Конференция

1999-MAY16-MAY20
2nd International Mammoth Conference

ABSTRACTS (4)

edted by Jelle W.F. REUMER & John DE VOS
DISCLAIMER

The Abstracts of the 2nd International Mammoth Conference,
published in this Volume of Conference Papers, are not published
for permanent scientific record and are therefore not to be considered
as publications in the sense of Article 8 of the International Code of
Zoological Nomenclature. They merely reflect the opinion of the
author(s) at the moment of submission, and they should not be cited
in scientific works without prior written consent of the author(s).

After abstract titles (P) means poster and (L) means lecture (oral contribution)

J.L PRADO, M.T. ALBERDI, B. SANCHEZ & B. AZANZA - DIVERSITY OF THE PLEISTOCENE GOMPHOTHERES FROM SOUTH AMERICA (L)

P.V. PUTSHKOV - MAMMOTH IMPACT ON MAMMOTH BIOME: CLASH OF TWO PARADIGMS (L)

M. RUSTIONI, M.P. FERRETTI, P. MAZZA, M. PAVIA& A. VAROLA - THE VERTEBRATE FAUNA FROM CARDAMONE: AN EXAMPLE OF MEDITERRANEAN MAMMOTH FAUNA (L)

M. SABLIN - NEW DATA ON FAUNA COMPOSITION OF LARGE MAMMALS OF THE CENTRE OF THE RUSSIAN PLAIN IN THE LATE WORM, RELATIVE ABUNDANCE DYNAMICS OF A NUMBER OF SPECIES (P)

A.V. SHER - THE IDENTITY OF THE TAMANIAN ELEPHANT (L)

A.V. SHER & A.M.LISTER - FOSSIL ELEPHANTS OF THE OLYORIAN AGE (NORTHEAST SIBERIA) AND THE EVOLUTION OF MAMMOTH LINEAGE IN EURASIA (P)

A. SIMAKOVA & A. MARKOVA - THE PECULIARITIES OF RUSSIAN PLAIN PLANT COMMUNITIES AND MAMMOTH DISTRIBUTION IN THE SECOND PART OF THE VALDAI GLACIATION (P)

John E. STORER - ENVIRONMENTS OF PLEISTOCENE BERINGIA: INFERENCES FROM CENOGRAMS (L)

S.A. SYCHEVA - THE ROLE OF LOCAL LANDSCAPES OF THE PERIGLACIAL FOREST-STEPPE ON THE RUSSIAN PLAIN IN THE FORMATION OF FOOD RESOURCES FOR MAMMOTH HERDS IN THE VALDAI TIME (P)

Keilchi TAKAHASHI and Keiko NAMATSU - MAMMOTH REMAINS IN THE JAPANESE ISLANDS

Alexel TIKHONOV - 50 YEARS ANNIVERSARY OF THE MAMMOTH COMMITTEE OF THE RUSSIAN ACADEMY OF SCIENCES (P)

Alexel TIKHONOV, Larry AGENBROAD & Sergey VARTANYAN - COMPARATIVE ANALYSIS OF THE MAMMOTH POPULATIONS ON THE WRANGEL ISLAND AND CALIFORNIA CHANNEL ISLANDS (L)

A.I. TOMSKAYA - ON THE TIME AND CAUSES OF MAMMOTH EXTINCTION IN YAKUTIA (P)

TONG Haowen - MAMMOTH AND OTHER CONTEMPORARY ELEPHANT FAUNA ON THE MAINLAND OF CHINA (P)

P. UKKONEN, H. JUNGNER, J. DONNER & J.P. LUNKKA - NEW RADIOCARBON DATES OF FINNISH MAMMOTHS (Mammuthus sp.) (L)

J. VAN DER MADE & A.V. MAZO - PROBOSCIDEAN DISPERSALS TOWARDS WESTERN EUROPE (L)

Hans VAN ESSEN - TOOTH MORPHOLOGY OF Mammuthus mendionalis (NESTI, 1825) FROM THE SOUTHERN BIGHT OF THE NORTH SEA AND THE NETHERLANDS (L)

Th. VAN KOLFSCHOTEN & Y. VERVOORT-KERKHOFF - THE PLEISTOCENE AND HOLOCENE MAMMALIAN ASSEMBLAGES FROM THE MAASVLAKTE NEAR ROTTERDAM (THE NETHERLANDS) (L)

S. VARTANYAN & V. PITUL'KO - LANDSCAPES, ANIMALS AND HUMANS OF THE SIBERIAN ARCTIC: THE PAST 30,000 YEARS (L)

Sergey A. VASIL'EV - FAUNAL EXPLOITATION, SUBSISTENCE PRACTICES AND PLEISTOCENE EXTINCTIONS IN PALEOLITHIC SIBERIA (L)

Nikolai VERESCHAGIN - THE PROBLEMS OF THE CONSERVATION OF THE PALEOZOOLOGICAL MONUMENTS OF THE QUATERNARY PERIOD (P)

Piotr WOJTAL - TAPHONOMY OF THE SITE KRAKOW SPADZISTA - B (L)

The species of the family Gomphotheriidae from South America range from the Middle Pleistocene (Ensenadan Land-mammal Age) to the latest Pleistocene, Lujanian Land-mammal Age (Alberdi & Prado 1995). These groups were descendants of a gomphothere stock from North America and arrived to South America during the 'Great Amencan Biotic Interchange' (Webb 1985), Based on vegetational distribution maps of South America during the glacial and interglacial phases, and the distribution of South American Gomphotheres, it appears that Cuvieronius inhabited an arid landscape, whereas Stegomastodon seems to have predominated in areas identified as savannah (Webb 1991 ). Cuvieronius hyodon seems to have been adapted to atemperate-cold climate, since in the intertropical zones it has been found only at the highest levels. On the contrary, in Chile it expanded to the littoral zone, that surely offered more similar living conditions, in terms of temperature, than the Andes corridor. This species would feed mainly on hard grasses, leaves, and seeds from bushy plants. Species of Stegomastodon would be better adapted to warm or temperate climatic conditions, because their most austral distribution does not surpass the 37њ S parallel in the Buenos Aires province (Tonni 1987). The frequency of Stegomastodon platensis diminishes in the Pampean Region by the latest Pleistocene, when environmental conditions became colder and drier. In lower latitudes, where Stegornostodon is more frequent, it would occupy savannahs orxerophytic pastures.

Simpson & Paula Couto (1957) considered that all South American forms must be included in one subfamily: the Anancinae, as there are only few and slight differences among them. We agree with this opinion and recognise in South America only two genera: Cuvieronius with only one species: Cuvieronius hyodon, and Stegornostodon with two species: Stegomastodon waringi and S. plotensis (Alberdi & Prado 1995). C. hyodon is geographically restricted to the Andean Region in Ecuador, Peru, Bolivia, Chile, and Northwest Argentina (Hoffstetteri 952, Casamiquela et ol. 1996). S. waringi was recorded in the Santa Elena peninsula in Ecuador (Hoffstetter 1952, Ficcarelli et 0/., 1996), and in the tropical zone of Colombia (Coireal Urrego 1981 ) and Brazil. S. platensis was recorded in the middle to latest Pleistocene in Argentina, especially the Pampean Region, and also dunngthe late Pleistocene of Uruguay (Mones & Francis 1973) and Paraguay (Cabrera 1929, Simpson & Paula Couto 1957). During the Pleistocene in South America two savannah corridors would have developed, one called 'the Andes route', or 'the height corridor' and another one the so-called 'East route' or 'plain corridor' (Webb 1978, 1985). Both corridors have special characteristics and very probably they constituted the main dispersion route for different mammal groups.
The Andes route was a more direct North-South connection and provided greater environmental homogeneity of non-woodland habitats in comparison with the East route. On the other hand, the low temperatures associated with the Andes route could have constituted a barrier for certain groups. These two corridors have conditioned the paleobiogeographic history of most North American mammals in South America. In fact, different models can be postulated for different groups depending on their capacity to produce distinct adaptive types along the duration of their dispersion process. In the case of the South American gomphotheres, the small Cuvieronius utilized the Andes corridor, whereas the large Stegomostodon dispersed through the East route (Figure I ).

We reconstruct the diets of Cuvieronius h/odon, Stegomastodon ptotensis and S. wanngi through isotopic analysis. Cuvieronius from Tarija indicates that they were almost exclusively mixed feeders. S. piatensis from the Middle Pleistocene of Argentina shows mixed feeding to browsing adaptations, and the same species from the Late Pleistocene in Argentina indicates atrend from mixed feeding to browsing adaptations. S. wanngi from the Peninsula of Santa Helena shows atrend of mixed feeding to grazing adaptations (Sanchez et o\. in press).

According to the 'steppe-tundra crash paradigm' mammoths and their satellites (woolly rhinos, musk-oxen, primitive bisons, horses of northern races) were stenobiotic cryoxerophiles living only understeppe-tundra extreme cryoarid conditions. The animals are considered to be killed by the outcomes of the Holocene warming, because of their inability to feed themselves in the established forests, swamps and tundras, and to withstand the weather-caused losses. Even steppe and forest-steppe vegetation is thought to become too monotonous for the efficient feeding of extinct herbivores. Guthrie ( 1990, etc.) claims that mammoths and their satellites were 'strict grazers' on 'arid grasses' making no significant use of woody plants, mosses and other dominant taiga and tundra vegetation. Because of this, he rejects the possibility of mammoths maintaining their pastures in the way modem elephants are doing. Sher ( 1995, etc.) explains mammoth survival throughout the interglacials staling that the climate then was rather unstable: short warm episodes alternated with very cold ones. Therefore, the permanent Arctic Ocean ice-shield did persist even in summer along the Siberian coast. Its cooling and drying influence maintained steppe-tundras in North-Eastern Siberia as well as pasture ecosystems westwards and southwards. Holocene warming, unlike the interglacial ones, was stable. Hence, it caused the disappearance of the mentioned ice-shield that resulted in the destruction of mammoth ecosystems and, subsequently, in the wave of extinctions.

