Palaeoecology of Southeast Asian megafauna-bearing sites from the Pleistocene and a review of environmental changes in the region more

by Julien Louys

Paleontology, Ecology, Community Ecology, Anthropology, Vertebrate Paleontology, Biogeography, Paleoenvironment, and Palaeoecology

Journal of Biogeography (J. Biogeogr.) (2010) ORIGINAL ARTICLE Palaeoecology of Southeast Asian megafauna-bearing sites from the Pleistocene and a review of environmental changes in the region Julien Louys1,2* and Erik Meijaard3,4 Research Centre in Evolutionary Anthropology and Palaeoecology, School of Natural Sciences and Psychology, Liverpool John Moores University, Liverpool L3 3AF, UK, 2School of Biological, Earth and Environmental Sciences, University of New South Wales, Sydney, NSW, 2052, Australia, 3People & Nature Consulting International, Vila Lumbung House No. 6, Jalan Raya Petitenget 1000X, Kerobokan, Badung 80361, Bali, Indonesia, 4School of Archaeology & Anthropology, Building 14, Ellery Cr., Australian National University, Canberra, ACT 0200, Australia 1 ABSTRACT Aim To reconstruct the palaeoenvironments of megafauna-bearing sites from Pleistocene Southeast Asia, and to describe general environmental changes in the region. Location Indochina and Sundaland, including Myanmar, Laos, Vietnam, Cambodia, Thailand, the Malay Peninsula, Borneo, Sumatra and Java. Methods This study reconstructs the habitat types of 25 Pleistocene sites in Southeast Asia through a synecological (community-based) method. This method speciﬁcally targets medium- and large-bodied mammals, and ecovariables that could be directly assessed from species lists were chosen. The methods allow the reconstruction of fossil sites as closed (continuous tree cover), mixed (heterogeneous tree cover) and open (very limited to no tree cover) through discriminant functions analysis of community guild structure. Results Four Pleistocene sites can conﬁdently be assigned to one of the three habitat types. Tam Hang (south), a Middle Pleistocene site from Laos, is classiﬁed as mixed. Ban Fa Suai, a Middle Pleistocene site from Thailand, is also classiﬁed as mixed. Trinil, a Middle Pleistocene site from Java, is classiﬁed as open. Lastly, Hang Hum II, a Late Pleistocene site from Vietnam, is classiﬁed as open. Insufﬁcient species are present in the fossil faunas of remaining sites to allow statistically conﬁdent habitat assignment. Nevertheless, conditional habitat assignments can be achieved, and these are largely congruent with other palaeoenvironmental evidence. Main conclusions Medium- and large-bodied mammals are the most frequently recovered mammals from Pleistocene sites in Southeast Asia. Previous palaeoenvironmental reconstructions of these sites have been hampered by this body size bias, as well as by limited site and faunal descriptions. However, our analysis demonstrates that reconstructions can still be achieved for megafauna-bearing sites in the region. The reconstructions suggest that through much of the Pleistocene, Southeast Asia had signiﬁcant areas of mixed habitats, and that the widespread distribution of rain forests, such as found today, was a relatively rare phenomenon. Keywords Discriminant functions analysis, ecovariable, mammals, megafauna, palaeoenvironment, palaeontology, Quaternary, Southeast Asia, synecology. *Correspondence: Julien Louys, School of Natural Sciences and Psychology, James Parsons Building, Liverpool John Moores University, Liverpool L3 3AF, UK. E-mail: j.louys@ljmu.ac.uk INTRODUCTION Southeast (SE) Asia (Fig. 1) is one of the most species-rich areas in the world (Myers et al., 2000), with, for example, ª 2010 Blackwell Publishing Ltd c. 13% of all mammal species occurring on about 1% of the world’s land mass. This species richness is probably the result of the region’s dynamic geological past (Holloway & Hall, 1998), and relatively stable environmental conditions, which www.blackwellpublishing.com/jbi doi:10.1111/j.1365-2699.2010.02297.x 1 J. Louys and E. Meijaard Figure 1 Southeast Asia, showing the locations of the fossil sites considered in this study. resulted in