was examined by eye and witha magnifying glass. For each bone or tooth, the state of fragmen-tation, the colour, and the surface state were noted. This lastparameter included any human-made alterations (e.g. traces of fire/combustion, knife marks, and percussion marks), biological alter-ations (digested or regurgitated bones, gnawing, edge damages,and punctures) and alterations caused by corrosion (polish,smoothing or patina). Finally, the artefact dimensions weremeasured with callipers.

4.3. ESR dating

ESR age calculations were performed for three teeth from theCave of the Monk (Table 2). Sample preparation was according tostandard techniques (Rink et al., 1994). Uranium concentrations inthe enamel, dentine, and collected sediment were determined bydelayed neutron counting (McMaster Nuclear Reactor); Thoriumand Potassium concentrations in the collected sediment weredetermined by instrumental neutron activation analysis (McMasterNuclear Reactor). The inside and outside of the enamel layers werestripped with a high-speed dental drill in order to remove theexternal alpha dose contribution (Table 3a). ESR measurementswere performed with a JEOL JES-FA100 X-band ESR spectrometerwith the following scan parameters: power 2.0 mW, modulationamplitude 0.5 mT, centre field 336.0 mT, scan width 5.0 mT, scanrate 0.167 mT/s, and time constant 0.1 s. Ages were calculated withROSY v2.0, which is an updated version of ROSY 1.41 and incorpo-rates one-group theory for beta particle transport (Brennan et al.,1997a). Alpha and beta dose rates to the enamel layers weredetermined by calculation from the radioisotope concentrations inthe enamel, dentine, and sediment (Table 3b). Gamma dose rateswere determined by in situ gamma spectroscopy near the samplelocations. The cosmic dose rate was determined assuming anoverburden of 20� 5 m (2.0 g/cm3) and was corrected for latitudeand altitude (Prescott and Hutton, 1994).

5. Results

5.1. Cave morphology

The Cave of the Monk, and notably the southern branch, isa natural gallery with a relatively constant diameter. In thissouthern corridor, the absence of blocks on the floor demonstratesthat there has been no rockfall in this gallery. On the other hand,there were many concretions, stalactites, stalagmites and pillars.Their evolution stage is very varied. Some are in the process of

Table 1List of paleontological artefacts from the Cave of the Monk.

Cave of the Monk Area 1 Area 2 Surface NtdR Cave of the Monk Area 1 Area 2 surface NtdR

Primate indet. 1 1 Suids indet. 10 4 14Macaca sp. 8 3 11 Sus sp. 150 119 21 290Macaca cf assamensis 1 1 Sus cf barbatus 1 5 6Macaca cf nemestrina 1 1 Sus cf scrofa 16 34 5 55Macaca cf andersoni 1 1 Total of Suids 167 168 30 365Pygatrix cf neamus 1 1 Bovidae indet. 28 33 1 62Hylobates sp. 1 1 Bos sp. 54 39 17 110Pongo cf pygmaeus 3 1 4 Bos cf gaurus 3 9 4 16Total of Primates 16 1 21 Bos cf javanicus 6 8 2 16Carnivora indet. 4 6 10 Bos cf sauveli 5 10 2 17Cuon cf alpinus 1 1 2 Bubalus cf arnee 1 2 3Cuon sp. 3 1 4 Pseudoryx sp. 1 4 2 7Canidae indet. 4 4 Total of Bovins 97 104 30 231Ursidae indet. 5 3 8 Cervidae indet. 21 27 5 53Ursus cf thibetanus 6 10 16 Cervus sp. 20 7 2 29Ursus cf malayanus 3 3 Cervus cf eldii 5 3 8Ailuropoda cf melanoleuca 4 3 7 Cervus cf unicolor 12 20 9 41Arctyonyx cf collaris 1 1 Cervus cf nippon 1 1Lutra sp. 1 1 Total of big Cervids 58 58 16 132Panthera cf tigris 1 1 2 Axis cf porcinus 4 2 2 8Hyaenidae indet. 1 1 Muntiacus sp. 42 40 4 86Total of Carnivora 28 31 59 Muntiacus cf muntjak 10 26 6 42Proboscidae indet. 19 2 21 Muntiacus cf vuquangensis 1 1Stegodon sp. 6 6 2 14 total of small Cervids 56 69 12 137Elephas sp. 14 6 20 Total of Cervids 114 127 28 269Total of Proboscideans 39 14 2 55 Naemorhedae indet. 47 116 15 178Perissodactyla indet. 10 10 Naemorhedus cf caudatus 2 2Rhinocerotidae indet. 102 109 6 217 Naemorhedus cf goral 2 2Rhinoceros cf unicornis 1 1 2 Capricornis cf sumatraensis 11 8 4 23Rhinoceros cf sondaicus 1 1 Capricornis sumatraensis cf kanjereus ? 1 1Rhinoceros cf sinensis 6 5 1 12 Total of Naemorheds 58 129 19 206Dicerorhinus cf sumatrensis 2 2 Artiodactyla indet. 659 535 57 1251Tapirus sp. 4 4 Total of Artiodactyla 1095 1063 164 2322Total of Perissodactyla 111 130 7 248Rodent indet. 154 131 6 291 Total number of tooth 1414 1462 191 3067Hystrix sp. 8 20 28 Bones 153 235 6 394Hystrix cf brachyura 11 18 12 41 Tympanic bones 23 30 3 56Atherurus cf macrorurus 1 1 Fragment of tooth indet. 112 56 14 182Muridae indet. 1 1 Fragment of dentine indet. 2 8 10Total of Rodents 173 171 18 362 Total 1752 1743 214 3709

Table 3(a) Enamel thicknesses and thicknesses removed from outside (in contact withsediment) and inside (in contact with dentine). (b) Mean dose rates to the enamelfrom internal and external sources (assumes early-uptake). All dose rates are mGy/ka.Gamma dose rates were determined by in situ gamma spectroscopy, at the location ofthe sample (BFS-G1-2) or less than 20 cm away (BFS1-P2). Beta dose rates for thesediment were determined from the radioisotope concentrations in bulk sedimentsamples. The beta and gamma dose rates from the sediment for BFS1-443 wereassumed to be the same as for BFS1-P2. The cosmic dose rate assumes an overburdenof 20� 5 m with density 2.6 g/cm3.

Panel (a)

Sample Enamel thickness [mm] Enamel removed

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173 165

being built whereas the deposit of a fine layer of clay indicates thatothers are fossils. Others, however, are being degraded – as forinstance a stalagmitic cascade situated at the junction of twopassages and largely smashed open by a pocket of corrosion whichhas also attacked the walls.

Measurements of the air temperature in different areas of thecave led to an evaluation of thermal regime. Temperatures variedbetween 14 and 18 �C in the entrance zone according to a dailycycle. The minimum temperatures were taken at the beginning ofthe morning and the maximum at the end of the afternoon, whichled to a rapid balancing with identical temperatures outside of thecave. Air currents generated by these changes in temperaturecovered the greater part of the length and breadth of the network.In the southern corridor, the air flow lost power just beyond theside passage and could no longer be felt beyond the curve in thenorth branch. This demonstrates that air currents exchanged the air

Table 2Dental samples selected for ESR dating. BFS-G1-2 was recovered from the sedimentexcavated in order to insert a gamma spectrometer; BFS-P2 was recovered in situfrom the wall profile; BFS1-443 was selected from teeth recovered during the initialexcavation.