Shers conception is based mainly on the properties ofRJss-Wurm Greenland ice layers. However, these layers do not indicate the real PJss-Wurm climate, being altered byfolding and other processes caused by the ice flow and bedrock influences. The non-altered Antarctic ice cores and deep sea Atlantic cores show that Riss-Wurm optimum climate was as stable as the Holocene one. The idea of preserving a permanent ice-shield over Siberian Arctic Ocean seas throughout interglacials cannot be reconciled with the evidence either. The real reasons of woolly mammoth and rhino persistence throughout interglacials (as well as throughout glacials) were their tolerance to the vast range of climates and their ability to maintain highly productive pasture ecosystems. Contrary to Guthrie's opinion, these giants were polyphagesthat ate various herbaceous and woody plants. They indeed held back forest and tundra vegetation by various direct and indirect influences (Putshkov 1997). Whatever the climate was, the activity of giants prevented the appearance of closed forests and ensured the predominance of grasses and herbs over mosses and shrubs as well as the high mosaicity of vegetation. The giants trampled on snow and broke the frozen snow crusts, thus contributing to the successful hibernation of smaller ungulates. Similarly, Pataeoloxodon and Dicerorhinus created vast meadows throughout the temperate forest zone. These meadows were used by giant deer, bisons, horses, etc. The ecosystem impact of the climate-resistant megaherbivores caused remarkable stability of pasture ecosystems throughout Pleistocene. The largest carnivores maintained the 'equilibrium' too, first of all by pressure on human populations. The Pleistocene crisis in the Palearctic included the same main processes as in other continental realms: ( I ) Liberation of humankind from the large carnivores control, resulting in the increase of human populations. (2) Removal of proboscideans and rhinos by man. (3) Drastic changes of plant communities as effects of this removal. With the extinction of giants, plant communities sustained by their activity were replaced, according to each climatic zone, with tundras, taiga or dense leaf-bearing forests; the lack of feeding and soil fertilizing activity of giants caused the substitution of meadow vegetation by plants of low nutritive qualities. Ungulates and carnivores could use 'mammoth paths' through thickets and snow no more. In steppes the vegetation became less mosaic and fires more frequent and intensive: for the lack of megagrazers more grass remained uneaten and dried up. (4) Extinctions ofherbivores dependent on environmental effects of giants. The mentioned changes and subsidiary reasons (competition from other herbivores; human hunting and burning activity; predation by wolves that became numerous due to the decline of larger predators, etc.) have caused the extinctions of giant deer, musk-oxen, primitive bisons, two antelopes in Central Asia, hydruntine ass, and horses of northern races, (5) Impoverishment of the large predators guild due to the impoverishment of large prey and, secondary to the competition from man (the case of lion, cave hyaena and leopard) or wolf (the case of dhole): wolves previously had been suppressed by larger predators. (6) Other secondary extinctions. Several Siberian soil mites became extinct due to the change of soil conditions after the cessation of the intensive grazing by megafauna. (7) Installation ofthe new equilibrium. Holocene ecosystems are different from Pleistocene interglacial ones first and foremost due to the lack of giant herbivore impact.

Due to the long coevotution ofPalearctic megafauna with man, the extinction process in Palearctic was long-lasting and less catastrophic as opposed to the Nearctic situation.

The vertebrates from Cardamone Cave were reported for the first time by Botti ( 1890). Following studies were from Vaufrey (1927), Guerin (1980) and Rustioni (1998). The rich assemblage comes from a karstic cavity infilling in the local Plio-Pleistocene calcarenites. The collection was dismembered in the years and partly dispersed to several institutions of Italy, while other specimens quoted in the literature are now lost. Although depleted, the fauna that survived the vicissitudes of the years is still enormously rich and this makes it very significant from different viewpoints. Diagrams of the bone representation of each species confirm an artificial selection of the specimens, possibly due to both improper recovery (which probably explains the absence ofmicromannmals and of other small vertebrates) and to the subsequent dismembering. Although aware that all the observations are affected by this unavoidable statistical error, the impressive abundance of the specimens (about 3700) and the fact that the all the species reported by Botti (1890) are still represented (with the exclusion ofFetis, Arvicota and A^us),
convinced the authors that the fauna was anyhow worth the consideration. Worthy of note is the extraordinariy high number of foxes, which may be the result of an attritional accumulation of carcasses in a long time lapse. The bones do not show evidence of carnivore activity, nor of transportation, nor ofbutchering, scamification, or of other human activity. Besides foxes, the cave provided remains of mammoth {Mommuthus primigenius), woolly rhino (Coetodonta antiquitotis), aurochs (Bos primigenius), red deer (Cervus eiaphus), horse (Equus ferus), wolves (Cams lupus), spotted hyaena (Croctrta crocuta), hedgehog (Erinaceus europaeus), hare (Lepus europaeus) and rabbit (OryctotogLK cuniculus). The assemblage as awhole is suggestive of relatively open landscapes under fairly cold climatic conditions.

The most significant occurrences from the stratigraphical viewpoint are the pachyderms, which seem to disperse in Italy during the last glacial. The woolly mammoth and, even more, the woolly rhino are extremely rare in Italy and many of the occurrences from Italy reported in the literature are due to misattribution. Scanty remains of both were found in northern and central Italy and isolated woolly rhino teeth were recovered in Apulia. The two species were never reported from any other part of southern Italy, The find from Cardamone is a confirmation of the occurrence of the woolly rhino in Apulia and the southernmost known occurrence of the woolly mammoth in Italy. The Cardamone cave also provided a rich collection of bird remains. The avifauna is dominated by Columbo iivia vel oenas. Significant occurrences are those offironta bernicta, atypical northern goose, very seldornly found at such low latitudes; Otis tetrox and 0. tarda, both indicative of open steppe pranes; Falconiformes are also well represented. A significant occurrence is that ofHahoetus albiciila, now sporadic in western Europe. The environment suggested by the bird assemblage from Cardamone is that of an open shore with sporadic nearby bushes, indicated by the presence of Bubo bubo, Buteo buteo and Cuculus cucuius minor. The occurrence of Mammuthus and Coelodonta in limited areas of northern Italy, in the central Tyrrhenian side of Italy and in the Salento peninsula and their absence in most of the Adriatic belt may attest to a completely different paleogeographic setting at the time of the arrival of the Cardamone fauna. The combination of an extensive sea-level fall and intense late tectonics in the Adriatic area may have caused the emersion of most of the Adriatic sea-bottom, giving rise to a vast wide open plain that acted as a corridor forthe pachyderms of the Balcan peninsula to the southeastern regions of Italy.

The conclusions that may be drawn from the Cardamone fauna is that some typical elements of the Mammuthus-Coelodonta Faunal Complex, and namely M. primigenius, C ontiquitatis, L ferus and C crocuto probably immigrated from the east and mixed to the' more temperate, autochthonous faunas characterized in particular by 6. primigenius, C. elaphus, Lepus europaeus and 0. cuniculus.

Over 25,000 fossil bone remains of mammals from 5 Late Paleolithic sites in the centre of the Russian Plain have been studied. The sites included Avdeyevo, Eliseyevichi, Kostenki, Sungir', and Yudinovo. The revision of species composition of Late Pleistocene theriofauna (Camivora, Artiodactyla) from the centre of the Russian Plain permitted describing new subspecies of wolf Canis lupus brevis KUZMINA et SABLIN, 1994 and polar fox Mopex logopus rossicus KUZMINA et SABLIN, 1993. New for this territory, subspecies of wolverine, Gulo gulo spelaea GOLDFUSS, 1 818, reindeer Rangifer tarandus guettardi DESMAREST, 1822, species of musk-ox Ovibos pollontis H. SMITH, 1827 and saiga Saiga boreolis TSCERSKYI, 1876 have been described. Cave bear Ursus speiaeus ROSEN., cave hyena Crocuto spelaea GOLDFUSS., boar Sus scrofa L„ roe Capreolus capreolus L, giant deerA^egatoceros giganteus BLUMENBACH and auroch Bos primigenius BOJANUS, indicated in the preceding studies were absent in the centre of the Russian Plain in the Late Pleistocene. Representatives of the families Canidae (wolf and Arctic fox), Cervidae (reindeer), and Bovidae (primitive bison and musk-ox) were predominant. The development of a climatic zone with extreme cold and arid conditions in the territory of the centre of the Russian Plain became the reason of predominance of cold-resistant tundra species of carnivorous and even-toed mammals by the end of the epoch. Occurrence of steppe forms, which were more warmth-loving, during summer migrations became extremely rare.
As awhole, the Late Pleistocene fauna of carnivorous and even-toed mammals inhabiting the centre of the Russian Plain was more cold-loving than had been assumed before.

 The elephant of the Tamanian Faunal Complex, usually called Mammuthus (Archidiskodon) meridionalis tomanensis, occupies an important position in the phylogeny of the mammoth lineage. According to the most common view, it is an intermediate between M. meridionalis and M trogontheni. The morphological change between these two species, however, is the greatest within the whole mammoth lineage, and some scientists have doubted that such an ancestry was sufficiently proven (Azzaroli 1977). The transitional character of the morphology of the Tamanian elephant has been used as an important proof of direct ancestral relations between the two species (or genera). The subspecies (founded by Dubrovo 1964) is based on the large series of elephant fossils from the Sinyaya Balka locality on the Taman' Peninsula (Azov Sea), excavated in 1912 by 1. Gubkin; some additional material was collected later by 1. Dubrovo (1953-1961), V. Zhegallo (1987-88) and others.

Earlier investigators of the Sin/aya Balka elephant collection (E. Belyaeva, N. Vereshchagin) inferred the heterogenerty of this sample, recognising 3-5 species of elephant, including 'Elephas meridionaHs, 'Ј'. trogonthen), and even 'Ј'. •primigenius. 1. Dubrovo (1963, 1964) made acomprehensive revision of the whole collection, and came to the conclusion that the large series of upper and lower M2 and M3 shows continuous variation in almost all quantitative parameters, and demonstrates clear unimodal distributions. From that she inferred that despite awide range of variation in many characters, the whole sample a) belongs to a single form, b) in all metric parameters lies within the variation of one species, Archidiskodon meridionohs, and c) represents an advanced form of this species. These conclusions have been unanimously accepted for more than 30 years.

