Records of the Western Australian Museum Supplement No. 57: 307-315 (1999).

Environmental change during the Quaternary in East Asiaand its consequences for mammals

Nina G. Jablonski 1 and Matthew J. Whitfort 2

1 Department of Anthropology, California Academy of Sciences, Golden Gate Park,San Francisco, CA 94118-4599, USA; email: njablonski@casnotes1.calacademy.org

2 Department of Anatomy and Human Biology, University of Western Australia,Nedlands, WA 6907, Australia

Abstract - Environmental changes during the Quaternary in eastern Asia hadprofound impacts on the distributions of mammals. Critical for many taxawere the southward latitudinal shifts of the tropical and subtropical zones,and decreases in the areal extents of these zones. In this study, a geographicinformation system (GIS) has been used to examine the spatial responses ofmammalian genera to environmental changes of the Pleistocene of China.Conditions of increasing seasonality during the Pleistocene favoured theevolution of large and well-insulated morphs in several mammalian families(e.g., Rhinocerotidae, Tapindae, Elephantidae), but this strategy was notsuccessful (in terms of long-term survival) in many cases.

The spatial responses of members of a single clade, the Catarrhini or OldWorld higher primates, are of particular interest. The main catarrhine generaof eastern Asia responded individually to conditions of increasing seasonalityduring the Pleistocene, and these responses reflected the life historyparameters of the genera involved. Apes were found to be the most sensitiveto conditions of increased seasonality, with large apes (Fongo, Gigantopithecus)retreating in advance of gibbons (Hylobates). Monkeys retreated southwardmore slowly and to a lesser extent than apes because of their abilities tosurvive in more highly seasonal environments and to produce offspring onrelatively abbreviated reproductive schedules.

These results provide a new basis for predicting the responses of primatesand other mammals to future environmental change.

INTRODUCTION

The record of environmental change during thePleistocene is rich and is yielding increasingamounts of detailed information, especially as newsources of data, from columns of continuousdeposition such as loess sequences and ice cores,are fully explored. These improvements in thepalaeoenvironmental record have beenenthusiastically welcomed by palaeontologistsbecause it is against these records that organismalevolution can be fully understood. Climate changeplaces myriad stresses on organisms, which differin their forcing response to these stresses(FAUNMAP Working Group 1996; Parsons 1993a,1993b; Roy et al. 1996; Webb and Bartlein 1992). It isnow widely recognized that natural populationsrespond to climatic change by latitudinal shifts inabundances and/ or geographic range boundaries(Huntley and Webb 1989; Roy et al. 1996; Webb andBartlein 1992). The sometimes extreme climaticfluctuations of the Pleistocene tended to destabilizespecies interactions within communities, andspecies tended to adjust themselves individually to

such stresses. Pleistocene communities were not justgeographically displaced versions of modem ones,they were fundamentally different in theircomposition because of the differential responses ofspecies to environmental change; hence their nameof 'non-analogue' communities (FAUNMAPWorking Group 1996).

Environmental changes during the Pleistocene inEast Asia were, to an extent, more extreme than inother continents because the local climatic effects ofthe Qinghai-Xizang (Tibetan) Plateau tended tomagnify, for much of eastern Asia, the orbitallyinduced climatic fluctuations associated withglacials and interglacials worldwide (Webb andBartlein 1992). Among the most dramatic of thesechanges was the abrupt increase, at about theGauss/Matuyama boundary, in aeolian dustdeposition occurring on the Loess Plateau due to anapparent northward shift and intensification of theSiberian High (Ding et al. 1997).

In this study, a geographic information system(GIS) has been used to visually superimposeand analyse data concerning environments and

308

Late Pleistocene

N.G. Jablonski, M.J. Whitfort

Middle Pleistocene

Holocene

permafrost

cold temperate

temperate

LegendIII

•~

warm temperate

[8l subtropical

m tropical

~ arid desert

B plateau

D nodata

- paleoshoreline

Figure 1 The major environmental zones of China in the Early, Middle, Late Pleistocene and Holocene.