Sample Taxon Unit

BFS1-G1-2 bovid IVBFS1-P2 rhino VBFS1-443 rhino VII

between the upper and lower network levels via small passageswhich go down at the edge of the corridors and chambers whenthey are not completely blocked. The temperature of the deep karst

From outside [mm] From inside [mm]

BFS1-G1-2 651� 41 46� 23 43� 22BFS1-P2 2260� 246 53� 26 40� 20BFS1-443 2186� 236 80� 40 61� 30

Panel (b)

Sample Enamel Dentine Sediment Cosmic

a b b b g

BFS1-G1-2 12.28 3.40 35.33 234.22 735.8� 73.6 23.0� 8.3BFS1-P2 0.00 0.00 61.56 90.45 815.2� 81.5 23.0� 8.3BFS1-443 0.00 0.00 26.45 87.94 815.2� 81.5 23.0� 8.3

Fig. 4. Northern profile of the test-pit done in the entrance of the Low chamber (Area 1).

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173166

beyond all seasonal fluctuation was assessed by measuring thetemperature of the water where it resurges, found to be 25 �C. Thiswas the same temperature as in the constriction, where no dailyvariation was in any case noted. Thus, this chamber is not influ-enced by changes of air, its thermal regime being that of the deepkarst, constant throughout the year.

The non-cemented soil also indicates the lower room asa unique setting in the cave. The soil surface is clayey. The topog-raphy is irregular and consists of juxtaposed shallow basins(10–20 cm) less than 1-m wide and bordered by a flattened ring ofcompact sediment. During the first exploration of the network, oneof these hollows was surrounded by porcupine quills.

5.2. Cave infilling

At several places in the network, ancient infilling is adhered tothe walls in the form of breccia. Most often it consists of conglom-erates of bedded sands and pebbles, indicating an ancient alluvialstage. In the shaft above the narrow opening to the south branch, thisconglomerate is overlain with fine sand and finely laminated silt.

The infilling of the lower chamber was examined as the depositlayers were removed during the excavation. The bedrock wasreached at a depth of 0.8–1 m. It has an irregular surface, with analtered cortex formed by a centimetre-thick black crust withalternating white-cream compact or powdery and black laminae.The mineralogical determinations indicated that it is mostlycomposed of phosphate.

The deposits consist of 3 stratified units (Fig. 4):

(i) The first unit is only found in the hollows in the rockysubstratum. It is 20 cm thick. It consists of calcareous granulesand pebbles arranged in horizontal beds sorted and inter-bedded with clayey red brown sand. The structure andcomposition of this deposit is the same as that of the brecciaconglomerate observed in the cave.

(ii) The second unit overlays the first deposit as well as the highestparts of the bedrock. The sediment is a yellow-brown to yellow(10YR 6/8) clayey sand which is more or less hardpan. X-raydiffraction determined that the major minerals are phosphateswhich in thin sections appears as either beige to yellowisotropic mass impregnating the deposit, or a fibrous isopaccoatings of hydroxylapatite. Subhorizontal layers of sedimentwhich is aggregated or rich in granules were also observed insome spots. Under the microscope, the lithology of the gran-ules is variable: pelite, sandstone and quartzite. Several piecesof gravel are scattered in the deposit and have a very welldeveloped phosphate-rich alteration rim.

(iii) The third unit contains the paleontological remains. It is0.2–0.5 m thick. The irregular and sharp lower limit indicatesan erosional contact. The deposit is a yellow-brown (7,5YR 4/6)to red brown (7,5YR 5/6), and some grey brown sandy claywith few pebbles. The general facies is a stratified lenticulardeposit. The lenses are thick at more than 10 cm. They arearranged in series of horizontal or slightly inclined conformably

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173 167

overlying lenses. The sets of lenses extend for several metresand their thickness reaches approximately 20–30 cm. Each setis separated by sharp or erosional surface.

In detail, each lens is formed by the superposition of two facies:The base and the body of the lenses consist of millimetre to

centimetre-thick clay aggregates. These aggregates are sorted andgrading, most often normal, occurs frequently. Examples of recur-rent grading have been observed in some lenses. Aggregation isclear at the coarse-grained basal part of the lenses and lessensprogressively towards the top. Associated with the coarser aggre-gates are small limestone pebbles which are smooth, somecompletely weathered. The clast morphology, small triangular orsquare plates, indicates that these stones are derived from thesurrounding rock that is fractured by a joint pattern. Dental remainswere preferentially associated with the coarse-grained parts oflenses. In thin sections, aggregates are heterogeneous, due tovariations in texture and colour. Also of note is the clay orientationobserved art the periphery of the aggregates.

The upper part of lenses is composed of compact and struc-tureless clay. This bed varies in thickness from 2 to 5 cm. It isespecially pronounced when it forms the roof of a series of lenses.The lower edge of these beds is diffuse whereas the top edge issharp and sometimes highlighted by a network of drying cracks.Locally, fine intercalated layers some centimetres thick can bedistinguished. These have wavy borders and stand out due tovariations in colour and texture. Clay orientation is also observedunder the microscope; clays are oriented on the periphery of grains(sand or bone) or according to the bedding plane and, in the case ofintercalated laminae, following the undulations of the sediment. Inthis structureless sediment, porosity is weak, and consists of cracksparallel to the bedding plane.

5.3. Sedimentary history of the cave

The cave infilling appears to have a complex history and severalphases can be distinguished. At the bottom of the sequence in theCave of the Monk, sands and pebbles are evidence, preserved in thebedrock depressions, of an ancient alluvial period, which is alsoindicated by the cemented conglomerate observed intermittentlyalong the walls of the cave.

The second unit was originally a deposit of fine material trans-ported by flow, as shown by several intercalated beds of granules.As is the case of the silt overlying the alluvial deposits in the well,this is evidence of deposition related to cave flooding at a timewhere the river flowed in a lower karst level. This is characterisedby the neomorphism of phosphates (mostly hydroxylapatite).These phosphates could be derived from animal guano, very likelybat, because of the occurrence of ditmarite in the mineralogicalarray which is an ammoniated phosphate of animal origin (Hill andForti, 1997). The diagenesis is pronounced. Among the phosphatesidentified are several mineralogical species that imply a neo-morphism process which led to the complete decarbonation of thesediment (tarakanite, leucophosphite and montgomeryite, cf.Karkanas et al., 2000). This type of transformation indicates thatbetween the deposition of this unit and the unit below it there wasa long hiatus.

The upper deposits are attributable to an accumulation ofsediment dug by burrowing animals. The round millimetric tocentimetric aggregates observed, commonly found in a karstcontext (Goldberg, 2001), are facies associated with burrow infill-ing (Karkanas, 2001). This interpretation is also supported by theoccurrence of grain size sorting of the aggregates. The strips ofdigging debris which accumulate when a burrowing animal digsshow a longitudinal sorting where the coarsest material ends up on

the edge of the rejected layer. The accretion of this debris then leadsto the superposition of material in progressively decreasing size(Lenoble, 2001).