My work with the Olyorian elephants, which lived in the North at the same time as the Tamanian, forced me to study the Sinyaya Batka collection myself. I concentrated on the last molars and used a new methodological approach, allowing more precise recognition of plate count and crown height (Sher& Garutt 1985), The results were fundamentally different from those obtained by 1. Dubrovo. The ranges of variation of the most important parameters were shifted towards higher plate counts and higher crown (index othypsodonty), and their distribution appeared bimodal. That was especially evident by comparison with the same parameters for/^l. meridionalis from Upper Valdamo and M. trogontherii from Sussenborn, kindly provided by Dr. A.M. Lister from his own measurements, which demonstrate much narrower variation and distributions close to normal. Variation in total plate number (P) excluding talons was shown on a histogram, combining upper and lower M3 for those
samples (Valdamo, Taman, Sussenborn). Even if we do not take into account one upper M3 from Sinyaya Balka with II plates, avery advanced morphology, and preservation different from most other teeth (PIN 1 249/72, not included in the statistics), the Tamanian sample still has a distribution markedly different from normal and is essentially double-peaked. Two statistically distinguishable groups can be seen. The first has P= 13-17 and most frequent values around 15-16. It is shifted from the Valdamo curve (P= I I -15, peak at 13) towards its more 'advanced' end. This part of the Taman' histogram, in its shape and position, closely follows the curve published by Dubrovo ( 1963), and seemingly fits her 'tamanensis concept. The second group has P= 17-19 with a sharp peak at 1 8 plates (this value is completely absent from Dubrovo's counts). In its position, this group corresponds to the less advanced' part of the Sussenborn /V\. trogontherii range, which has a peak at 19.

A similar pattern is observed in the hypsodonty index (Hl=max. crown height/max, crown width, %) curve for upper M3. The Taman' sample shows one clear peak at Hl= 130-140, which corresponds to the 'advanced' wing of the Valdamo distribution and closely fits the values calculated from Dubrovo's 'width to height' index for 'tamanensis'. The other peak in our Taman' sample (around Hl= 170) almost reaches most frequent values for Sussenborn, but in general overlaps with the less advanced' wing of M trogontherii distribution. Evidently, the reason for disagreement with previous observations is that some teeth were earlier interpreted as being complete, while in fact a few anterior plates were lost through wear - avery common error before certain rules of observation of the crown base and root system were suggested (Sher& Garutt 1985). Correspondingly, underestimation of plate count resulted in the fact that the plates, measured for maximum crown height, were actually not the highest in the crown, which led to underestimation of hypsodonty index. From the bimodal distribution of these two important characters I assume that the Sinyaya Balka collection is not homogeneous. Two groups of last molars with a different level of evolutionary advancement are present here (ignoring PIN 1249/72 specimen). Other characters show rather good correlation with these two groups. For example, enamel thickness for the first group (with the lower P and HI) is usually more than 3 mm, while for the second it ranges mostly from 2.0 to 3.0 mm. Lamellar frequency on upper M3 in the first group ranges from 4.7 to 5.5, while in the second it is most commonly higher than 6.0 (5.6-6.8).

According to the morphological parameters of the last molars from Sinyaya Balka, the first group represents advanced meridionalis, and it is quite reasonable to classify 'it under Mommuthus {orArchidiskodon) meridionaiis tamanensis Dubrovo, 1964. From my view, it is only necessary to exclude more advanced teeth of the second group from the type series of this subspecies. The second group could represent one of the earliest occurrences ofM. trogontherii with relatively primitive morphology. The heterogeneity of the Sinyaya Balka dental sample can be interpreted in two ways. The first is to assume that the material is a mixture of fossils from sediments of different ages. Even if we ignore the evidently alien single upper tooth, mentioned above, such a possibility cannot be excluded, Sinyaya Balka is a secondary deposit of mud flow origin, further complicated by landslides. Let us assume that the more primitive ('tamanenslS') subsample is ofJaramillo age, as it is assumed now. In this case the advanced subsample (the 'primitive trogontherii') could come from the sediments corresponding to the very end of Matuyama Epoch (Vangengeim et at. 1991). This assumption still leaves the opportunity of direct ancestry open. However, it finds no confirmation by geological evidence (Vangengeim et aL, 1991 ). The second interpretation is that the Sinyaya Balka sample is not mixed, but includes fossils of two contemporaneous forms of elephants. That would mean that the trogontherii morphology appeared in Europe much earlier that was thought, and practically excludes the idea of autochthonous origin ofM trogontherii from the late forms of/VI, meridionalis. Whichever hypothesis is confirmed by further research, I believe that the whole Tamanian sample can no more be interpreted as a single fossil population transitional from M. meridionalis to M trogontheni, which by itself makes the concept of direct ancestry less convincing.

Azzaroli, A„ 1977 - Evolutionary patterns ofVillafranchian elephants in Central Italy - Atti Accad. Lincei, Mem., Cl. Sc. Fiz., Mat., Nat. Ser. 8, 14 (lla): 149-168
Dubrovo, I.A., 1963 - New data on the Tamanian Faunal Complex of vertebrates - Bull. Moscow Society of Naturalists, Geol. Section 38 (6): 94-99 (in Russian)
Dubrovo, I.A., 1964 - Elephants of the genus Archidiskodon in the USSR territory - Paleontologicheskiy Zhumal 3: 82-94 (in Russian)
Sher, A.V. & Garutt, V.E„ 1985-New data on morphology of elephant molars - Doklady AN SSSR 285 ( I ): 221 -225 (in Russian) - English translation, 1987 - in: Transactions Doklady of the U.S.S.R Academy of Sciences, Earth Science Sections 285 (1): 195-199
Vangengeim, E.A„ Vekua, M.L, Zhegallo, V.I., et al., 1991 - Position of the Tamanian Faunal Complex in stratigraphic and magnetochronological scales - Bulletin of the Commission on Quaternary Research, USSR Academy of Sciences 60: 41 -52 (in Russian)

 The Olyorian Land Mammal age has been recognised on the basis of a series of local faunas coming from fine-grained deposits widely spread overthe vast lowlands in Northeast Siberia (Sher, 1987). Chronologically, it covers a rather long time span, including most of the Early Pleistocene and the early part of the Middle Pleistocene. The corresponding stratigraphic unit is subdivided into the Early Olyorian (Chukochyan) and the Late Olyorian (Akanian); the boundary between them coincides very closely with the Matuyama/Brunhes magnetic reversal. Faunas very similar to the Olyorian have also been discovered in the Yukon Territory and Alaska, so this mammalian age may be applicable to the whole Beringian Realm. An important paleoecologicaJ feature of the Olyorian is that these sediments accumulated undersevere conditions. Cold and dry continental climate promoted predominance of arctic grassland communities Ctundra-steppe', or 'arctic steppe'), limited arboreal vegetation, and continuous permafrost. The Olyorian fauna in general is a peculiar fauna of cold-adapted grazers, which evolved much earlier than the appearance of the first elements of 'cold' fauna in the temperate latitudes of Eurasia in the course of glacial climatic cooling.

Mammoth fossils have been known from Olyorian sediments since long ago (Sher 1971 ), but their taxonomic and phylogenetic interpretation was difficult because of their quite advanced morphology, evidently contradicting existing models of mammoth phylogeny. This problem could not be solved without the collection of additional fossil material in clear stratigraphic position, and detailed comparative analysis ofthis material with the classical European collections. The studied collection of mammoth teeth from sedimnets of various ages in Northeast Siberia includes almost 300 last molars. Also, the development ofne\w methodological approaches in the study of fossil elephant dentition was necessary (Sher & Garutt 1985, Lister & Joysey 1992); this sometimes triggered the revision of material from previously studied key localities. For comparison we preferred to use the dental samples from one locality or a close group of similar sites studied by one of the writers personally (e.g., SOssenborn sample forM trogontherii).

The early form ofOlyorian mammoth comes from Chukochyan sediments; the best specimens, on which the description ofthis form is based, are related to the lowermost visible part of the section in the well-studied type-area of the Olyorian (Bolshaya Chukochya River, Kolyma Lowland). This part (about 10 m thick) is reverse-magnetised and lies below the normal event correlated with the Jaramillo subchron (ca 1.0 My). Based on sedimentation rate, the minimum estimate of the eaHiest possible age forthe zone, from which the fossils come, is about 1.2 My. The last molars of early Olyorian mammoth (EOM) are relatively small in size (maximum crown width usually <100 mm), have rather high lamellar frequency (6-8), and moderately thin enamel (1,5-2.5 mm). Together with the high crown, these features make the EOM dentition look like rather advanced forms of the mammoth lineage. In total number of plates in the crown (19-21 plates on upper M3, 20-22 plates on lower m3
excluding talons) and index othypsodonty EOM last molars are very similarto those ofA^ammuthus trogontherii from Sussenborn. They also have avery similar pattern and depth of plate dissection, usually represented on the occlusal surface as a figure of three more or less equal ovals, and not common among more advanced mammoths. Their relatively high lamellar frequency results in part from a small crown size (Lister & Joysey, 1992); corrected tor-size, it is lower than in M. trogonCheni on uppers, and equal to it on lowers. Similar scaling may be partially responsible for relatively thinner enamel of EOM. A partial fore limb find suggests that EOM had aserial carpal arrangement. Thus, except for a little smaller size and related features, morphological characters of EOM are statistically indistinguishable from typical European M. trogontherii, which existed almost 500 ky later.

Fossils of the late Olyorian mammoth (LOM) come from Akanian sediments. The lowest level they were found is immediately above the Matuyama/Brunhes boundary, but they are common in the whole Upper Olyorian. The LOM last molars have high crown with 22-24 plates; all their morphological characters are similarto those of/VI. primigenius. Compared with Late Pleistocene mammoths of the same region, the LOM dentition look only slightly less advanced. Their lamellar frequency, corrected forthe crown size, is on the average only 0.3-0.4 units lower than in the late M. primigenius, and the enamel (1.3-1.8 mm) only 10-12% thicker than in the latter. The plate dissection typical for EOM is less frequently observed in LOM, with shallower clefts, but in general enamel pattern the occlusal surface is of the same type and looks more regular than in late M. primigenius. Except for these minor differences, LOM dentition has typical 'pnmigenius' morphology. This morphology can hardly be observed in European record until as late as 0.3-0.2 My, i.e. 400-600 ky later than in Northeast Siberia.