many speciation opportunities and relatively few extinction events (e.g. Whitmore, 1987; Meijaard, 2004; Louys, 2007, 2008; Louys et al., 2007). In fact, few taxonomic groups disappeared from SE Asia or became extinct in the Pleistocene, although some that did include proboscideans (Stegodon and Palaeoloxodon), carnivorans (Crocuta and Pachycrocuta) and hippopotamus (Hexaprotodon) (Louys et al., 2007). The faunas of SE Asia are subdivided into a number of zoogeographic subregions, of which two, the Indochinese and Sundaic (sensu Lekagul & McNeely, 1988), are the subject of the present study. The biogeography of these subregions remains at best partially understood; this is partly a result of the intense geological activity in this highly fragmented region, which has led to complex patterns of species diversity (e.g. Heaney, 1984, 1986; Ruedi, 1996), as well as of the limited information currently available on palaeo-faunas and -ﬂoras (Woodruff & Turner, 2009). 2 Some recent research has, however, begun to shed light on the palaeobiogeography, palaeoecology and palaeontology of the SE Asian region (e.g. Heaney, 1991; van den Bergh et al., 2001; Tougard, 2001; Bacon et al., 2004; Meijaard, 2004; Zeitoun et al., 2005; Storm & de Vos, 2006; Louys, 2008; Meiri et al., 2008; Westaway et al., 2009a,b). Owing in large part to the presence of hominins in the region from the beginning of the Pleistocene, some attempts have been made to understand regional palaeohabitats and palaeoenvironments. One of the best sources of information on palaeoenvironments, particularly as it applies to early humans, is that of the mammalian fauna associated with them, and SE Asia is no exception. In fact, the use of fossil faunas from SE Asia to infer palaeoenvironments and palaeobiogeography enjoys a long history (e.g. Dubois, 1908; Colbert, 1943). Until recently, however, all attempts at palaeoenvironmental reconstruction in the region have been qualitative and based on the presence Journal of Biogeography ª 2010 Blackwell Publishing Ltd Palaeoecology of Southeast Asian megafauna-bearing sites or absence of one or a small number of species (e.g. Medway, 1977; de Vos et al., 1994; van den Bergh et al., 2001). Although such studies can be quite informative, they fail to make use of the faunal community preserved in an integrated way. Multivariate synecological analyses, on the other hand, rely on comparisons between faunal communities in order to reconstruct palaeoenvironments. These methods examine the presence or abundance of taxonomic or ecological groups within a community, and allow the association of differences in representation with particular habitat types (e.g. Reed, 1997; Sandrock et al., 2007; Louys et al., 2009). A synecological analysis on SE Asian sites was recently carried out by Tougard & Montuire (2006), who attempted to reconstruct Pleistocene environments of the region using cenograms (rank-ordered graphs of body sizes in a community). However, because cenograms rely on the full spectrum of body mass distributions to reconstruct palaeoenvironments (see Legendre, 1986, 1989 and Travouillon & Legendre, 2009 for a discussion of this methodology), the heavy bias towards larger mammals in SE Asian fossil deposits makes the use of this method problematic for the region. This bias is in part due to many SE Asian excavations having being undertaken at the beginning of the last century, and therefore before modern excavation techniques, which screen and retain remains of smaller species, were developed. In the case of modern excavations, taphonomic bias (see, for example, Bacon et al., 2008a), as well as a dearth of micromammal specialists in the region, continue to produce a large-bodied bias. This bias hampers palaeoenvironmental reconstructions in the region, such that Tougard & Montuire (2006) were only able to assign a single palaeohabitat with any conﬁdence to a single fossil fauna out of the seven they examined, that of Thum Wimam Nakin, Thailand. Analytical methods that seek to reconstruct the palaeoenvironments of SE Asian sites must be equipped to deal with a large