The orientation of clay minerals observed under the microscopesuggests that the deformed sediment is in a plastic state. The ill-defined bedding and the flattening observed (indicated by the wavyedges of the intercalated fine layers) are also evidence of thisdeformation. The compaction and the fissural porosity in accor-dance with the bedding plane suggest that it is a trampled horizon(Goldberg and Whitbread, 1991). So, these compact beds areformed in an originally aggregated material.

The association of aggregate lenses and beds of clay thusrepresent episodes of burrowing (production of aggregates) fol-lowed by a period of occupation of a burrow (formation of a tram-pled horizon). The succession of many lenses thus means that therewere many burrowing events. The size and morphology of thedigging can be determined from the geometry of the erosivecontact between each set of lenses. These are suggestive of shallowbasins extending for about 1 m. This morphology is identical to thetroughs which presently structure the ground of the lowerchamber. The sediment rings which delimit the troughs on the soilsurface represent the strips of accumulated debris which havesecondary been compacted by trampling. The stacking of a series oflenses indicated multiples phases of digging where the debris filledin the existing topography.

These lairs are significantly smaller than those described forbears (Fosse et al., 2001). They were thus made by a medium sizedanimal. The thorns of porcupines found around the lairs during thediscovery of the room suggest that this rodent is the animal whomade the diggings. This interpretation is supported by the presentdescriptions of the Hystricidae lair. Brain (1981: 112) reporteda similar organisation of burrows and a central area flattened bytrampling and used for sleeping: ‘‘Inside the lair was a circularraised piece of clean ground well consolidated by the porcupinesthat had lain there sleeping’’.

Two other arguments also support this interpretation. Firstly,this area of the cave is exceptional in that the temperature does notvary, making it attractive for occupation by an animal. Secondly,unlike other galleries where puddles of water are present at sometimes during the year, the absence of structures, on the ceiling orfloor, is evidence that dripping is rare in this chamber, which wouldalso make it a more favourable place for a lair.

The sedimentological study indicated that the fossil deposits areassociated with the use of the lower chamber as an animal lair. Theformation of these fossil deposits was late in the history of thekarst. Indeed, it was at the moment that the network became fos-silised, meaning that there was no longer any alluvial depositiontaking place. The cave has probably been frequented for a long timejudging by the large number of beds which have remodelled theupper part of the room and which reach a thickness of 0.5 m. Thisuse of the cave appears to have continued to the present day, asporcupine quills are present on the floor.

5.4. Taphonomic study

The taphonomical analysis provides information concerning theaccumulation of faunal remains of the Cave of the Monk. It allowsthe fossils to be placed in history and in the evolution of the caveand offers arguments in the determining of the accumulating agentand the modalities of the deposits. The fossil material is made of12% of bone (most are pieces of 1.5 cm long) and only 11% of thedental remains are complete teeth.

A large majority are dental fragments made of enamel anddentine, but some are only enamel (Table 4). The size of thesefragments obviously varies greatly depending on the taxonomical

Table 4Distribution of material (tooth, dentine, bone).

Teeth Bone

Complete FragmentsEnamel Enamel & dentine769 1958

297 27273024 472total 3496

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173168

origin of the remains. Nevertheless these are only rarely larger than3 cm.

The majority of bone remains are small splinters which aredifficult to identify. As for the teeth, their size is often less than3 cm. The largest bone artefact found, a fragment of large mammallong bone diaphysis, is not bigger than 5.5 cm. Most of the bonesare complete or almost complete and not very diverse. For the mostpart, these are: petrus bone, phalanx, metapode and calcaneum(Table 5).

No traces of human-made alterations on the bone are present,whether from burning, cutting or percussion marks. Similarly, nomodifications due to the action of carnivores (teeth marks, perfo-rations, grooves, crenulated edges, or traces digested bone) werefound. Alterations of biological nature were found in abundance,however, and are due to gnawing. Typically, this gnawing is seen asflat to concave plurimillimetric to centimetric facets, bearinggroups of striations (Fig. 5). Small numbers of these striations aregrouped within shallow grooves, which are parallel to each otherand sometimes slightly curved. Grooves are several millimetreswide and their bottom is slightly concave in a transverse plane.

From the size and morphology of these grooves these could beidentified as gnawing marks made by a porcupine (Pei, 1938; Brain,1981). The list of fauna from the site includes two taxa which couldbe the culprits: Hystrix brachyura and Atherurus macrorus.

Almost all of the bone fossils (78%) display gnawing marks.Where exhaustive, gnawing causes modification of the bone shape.Two noteworthy morphologies are produced by exhaustivegnawing:

- small compact prisms of bone, the outside of which is formedexclusively of subvertical facets. These are 1–2 cm thick andapproximately the same width with a length shorter than 3 cm.These pieces can have been completely shaped by gnawing(Fig. 5b), but in some cases the non-altered bone in which theprism was sculpted can be seen at the bottom or at the topsurface. The ridges are arranged in a radial pattern. Some facetsno more than a few millimetres wide show evidence ofprevious gnawing.

- flat splinters with crenulated edges. Their width and thicknesscan reach several centimetres whereas their width is nevergreater than 0.5 cm. The facets forming the edge of the piecesare slightly inclined with respect to the flat plane. The ridgesare also arranged radially.

The teeth have also been gnawed. In general, the surface of thedentine has been nibbled at but occasionally the tooth has beenreduced to its crown. In rare cases, even the enamel has been

Table 5Proportion of the different bones.

Petrosal Phalanx Metapode Calcaneus

N 71 9 5 5% 15.4 1.9 1.1 1.1

attacked. These traces are often found in the area surrounding thefracture point suggesting that broken teeth were gnawed. Half ofthe teeth were gnawed, as well as 21% of the teeth fragmentswithout any obvious differences between the different categories oflarge mammals (Ungulates, Carnivores and Primates). The propor-tion of Rodent teeth that had been chewed was, however, very lowat only 2.2%.

Very few alterations were noted after the gnawing on the bonesor teeth. Thus, gnawing marks were not interrupted by fractures.

The modifications identified concern the surface patina of theremains. Most of the bones are light coloured and smooth. A smallfraction, less than 5% of the bone remains, however, is distinguishedby their brown colour and a polish which has also blurred thegnawing marks. Some of the teeth, especially molars of Suidae andseveral fragments of walls from teeth of Muntjac, also have thesame brown colouration and a smoothing out of the relief caused bygnawing.

The alteration pattern can be detailed by the observation onbone and teeth included in thin sections of sediments. The alter-ation intensity is variable and affected both bone and dentalremains. It is characterised by 1) cracking, 2) corrosion holes alongthe walls, 3) the development of brown spots in a dendriticnetwork and 4) the appearance of darkening zones with a loss ofbirefringence. No preferential alteration of bones versus teeth canbe recognized.