distributions of mammals in the past. Ourpalaeoenvironmental data have been compiledfrom several sources and are based mainly onthe distribution of vegetation types (as judgedmostly from pollen records), soils, permafrostand loess (An et al. 1991; Cui and Song 1991; Li1991; Winkler and Wang 1993; Zhou et al. 1991).Our attention in this study was concentrated onChina because it displays the highest densityand most even distribution of

palaeoenvironmental and palaeontological data.The boundary dates used for the subdivisionsof the Pleistocene are those applied by mostChinese geoscientists and by an increasingnumber of stratigraphers elsewhere, whorecognize the Gauss-Matuyama palaeomagneticreversal at 2.5 Myr as the Plio-Pleistoceneboundary (Ding et al. 1997; Partridge 1997). TheEarly Pleistocene is then taken as 2.5 Myr to780,000 yr, the Middle Pleistocene from 780,000

Mammals of the East Asian Quaternary

to 128,000 yr, and the Late Pleistocene from128,000 to 11,000 yr. The Holocene is taken asthe span from 11,000 yr to the present.

ENVIRONMENTAL BACKGROUND

The Early Pleistocene of China was characterizedby generally warmer and more humidenvironmental conditions of greater homogeneitythan those that obtained later in the Pleistocene orin the Holocene. At this time (Figure 1), thesubtropical zone covered a broad east-westexpanse (including the area over the proto-TibetanPlateau) and there existed a large tropical zonesouth of it. There is evidence of permafrost overthe Tibetan Plateau and in northeastern China bythe later part of the Early Pleistocene (Zhou et al.1991). By the Middle Pleistocene, the subtropicaland tropical zones had shifted south- andeastward, and the Tibetan Plateau had emerged asa dis tinct environmen tal zone. By the LatePleistocene the subtropical and tropical zones hadmigrated even farther southward and wereconsiderably reduced in area, the Loess Plateau(see Figure 3) had increased in size due to morewidespread aeolian deposition of loess under coldand dry conditions, and permafrost covered muchof Tibet and the northeast. A broad area ofcontinental shelf was also exposed during much ofthis time, permitting extensive exchanges ofmammals between the mainland and the islandenvironments of Hainan and Taiwan. Conditionsat the Last Glacial Maximum (not depicted inFigure 1) were even more extreme, withelimination of the tropical zone entirely from theChinese mainland.

For mammals, marked increases inenvironmental seasonality at all latitudes,increasing environmental heterogeneity andfragmentation, an increasing potential for physicalisolation of populations as a result of habitatfragmentation, and changes in the configuration ofbiogeographic corridors were the most importantconsequences of Pleistocene environmentalchanges (Ferguson 1993; Jablonski 1993). Theeffects of this reorganization were pronounced. Thereduction in the size of the subtropical zone wasmarked. This zone, which had extended across thebreadth of the continent in the Tertiary, lost itswestern flank and was shifted southward over thecourse of the Pleistocene, with predictable changesin mammalian distributions. This trend becamedramatically exaggerated during the climaticfluctuations of the Late Pleistocene, when itappears that the zone was entirelyeliminated from China In connection withbiogeographic of new

309

distribution was another trend which characterizedthe East Asian Pleistocene. The most famousexample here is that of the Qin Ling Mountains,lying directly to the south of the Loess Plateau.Uplift of these mountains from the mid-Pleistoceneonward put an end to the north-south exchange offaunas that had occurred up to that time. Thisbarrier not only prevented the southwardmigration of thermophilic species during glacialphases, but also prevented the northwardmigration of cryophilic species during interglacialsand at the end of the Pleistocene. A further specificimportant consequence for mammals of thedramatic environmental changes of the lateTertiary and Quaternary was the creation of nichesin which archaic taxa could survive and newspecies could evolve. The most important of thesesites for mammals was the Heng Duan Mountainsin northeastern Yunnan that, with their deeplydissected north-south trending valleys, provided alast resting place for palaeoendemics and a cradlefor diverse neoendemics. The final importantconsequence - related to the last - was that theTibetan Plateau itself and adjacent mountainranges became a refuge for cold-loving species.These areas acted as a staging grounds of sorts forsuch organisms to invade lower-lying areas innorthern China during the colder phases of theQuaternary and to return to during theinterglacials. This last point had as importantconsequences for plants as for animals, because itcreated numerous local sources of montane plantsand animals from which cryophilic species couldemigrate during glacials and retreat to duringinterglacials (Ferguson 1993).

The Holocene amelioration, peaking at 6000 yrBP, saw a major warming and northward re­expansion of the tropical and subtropical zones,the emergence of large tracts of arid desert northand northeast of the Tibetan Plateau, and a marinetransgression.