There are good reasons to see continuity between the early and late Olyorian mammoths. The evolution of 'primigenius' morphology in LOM is correlated with important environmental and faunal changes in Northeast Siberia around the Matuyama/Brunhes boundary (Sher 1997). Fossil teeth from later Middle Pleistocene sediments demonstrate generally intermediate morphology between LOM and the late Pleistocene mammoths, but because of rather low amplitude of morphological change, their characters are within the range of variation of those for M. primigenius. This evidence unquestionably suggests continuous existence of a northern lineage of mammoths, in which each evolutionary grade was achieved much earlier than in the middle latitudes. Since the commonly accepted main cause of rapid evolution in the mammoth lineage in Eurasia was climatic cooling and the spread of boreal grassland, it is not surprising that on the vast Arctic plains, where this environment was established much earlier, progress in mammoth adaptation ran even faster.

The most striking morphological shift in the European lineage seems to occur between hA. mendionalis and M. Vrogontheni. The broad, low-crowned molars with few and widely positioned plates of the former and highly hypsodont teeth with closer packed plates of the latter are a reflection of different shapes of the skull - long and low in M. meridionalis and supposedly short and high in M trogontheni. In many respects, morphological distance between these two species is incomparably more than between M. trogontheni and woolly mammoth (Lister 1996). Azzaroli ( 1977) believed that the late forms of European M. meridionaHs demonstrated extreme specialisation not in the trogontherii direction, and M. trogontheni could not be their immediate descendant. This view is supported by some recent data on co-occurrence of dentition with advanced 'mendionalis' and 'trogontheni' morphologies in some European localities (Lister 1996 and this volume). The knowledge that 'trogontherir morphology already existed in northern Eurasia at the time when M. mendionalis flourished in southern Europe, gives strong support to these ideas. Thus, we forward a hypothesis of northern origin ofM trogontheni from an earlier meridionalis stock and its subsequent dispersal to temperate latitudes. Similarly, concerning the further history ofM primigenius, we think that a model of the southward dispersal, can also be applied. An apparent overlap in the temporal range of late 'trogontheni" and early 'primigenius' in Europe (Listerthis volume) is consistent with this idea.

The PALEOFLORA database in Paradox V.4 has been developed for Late Pleistocene palynological materials. It includes pollen data of 64 sections, corresponding to the late Valdai glaciation (24 -10 k/BP). The peculiarities of plant communities were elucidated on the base of these materials (Markova & Simakova 1998). An analysis of subfossil pollen spectra reveals their sufficient correlation to the composition of plant communities (Grichuk 1946). Thus, non arboreal pollen percentage in excess of 40% ofthe total pollen and spores quantity implies the presence of open ecosystems (steppe and forest-steppe) at the time when the sampled deposits were formed. The PALEOFLORA database permits to construct a series of maps showing locations of the different plant species and also plant communities using ARC/INFO and ARC/VIEW cartographic software forthe following intervals ofthe late Valdai: 24-21,21-17, 17-15, and 15-10 kyBP.

During the late Valdai, a radical restructuring ofthe vegetation zones took place. Typical arctic and northern taiga species expanded their ranges towards the south as far as 47њ N, while steppe species penetrated to the north and west and reached 62њ N. The broad-leaved trees survived within very limited refugia in the south ofthe Russian Plain and in the Carpathians. Mosaic periglacial landscapes occupied the major part ofthe Russian Plain during this time and featured combination of open plant communities and small local areas of forest and bush vegetation. At present, abundant material from the Late Pleistocene mammal sites ofthe Former Soviet Union has been accumulated in the form ofthe PALEOFAUNA database (Markova et o\. 1995). This database indudes information on about 300 Mammuthus primigenius sites.

We decided to make an attempt to examine mammoth distribution during the late Valdai in the context of herb communities distribution. During the period of 24-21 kyBP open vegetation communities prevailed over most of the Russian Plain (from the ice sheet margin to the Black Sea coastal regions). These data are based on pollen material from 40 sections. The mammoth range at that time was broad enough on the Russian Plain, excluding only the Black Sea coastal regions. Practically all the locations ofthe mammoth remains are within the areas of dominant herb and grass communities. The material of 46 mammal localities and 30 sections with palynolodgical data correspond to the maximum cooling ofthe Valdai glaciation (21-17 kyBP). The herb communities prevailed also on the Russian Plain at that time. Its northern limit was at 62њN. Mammoths were widely distributed during this inten/al (Fig. I ). Later, during the next two millenniums, the mammoth range underwent a definite reduction. However, the open plant communities were still widespread. The reduction ofthe mammoth range could be explained by the pressure of Paleolithic hunters and possibly also by the very beginning ofclimating warming. During the period of 15-10 kyBP the northern limit of open plant communities shifted southward. The end of this interval was marked by the onset of global warming and the beginning ofthe formation of a forest zone which was unfavourable for large herbivorous mammals, including mammoth.

Grichuk, V.P., 1946 - To the history of vegetation of European part of USSR during Quaternary - Trudy instituta Geografii, Vyp. 37:249-266
Markova, A.K., Smirnov, N.G., Kozharinov, A.V., Kazantseva N.E., Simakova, A.N. & Kitaev, L.M„ 1995 - Late Pleistocene distnbutuion and diversity of mammals in Northern Eurasia - Paleontologia i Evolucio 28-29: 5-143
Markova, A.K. & Simakova, A.N., 1998 - Distribution of mammals' and plants' indicator species at the second part of Valdai' Glaciation - lzvestia RAS, Seriageograficheskaia3: 49-61

Reconstruction of Pleistocene environments ofBeringia, including the Mammoth Steppe, has usually been based on ecological
requirements of individual plant and animal species. In the standard approach environmental parameters favored by living populations or close relatives are projected backto fossil communrties. The method I use in this paper, analysis ofcenograms, provides an independent analytical tool that is not based on attributes of particular species in the community. Cenograms are graphs of body mass in mammalian communities, ranked by size. The distribution of mammalian body mass shows a remarkable correlation with two important environmental factors. Aridity is inferred from the slope of medium-sized species and openness of plant formations is correlated with the size gap between medium and small species. Inferences made from cenogram analysis are largely free from both taxonomic and taphonomic influences.

Cenograms have not previously been used to infer environmental parameters for arctic and subarctic faunas. I constructed graphs for modem Yukon forest, taiga, and tundra faunas to augment data from lower-latitude assemblages available in the literature. The present-day Yukon boreal forest fauna produces cenogram statistics that indicate semiarid climate and vegetational formations about as open as in more southern forests. Taiga and tundra both fall into the arid part of the moisture gradient, with tundra at an extreme end. Taiga and tundra are also extremely open habitats.

Cenograms for seven Quaternary faunas ofBeringia, ranging frorn early Pleistocene to latest Wisconsinan, all indicate andity comparable with taiga conditions and openness similar to tundra. No forest-habitat faunas were detected. Arid climate and very open habitats probably appeared in Beringia by the earliest Quaternary and persisted for a major part of the Pleistocene even at times other-than glacial maxima. Mammalian diversity is very low in Beringia today and was even more restricted in the Pleistocene among small and medium-sized groups. Reduced diversity is probably due to high actual evapotranspira-tion and other factors including glaciation and permafrost substrate. Low diversity in Pleistocene faunas of Beringia is not an airtifact of inadequate sampling. Several faunas have been collected comprehensively enough to indicate actual diversity among small and medium-sized mammals.

The normal renewal of great mammoth herds needed a highly productive herbaceous vegetation that is associated with the formation of fertile meadow soils. The notion of extremely severe conditions of the periglacial zone of the Russian Plain disagrees with a real possibility of occurrence of such plant communities and associated soils. This inference was made from the reconstruction of zonal conditions of the Valdai loess formation on watersheds of the Russian Plain, where extreme environments dominated, and weakly soddy primitive loess soils unfavorable for the development of productive meadow-steppe vegetation, occurred (Anon. 1993). This contradiction is removed in studies of the evolution of local landscapes of the Russian Plain in the Late Pleistocene.

We traced the succession of transformations of local landscapes including watershed fragments of the Russian Plain, large balka systems, and valley slopes from the Mikulino Interglacial (135-1 10 ky) to the Bryansk Interstadial (33-23 ky) (Sycheva 1996, 1997). In the Mikulino Interglacial the watersheds were most dissected. Balkas substantially differentiated the vegetation and soil cover. With the advent of the Valdai Glaciation the denudation of watershed and slope soils began to progress rapidly. The burial of the linear erosional network, concurrent with its transformation into the system of enclosed forms (flat-bottom depressions), began. In the subsequent eariy- and mid-Valdai time precisely these numerous flat-bottom hollow landscapes were favourable forthe renewal and maintenance of productive plant communities, because they accumulated additional moisture and nutrition necessary forthe formation of fertile meadow soils. At the local level, in passing from the interglacial to the glacial, watershed areas underwent considerable transformations and were covered by a network of flat-bottom
depressions and hollows with tall grass meadows, which were the major food resource for mammoth herds. The landscapes of low river terraces with similar ecological conditions played an important role in the formation of pastures for mammoths.

In the sediments filling buried balkas we have distinguished two to four interstadial paleosols of the early and middle Valdai (Sycheva 1997, 1998). Paleosols are of the same genesis: they are referred to the group of meadow calcium-humus soils (Glazovskaya 1972). The humus content of them (up to 1.1%) is high for fossil soils, whereas in loesses it is ten times lower. The ratio othumic to fulvic acids (Ch/Cf) is more than I, which substantiates their meadow-steppe genesis (Table 1). Modem analogues of these paleosols, the meadow chemozemic soils, are characterised by considerable values of the biological cycle of ash content from 1500 to 3000 kg/ha.y (Titlyanova 1979, 1 997). Their humus content is higherthan that ofchemozems, constituting 180-300 cent/ha (Glazovskaya 1972, Titlyanova 1997). Such values were, most likely, characteristic of the Valdai meadow soils of river terraces, flat-bottom hollows, and other depressions with a highly productive herbaceous vegetation that served as a food resource for mammoth herds. The study of the evolution of local landscapes and their components (topography and soils) ofthe Valdai periglacial zone ofthe Russian Plain enabled the explanation forthe occurrence oftall grass
meadows that served as pastures for great mammoth herds.