5.5. Site age

The age of the site is established by ESR dating. The samplelocations for the teeth are shown in Fig. 6 and elemental concen-trations in Table 6. Two model ages are calculated: an early-uptake(EU) age and a linear-uptake (LU) age, as is typical for ESR dating(e.g., Grun and McDermott, 1994). For early-uptake, it was assumedthat uranium entered the tooth immediately after burial, while inthe case of linear-uptake, it was assumed that uranium entered thetooth at a constant rate over time. The dose rates for the EU modelare given in Table 6b, and the calculated ESR ages are given in Table 7.Note that the EU and LU ages agree, for each sample, within theuncertainties, due to the very low uranium concentrations in thedental tissues. The dose rate is dominated by the gamma contribu-tion from the sediment (70–86% of the total dose rate); radioactiveequilibrium has been assumed. A �10% uncertainty was applied tothe measured gamma dose rate.

The LU ages range from 41.6� 4.9 ka and 37.6� 3.8 ka forBFS1-443A and BFS1-P2, respectively (level 7) to 15.9�1.4 ka forBFS1-G1-2. The site was therefore formed over an age range of atleast 19,000 to 32,000 years. However, the ages have been calcu-lated with the assumption that at least 30 cm (the approximaterange of gamma radiation from natural sources) of overburden wasdeposited immediately following burial of each tooth. Given theshallowness of the site and the discrepancy between the age ofBFS1-G1-2 and the older teeth, this assumption is not warranted.

An estimate of the effect that slowly accumulating sedimentwould have on the ESR ages was conducted. It is well known thata 30 cm radius of sediment provides approximately 90% of theinfinite matrix gamma dose rate (Brennan et al., 1997b). Numericalcalculations show that a 5 cm overburden (along with the

Indet. fragment> 3 cm Indet fragment< 3 cm Total

18 354 46239 76.6 100

Fig. 5. Gnawing marks. (a) gnawed root of molar of Sus cf scrofa; (b) gnawed diaphysis; (c) polygonal gnawed piece of bone.

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173 169

semi-infinite layer underneath the sample) provides approximately79% of infinite matrix dose rate. The percentage of the infinitematrix dose rate rises rapidly with the thickness of the overburden.At 10, 15, and 20 cm, the dose rate is 88%, 93%, and 96%, respectively.These results are essentially identical to those reported by Aitkenet al. (1985).

Although these calculations are crude, they illustrate animportant point: even a shallow planar layer of overburden resultsin nearly as much dose rate as an infinitely thick overlying layer ofsediment. If the average sediment thickness over time was half thatof the current thickness, the ages could be older by w8%. As theuncertainty in the calculated ages is approximately�10%, any errorin the age due to slow sediment accumulation will already beaccounted for. Therefore, no correction to the ages is necessary dueto the shallowness of the site.

5.6. Time averaging

The remains found by excavations were preferentially associ-ated with layers of coarse aggregates and stones, which indicatesthat they have filled in basins made by a burrowing animal. Thesedimentological study shows evidence that numerous beds weredug between which there were periods during which the den wasoccupied as shown by the development of a trampled surface. Thistype of functioning implies a long formation time. It is possible toestimate this time length by relating the number of remains whichmay be found at the site with the rate of introduction of remainsestablished by studies on modern porcupine dens. In the Cave ofthe Monk, about 3500 dental fragments were found in 5 m2. Thesurface of the lower chamber that has a surface structure withhollows and mounds, and thus functioned as a den, is slightly largerthan 50 m2. Thus, slightly less than one tenth of the deposit wasexcavated, and that a complete collection of all material would beabout 35,000 dental remains. The site studied by Brain (1981) givesan example of the time necessary to accumulate remains in the

cave. Teeth are introduced into the den in the form of pieces ofcranial mandible. The figure given by this author for the accumu-lation of these pieces is 80 per century. A minimalist hypothesis of 8teeth per cranial fragment leads to a figure of 640 teeth beingintroduced per century. Thus the time necessary to introduce35,000 dental remains in a den is more than 5000 years. Consid-ering that the cave was probably not occupied continuously, thetime necessary to form the site is in the order of one or several tensof thousands of years.

This estimate is consistent with the dating of the site. The ESRdating indicates a time range from 19 to 32 ka.

6. Discussion

6.1. Site formation

The particular nature of Ailuropoda–Stegodon assemblages inkarst context, with an under-representation of bones versus theabundance of dental elements, does not allow the consideration ofall the criteria proposed to characterize the responsible agent of theaccumulated assemblage. Hence, the criteria based on post-cranialelements to determine a Hyena den, such as the occurrence ofdigested bones or the under-representation of long bonesextremities (e.g. Marean, 1991; Brugal et al., 1997), cannot be usedin the paleontological sites from Southeast Asia.

The simplest criterion that can be used to identify a hyena den isthe Carnivores/Ungulates ratio proposed by Cruz-Uribe (1991).Usually, this ratio is calculated using the minimum number ofindividuals (MNI). Here, however, due to the very fragmentary stateof the collected remains, the MNI would have led to an underesti-mate of the ratio. The value for a hyena lair is above or equal to 20%.Instead, the number of taxonomically determined remains (Ntdr) atthe genus level was used. This does not modify the number ofCarnivores represented in the assemblage. In the Cave of the Monk,this ratio is 2.3% and does not significantly change from one area to

Fig. 6. Northwestern profile of the test-pit done in the end of the Low chamber (Area 2) with the paleontological assemblage units and the location of the sample gathered for ESRdating.

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173170

another in the chamber (Table 8). This result indicates that the sitecannot be qualified as a Hyena den which is additionally confirmedby the very low number of remains of this Carnivore and theabsence of Hyena cub milk teeth. In fact, the assemblage has onlyone unique Hyena tooth. Moreover, the lack of gnawing marks,digested remains and of whole or split up coprolites attributable tothis Carnivore, reinforces this interpretation.

On the basis of the same criteria, the other Carnivores identifiedon the site can be disqualified. Among Carnivores that can gatherbig mammal bones, only two Tiger dental fragments, and six Cuondental fragments were present out of a total of several thousandremains. Among Carnivores, Bear were the most significant withtwenty-seven determined remains, including sixteen attributed tothe Asian black bear.

Many agents can contribute to the accumulation or the alter-ation of remains and consequently disturb attempts to reconstruct

Table 6ESR analytical data (DE¼ equivalent dose; DL¼ detection limit, assumed zero). Neutrconcentrations for dental tissues and sediment samples. The uncertainty in the water co

Sample DE [Gy] Enamel Dentine

U [ppm] U [ppm]

BFS1-G1-2 16.16� 0.60 0.1� 0.1 2.64� 0.10BFS1-P2 36.06� 0.83 <DL 11.35� 0.10BFS1-443 38.09� 2.75 <DL 4.79� 0.10

the original environment (Lyman, 1994). From the regional point ofview it is interesting to observe the variability in finds of someremarkable species. Thus the absence of Hyenidae is noted in somesites, and suggests that the hyena lair hypothesis, held by manyauthors, is weak. The sites in the neighbourhood of Koloshan inSouthern China are a good example, even though this series is saidto be one of the richest in Asia (Young and Liu, 1950). Similarly, inthe Tham Khuyen cave in Vietnam (Cuong, 1992; Schwartz et al.,1994), and in all the sites of Lang Son, Tam Pa Loi and Houec Oi inLaos (Beden et al., 1972) or the Wuyun cave, Guangxi (Chen et al.,2002) no Hyena or Tiger were found. A close relative of the hyena,the cuon is considered as the potential bone gatherer in the study ofthe Cave of the Gigantopithecus (White, 1975). The Cuon waspresent at Ban Fa Suai, but this Carnivore is not systematicallyidentified in Southeast Asian paleontological sites; for example itwas absent at Tham Wiman Nakin (Tougard, 1998). The Bear is

on activation analysis was used to determine uranium, thorium, and potassiumntent assumed to be �5.0%.