Before we can examine changes in thedistributions of mammals against this backdrop, itis important to recognize that there are severalfactors that affect the potential accuracy of ourinformation on the geographical distribution offossil taxa, and that will influence the biologicalconclusions drawn from observed changes inanimal distributions through time. Collecting andsampling biases are potentially very significant. Inour choice of the relatively well-surveyed countryof China as the focus of this study, we havereduced, but not eliminated, these influences. Alsopotentially important, are various taphonomicbiases such as 'non-preserving' or poor-preservingenvironments such as tropical forests, selectivedestruction of species or age classes after death,and mortem removal of animals from the life

(Roy et al. 1996).

310

• Rhinoceros (9) + Dlcerorhinus (6) X Coe/odonta (7)

N.G. Jablonski, M.J. Whitfort

Middle Pleistocene

• Rhinoceros (25) + Dlcerorhinus (10) X Coe/odonta (6)

• Rhinoceros (19) + Dlcerorhinus (4) X Coelodonta (30) • Rhinoceros (2) + Dlcerorhlnus (1) X Coe/odonta (1)

Figure 2 The distributions of Rhinoceros, Dicerorhinus and Coelodonta in the Early, Middle, Late Pleistocene andHolocene.

CHANGES IN THE DISTRIBUTIONS OFSELECTED GENERA OF MAMMALS

The environmental changes of the Pleistoceneare correlated with a series of profound changesin the diversity and distribution of mammals inChina. Interesting individual stories can be toldfor all genera, but we shall concentrate here on afew illustrative genera of larger, hoofed mammalsand the genera of higher primates. Data on thedistributions of fossil mammals were drawn fromthe Eurasian Fossil Mammal Database, whichcomprises point data for the distribution in timeand space for fossil mammal species in Eurasiaduring the Cenozoic. The database consists, inpart, of the location, and geological andnumerical age of mammalian fossils. One record

in the database consists of the occurrence, at onespecific time and place, of one species, and is thusbest referred to as a 'species occurrence'. Thenumber of actual specimens of a species thatoccurred at a specific time and place did notinfluence the number of records; the database wasdesigned only to record the occurrence of aparticular species at a particular place at aparticular time.

The distributions of several large mammalsshifted southward and appear to have becomeless dense during the Late Pleistocene. Lookingat the genera of one family, the Rhinocerotidae,we see this pattern followed by Rhinoceros, thegenus representing the modem Javan rhinoceros(Figure 2). Dicerorhinus, another thermophilic

Mammals of the East Asian Quaternary

• Equus (24)

Late PleistoceneI' --~_.~.

I/////

o· ,

• Equus (42)

Middle Pleistocener+ I r--+-

I

• Equus (17)

Holocene

• Equus (1)

311

Figure 3 The distributions of EqulIs in the Early, Middle, Late Pleistocene and Holocene plotted relative to loessdistribution (vertical hachures) and glacial cover (oblique hachures). Note the strong and consistentassociation of Equus with the grassland environment of the Loess Plateau.

genus of the Rhinocerotidae, met a different fate:its southward migration during the Pleistoceneappears to have been blocked by the newlyformed Qin Ling Mountains. With its exitblocked, the genus became extinct in China andsurvived only in those tropical parts ofSundaland, such as Sumatra, where it hadpreviously established itself (Croves 1983;Jablonski unpublished data). (This effect may, inreality, have been less pronounced than itappears. According to c.P. Croves [pers. comm.1998], Dicerorlzinus is almost certainly para­phyletic and one of the northerly speciesrepresenting the genus in Figure 2, "D."choukoutienensis, is probably more closely related

to Coelodonta.) As was the case in manymammalian lineages, the rhinocerotid genus whichmost successfully withstood the increasinglyintense cold, dry conditions and the heightenedseasonality of the later Pleistocene was the giant,well-insulated Coelodonta, the woolly rhinoceros.For this genus and other cryophils, their numberof occurrences peaked in the Late Pleistocenebefore they disappeared abruptly from the faunalrecord in the early Holocene.

Some of the real benefactors of the environmentalchanges of the Pleistocene of East Asia were thetrue horses and gazelles which could successfullyexploit the newly evolved grassland and steppeniches of the Tibetan and Loess Plateaux (Figure 3).