body size bias. Synecological analyses based on the proportions of species classiﬁed into ecological categories have been shown to be useful in reconstructing past environments (e.g. Reed, 1998; Mendoza et al., 2005). In addition, these analyses can be conducted using only the large-bodied subset of the entire fossil community, and as such do not suffer from the limitations affecting cenogram analyses. Furthermore, because SE Asia has suffered relatively few Pleistocene extinctions, modern faunal communities provide good comparative models for extinct ones. However, these types of analyses are subject to taphonomic limitations (Soligo & Andrews, 2005) as well as requiring the preservation of a minimum number of species (Louys et al., 2009). For example, Louys et al. (2009) calculated that to differentiate with conﬁdence between closed and mixed habitats, a site must contain at least 32 species, between mixed and open at least 12 species, and between closed and open at least eight species. Unfortunately, the target of 32 species is achieved in only a few sites in SE Asia. Nonetheless, classiﬁcations based on fewer species can provide an indication of environments present, especially when they are placed within a larger palaeoecological context (Louys et al., 2009). Palaeoenvironments of most SE Asian sites from Journal of Biogeography ª 2010 Blackwell Publishing Ltd the Pleistocene will therefore be proposed based on a synecological analysis of medium- and large-bodied mammals (‡ 1 kg, termed here ‘megafauna’) examined in the context of other palaeoecological information. The results presented herein represent the ﬁrst palaeoenvironmental reconstructions for many SE Asian sites on the basis of multivariate synecological methods. Furthermore, the ecological variables chosen for this analysis not only allow such reconstructions to be achieved relatively ‘materials-free’ for this understudied region, but also provide insights as to the evolution of faunal community structure concomitant with environmental changes. MATERIALS AND METHODS Modern species lists and habitat descriptions Multivariate synecological approaches to palaeoenvironmental reconstruction rely on modern faunal communities as a basis of comparison with fossil faunas. We used the species lists from 25 Asian nature reserves and national parks, where possible taxonomic bias in those lists had been controlled for (Louys et al., 2009). Each protected area represents one of three habitat types. Closed habitats are represented by protected areas with complete forest cover; mixed habitats have heterogeneous forest cover – signiﬁcant treeless areas can be present in these protected areas; and open habitats are almost completely bereft of trees (Louys et al., 2009). Species used in this analysis are listed in Appendix S1 in the Supporting Information. The geographical positions of the natural protected areas can be found in Louys et al. (2009). Fossil sites The palaeontological literature was examined for faunal lists from select Pleistocene SE Asian sites. Species from the following sites were included: Myanmar (Colbert, 1943): Mogok Caves, Irrawady beds; Thailand (Pope et al., 1981; Tougard, 1998; Zeitoun et al., 2005): Thum Wiman Nakin, Thum Phra Khai Phet, Kao Pah Nam, Ban Fa Suai; Laos (Fromaget, 1936; Arambourg & Fromaget, 1938): Tam Hang (north), Tam Hang (south) [previously referred to as Tam Nang (e.g. Arambourg & Fromaget, 1938); however, this was a typographical corruption of Tam Hang (A-M. Bacon, CNRS Paris, pers. comm., 2009; see also Bacon et al., 2008a)]; Vietnam (de Vos, 1983; Olsen & Ciochon, 1990): Keo Leng, Tham Khuyen, Tham Hai, Lang Trang, Tham Om, Hang Hum ´ I & II; Cambodia (Beden & Guerin, 1973): Phnom Loang; Malay Peninsula (Medway, 1972): Kinta Valley; Borneo (Medway, 1960, 1972): Niah Caves; Sumatra (de Vos, 1983): Lida Ajer, Sibrambang; Java (Hooijer, 1975; de Vos, 1983; Aimi & Aziz, 1985; van den Bergh et al., 1996, 2001): Ci Saat, Trinil, Kedung Brubus, Kali Glagah, Punung. Geographical locations of these sites are presented in Fig. 1; species used in the analysis are listed