Sediment Water content

U [ppm] Th [ppm] K [wt %] [% dry wt]

3.39� 0.10 7.00� 0.40 1.27� 0.03 35.4� 5.03.49� 0.10 9.80� 0.60 1.43� 0.04 24.8� 5.0n/a n/a n/a n/a

Table 7Early-uptake (EU) and linear-uptake (LU) ESR ages.

Sample EU [ka] LU [ka]

BFS1-G1-2 15.5� 1.4 15.9� 1.4BFS1-P2 36.4� 36 37.6� 3.8BFS1-443 41.0� 4.8 41.6� 4.9

Table 8Ratio of Carnivores/Ungulates.

Area 1 Area 2 Total in the cave

Ungulates 1206 1193 2570Carnivores 28 31 59Ratio Canivores/Ungulatesþ Carnivores 2.3% 2.5% 2.2%

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173 171

omnipresent in the sites and could theoretically be a potentialgatherer of Ailuropoda–Stegodon complex bones. Nevertheless, aswith other Carnivores, this taxon is often represented by only a fewremains which is too low compared to the amount usually found ina lair (Cruz-Uribe, 1991).

Studying an assemblage originating from a drugstore collection,Pei (1935: 425) argued that the abundance of suids remains asso-ciated with large mammals and shells (paludina sp.) excludea natural trap and implies human involvement in the constitution ofassemblages. von Koenigswald (1952: 299) also noted thatnumerous suid dental remains characterised the material found inChinese drugstores: ‘‘There were no horse, but many of the porcu-pine and Sus’’. The prevalence of boar was also observed at LangTrang, but the authors did not reach any firm conclusions regardingthis discovery (Long et al.,1996). In the Thai sites studied by Tougard(1998), the frequency of Sus was 13% at Tham Phra Kai Phet and 8% atTham Wiman Nakin. In both cases, Cervids and Bovines were morecommon, which is different from the Cave of the Monk where Suidsaccounted for one third of all the Artiodactyls (the most commontaxa found) collected. In an area at the back of the cave chamber,where the record is the most complete, the distribution of identifiedremains is the following: 0.5% Primates, 1.4% Carnivores, 1.3%Proboscidians, 6% Perissodactyls, 56.1% artiodactyls (including 34.1%Suidae, 21.5% Bovinae, 25.1% Cervinae and 19.2% Caprinae) and8.8% Rodents. However, this abundance of Suids remains cannot beimputed to a human action, in consideration to the absence ofanthropic marks (burned bones, cutmarks, and percussion marks)and to the lack of other human artefacts in the site. The besthypothesis to explain this abundance of Suids can be found in toothmorphology. Suids wear bunodont-shaped teeth that are moreresistant once roots are eaten than the selenodont-shaped teeth ofother Artiodactyls. These teeth can easily be split up in small frag-ments once the root is gnawed, making their collection in the fieldand their specific identification difficult.

6.2. Comparison to other sites of the Indochinese province

The paleontological criteria indicate that this site, as others inthe area, is not a carnivore accumulation. The sedimentologicalstudy concludes that an animal lair, most probably of a porcupine,was present. The taphonomic approach corroborates this inter-pretation because the principal actor who modified the bones wasindeed a porcupine. Next, the question arises as to the role of thisanimal in the formation of other sites of the Indochinese province.

The recognition that the site contains the deposits of a lairsuggests that the bones in the cave have a purely biological rela-tionship. Furthermore, this interpretation takes into account thefact that the bones were transported into the deep karst context.The porcupine is an animal renowned for accumulating faunalremains in its dens (Brain, 1981). Its contribution to the formationof paleontological sites has been especially studied in South Africaand the Middle-East (Alexander, 1956; Maguire, 1976; Maguireet al., 1980; Brain, 1981; Rabinovitch and Horowitz, 1994). Alex-ander (1956) observed evidence that objects as much as 2 kg hadbeen moved. Thus, all of the bone remains discovered at the site inthe lower chamber of the Cave of the Monk were accumulated bythis animal.

Studies describing present-day alterations to bone fossilscarried out in South Africa and the Middle-East, however, did notreport similar measurements to those found in Southeast Asia.None of these studies report an almost complete disappearance ofthe post-cranial skeleton. The example of the Cave of the Monk isnot unique, however. A number of authors have reported theoccurrence of gnawing on bone fossils from caves in the tropicalFar-East (eg. Patte, 1928; Pei, 1935, 1938; Bien and Chia, 1938; vonKoenigswald, 1938–39; Young and Liu, 1950; De Vos, 1984; Tougardand Ducrocq, 1999). Some of these authors argue that thedestruction of bones in these sites could be explained by a differ-ential dissolution (e.g. Tougard, 1998). This hypothesis cannot becorrect because some bones are preserved, even if reduced, andtheir surfaces do not shows dissolution marks. This is confirmed bymicroscopic observations of sediment thin sections, where analteration of bones and teeth is noted but is not different betweenthese two categories. The preservation of bone fragments is docu-mented in other Southeast Asian caves as well, and it is alsonoteworthy that the same prism shapes or crenelated splintersdescribed here were previously observed by Pei (1938) and Youngand Liu (1950) in paleontological material collected in caves inSouth China.

The complete transformation of bones by gnawing by a porcu-pine is thus not unique to the Cave of the Monk, but is a widespreadphenomenon throughout all of Indochina. No ethological study ofAsian porcupine can yet illustrate this phenomenon. One hypoth-esis explains this characteristic by the lack of available bone in thetropical forest environment compared to the semi-arid environ-ments, where the available present-day studies were carried out.This shortage would explain the increased consumption of bonescollected by porcupines in Southeast Asia.

6.3. Biochronology and ESR dating

Due to the lack of precise chronological data for most of thefaunal complex Ailuropoda–Stegodon, it is difficult to improve thedating. The remaining question concerns the definition of sucha complex. Ailuropoda, Pongo and Stegodon are not systematicallypresent at each site said to belong to this regional assemblage.Among 29 Chinese sites described by Kahlke (1961), 7 provide only2 taxa (Yenchingkuo, Hsiachungchiawan, Maba, and Shaochin) or 3taxa (Hoshantung cave, Hsinsuehchungtsun, and Newshuishan)among Ailuropoda–Pongo–Stegodon. At Tongzi and Panxian Dadong(Bakken et al., 2004) in China, and at Lang Trang (Long et al., 1996),Tham Om, Keo Leng, and Tham Khuyen (Cuong, 1992) in Vietnam,all of the taxa are present together but Pongo is lacking fromChinese sites such as Guanyindong, Xuetangliangzi, Longtandong,Gongwangling (Dong et al., 2000) and also from Maba and Yan-jinggou. In Tham Wiman Nakin (Tougard, 1998) in Thailand, Steg-odon is missing. In Vietnam, Ailuropoda is missing at Tham Om andTham Hai I and II (Cuong, 1992) and Stegodon is the only one of theassemblage at Ham Hum I. In Vietnam, Pongo is the only species atHam Hum II and there are none of these taxa at Ma U’Oi (Baconet al., 2004).