312

Early Pleistocene

• Mscaca (11) + ProcynocephaJus (4) X RhInoplthecvs (3)

Late Pleistocene

• Mscaca (32) + ProcynocephaJus (0) X RhInopithscvs (5)

N.G. Jablonski, M.J. Whitfort

Middle Pleistocene

• Mscaca (23) + ProcynocephaJus (0) X Rhinoplthecvs (3)

Holocene

,

• Mscaca (12) + ProcynocephaJus (0) X Rhinoplthecvs (1)

Figure 4 The distributions of Macaca, Procynocephalus, and Rhinopithecus in the Early, Middle, Late Pleistocene andHolocene.

This is best illustrated by the genus Equus, which,by Late Pleistocene times, was confined mostly tothe Loess Plateau, an area now recognized to havebeen consistently covered with grasses and herbsfor at least the last 100,000 yr, and perhaps longer(Sun et al. 1997).

The catarrhine or higher Old World primatespresent an interesting set of case historiesconcerning the effects of Pleistocene environmentalchange, because the entire assemblage comprisesboth relatively stenotopic and eurytopic genera andis considered part of China's so-called OrientalFauna. The Old World monkeys of the generaMacaca, Procynocepllalus and Rllinopitllecus togethershowed a noticeable south- and eastward shiftduring the Pleistocene, mirroring the Rllinoceros

pattern quite closely (Figure 4). Apes reactedsomewhat differently (Figure 5). The so-calledlesser apes, the gibbons (genus Hylobates),remained quite abundant in the tropical zonethroughout the Pleistocene, while the large apes,the orangutan (genus Pongo) and Gigantopitllecus,appeared to retreat southward earlier in thePleistocene, and were eliminated from China bythe end of the Pleistocene. If number of occurrencescan be counted as some measure of evolutionarysuccess, the smallest-bodied ape genus, Hylobates,was the most successful of all, with distributionsmaintained in the southern portion of thesubtropical zone through the Late Pleistocene, andsurviving the Last Glacial Maximum into theHolocene.

Mammals of the East Asian Quaternary

Early Pleistocene

• Hylobates (2) + Pongo (1) X Gigantopithecus (4)

Late Pleistocene

• Hylobates (32) + Pongo (7) x Gigantopithecus (0)

Middle Pleistocene

• Hylobates (8) + Pongo (8) X Gigantoplthecus (3)

• Hylobates (24) + Pongo (0) X Gigantopithecus (0)

313

Figure 5 Distributions of Hylobates, Pongo and GigantopithecJls in the Early, Middle, Late Pleistocene and Holocene.

The changing distributions of the cercopithecidgenera Macaca and Rhinopithecus show a fewsignificant departures from those of the apes. Firstly,the distributions of the monkey genera extendfarther northward than those of the apes for all timeperiods. Interestingly, however, the distribution ofmonkeys only extends minimally north of thesubtropical zone at any time. Secondly, the degreeof southward compression of distributions duringthe Late Pleistocene and Holocene would appear tobe less pronounced for monkeys than for apes. Inthis connection, the contrast between the changingpatterns of distribution of Macaca and Hylobatesthrough time is particularly given that thetwo genera are of (approxill1ate

DISCUSSIONThe few examples considered here suggest that

mammalian taxa responded individually to thesevere environmental fluctuations of thePleistocene. This is illustrated most clearly by thecatarrhine primate genera, whose distributionsshifted at different times and at different rates inresponse to changes in Quaternary environments.This finding is in accord with that of theFAUNMAP Working Group (1996), who showedthat mammalian species responded individually tolate Quaternary climatic fluctuations in NorthAmerica. This result also helps to account forobservations of living faunas in North America andAustralia that indicate that species and

are individualistic in spatial

314

distribution and that communities are notrestricted to specific combinations of species(Brown and Kurzius 1989; Morton et al. 1994). Theconclusion that mammalian communities arecontinually and unpredictably emergent, is oneshared by the present study.

The evolution of larger body sizes has occurredin many lineages of mammals living in seasonal orunpredictable climates through time and was oneof the most common strategies employed bymammals to contend with the climatic fluctuationsof the Pleistocene (Owen-Smith 1988; Zeveloff andBoyce 1988). A larger body is more efficient it itsutilization of food energy because its smallersurface area relative to volume makes possible lessloss of energy through dissipation of heat at thebody's surface. Larger herbivores tend to eat lessfood or lower quality food per day, as a proportionof body mass, than small herbivores, and largeanimals lose condition more slowly on asubmaintenance diet than do smaller animals(Owen-Smith 1988). Finally, stored fat reservesbecome a greater fraction of body mass as sizeincreases. These factors helped mammals to copewith extreme seasonal fluctuations in foodavailability such as those that occurred during thePleistocene.