in Appendix S2. These sites were chosen as they all occur on continental SE Asia – that is, they are currently on the 3 J. Louys and E. Meijaard continent, or were connected to continental areas during Pleistocene periods of lower sea-level. Faunal communities from continental sites have similar faunas, body size ranges and diversity to the modern communities sampled. Oceanic parts of SE Asia that possess island communities, such as Sulawesi and the Phillipines, are therefore not included in this analysis. Discriminant functions analysis To assign extinct taxa to ecological variables that have previously been published (e.g. Reed, 1998; Mendoza et al., 2005) requires access to original fossils, or at least to detailed descriptions of individual specimens. It was not possible to get access to all SE Asian fossil faunas within the scope of this study, and for some sites this fossil material is incomplete. Detailed anatomical descriptions of fossils from SE Asia are rare, and in some cases descriptions of fossil material by early excavators were either perfunctory or even non-existent. The variables deﬁned in this study have been chosen such that they can be determined, in almost all cases, directly from species lists themselves. They are based where possible on extant members of the fossil species. In the case of extinct species or where information was missing for extant species, they are based on the most closely related taxa, within the same genus wherever possible. Because of this, these variables cannot be considered ‘taxon-free’ or ‘phylogeny-free’. To truly achieve such a designation, extinct fauna in such an analysis would need to be subjected to a taxon-free ecomorphic study or similar treatment in order to determine palaeobiology. For reasons mentioned above, this cannot be done for many SE Asian sites. In any case, as far as synecological studies are concerned, we eschew the use of the terms ‘taxon-free’ or ‘phylogeny-free’, and prefer instead to refer to the variables as ecological (‘ecovariables’). All ecological information for extant species was obtained from Nowak (1999). Ecological variables were determined for each species using three basic categories. First, each species was placed in one of ﬁve body size classes according to average body mass (cf. Andrews et al., 1979; Gagnon, 1997; Andrews & Humphrey, 1999; Soligo & Andrews, 2005), and as such are comparable. The body size categories are: tiny (A) < 1 kg; small (B) 1– 10 kg; medium (C) 10–45 kg; large (D) 45–180 kg; and very large (E) > 180 kg. Taxa belonging to body weight category A are excluded from further analysis here: they are not adequately represented in faunal lists from Pleistocene sites in SE Asia. In addition, small mammals show increased levels of endemism relative to larger mammals, in particularly those occurring in montane regions of SE Asia (e.g. Heaney, 1984, 1986). As the modern and fossil sites examined in this study largely reﬂect lowland (as opposed to montane) communities, tiny-bodied (body size category A), endemic taxa are again not applicable in this study. The second category was that of trophic guild: speciﬁcally, primary (P) and secondary (S) consumers (as in, for example, Soligo & Andrews, 2005). Secondary consumers are hyper4 carnivores, carnivores, piscivores, omnivores, insectivores, as well as any other trophic guild in which non-plant matter makes up a signiﬁcant part of the diet. Primary consumers are those whose intake of non-plant matter is very small or nonexistent. The ﬁnal category groups species on the basis of their arboreality. Those species spending some to all of their time in trees are classiﬁed as arboreal (A), and thus this category includes scansorial species. Terrestrial species (T) are deﬁned as those that spend no time in trees (i.e. strictly terrestrial). Again, where information on arboreality was not available, it was estimated by comparison with closest related taxa. Note that the term ‘arboreal’, as applied herein, implies potential, not strict, arboreality. Although some species may have been incorrectly classiﬁed (owing to the