The chronological range of the complex Ailuropoda–Stegodon isquite wide as mentioned by Esposito et al. (2002). For Tham WimanNakin the spread of the ages is important: from 8 ka to 350 ka. In

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173172

China, the range of the dating record is also quite large. Ma andTang (1992) provide dating of about 8 ka for Stegodon in (Jinhua,Zhejiang), but in Huanglong cave (Hubei) the dating is 103�1.6 kaand 44�12.5 ka (Wu et al., 2006). According to the work of Bakkenet al. (2004), in Guanyindong where Ailuropoda and Stegodon areboth present, the ages range between 240 ka and 57� 3 ka. InTongzi, Pongo, Ailuropoda sp. and Stegodon orientalis date from113�11 ka to 181þ11�9 ka, and in Maba from 135 ka to 129 ka(Han and Xu, 1989). According to Dong et al. (2000), S. orientalis andAiluropoda have been dated to 150–190 ka in Longtandong. Thebreccia in the roof of the cave of Ma U’Oi (Bacon et al., 2006) hasbeen dated between 193�17 ka and 49� 4 ka but no specific taxonof the complex Ailuropoda–Stegodon is present. The Pleistocenefaunas that globally include the complex Ailuropoda–Stegodon aredifficult to subdivide based on the long temporal ranges of manytaxa and a reduced number of genera in comparison to faunas fromtemperate north China (Rink et al., 2008), but variability in thecomposition of the assemblages was pointed among the sites datedby these authors.

As has been shown here, the duration and the definition of thecomplex Ailuropoda–Pongo–Stegodon will need to be more strictlydefined. This is an important point because the duration for siteformation has two consequences on the significance of theassemblage collected. First, this time frame can be longer thanperiods of climatic fluctuation which could have led to changes inthe environment. Consequently, if the assemblage is considered asa homogeneous group, then species which actually lived in verydifferent biozones could be found associated in the same assem-blage. The identification of the porcupine as the main accumulationagent of paleontological sites in Southeast Asian caves supportsJ. De Vos’ hypothesis and indicates that the composed series cannotbe considered as a homogeneous reference set from bio-chronological and especially paleoenvironmental point of view.

This study answers the statement that there exists limitedinformation on taphonomic and paleoecological factors affectingdeposition and preservation. It also documents that the chrono-logical resolution of the assemblages is still too coarse to make morethan preliminary predictions about causes of faunal extinction inSoutheast Asia (Louys et al., 2007). The data gathered in the Cave ofthe Monk show an alternating series of two assemblages that replaceeach other through time (Fig. 6). This is not a unique case for Ailur-opoda–Stegodon complex. In China, Bakken et al. (2004) alreadyshow the evolution of four different assemblages defined beforea lower level older than 261�31 ka (LU) and after an upper layeryounger than 156�19 ka (LU) in Panxian Dadong. Nevertheless ifStegodon is present in the four groups defined in Panxian Dadong,this is not the case in the Cave of the Monk where Stegodon andElephas are never together. This fact offers new perspective on thesignificance of Pleistocene Southeast Asian assemblages.

The kind of site formation processes underlined at the Cave ofthe Monk can be assumed to be a dominant one in Southeast Asia.Consequently, the paleontological assemblages from this area donot provide a good data set for paleoenvironmental reconstructionor biochronological purposes when considered as homogeneousassemblages. On the other hand, considering the site formation andthe resulting deposit organisation shows that a microstratigraphicapproach can be realized and that this approach can provide a dataset with which accurate biochronology and paleoenvironmentalreconstruction can be attempted.

7. Conclusion

Both geological and taphonomic studies carried out on theinfilling and fossil material at the Cave of the Monk illustrate therigour needed to analyse the paleontological sites of Southeast Asia

for further ecological purposes. The nature of the Ailuropoda–Stegodon complex collected at Ban Fa Suai I could be specified, andby direct dating on tooth enamel we could describe the timing ofsite deposition. The deposition mechanism was identified, eventhough the number of remains was statistically weak. Indeed, datafrom the Cave of the Monk excavations are the most precise amongthe sites known in Southeast Asian, which are mostly fossiliferousbreccia. The data set and analyses show that the deposition of thepaleontological material required at least 19,000–32,000 yearsaccording to the ESR dating, and may have required even more. Thisdating gives a very young chronological range for the presence ofa real Ailuropoda–Stegodon assemblage, as Stegodon, Ailuropoda andPongo have been found in the site.

It is hoped that the painstaking excavation at the Cave of theMonk in Ban Fa Suai, which integrated taphonomy, geology and ESRdating, will illustrate the need for in-depth study of new paleon-tological cave sites in Southeast Asia.

References

Aitken, M.J., Clark, P.A., Gaffney, C.F., 1985. Beta and gamma gradients. NuclearTracks 10, 647–653.

Alexander, A.J., 1956. Bone carrying by porcupine. South African Journal of Science52, 257–258.

Bacon, A.-M., Demeter, F., Schuster, M., Long, V.T., Thuy, N.K., Antoine, P.-O., Sen, S.,Nga, H.H., Huong, N.T.M., 2004. The Pleistocene Ma U’Oi cave, northern Vietnam:palaeontology, sedimentology and paleoenvironments. Geobios 37, 305–314.

Bacon, A.-M., Demeter, F., Rousse, S., Long, V.T., Duringer, P., Antoine, P.-O.,Thuy, N.K., Mai, B.T., Huong, N.M., Dodo, Y., Matsumura, H., Schuster, M.,Anezaki, T., 2006. New palaeontological assemblage, sedimentological andchronological data from the Pleistocene Ma U’Oi cave (northern Vietnam).Palaeogeography, Palaeoclimatology, Palaeoecology 230, 280–298.

Bakken, D., Schepartz, L., Miller-Antonio, S., Yamei, H., Weiwen, H., 2004. Taxonomicabundance at Panxian Dadong, a middle Pleistocene cave in south China. AsianPerspectives 43, 334–359.

Bakken, D., 1997. Taphonomic parameters of Pleistocene hominid sites in China.Bulletin of the Indo-Pacific Prehistory Association 16 (3), 13–26.

Beden, M., Guerin, C., 1973. Le gisement de vertebres du Phnom Loang (Province deKampot, Cambodge). Faune pleistocene moyen terminal (Loangien). T.e.d.d.l’ORSTOM, 6–97. Paris.

Beden, M., Carbonnel, J.-P., Guerin, C., 1972. La faune du Phnom Loang (Cambodge):comparaison avec les faunes pleistocenes du Nord de l’Indochine. Archivesgeologiques du Vietnam 15, 113–122.