The differences in patterns of response toenvironmental change demonstrated by thevarious primate genera are of great interestbecause they can be traced to known aspects of thebiology of those forms. It is now widely recognizedthat changes in patterns of climatic seasonalityhave profound effects on animals and especially onthose mammals like primates with long generationtimes and energy- and time-intensive modes ofraising offspring. Among the most interestingfindings of this study is that, for apes, evolution oflarger body sizes was not a key to evolutionarysuccess as it was, at least temporarily, for manyother mammalian lineages. The life historyparameters of apes evolved under the relativelypredictable and stable climatic regimes of theMiocene and are exquisitely suited to them (Kelley1997). The animals' life spans are long, theirgestation times are long, their interbirth intervalsare long, and females begin to reproduce at arelatively advanced age. Apes also possess verylarge brains relative to their body size, and suchmetabolically costly organs require nutrient- andenergy-rich diets in order to develop and bemaintained. The absolutely greater metabolicdemands of large body size, coupled with thedemands of a large brain, proved too great forGigantopithecus and, to a lesser extent, for Pongounder conditions of increasing seasonality duringthe. PI~istocene. The orangutan managed to escapeextmcbon, maintaining its distribution in relativelyunseasonal equatorial tropical forests. The

N.G. ]ablonski, M.]. Whitfort

orangutan's strategy has proven not to be aneffective formula for survival. With its reliance onregular supplies of high-energy tropical foodsources, the species was (and still is) at risk. Thisrisk has been exacerbated by the species' longgestation period and very long interbirth intervalsthat made it impossible for it to survive in areaswith anything but predictable environments withrelatively little seasonal fluctuation in availabilityof easily-digested, energy-rich foods (Jablonski1997).

The relatively greater success of the monkeysduring the Pleistocene had to do with their abilitiesto exploit a wide variety of plant foods and toproduce offspring more quickly, because of shortergestation times and shorter interbirth intervals.

In the rapidly fluctuating environments of thelatest Pleistocene and early Holocene in Asia, theresponse of catarrhine genera was linked to acomplex constellation of factors including nichebreadth, gestation time, length of interbirthinterval, relative brain size, and other factors stillto be elucidated. As we learn more about the lifehistory parameters of other mammalian genera, amore complete understanding of their reactions toenvironmental change will also be forthcoming.This knowledge will significantly aid our abilitiesto predict relative levels of species vulnerability inresponse to the environmental fluctuations of thefuture.

ACKNOWLEDGEMENTS

This research was funded through the Large andSmall Grant Schemes of the Australian ResearchCouncil, to which we extend deep thanks. NolaRoberts-Smith and Paul Ottaviano are thanked fortheir assistance in developing GIS coveragesduring early phases of this project. Jane Bailey isacknowledged for her expert maintenance of theEurasian Fossil Mammal Database, and Xu Qinqifor his critical examination of the database fortaxonomic and other errors. The authors also wishto thank Colin Groves for his constructivesuggestions and for information on thedistributions of past and present species ofrhinoceroses.

REFERENCESAn, Z.s., Wu, X.H., Lu, Yc., Zhang, D.E., Sun, X.J., Song,

G.R. and Wang, W.M. (1991). Palaeoenvironmentalchanges of China during the last 18,000 years. In Uu,TS. (ed.), Quaternary geology and environment in China:228-236, Science Press, Beijing, China.

Brown, J.H. and Kurzius, M. (1989). Spatial and temporalvariation in guilds of North American granivorousdesert rodents. In Morris, D.W., Abramsky, Z., Fox,B.J. and Willig, M.R. (eds), Patterns in the structure of

Mammals of the East Asian Quaternary

mammalian commU/lltles: 71-90, Special Publications,Museum, Texas Tech University, No. 28.

Cui, Z.J. and Song, CQ. (1991). Quaternary periglacialenvironment in China. In Liu, TS (ed.), Quaternary

and environment in ClllIlIl: 78-85, SciencePress, Beijing, China.