Anonymous, 1993 - Evolution of landscapes and climates of Northern Eurasia. Late Pleistocene - Holocene elements of prognosis - Nauka, Moscow: 102 pp.
Glazovskaya, M.A., 1972 - Soils ofthe World - lzd. MGU, Moscow
Sycheva, S.A., 1996 - Evolutionary analysis ofthe Pleistocene buried small erosional landforms - Geomorphology 3:31-38
Sycheva, S.A., 1997 - Evolution of balka system in the climatic cycle 'glacial-interglacial-glacial' - Geomorphology 2:100-1 I I
Sycheva, S.A., 1998 - New data on the composition and evolution ofthe Mezin Loess-Paleosol complex in the Russian Plain -Soil Science 10: 1177-1189
Titlyanova, A.A., 1979 - Biological nitrogen and ash elements cycle in herbaceous biogeocoenoses - Nauka, Novosibirsk: 152 pp.
Titlyanova, A.A., 1997 - Biological carbon cycle in herbaceous biogeocoenoses - Nauka, Novosibirsk; 2.22 pp.

Table I Some chemical properties of the Mikulino Interglacial and Valdai Interstadial paleosols.
 

Paleosols		Hori- son	Depth, m	Humus %	A1203, %	Fe203 %	Ca0, %	Ch/ Cf
Val- dai	Bryan- skaya	Al	2.1	0.82	13.0	7.2	5.4	*
		Bca	2.7	0.70	8.6	6.4	7.4	0.8
	Aleksand- rovskaya	Al	3.1	0.53	14.0	7.0	2.7	1.3
		B	3.5	0.43	12.8	5.7	2.8	0.5
	Strelets- kaya	Al	3.8	0.95	7.7	5.8	5.6	1.1
		BC	4.6	0.22	10.1	4.9	1.8	0.4
	Kukuev- skaya	Al	4.9	1.20	12.3	6.0	1.9	2.2
		C	5.9	0.04	6.9	5.0	1.3	0.8
Mikulino		Al	4.7	1.25	7.0	4.8	1.6	* no facts
		A2	5.0	0.20	5.5	3.2	1.0
		Bt	5.2	0.14	6.5	4.2	1.2

 Animals classified as mammoths first appeared in sub-Saharan Africa during the early Pliocene. After that, mammoths migrated to Eurasia, it evolved into M. meridionolis in the Pliocene, /VI. trogontherii in the eariy to middle Pleistocene, and M. finmigenius in the late Pleistocene (Maglio 1973). Eventually they crossed eastward and southward of Europe, and through Asia and entered North America by at least 1.7 My ago (Webb et ol. 1989). These studies were mainly based on specimens from Africa, Europe, and America, Recently Lister (1996) summarised the evolution and taxonomy of Eurasian mammoths, but he made no mention about Chinese and Japanese material. In fact, many fossils of mammoths have been obtained from the Japanese islands, and they are very important in any consideration of the evolution and migration of this group of proboscideans. In the history of the study of the Japanese mammoths, since 'Euelephas' protornammonteus and 'Ј.' trogontherii were described by Matsumoto (1924), many specimens have been described. These specimens have been assigned to different genera and species overtime by various researchers. As a result, Japanese mammoths have been divided into two groups. One is M. primigenius in the late Pleistocene and other is the M. meridionaUs - M. trogontheni transition type in the early Pleistocene. The taxonomy of the latter has been confused until now.

According to my observation on the molars of Japanese mammoths, in fact 3 species have existed there. Type I occurred in Hokkaido, and is obviously identified as M. primigenius according to the same criteria used by former researchers. Type 2 has been considered as an M. meridionaUs - M. trogontherii transition stage. Their molars are of middle size: the molar crowns are narrow with more than 19 ridge-plates on M3; the plates are relatively thin, not closely spaced, with a lamellar frequency of 6 to 7 for M3; posterior loops, although present on some molars, are often absent; worn enamel figures are parallel in the medial and distal parts of the enamel loops; the enamel is weakly ribbed externally with fine and wrinkles, 2.2-2.8 mm thick; slightly worn ridges almost show a ring-loop-ring pattern, sometimes loop-ring-loop. These characters are considered as representing the M. trogontherii stage although they are different from the typical morphology ofM trogontherii in Europe.
Type 3 has more primitive characters than type 2. There is only one specimen, from Kagawa Prefecture. It has strong folding even in the median portion of strongly worn plates. This specimen could easily have been interpreted as representing the M. mendionalis stage.

Lister ( 1996) implied that M trogontherii originated as two lineages that coexisted before the extinction of/VI. meridionolis (c. 1.2 - 0.6 My). Typical remains ofM trogontherii are dated as c. 0.7 - 0.5 My in Europe. Japanese specimens indicate that M trogontherii had already appeared in the early Pleistocene (c. I - 0.7 My) of the Japanese Islands. This is avery important finding when considering the migration of mammoths.

 The mammoth research is very important in Russia since the 19th century. Numerous expeditions with this purpose were supported by the Russian Academy of Sciences. At the beginning ofthe 20th century some unique mammoth carcasses were found in the permafrost of Siberia. But the idea to organize a special scientific committee on mammoth research come only at the middle ofthe 20th century. The Mammoth Committee was established in 1947 by the Russian Academy of Sciences with the purpose to organize an expedition to Taimyr, where awhole mammoth skeleton was found. Later, in 1948, by the decision of persons like Beria and Molotov, the Mammoth Committee was transformed into a scientific Committee that gathered all information about outstanding paleontological finds connected with the mammoth and the so-called mammoth fauna. During 50 years our committee was involved in the most famous expeditions searching for the mammoth fauna remains. Carcasses of mammoth calves, Pleistocene bison, Pleistocene horses and twelve complete skeletons of mammoth, woolly rhinoceros, and other extinct mammals were found. At present the Mammoth Committee is a large scientific organization uniting nearly 100 specialists of different branches of science from 14 countries. a and Taimyr, together with foreign specialists. Almost ever/trip was prized with new interesting finds. We hope that in the

The special interest of our organization is the Arctic zone of Siberia, where the most unique paleontological treasures can be found. During the 90's 7 expeditions were conducted to the Novosibirskie Islands, Wrangel Island, the North ofYakuti future our international cooperation will be fruitful for further mammoth research.

 At the end ofthe Pleistocene the range of distribution ofthe woolly mammoths (Mammuthus primigen/us) appreciably moved on the extreme northern portions of Eurasia. At the boundary ofthe late Pleistocene and early Holocene the distribution was divided into isolate groups. In the early Holocene the last mammoth populations were present on the Taimyr Peninsula and on Wrangel Island, in the Siberian Arctic Ocean. In the latter location, mammoths survived to the middle Holocene (3-4 thousand years ago). The existence of this isolated population on the Arctic island resulted in appreciable morphological changes, such as reduction of size, narrowing of molars, etc.

These morphological changes allowed the description of a new mammoth subspecies - M prirnigenius vrangeliensis GARUTT, AVERIANOV et VARTANYAN, 1993. In Northern America the range of distribution ofthe Columbian mammoth (Mammuthuscolumbi) also changed appreciably at the end ofthe Pleistocene, producing separate, isolated populations. One of these isolate groups was established on the California Channel Islands, probably dunng Pleistocene sea level lowering, when an ancient large island was near shore. Long isolation on the island resulted in appreciable size reduction and a series of other morphological changes. The species Hephas exilis STOCK et FURLONG, 1928 was proposed for these small mammoths. Now, the majority of scientists refer to the species as Mammuthus exilis.

Despite huge differences in the paleogeopgraphy and environments ofthe two island populations it is of great interest to make comparisons ofthe species. To begin, comparisons ofthe morphological features ofthe skeleton and dentition were made, to reflect the adaptations of habitation in the separate environments. It is also necessary to take into account the different temporal frameworks for the existence ofthe mammoths in both territories. For example, it has been established that the separation of Wrangel Island from the mainland took place at the boundary between the Pleistocene and the Holocene. The separation of Pleistocene and Holocene mammoth remains on Wrangel is possible only by radiocarbon dating. In the last few years, more then 100 dates have been processed, which is insufficient, as the majority of the bones of the skeleton have not been dated. On the California Channel; Islands, we do not yet know the time of arrival, nor the rate of shrinkage of the mammoths, Prior to March, 1999, only 21 radiocarbon dates have been produced, some beyond the 40,000 year limit' of radiocarbon dating.

No complete crania of mammoths have yet been found on Wrangel Island, so comparison with California Channel Island mammoths is limited to postcranial bones, dentrtion and tusks. Collections of postcranial remains from Wrangel are limited to three femora, five tibiae, and three pelvi. None of these bones have yet been dated. One complete Wrangel pelvis has a width of 1300 mm; the California Island pelvi are appreciably smaller, varying from 540 to 942 mm. Three holocene femora from Wrangel Island have lengths of 836 mm (subadult with unfused epiphyses), 985 mm and 1010 mm. Mammoths from the Channel Islands had smaller lengths for this bone varying from 590 to 842 mm (n = 14). The length of tibiae from adult specimens from Wrangel Island range from 431 to 519 mm (n = 5). Tibiae from the Channel Islands did not exceed 505 mm (n = I ), some were 400 - 500 mm (n = 4): and the majonty were between 300 and 400 mm (n ^ 1 1).

Two Holocene tusks from Wrangel Island have large curvature, with lengths of 255 and 280 cm (01); they have diameters at the alveoli of 1 20 and 121 mm. Channel Island tusks have lengths of the outside cuivature of 9-143 cm (n = 5) and diameters at the alveoli of 32-105 mm (n = 8). The Holocene Wrangel Island molars have a narrow crown (46-74 mm width for M/3) and a frequency of enamel plates (average of 10.1 in M/3) slightly exceeding Pleistocene Siberian mammoths. This is related to feeding on bush branches, the main food supply on Wrangel Island during the Holocene. The width of M/3 for Channel Island mammoths averages 63.5 mm (a range of 45-79 mm; n = 21 ); with a lamellar frequency of 8.1 (ranging from 6-10; n = 21 ). This represents an increased lamellar frequency of 20% for M. exilis.

By comparison of bone parameters it isapparent that the American mammoths generally had a smaller size although some individual bones (tibia for example) are quite comparable. Ourconcluisions indicate that the Channel Islands mammoths are dwarfed, whereas the Wrangel mammoths are no longer considered to be dwarfs. It is probable that the major difference between the two populations is the length of isolation. For Wrangel Island it was not more than 6,000 - 7,000 years. Also, during the winters, Wrangel Island was connected to the mainland by ice , allowing migration in both directions. Forthe Channel Islands the time of isolation is not fully known, but it apparently exceeds 40,000 years ago. The Channel Islands apparently were colonized by M columbi, v^hich swam the 6-9 km strait separating the ancient island from the mainland during Pleistocene sea level lowering. Once on the island, selective forces selected for smaller animals, resulting in new species animals (M exilis) wrth 150-180 cm shoulder height. It is possible for additional migrations ofM columbi from the mainland, but no dwarf nnammoths have been discovered on the Califomina coast. Holocene mammoths from Wrangel Island ranged from 18010 230 cm shoulder height and probably corresponded to the last Late Pleistocene populations in northern Siberia.

