Abstract: The fossiliferous deposits at Choukoutien, Loc. 1, where Peking Man lived, are some 40 metres in thickness, and may be divided into 13 layers from top to bottom. Their dates are summarized in Table 1. According to the dates, Layer 3 would be correlated with Stage 8 in the record of deep-sea sediments and Layer 10 would correspond to Stage 12. Just as Kahlke and Chow Benshun (1961) said, "There in no real break in the continuity of the record as well as in the composition of the fauna., ? So the Peking Man once experienced three cold stages (i.e. Stages 12, 10, 8) and two warm stages (ie. Stages 11, 9). Fossil fauna lists of the different layers of Loe. 1 (vertebrate) were given by Kahlke et al (1961) and Chia Lan-po (1978) (Table 2). In order to correlate the Choukoutien sequence associated with Homo erectus pekinensis (i.e. Layers 3—10) with the O18 record of deep-sea sediments (Fig. la), we have used cluster analysis and other mathematical techniques. As a consequence, there are two groups in Layers 3—11. The first group is composed of Layers 5, 8—9. The other layers belong to the second group (Fig. 2). In comparison with Layers 3, 6, 7, Layers 11, 10, 4 have longer distance from the first group (i.e. Layers 5, 8—9) (Fig. 3 or Fig. Id). Li Yanxian and Ji Hongxiang (1980) held that Layers 8一9 (38 forms in eluding ostrich) are characterized by forest animals, with an abundance of Carnivora, a scarcity of Rodent!a and an absence of grassland Rodentia. This suggests that the climate was warm and humid at the time. They also said that Layer 5 contained 29 forms, but still with more forest animals than grassland ones. Carnivora are aboundant. In addition there is an aquatic species in the fossil list (i.e. Trogontherium cuvieri). The rich finds have been made in the calcareous tufas or travertines. Lithological and palaeontological evidences suggest that the climat-e was warm and damp at the time. These layers are exactly the same as the first group in Fig. 2. They show a forest landscape. Li et al (1980) held that Layer 11 contained 29 forms excluding Honio erectus pekinensis, of which the ratio of grassland animals to forest animals is 55 : 45. In Layer 10 were recognized 28 forms (including ostrich), here the grassland animals are predominant over forest ones. They pointed out that Layer 4 gives 21 forms, including ostrich in it. Here Carnivora decreases while Rodentia increases in number. There are more grassland fossil animals than forest ones in Layer 4 with aridadapted fossil animals as the predominant forms. Just as mentioned above, the average distances between Layers 11, 10, 4 and the first group are much longer than those of the others (Fig. 3). Layers 11, 10, 4 show a predominantly grassland landscape, though there was some forest in the Choukoutien area then. The other layers in Fig. 2 consists essentially of intermediate types. Layers 3 is close to Layers 11, 10, 4 and shows a grasslandforest landscape as Li et al (1980) said. Layers 6一7 are close to the first group in Fig. 2. Li et al (1980) held that in Layers 5—9 forest animals exceed grassland animals and aquatic animals outnumber arid-adapted animals. Therefore the result of cluster analysis supports the view of Li Yanxian and Ji Hongxiang (1980).
According to Table 3 we calculate average of distances between each layer and Layer 5 and those between each, layer and Layers 8—9. Thus we get Fig.3 (or Fig. Id). Obviously, Layers 3—4 correspond to O18 Stage 8 and Layer 5 to Stage 9. So Layers 3—5 represent a completed glacial cycle, that is cycle D. Layers 6—7 would be correlated with Stage 10 and Layers 8—9 with Stage 11. So Layers 6一9 correspond to another completed glacial cycle, that is cycle E. Layer 10 would be correlated with Stage 12 or the late part of glacial cycle F. Therefore Peking Man's time includes at least three cold stages, i.e. Layers 3― , 6—7, 10； and two warm stages, i.e. Layers 5, and 8—9. Peking Man's time corresponds to glacial cycles D-F. On Palaeontological evidence, the Peking Man^ time would correspond to Holstein in Europe, i.e. glacial cycles E-F (Kukla. 1977). Therefore the correlation of the Asian fauna with the European fauna is roughly correct.
Marmot a bobdk, M. complicidens, Ochotona, Gulo, Ursus spelaeus and Coelodonta etc. are generally recognized as glacial elements. They are reported both from Layers 4, 6—7, 10—11 which represent the cold stages and from Layers 8—9 which represent the warm stage. Macaca, Rhizomys, Hystrix, Acinonyx, Ursus thibetanus, Palaeoloxodon namadicus and Bubalus etc. are generally recognized as interglacial elements, Under similar circumstances, they occur both in Layers 5, 8—9 which indicate warm climate and in Layers 3— , 6一7, 10一11 which represent cold climate. The coexistence of the glacial elements with the interglacial elements in Choukoutien sequence indicates that the climate of the three cold stages was not very cold and the climate of two warm stages was not very warm during Peking Man's time. Therefore the climate during Peking Man's time as a whole was similar to that of North China today. Kukla (1977) said, 4 4 Marine molluscs and foraminifers of Holstein epicontinental seas point to relatively chilly waters, similar to those of the present North Sea.'' Therefore the climatic changes of Asia was the same as Europe then.
The senior author Xu Qinqi (1980) held that the distribution of winter insolation at 35°N (Fig. lb) is critical to the variation of the mean annual temperature of the world. In Fig. lb the warm peaks in Stages 9,11 are not very high and they are obviously lower than those in Stage 7 or 15. So the climate would not be very warm during Stage 9 or 11. This is the fundamental cause why the climate of the two warm stages during Peking Man's time was not very warm. On an equal basis, the cold valleys in Stages 8, 10, 12 are not very low. So the climate would not be very cold during these stages. This is the fundamental cause why the climate of the three cold stages during Peking Man 7s time was not very cold.

Key words: palaeoclimate; Choukoutien Loc. 1; mammalian faunas