Ding, Z.L, Rutter, N.W. and Liu, TS (1997). The onsetof extensive loess deposition around the GIMboundary in China and its palaeoclimaticimplications. Quaternary lnternatlonal 40: 53-60.

FAUNMAP Working Group (1996). Spatial response ofmammals to late Quaternary environmentalfluctuations. Science 272: 1601-1606.

Ferguson, D. (1993). The impact of late Cenozoicenvironmental changes in East Asia on thedistribution of terrestrial plants and animals. InJablonski, N.G. (ed.), Evolving landscapes and evolvingbiotas of East Asia since the mid-Tertiary: 145-196,Centre of Asian Studies (University of Hong Kong),Hong Kong.

Groves, c.P. (1983). Phylogeny of the living species ofRhinoceros. Zeitschrift flir zoologisches Systematik undEvolutionforscllllng 21: 293-313.

Huntley, B. and Webb, T. 1Il (1989). Migration: species'response to climatic variations caused by changes inthe Earth's orbit. Journal of Biogeography 16: 5-19.

Jablonski, N.G. (1993). Quaternary environments and theevolution of primates in East Asia, with notes ontwo new specimens of fossil Cercopithecidae fromChina. Folia Primatologica 60: 118--132.

Jablonski, N.G. (1997). The relevance of environmentalchange and primate life histories to the problem ofhominid evolution in East Asia. In Jablonski, N.G.(ed.), The changing face of East Asia during the Tertiaryand Quaternary: 461-474, Centre of Asian Studies(University of Hong Kong), Hong Kong.

Kelley, J. (1997). Paleobiological and phylogeneticsignificance of life history in Miocene hominoids. InBegun, D. (ed.), Function, phylogeny and fossils:Miocene hominoid evolution and adaptations: 173-208,Plenum Press, New York, N.Y., USA.

Li, W.Y. (1991). Vegetation and environment of theQuaternary in China. In uu, TS. (ed.), Quaternarygeology and environment in China: 151-157, SciencePress, Beijing, China.

Morton, S.R., Brown, J.H., Kelt, D.A. and Reid, J.R.W.

315

(1994). Comparisons of community structure amongsmall mammals of North American and Australiandeserts. A ustralian Journal of Zoology 42: 501-.525.

Owen-Smith, R.N. (1988). Megaherbivores. CambridgeUniversity Press, Cambridge, U.K.

Parsons, P.A. (1993a). The importance and consequencesof stress in living and fossil populations: from life­history variation to evolutionary change. TheAmerican Naturalist 142(supplement): 55-520.

Parsons, P.A. (1993b). Stress, extinctions andevolutionary change: from living organisms tofossils. Biological Reviews 68: 313-333.

Partridge, T.C (1997). Reassessment of the position ofthe Plio-Pleistocene boundary: is there a case forlowering it to the Gauss-Matuyama palaeomagneticreversal? Quaternary International 40: 5-10.

Roy, K., Valentine, J.W., Jablonski, D. and Kidwell, S.M.(1996). Scales of climatic variability and timeaveraging in Pleistocene biotas: implications forecology and evolution. Trends in Ecology & Evolution11: 458-463.

Sun, X.J., Son, C.Q., Want, F.Y. and Sun, M.R. (1997).Vegetation history of the Loess Plateau of Chinaduring the last 100,000 years based on pollen data.Quaternary International 37: 25-36.

Webb, T. III and Bartlein, P.J. (1992). Global changesduring the last 3 million years: climatic controls andbiotic responses. Annual Review of Ecology andSystematics 23: 141-173.

Winkler, M.G. and Wang, P.K. (1993). The Late­Quaternary vegetation and climate of China. InWright, H. E. (ed.), Global climates since the LastGlacial Maximum: 221-264, University of MinnesotaPress, Minneapolis, Minn., USA.

Zeveloff, 5.1. and Boyce, MS. (1988). Body size patternsin North American mammal faunas. In Boyce, MS(ed.), Evolution of life histories of mammals: 123-146,Yale University Press, New Haven, Conn., USA.

Zhou, Y.W., Qiu, G.Q. and Guo, D.X. (1991). Changes ofpermafrost in China during Quaternary. In uu, TS.(ed.), Quaternary geology and environment in China:86-94, Science Press, Beijing, China.

Manuscript received 19 December 1997; accepted 25 May1998.