limited amount of ecological information available for many extant SE Asian mammals), the use of multivariate analyses ensures that, unless the majority of species are incorrectly classiﬁed, ecologically meaningful results will still be obtained. Sixteen ecological categories were established by combining the remaining four body weight categories, the two trophic categories and the two locomotor categories (see Appendix S3). However, two of these categories are unﬁlled by either extant or extinct species considered herein (very large, arboreal primary and secondary consumers). Therefore, a total of 14 ecological categories were used throughout the analyses. Species from each nature reserve and national park discussed in Louys et al. (2009) were therefore allocated to these categories, and percentage representation was calculated (see Appendix S1 and Table 1, respectively). These ecological variables were subjected to a principal components analysis (PCA) in order to verify that the habitat groupings deﬁned by Louys et al. (2009) were reﬂected by the ecological categories deﬁned above. This analysis was run on past software (Hammer et al., 2001). In addition to conﬁrming groupings, an examination of the loadings for each of the components allows a determination of which ecological variables characterize each habitat type. Changes from one habitat type to another should be reﬂected in changes to these loadings. A canonical discriminant functions analysis (DFA) calculates the relationships between variables belonging to pre-deﬁned groups in order to assign unknown variables to those groups. The most often used variables in palaeoecological analyses are either taxonomic (e.g. Sandrock et al., 2007; Louys et al., 2009) or ecological (e.g. Reed, 1998; Mendoza et al., 2005). Taxonomic variables are not used in the following study. Although a range of habitat types are represented by the natural protected areas discussed above, those reserves and parks sampled from SE Asia are exclusively closed. In contrast, open habitats are represented exclusively by Chinese parks and reserves. Therefore, in order to avoid a potential zoogeographical bias in the analysis, ecological variables were preferred over taxonomic. Fossil taxa from the 25 sites in SE Asia were likewise deﬁned by the ecological variables (see Appendix S2 and Table 2). Percentages were standardized for both modern and fossil lists Journal of Biogeography ª 2010 Blackwell Publishing Ltd Palaeoecology of Southeast Asian megafauna-bearing sites Table 1 Proportional representation of medium- and large-bodied mammalian communities found in select national parks and reserves in Asia, broken down according to ecological guilds. B, small; C, medium; D, large; E, very large; A, arboreal; T, terrestrial; P, primary consumer; S, secondary consumer. Reserve/Park Ailaoshan Changshanerhai Gaolingong Maolan Medong Sai Yok Khao Yai Kaeng Krachan Erawan Kerinci Seblat Dinghushan Fanjingshan Wuyishan Funiushan Nujiang Tianmushan Yellow River Maojingba Yunwushan Longqi Poyang Baima Xueshan Bogdhad Jiuzhaigou Altun Mountain Classiﬁcation Closed Closed Closed Closed Closed Closed Closed Closed Closed Closed Mixed Mixed Mixed Mixed Mixed Mixed Mixed Mixed Mixed Mixed Mixed Open Open Open Open BPA 0.14 0.09 0.07 0.11 0 0.14 0.09 0.09 0.09 0.16 0 0.09 0.03 0 0.09 0 0 0 0 0.05 0 0.13 0 0.1 0 BPT 0.07 0.18 0.1 0.16 0.05 0.06 0.09 0.09 0.09 0.09 0.06 0.09 0.13 0.08 0.09 0.07 0.25 0.08 0.13 0.14 0.14 0 0 0 0.13 BSA 0.14 0.14 0.14 0.11 0.05 0.11 0.16 0.15 0.13 0.18 0.13 0.09 0.1 0.17 0.15 0.11 0.25 0.08 0.13 0.14 0.29 0.13 0.11 0.1 0.06 BST 0.1 0.09 0.1 0.16 0.16 0.08 0.13 0.03 0.13 0.09 0.31 0.18 0.2 0.21 0.18 0.21 0.25 0.25 0.38 0.14 0.14 0 0.11 0.1 0.13 CPA 0.1 0.09 0.14 0.05 0.16 0.08 0.03 0.09 0.09 0.04 0 0.09 0.07 0 0.06 0.04 0 0 0 0.05 0 0.13 0 0 0 CPT 0.21 0.05 0.1 0.26 0.26 0.08 0.06 0.09 0.07 0.04 0.13 0.15 0.13 0.21 0.12 0.18 0 0.08 0 0.1 0.07 0.25 0.11 0.2 0.13 CSA 0.14 0.14 0.1 0 0.05 0.11 0.13 0.09 0.11 0.13 0.13 0.09 0.13 0.04 0.03 0.14 0 0 0 0.1 0.14 0.13 0 0 0 CST 0 0 0.03 0 0 0.03 0.06 0.09 0.02 0.09 0.06 0.09 0.07 