Late Pleistocenic mammoth populations were widely spread in Yakutia, the evidence of which are numerous burial places of mammoth remains, fossils and carcasses. Research of the floristic contents of the mammoths' gastric and intestinal tract allowed us to identify the food spectrum ofMammuthus primigenius. It turned out to be rather wide - ranging from branches, rind, bushes and shrub leaves, larch needles, grass to moss, with sedge and cereals prevailing. Five types of sedge were defined among seeds, frurts, and masses of integument tissue, and fibro-vascular bundles. The nutritional qualities and the ecology of plants, of which microelements were found in the mammoths' gastnc-intestinal tracts reveal the following: ( I ) grasses, growing in damp places have a higher nutritional value and preserve in winter most of their green mass; (2) needles, larch branches, leaves of some willows, birch, poplar, aspen, alder and horse-tail contain considerable protein and vitamins; (3) mosses, which have a lower nutritional value, are eaten by deer during fasting. Almost all the plants, eaten by mammoths, are in ration of modern representatives of mammoth fauna: horse, deer, elk etc. Horses, which, like mammoth, find food by themselves all year round, pasture mostly on swamped and shrub pastures in winter. Swamped pastures, that contain the largest phytomass and possess valuable nutritious and biological properties, are considered to be one of the best for horse pasturing. The associations Arctophila fulva, Eriophorum vaginato, L angustifolium, Ј. polystochion provide winter feeding for horses. Among shrub pastures the best are sedge-and-cereal-and-motley grasses willow-beds with Eonus -palustra, as well as Viluisky sedge growing places (Carex wiluica) wrth Salix pulchrai straightly growing sedge and Eriophorum vaginata. Their high feeding qualities are due to shrub leaves and grass plants rich in proteins and vitamins, preserved in considerable quantities in winter and autumn, as well as leaves of some willows and lower parts of grasses that are covered by snow. Horse and deer, as well as woolly rhinoceros being 'mammoth companions', had mostly the same feeding basis in the late Pleistocene. But large animals needed considerably more food. Due to lower anthropogenic pressure at that time, pastures productivity was higher than it is now. Mammoths ate meadow grasses in snowless seasons, and used the tusks to find fallen leaves, needles and grasses wrth green stems and sprouts in the snow. The most favourable time for mammoths was the end of the first interglacial period and the first glacial period, when summer was cool and wet, when forest areas were reduced, and swamps, swamped meadows and shrubs abounded. This is confirmed by numerous researches and datings of mammoth burial places, that date back to 44,000 - 26,000 kyBP - the second interglacial period, the time when cool low productivity steppes expanded. At that extra-arid and markedly continental period, mammoths suffered from a strong deficiency in green forage, and due to a shortage of vitamins and nutritious elements mammoths failed to resist winter frosts.

We may come to the conclusion: mass extinction of mammoths in Yakutia began in the dryest and coolest period - the Karginian interglacial. The main reason for the extinction was the disturbed forage balance, in addition to human hunting and a low fertility of the mammoths.

In the late Late Pleistocene, elephants still were quite abundant in China, mammoth (Mammuthus primigenius) is the most prosperous group, which mainly distributed in north-east China; other contemporary elephants are as follows: paleoloxodonts (Paleoloxodon naumanni) in north China, relics ofstegodonts (Stegodon orientalis) in south China, and E-lephas in many parts of the country.

Taxonomically, Mammutj"ius pnmigenius is the best verified group, because of its easily identified teeth and its unique geographical distribution; Stegodon onentalis is also a species widely accepted. As to Palaeoloxodon, there are always questions, some disagreements existed fora long time, from the 50's, such as different opinions proposed by Pei (1958), who thought Polaeoloxodon to be almost the same as Qephas, the Late Pleistocene elephant material from north China also should be put into the genus Etephos. In the most recent publications (Shoshani & Tassy 1996), Palaeoloxodon was combined into E/ep/ias. From this point of view three genera and species existed in China dunng late Late Pleistocene: Mammuthus primigenius, Stegodon orientals and Elephas maximus. Chronologically, mammoth had a very short history in China, it existed during the time span between 40,000 yBP and 12,000 yBP, it was really a 'passer-by' in this territory. Stegodont originated in China in the Pliocene, its golden time was Middle and Late Pleistocene, and it even survived into Holocene (5,000 yBP) but one record only (Ma & Tang 1992). E-iephas immigrated to China from late Pliocene, its localities were widely distributed in China during the Pleistocene, and it also survived into the Holocene. More than 10 Holocene localities have been reported, and today we also have a small group of extant Bephas near the south-western boundary.

Geographically, mammoth is only limited to the north-eastern part, and some rnatenal was also found on the bottom of the northernmost part of the Yellow Sea. E/ephos is the most widely distributed group during the Late Pleistocene. From North to South more than 25 localities have been reported in formal publications. Stegodont is also popular during the Late Pleistocene, more than 20 localities have been reported, but only limrted to south China.

In the late Late Pleistocene, 3 ecological zones can be recognized : the north-eastern China zone (up to 38њ N), which is typical oftheMammuthus-Coelodonta fauna; the northern China zone (between 38њ N and 28њ N) is represented by Coelodontci-Bephos (Po/aeotoxodon); and the southern China zone (south of the Yangtze River), which is dominated by Stegodon-Etephas. It can be seen that the environment changed gradually from North to South, each ecological zone shares at least one kind of animal with the neighbouring zone, the north-eastern China zone shares Coelodonta with the northern China zone. On the other hand, the northern China zone shares Elephas with the southern China zone.

There are only nine mammoth bone finds from Finland. Three of them have been radiocarbon dated earlier (Donnereta/. 1979). The molar from Espoo, southern Finland (Metzger 1921) was older than could be measured with radiocarbon dating 043,000 BP). The humerus found in Helsinki Herttoniemi (Donner 1965) yielded surprisingly young ages: Pear-son et o\. ( 1965) reported an age of 9,030 ± 65 BP, but later Donner et al. ( 1979) dated the humerus at 15,500 ± 200 BP. The femur from Lohtaja, western Finland (Okko 1949) gave an age of 25,200 ± 500 BP. The latter dates have puzzled geologists and palaeontologists since such young ages are neither in accordance with the glacial history ofFennoscandia norwrth the faunal history of the mammoths. New radiocarbon ages of dwarf mammoths from Wrangel Island in Northern Siberia (7,000 - 4,000 BP; Vartanyan et al. 1993, Vasil'chuk et al. 1997) arose again interest on mammoth datings in general. Lepiksaar ( 1992) had already suggested earlier that the finds in Helsinki and in Kunda, Estonia (9,780 ± 260 BP) could represent a late, isolated
mammoth population. In this research project, all relevant specimens were dated or redated with the AMS- technique. The results elucidate both the history ofthe mammoths in Fennoscandia, and the glacial history of the area.

The molar from lijoki, Northern Finland (Hoirn 1904, Korvenkontio 1915) gave an age of 31,970 ± 950 BP. The femur found in Lohtaja, western Finland (Okko 1949), was now dated to 24,450 ± 390 BP, which is similarto the earlier dating. Only one piece of tusk has been found in Finland. The tusk from Haapajarvi (Donner 1965) was dated to 28,740 ± 670 BP. The molars from Nilsia (Malmgren 1874-1 875, Ramsay 1900, Korvenkontio 1915) and Helsinki Todio (Korvenkontio 1915) yeilded ages of 22,420 ± 315 BP, and 23,340 ± 350 BP, respectively. The molar found in Espoo was not redated. The rib from Pohja BrOdtorp (Ramsay 1897; Rosberg 1901) and the piece of humerus from Tuulos (Rosberg 1924) have not yet been relocated.

Similar dates as those obtained in Finland have been recorded in Denmark (Aaris-Sorensen etal. 1990), Sweden (Berglund et al. 1976), and Norway (Heintz et ol. 1979, Follestad & Olsson 1979), but the reliability ofthe datings was questioned by the authors ofthe original papers. Lepiksaar (1992) noted, however, that these radiocarbon datings may indicate ice-free areas in Scania and the west coast of Sweden immediately prior to the Late Weichselian glacial maximum at about 20,000 BP.

New dating results of mammoth bones from Finland as well as the earlier dating of a reindeer antler from Tornio, northern Finland (34,300 +/- 2000/1450 BP: Siivonen 1975) are interesting since the number of radiocarbon dates older than the Late Weichselian glacial maximum indicate that there was probably a larger ice-free area in Finland and Sweden during a Middle Weichselian nonglacial interval than generally assumed. In addition, the new dates give invaluable information on the timing ofthe Scandinavian Ice Sheet as it advanced over the Finnish territory to its maximum position during the Late Weichselian.