Behrensmeyer, A.K., 1988. Vertebrate preservation in fluvial Channel. Palae-ogeography, Palaeoclimatology, Palaeoecology 63, 183–199.

Bien, M.N., Chia, L.P., 1938. Cave and Rock-shelter deposits in Yunnan. Bulletin of theGeological Society of China 18, 325–347.

Brain, C., 1981. Porcupines as bone collectors in African caves. In: Brain, C. (Ed.), TheHunters or the Hunted? The University of Chicago Press, Chicago, pp. 109–117.

Brennan, B.J., Rink, W.J., McGuirl, E.I., Schwarcz, H.P., Prestwich, W.V., 1997a. Betadoses in tooth enamel by ‘‘one-group’’ theory and the ROSY ESR dating soft-ware. Radiation Measurements 27, 307–314.

Brennan, B.J., Schwarcz, H.P., Rink, W.J., 1997b. Simulation of the gamma radiationfield in lumpy environments. Radiation Measurements 27, 299–305.

Brown, P., 2001. Chinese Middle Pleistocene hominids and modern human origin inEast Asia. In: Barham, L., Robson Brown, K. (Eds.), Human Roots – Africa andAsia in the Middle Pleistocene. Western Academic & Specialist Publishers,Bristol, pp. 135–148.

Brugal, J.-P., Fosse, P., Guadelli, J.-L., 1997. Comparative study of bone assem-blages made by recent and Pleistocene hyenids. In: Hannus, L.A.,Rossum, L., Winham, R.P. (Eds.), Proceedings of the 1993 Bone ModificationConference, Hot Springs, South Dakota, vol. 1. Archaeological Laboratory,Sioux Falls, pp. 158–187. occasional publications.

Chen, G.J., Wang, W., Mo, J.Y., Huang, Z.T., Tiang, F., Huang, W.W., 2002. Pleistocenevertebrate fauna from Wuyun cave of Tiandong county, Guangxi. VertebrataPalasiatica 40, 42–251.

Ciochon, R., Long, V.T., Larick, R., Gonzales, L., Grun, R., De Vos, J., Yonge, C., Taylor, L.,Yoshida, A.H., Reagan, M., 1996. Dates co-occurence of Homo erectus andgigantopithecus from Tham Khuyen cave, Vietnam. Proceedings, NationalAcademy of Science U.S.A. 93, 3016–3020.

Cruz-Uribe, K., 1991. Distinguishing hyena from hominid bone accumulations.Journal of Field Archaeology 18, 467–486.

Cuong, N.L., 1992. A reconsideration of the chronology of hominid fossils in Viet-nam. In: Akazawa, T., Aoki, K., Kimura, T. (Eds.), The Evolution and Dispersal ofModern Human in Asia. Hokusen-Sha, pp. 321–335.

De Terra, H., 1938. Preliminary report on recent geological and archaeologicaldiscoveries relating to early man in Southest Asia. Proceedings of the NationalAcademy of Science U.S.A. 24 (10), 407–413.

V. Zeitoun et al. / Quaternary International 220 (2010) 160–173 173

De Vos, J., 1984. Reconsideration of Pleistocene cave faunas from South China andtheir relation to the faunas from Java. Courrier Forschungsinstitut Senckenberg69, 259–266.

Delange, P., 1997. Etude d’un karst nord thaılandais: le massif de Doi Chiang Dao.Speleologie 162, 49–53.

Dong, W., Jin, C., Xu, Q., Liu, J., Tong, H., Zheng, L., 2000. A comparative analysis onthe mammalian faunas associated with Homo erectus in China. Acta Anthro-pologica Sinica 19, 246–256.

Dunkley, J., 1995. The Caves of Thailand. Speleological Research Council Ltd., Sydney.Esposito, M., Reyss, J.L., Chaimanee, C., Jaeger, J.J., 2002. U-series dating of fossil

teeth and carbonates from snake Cave, Thailand. Journal of ArchaeologicalScience 29, 341–349.

Fosse, P., Jauzion, G., Maksud, F., Quettier, D., Quettier, R., Rouch, P., Besson, J.-P.,2001. Ursides pleistocenes des Pyrenees: elements de paleontologie et depaleobiologie. Bulletin de la Societe Prehistorique Ariege-Pyrenees 56, 103–138.

Fromaget, J., 1936. Sur la stratigraphie des formations carbonates de la chaıneannamitique septentrionale et sur l’existence de l’homme dans le Quaternaireinferieur de cette partie de l’Indochine. Comptes Rendues. Academie desSciences Paris 203, 738–741.

Ginsburg, L., Ingavat, R., Sen, S., 1982. A middle Pleistocene (Loangian) cave fauna innorthern Thailand. Comptes Rendues, Academie des Sciences Paris 294, 295–297.

Goldberg, P., Whitbread, I., 1991. Micromorphology study of a Bedouin tent floor. In:Goldberg, P., Nash, D.T., Petraglia, M.D. (Eds.), Formation Processes in Archae-ological Context. Prehistory Press, Madison, pp. 165–188.

Goldberg, P., 2001. Some micromorphological aspects of prehistoric cave deposits.Cahiers d’Archeologie du CELAT. Serie Archeometrie 10, 161–175.

Granger, W., 1938. Medicine bones. Natural History 42, 264–271.Grun, R., McDermott, F., 1994. Open system modelling for U-series and ESR dating of

teeth. Quaternary Geochronology (Quaternary Science Reviews) 13, 121–125.Han, D., Xu, C.,1989. Quaternary mammalian faunas in south China. In: Wu, R., Wu, X.,

Zhang, S. (Eds.), Early Humankind in China. Science Press, Beijing, pp. 338–391.Hill, C., Forti, P., 1997. Cave Minerals of the World, second ed. National Speleological

Society, Huntsville.Kahlke, H.D., 1961. On the complex of the stegodon–ailuropoda fauna of southern

China and the chronological position of Gigantopithecus blacki V. Koenigswald.Vertebrata PalAsiatica 2, 83–108.

Karkanas, P., Bar-Yosef, O., Goldberg, P., Weiner, S., 2000. Diagenesis in prehistoriccaves: the use of minerals that form in situ to assess the completeness of thearchaeological record. Journal of Archaeological Science 27, 915–929.

Karkanas, P., 2001. Site formation processes in Theopetra Cave: a record of climaticchange during the late Pleistocene and early Holocene in site formationprocesses in Theopetra cave. Geoarchaeology 16, 373–399.

Kiernan, K., 1991. Tropical mountain geomorphology and landscape evolution innorth-west Thailand. Zeitschrift fur Geomorphologie 35, 187–206.

von Koenigswald, G., 1938–1939. The relationship between the fossil mammalianfaunae of Java and China, with special reference to early man. Peking NaturalHistory Bulletin 13, 293–298.

von Koenigswald, G., 1952. Gigantopithecus blacki von Koenigswald, a giant fossilhominid from the Pleistocene of South China. American Museum of NaturalHistory, Anthropological Papers 43, 295–325.

Kurten, B., 1956. The status affinities of Hyaena sinensis Owen and Hyaena ultimaMatsumoto. American Museum of Natural History, Novitiates 1764, 1–48.