Aaris-S0rensen, K., Peter-sen, K., Strand, K. & Tauber, H., 199O - Danish Finds of Mammoth (Mommuthus primigenius (Blumenbach)). Stratigraphical position, dating and evidence of Late Pleistocene environment - DGU Sene B 14: 44 pp.
Berglund, B.E., Hakansson, S. & Lagertund, E„ 1976- Radiocarbon-dated mammoth (Mammuthus primigenius Blumenbach) finds in South Sweden - Boreas 5: 177-191
Donner, J., 1965 - The Quaternary of Finland - in: Rankama, K. (ed.) - The Quaternary 1: 199-272
Donner, J., Jungner, H. & Kurten, B., 1979 - Radiocarbon dates of Mammoth finds in Finland compared with radiocarbon dates ofWeichselian and Eernian deposits - Bull. Geol. Soc. Finland 51: 45-54
Follestad, B. & Olsson, 1. U., 1979 - The "C age of the "Toten" mammoth, eastern Norway - Boreas 8: 307-312
Heintz, N., Games, K. & N/dal, FL, 1979 - Norske og sovjetiske mammutfunn i kvartaergeologisk perspektiv (English summary: Norwegian and Soviet-Russian Mammoth-finds in Quaternary Geological perspective) - in: Nyland, R„ Westin, S., Hafsten, U. & Gulliksen, S. (eds.) - Fortiden i s0kelyset - pp. 209-225. Trondheim
Holm, G., 1904 - Kindtand av mammut funnen vid ljo alfi Osterbotten i Finland - Geol. F5r. F5rh. 26: 238
Korvenkontio, VA, 1915 - Ein Mammutzahn-Fund in Helsingfors - Fennia 35/9: 1-15
Lepiksaar, J„ 1992 - Remarks on the Weichselian megafauna (Mommuthus, Coelodonto and Bison) of the 'intraglacial' area around the Baltic basin - Ann. Zool. Fennici 28: 229-240
Malmgren, A.J., 1874-75 - 0m mamrnut-fyndens forekornst och utbredning samt vilkoren for detta djurs fomtida existens - Ofversigt af Finska Vet. Soc. F5rhandl. 17: 139-154
Metzger, A., 1921 - Skelett av subfossil delfin frin Karsby i Tenala Kindtand av mammut frSn Esbo - Medd. Soc. Fauna Flora Fennica 48: 126
Okko, V., 1949 - Suomen Geologinen Yieiskartta, Lehti B 4, Kokkola Maalajikartan selitys - Geologinen Tutkimuslaitos, 108 pp.
Pear-son, F.J.Jr., Davis, E.M., Tamers, M.A. & Hohnstone, FLW., 1965 - University of Texas F^adiocarbcon Dates. Ill - Radiocarbon 7:296-314
Ramsay, W., 1 897-Ett refben af mammut antraffats i POJO - Fennia 15/1: 14-15
Ramsay, W„ 1900 - Finlands geologiska utveckling frSn istidematill v3.ra dagar. Helsinki, 55 pp.
Rosberg, J.E„ 1901 - Ett mammutfynd i den s.k. Br5dtorp-asen - Fennia 18/8: 1 -8
Rosberg, J.E., 1924 - Ett fossilt benfynd i Tuulos - Fennia 44/4: 1 -8
Siivonen, L, 1975 - New results on the history and taxonomy of the mountain, forest and domestic reindeer in Northern Europe- Proc, First Int. Reindeer and Caribou Symp., Fairbanks 1972 - Biol. Papers Univ. Alaska, Special Rep. 1:33-41
Vartanyan, S.L, Garutt, V.E. & Sher, A.V., 1993 - Holocene dwarf mammoths from Wrangel Island in the Siberian Arctic - Nature 362: 337-340
Vasil'chuk, Y., Punning, J.-M. & Vasil'chuk, A., 1997 - Radiocarbon ages of mammoths in northern Eurasia: implications for population development and late quatemal environment - Radiocarbon 39: I -1 19

Dispersals of proboscideans have attracted much interest from palaeontologists. The 'Proboscidean Datum Event' has been subject of many studies. After a long period of isolation Africa and the Indian Subcontinent became connected with Eurasia. This may have happened during a sea level low stand 21 My ago. There was some faunaJ exchange between Africa and India, but most ofthe dispersaJswere from Eurasia to the southern (sub)continents. Only later, dispersals to the North became more common. This was around 17.5 My ago. It has been hypothesised that Europe then had atropical climate, alternatively, the European climate may have been simply warmer or less seasonal than in the present. The dominant direction of dispersals may be indicative of climate at a particular time. Tropical, less seasonal or warm climates may have occurred at several moments in Europe. It seems likely that Gomphotherium reached the Indian Subcontinent during the first wave offaunal exchange, some 21 My ago. The only large mammal of African origin that seems to have reached Europe then is the anthracothere Brachyodus. The arrival of Gomphotherium in Europe seems to have been late in MN3 or early in MN4, more or less coincident with dispersals ofthe creodont Hycinailouros (with previous records in Africa and the Indian Subcontinent) and the cricetid
Democricetodon (ofAnatolian origin). This may have been around 17.5 My ago. Gomphothenum reached also America with a delay.

Deinotherium seems to have reached the Indian Subcontinent some 21 My ago, but entered Europe during MN4, some 1 6.5 My ago. This happened during a massive dispersal event, when also BunoUstriodon, Chalicotherium, vanous species ofthe tragulid genus Dorcatherium, and vanous rodents dispersed to the North. The dispersals to the North
were more important than those to the South and this is the time when Europe has been interpreted to have a tropical climate. The migrations first ofGomphotherium and later of Deinotherium might reflect the tendency towards the climaticaJ optimum. For some authors, the oldest European record ofZygolophodon is at the base of MN4. From MN6 onwards, the record is indisputable. The origin ofthis genus is in Africa. Oxygen isotope records are interpreted as indicating a major and definitive decrease in global temperature during the early Middle Miocene. Nevertheless, events with multiple dispersals to the North continued to occur. The first European record of Tetralophodon is late in MN7+8. The oldest record ofAnancus in Europe is in MN 12 in localities with an estimated age of about 7 My. Some authors supposed an evolution from Gomphotherium to Tetiviophodon to Anancus, whereas at present a more dominant view is that they represent different dispersals into Europe.

The oldest European record ofMammuthus is in MN 17, estimated at about 2.5 My. The dispersal ofMommuthus, of African origin, coincides with the dispersals to Europe ofEquus, of Amercian origin and the cervids Eudadoceros and Do/no, of Asian origins. The dispersal ofEquus is traditionally related to afirst cold phase, resulting in a sea level low. Around this time, deer reached forthe first time the Indian Subcontinent and the North of Africa. It has been hypothesised that cervids have been limited for millions of years in their southward expansion by an area that was too arid forthem to cross. The arid belt that runs at present from the North of Africa to the Middle East and continues into central Asia may have been present in some form for a long time, but it may have been iriore permeable to mammal dispersals 2.5 My ago. At this moment the dominant cyclicity ofthe climate changed from about 20 ky to 40 ky. The latter cycle is caused by variations in the obliquity ofthe axis of rotation ofthe earth and affects seasonality. The oldest European record ofE-lephas (Palaeoloxodon) is around I My old. The subgenus seems to have evolved in Africa some 3 My ago. About I My ago the dominant cyclicity ofthe climate changed from some 40 ky to 100 ky. The latter cycle is caused by variations in the periodicrty ofthe orbft ofthe earth. This resumed in an annual temperature cycle that, depending on the 20 ky precession cycle, may have been coincident or contrary to the seasonal cycle. Palaeoloxodon may have had a continuous record in Southern Europe and its geographical range expanded during the interglacials into central Europe.

The proboscideans that dispersed towards Europe, were predominantly of southern origins. In particular during the Miocene, the proboscideans dispersed northward accompanied by a number ofothertaxa. These dispersaJs cannot be explained as the simple result of high global temperatures, seasonality, or other single factors. The effect of climate on the vegetation is probably a limrting factor. The latest Pliocene and Pleistocene proboscideans were clearly adapted to differnt diets than the Miocene proboscideans and may therefore have been less limited by cold and/or seasonal climates.

Size extremes within the North Sea sample are far apart. This may reflect size increase beyond individual variability, as probably also occurred between the Olivola and Tasso faunal units (Azzaroli 1977), The morphological features of the majority of specimens fall within the range of the Italian sample and therefore probably also date from the Late Tiglian. The very low HI of some specimens leaves open the possibility of a Middle Tiglian age (TC4b - ?TCI ), which would imply derivation from the Smith's Knoll / IJmuiden Ground Fms, (see also: Van Kolfschoten & Laban 1995, Van Essen & Mol 1996). As the Italian skeletons referred to the Farneta faunal unit imply further size increase afterthe end of the Tiglian (Azzaroli 1977), some of the largest specimens of the sample may relate to post-Tiglian (Waalian) deposits of the Yarmouth Roads Fm. On the basis of the HI in (upper) M3, others are morphologically equivalent to the material from Dorst (NL), which implies derivation from stilt younger deposits within the Yarmouth Roads Fm. To date, the Oosterschelde has yielded few complete specimens. The accompanying fauna indicates an age greater than that of the Tegelen type locality from the Tiglian C5 (Van Kolfschoten & Van der Meulen 1986). There are faunistic arguments for contemporaneity with the Chilhac assemblage, France (± 1.9 Ma, Boeuf 1983, De Vos et a/.. 1998, Reumer et aL 1998), provided Anancus and Mammuthus are from the same level.

Most Maasviakte specimens are incomplete. Their HI is generally nearthe Valdamo Superiore mean. Some of the Bavelian molars from Dorst (Van Kolfschoten 1990) and Oosterhout seem to have remained indistinguishable from the 'typical form' of the Valdamo Superiore specimens, but others are advanced in HI or LF and reminiscent of specimens from Imola, N. Italy (Azzaroli & Berzi 1970). M2. and M3 have comparatively low modal ET values, Other inland sites include the Tegelen type locality, with scanty remains ofM. mendionalis studied by Guenther ( 1986). The stratigraphic provenance of most other inland specimens has been uncertain so far, but the material seems indi-cative of both earlier and later forms,