Lenoble, A., 2001. Etude geoarcheologique. In: Beauval, C. (Ed.), Les Rochers deVilleneuve, rapport de fouille programme. SRA Poitou-Charentes, pp. 17–27.

Long, V.T., De Vos, J., Ciochon, R.S., 1996. A fossil mammal fauna of Vietnam (LangTrang caves), compared with fossil and recent mammal faunas of southeastAsia: their geographical implications. Bulletin of the Indo-Pacific PrehistoryAssociation 14, 101–109.

Louys, J., Curnoe, D., Tong, H., 2007. Characteristics of Pleistocene megafaunaextinctions in southeast Asia. Palaeogeography, Palaeoclimatology, Palae-oecology 243, 152–173.

Lyman, R.L., 1994. Vertebrate Taphonomy. Cambridge Manuals in Archaeology.Cambridge University Press, Cambridge, 524 pp.

Ma, Ancheng, Tang, Huliang, 1992. On discovery and significance of Holocene ailur-opoda–stegodon fauna from Jinhua, Zhejiang. Vertebra PalAsiatica 30, 295–312.

Maguire, J.M., Pemberton, D., Collett, M.H., 1980. The Makapansgat limework greybreccia: hominids, hyaenas, hystricids or hillwash? Palaeontologica Africa 23,75–98.

Maguire, J.M., 1976. A taxonomic and ecological study of the living fossil Hystricidaewith particular reference to Southern Africa. PhD thesis, University of theWitwatersand, Department of Geology, pp. 530–655.

Marean, C.W., 1991. Measuring the post-depositional destruction of bones inarchaeological assemblages. Journal of Archaeological Science 18, 677–694.

Matthew, W., Granger, W., 1923. New fossil mammals from the Pliocene ofSzechuan, China. Bulletin, American Museum of Natural History 48, 563–598.

Munsell Inc., 1954. Soil Color Chart. Munsell Color Company Inc., Baltimore.Patte, E., 1928. Comparaison des faunes de mammiferes de Langson (Tonkin) et du

SE Tchouen. Bulletin de la Societe Geologique Française 28, 55–63.Pei, W., 1935. Fossils mammals from the Kwangsi caves. Bulletin of the Geological

Society of China 14, 413–425.Pei, W., 1938. Le role des animaux et des causes naturelles dans la cassure des os.

Palaeontologica Sinica 118, 1–61.Pei, W., 1957. The zoogeographical divisions of Quaternary mammalian faunas in

China. Vertebrata Palasiatica 1, 9–24.Pope, G., Frayer, D., Liangcharoen, M., Kulasing, P., Nakabunlung, S., 1981.

Palaeoanthropological investigations of the Thai-American expedition innorthern Thailand (1978–1980): an interim report. Asian Perspectives 21,147–163.

Prescott, J.R., Hutton, J.T., 1994. Cosmic ray contributions to dose rates for lumi-nescence and ESR dating: large depths and long-term time variations. RadiationMeasurements 23, 497–500.

Rabinovitch, R., Horwitz, L., 1994. An experimental approach to the study ofporcupine damage to bones: a gnawing issue. Artefacts 9, 97–118.

Rink, W.J., Schwarcz, H.P., Grun, R., YalçinkayaTaskiran, H., Otte, M., Valladas, H.,Mercier, N., Bar-Yosef, O., Kozlowki, J., 1994. ESR dating of the last interglacialmousterian at Karaın Cave, southern Turkey. Journal of Archaeological Science21, 839–849.

Rink, W.J., Wei, W., Beken, D., Jones, H.L., 2008. ESR geochronology of Ailuropoda–Stegodon fauna and Gigantopithecus in Guangxi province, southern China.Quaternary Research 69, 377–387.

Schepartz, L.A., Stoutamire, S., Bakken, D.A., 2001. Taphonomy of Stegodon orientalisat Panxian Dadong, a Middle Pleistocene Site in Guizhou, South China. TheWorld of Elephants. International Congress, Rome, pp. 243–246.

Schepartz, L., Bakken, D., Miller-Antonio, S., Paraso, C., Karkanas, P., 2003. Faunalapproaches to site formation processes at Panxian Dadong cave, Guizhou,China. In: Chen, Shen, Keates, S. (Eds.), Current Research in Chinese PleistoceneArchaeology. British Archeology Research Monographs, pp. 70–94.

Schlosser, 1903. M. Schlosser, Die fossilen Saugethiere Chinas nebst einer Odon-tographie der recenteen Antilopen. Abhandlingen Bayerische Akademie Wis-senschaft 22, 1–221.

Schwartz, J., Long, V.T., Nguyen, L.C., Kha, L.T., Tattersall, I., 1994. A diverse hominoidfauna from the late middle Pleistocene breccia cave of Tham Khuyen, SocialistRepublic of Vietnam. Anthropological Papers of the American Museum ofNatural History 73, 2–11.

Simons, E., Ettel, P., 1970. Gigantopithecus. Scientific American 222, 76–85.Tougard, C., Ducrocq, S., 1999. Abnormal fossil upper molar of Pongo from Thailand

Quaternary climatic changes in Southeast Asia as a possible cause. InternationalJournal of Primatology 20, 599–607.

Tougard, C., Montuire, S., 2006. Pleistocene paleoenvironmental reconstructionsand mammalian evolution in South-East Asia: focus on fossil faunas fromThailand. Quaternary Science Reviews 25, 126–141.

Tougard, C., 1998. Les faunes de grands mammiferes du Pleistocene moyen terminalde Thaılande dans leur cadre phylogenetique, paleoecologique et bio-chronologique. These de dosctorat, universite de Montpellier.

Voorhies, M., 1969. Taphonomy and Population Dynamics of an Early PlioceneVertebrate Fauna, Knox County, Nebraska. In: Contributions to Geology SpecialPaper, vol. 1. University of Wyoming, Laramie.

Wanpo, H., Ciochon, R., Yumin, G., Larick, R., Qiren, F., Schwarcz, H., Yonge, C., DeVos, J., Rink, W., 1995. Early Homo and associated artefacts from Asia. Nature378, 275–278.

White, T.D., 1975. Geomorphology to palaeecology: Gigantopithecus reappraised.Journal of Human Evolution 4, 219–233.

Wu, Xianzhu, Liu, Wu, Gao, Xing, Yin, Gongming, 2006. Huanglong cave, a new latePleistocene hominid site in Hubei Province, China. Chinese Science Bulletin 51,2493–2499.

Young, C.C., Liu, P.T., 1950. On the mammalian fauna at Koloshan near Chingking,Schezuan. Bulletin of the Geological Society of China 30, 43–90.

Zeitoun, V., Seveau, A., Forestier, H., Thimas, H., Lenoble, A., Laudet, F., Aantoine, P.-O., Debruyne, R., Ginsburg, L., Mein, P., Winayalai, C., Chumdee, N., Doy Asa, T.,Kijngam, A., Nakbunlung, S., 2005. Decouverte d’un assemblage faunique aStegogon–Ailuropoda dans une grotte du Nord de la Thaılande (Ban Fa Suai,Chiang Dao). Comptes Rendues Palevol 4, 255–264.