Azzaroli, A., 1977 - Evolutionary patterns ofVillafranchian elephants in Central Italy - Atti dellaAcadernia Nazionale dei Lincel, Memorie, Classe di Scienze fisiche, Ser. VIII, Vol. XIV, Sez. II' (Fasc. 4): 149 - 168
Azzaroli, A. & Berzi, A„ 1970 -On an Upper Villafranchian Fauna at Imola, Northern Italy, and its correlation with the Marine Pleistocene Sequence of the Po Plain - Palaeontografia Halica 66: 1-12
Boeuf, 0„ 1983-Le site Villafranchien de Chilhac (Haute-Loire), France. Etude paleontologique et biochronologique - Dissertation, Paris
De Vos, J„ Mol, D. & Reumer, J.W.F., 1998 - Eariy Pleistocene mammalian remans from the Oosterschelde or Eastern Scheldt (province ofZeeland, The Netherlands) - in: Van Kolfschoten, T. & Gibbard, P.L. (eds.) - The Dawn of the Quaternary. Proceedings of the SEQS-EuroMam symposium, Kerkrade 1 6-21 June 1996 - Mededelingen Nederiands Instituut voor Toegepaste Geowetenschappen TNO 60: 173-186
Guenther, E.W„ 1986 - Funde von Archidiskodon meridional'is und von Trogonthenum cuvien aus den interglazialen Tege-len-Schichten - Quartarpalaontologie 6: 53-65
Hooijer, DA., 1953 - On dredged specimens ofAnanais, Archidiskodon, and Equus from the Schelde estuary, Netherlands -LeidseGeologischeMededelingen 17:185-201
Hooijer, D.A., 1984 - Mammulhus meridionolis (Nesti) and M. armeniacus (Falconer) from the North Sea - Proceedings van de Koninklijke Nederiandse Akademie van Wetenschappen B 87 (3): 335-359
Reumer, J.W.F., Van Veen, J.C„ Van der Meulen, AJ„ HordiJk, L.W. & De Vos, J., 1998 - The first find of small mammals (Desmaninae, Arvicolidae) from the Early Pleistocene Oosterschelde fauna in The Netherlands - Deinsea -4: 41 -45
Fatten, LM.FL, 1909 - Die diluvialen Saugetiere derNiederiande - Dissertation, Utrecht
Van Essen, H. & Mol, D., 1996 - Plio-Pleistocene Proboscideans from the Southern Bight of the North Sea and the Eastern Scheldt (The Netherlands) - in: Shoshani, J. & Tassy, P. (eds.) - The Proboscidea; Trends in Evolution and Paleoecology - pp. 214-224. Oxford University Press, Oxford
Van Kolfschoten, T„ 1990 - The Early Biharian mammal faunas from Bavel and Dorst-Surae - Quartarpalaontologie 8: 265-272
Van Kolfschoten, T. & Laban, C., 1995 - Pleistocene terrestrial mammal faunas from the North Sea - Mededelingen Rijks Geologische Dienst 52: 135-151
Van Kolfschoten, T. & Van der Meulen, A.J., 1986 - Villanyan and Biharian mammal faunas from The Netherlands -Mernorie della Societa Geologica Italiana 31: 191 -200

Fauna 0, with Mimomys reidi, Mimomys tigliensis, Alces cf. gallicus, Eudadoceros sp. and a small cervid in the size category of Cervus rhenanus has an Earty Pleistocene age.

Fauna I, late Early- to early Middle Pleistocene in age, consists of a variety of species such as Mimomys sovini, Trogontherium cwien, Ursus aff. deningeri, Mammuthus meridionalis, Dicerorhinus etruscus brachycephalus, Hippopotamus antiquus, cf. Megoloceros verticomis and possible other Megaloceros species, Damo damo, Cervus elophus, Cervalces latifrons, Soergelia minor and Praeovibos cf. priscus. The smaller mammals assigned to Maasviakte Fauna. I correspond in many aspects with the fauna from Zuuriand level —28 to 36 with a late Early Pleistocene (Bavelian) or early Middle Pleistocene (eariy Cromerian) age. In both faunas a small Mimomys species and the advanced larger Mimomys savini are present and Allophoiomys is absent. Most of the large mammals of Fauna I have an evolutionary stage in between the Earty Pleistocene (Tiglian) mammals from Tegelen and the Middle Pleistocene (Late Cromenan) ones from Mosbach. The presence of e.g. Hippopotamus major indicates that Fauna I dates from an interglacial phase just before or afterthe transition of the Early to the Middle Pleistocene. Fauna I has an age which is most probably comparable to that of faunas such as Untermafifeld (Germany).

Fauna II, dates from the Late Pleistocene (Eernian? and Weichselian). The assemblage is composed of species (e.g. Mammuthus primigenius, Coelodonta antiquitotis, Megaloceros giganteus, Rangifer tarandus) that indicate 'glacial' conditions and a 'Mammoth Steppe' environment. Apart from the 'cold stage' elements of the Maasviakte Fauna II there are also remains of the straight-tusked elephant Polaeoloxodon antiquus and the fallow deer Damo damo that are assigned to the same assemblage. These remains might date from a warmer episode in the time range covered by the Maasviakte Fauna II assemblage.

Most of the specimens collected on the Maasviakte are from species assigned to Fauna III with a Holocene age. Domesticated animals are well represented in the fauna, in particular cattle. Many bones show cut-marks or are fragmented. Various types of bone-artefacts have been collected such as a fishhook, more than 20 harpoons, a comb made out of bone, and a flute.

1. Pollen and oxygen isotopic data indicate that the climate of the Sartan glaciation (the last glacial stage in Siberia) was extremely harsh, and that at this time the arctic territories were not a favourable natural habitat for the mammoth fauna. Mammoths and other large animals characteristic forthe late Pleistocene fauna) complex of Northern Asia populated the territory of Wrangel Island, as well as otherterritories of the Asian Arctic. Our data show that they were not numerous and migrated across extensive areas. Seasonal migrations were typical, and the northern portion of the habitat was occupied only in the summer time. Keeping in mind the variability of the environment forthat time, it is possible to suggest that significant changes in the composition of species populating the shelf area and Wrangel Island took place within a thousand-year span, and the same is true for morphology, migration routes, and winter stay. It is entirely possible that many high arctic areas were not visited at all by large land mammals during some intervals.

2. The interval from approx, 12,000 to 8,000 y ago was a critical one forthe late Pleistocene faunal complex. After 12,000 y ago the terminal Pleistocene warming (the Holocene optimum forthe Asian Arctic is found to have been 10 - 9,000 y ago) and the oceanic transgression cut off the northern portion of the mammoth habitat, to comprise a narrow strip including the present continental plain north of the Polar circle and a portion of the continental shelf area that was not transgressed at that time. However, the relative numbers of mammoths populating this 'strip' significantly exceeds that of the Sartan interval. This may be explained by better conditions of food resources, improved considerably by the north-ward shift of southern tundra vegetation, especially of shrubs. The latter were an extremely important component of the winter diet of mammoths. The reduction of the mammoth habitat in a southerly direction may have been caused by the formation of a continuous taiga zone, which was most likely impenetrable for these animals, and by the appearance of forest tundra landscapes with thick moss cover (this landscape can be found currently in Northern Yakutia).

3. Mammoths populated the shelf area of the East Siberian Sea priorto at least 7,500-8,000 y ago, from where they penetrated to Wrangel island already as the dwarf form. In the Taimyr, mammoths became extinct not earlier than 10,000-9,500 y ago. Therefore, bone remains of Holocene mammoths could possibly be found in Western Chukotka and on the New Siberian Islands. It is now known on the basis of radiocarbon dating that there were several Holocene refuges of Pleistocene fauna species in the Arctic: bison (9,500 y ago, Wrangel Island), musk-ox (up to 2,700 y ago, Taimyr), and horse (up to 3,000 y ago, Taimyr).

4. Inasmuch as a disunity of the mammoth habitat, as well as disunity of the habitat of another animals favoring open landscapes (other than reindeer, which can survive even in taiga), was determined by the formation of the continuous taiga zone, it is possible that some isolated populations of mammoth may had persisted in Siberia to the south of the taiga zone at least to 12,000-10,000 y ago (or even later), as has been found for some another animals of the Mammoth complex (bison, horses). Such an assumption receives support from carbon-dated Paleolithic sites of Southern Yakutia, although the bone remains themselves are not dated. On this basis, we can conclude mammoths continued to survive near the Pleistocene-Holocene boundary in isolated refuges that still provided them with more or less favourable natural conditions. However, the disunity of the habitat made the existence of this species fragile and susceptible to negative changes of the environment and especially to 'human pressure'. Isolated populations did not have an opportunit/ to migrate, or have exchanges with the remaining portions of similar habitat. Most likely, mammoths became extinct throughout most of their refuges only in the Holocene due to over-hunting. It could be suggested that this refuge strategy, workable for survival during periods of global warming, was successful as well for many of the late Pleistocene species during the last interglaciation, but failed during the Holocene because of wide human dispersal.

5. For at least the past 30,000 y the human habitation in Eastern Siberia (the lkhine, Ust-Mil 2, Berelekh sites, and others) was determined by the distribution dynamics of the main prey species, the most important of which was supposedly mammoth. The arctic territories were possibly not populated by humans during some periods of the Late Pleistocene, ortheywere occupied seasonally ortemporarily. In all probability, most ofthese territories became populated around the Holocene boundary. It is also possible that landscape dynamics and related changes of the composition of hunting species brought significant changes to human behavior. Survival strategies were modified on atime-continuum basis. This human response to unfavourable environmental changes (which were definitely unfavorable in the view of Siberian aboriginal people who have lost major
traditional food sources) could be archeologically recognized as the appearance of micro-prismatic technologies that very rapidly spread across Eastern Siberia. They appear approx. 9,000 y ago, and it takes them less than 3,000 y to get to the Taimyr, Chukotka, Southern Yakutia, and Zhokhov Island. Finally, rt is possible to maintain that human migrations from the Old to the New World across the Bering Land Bridge was strongly determined by the spatial and chronological dynamics of habitat of the mammoth and its satellite species.

About 6% of identifiable bones were gnawed by carnivores. Carnivore gnaw-marks are similar in morphology and size to marks made by wolf and spotted hyena. About 0.6% of the mammoth bones show visible teeth marks made by rodents. Most marks are situated on long bones and ribs. Cut marks are very rare, with only a very few bones showing the clear incisions that may have resulted from stone implements. Some bones (0.75% of identifiable elements) show marks that may be the result oftrampling, perhaps indicating the bone deposit was visited several times by mammoths before final burial. About 8% of the identifiable bones have marks that appearto be root-etching, a common bone-surface modification in wet grasses or sedges. This is expectable, because the mammoth bones had been deposited in tundra or tundra-steppe, and the bones were buried in wet, fine-grained sediments. Most bones are in weathenng stage 0 (no signs of decay). Only 8.2% of the total assemblage show the first signs of weathering (Behrensmeyer stage I ), and only 0.001 % (9 bones) show Behrensmeyer stage 2 and 3. Some bones therefore appear to have been exposed longer than others on ground surfaces.

Eariierthe bone accumulation has been interpreted as the remains oftwo orthree possible dwellings, disturbed by solifluction. But after new excavation appeard new interpretation these remains. The site may be the remains of a mass-dnve of mammoths by Ice-Age humans, who later butchered the animals. It is possible that this deposit is a noncultural accumulation. But an assemblage of so many bones from 71 individuals in avery smail spatial area (about 150 square meters) suggests a place where a prolonged process of bone accumulation occurred, and not a location where a single event took place. The data do not permit to decide ifSpadzista reflects human hunting on mammoth herds or on individual animals, or if the bones resulted from natural mortality of mammoths and the subsequent utilisation of the animal carcasses by PaJeoIrthic man.