0.13 0.12 0.11 0.25 0.17 0.25 0.05 0.14 0 0 0 0.06 DPA 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 DPT 0 0.05 0 0 0.05 0 0 0 0.02 0 0 0 0 0 0.03 0 0 0 0 0.05 0 0.25 0.33 0.3 0.19 DSA 0 0.05 0.03 0.05 0.05 0.03 0.03 0.03 0.02 0 0.06 0.03 0.03 0.04 0.03 0.04 0 0.08 0 0.05 0 0 0 0.1 0 DST 0.03 0.05 0.07 0.05 0.05 0.11 0.09 0.09 0.09 0.04 0.06 0.09 0.07 0.08 0.09 0.07 0 0.17 0.13 0.1 0 0 0.11 0 0.19 EPT 0.03 0.09 0.07 0 0.11 0.14 0.09 0.15 0.11 0.09 0 0 0 0 0.03 0 0 0 0 0 0 0 0.22 0.1 0.13 EST 0.03 0 0.03 0.05 0 0.03 0.03 0.03 0.02 0.04 0.06 0.03 0.03 0.04 0 0.04 0 0.08 0 0.05 0.07 0 0 0 0 by taking the square-root of the raw values. The standardized values from modern habitats were used to deﬁne the discriminant functions, calculated using spss Standard Version, Release 11.0.1 (http://www.spss.com). The functions were cross-validated by leaving one reserve or park out, re-running the analysis using the remainder of the protected areas, and classifying the omitted reserve. The discriminant functions were then used to classify all Pleistocene sites. To assess the accuracy of palaeoecological predictions for Pleistocene sites on the basis of fossil fauna compositions, we compared these results with palaeoecological reconstructions obtained in other ways and available in the scientiﬁc literature. RESULTS Species analysis The results of the PCA (Fig. 2) conﬁrmed the characterization of the three habitat types as established in Louys et al. (2009) on the basis of the ecological variables deﬁned in this paper. In particular, the ﬁrst two components [explaining 43% and 19% of the variation, respectively (Table 3)] best separated the different habitat types. Negative values of PC 1 were generally indicative of closed and open habitats, with the more negative values associated with open habitats. Positive Journal of Biogeography ª 2010 Blackwell Publishing Ltd values for this component were generally indicative of mixed habitats. PC 2 most clearly differentiated closed habitats from mixed and open habitat types, with the latter values negative or close to zero. An analysis of the biplot and loadings for these two components (Fig. 2; Table 3) suggested that the variable DPT (large terrestrial primary consumers) helped separate open habitats; BPT (small terrestrial primary consumers), BSA (small arboreal secondary consumers), BST (small terrestrial secondary consumers), CST (medium terrestrial secondary consumers) helped separate mixed habitats; and BPA (small arboreal primary consumers), CPA (medium, arboreal primary consumers), CSA (medium arboreal secondary consumers) helped separate closed habitats, whereas DPT (large terrestrial primary consumers) and DST (large terrestrial secondary consumers) separated mixed/open from closed habitats. An examination of the PCA results showed that open habitats, such as those represented by grasslands and deserts, were characterized by the smallest number of ecological variables. The variables separating open habitats from all other habitats were large, terrestrial primary consumers (Fig. 2; Table 3). These results were congruous with that expected: that open habitats would have a greater relative proportion of larger, terrestrial primary consumers, especially grazers. This variable was also a selected ecological variable with regard to 5 J. Louys and E. Meijaard Table 2 Proportional representation of medium- and large-bodied mammalian communities found in select Pleistocene fossil sites in Southeast Asia, broken down according to ecological guilds. B, small; C, medium; D, large; E, very large; A, arboreal; T, terrestrial; P, primary consumer; S, secondary consumer. No. megafauna species 27 20 31 25 19 21 24 11 9 21 23 18 8 22 18 20 29 8 34 7 11 8 19 13 36 Pleistocene site Tham Khuyen Tham Hai Tham Om Hang Hum 1 Hang Hum 2 Lang Trang Keo Leng Kali Glagah Ci Saat Trinil Kedung Brubus Punung Kinta Valley Niah Lida Ajer Sibrambang Thum Wimam Nakin Thum Phra Khai Phet Ban Fa Suai Kao Pah Nam Phnom Loang Mogok Caves Irrawaddy beds Tam Hang (north) Tam Hang (south) BPA 0.04 0 0 0 0 0.05 0.04 0 0 0.05 0.04 0.11 0 0.14 0.22 0.2 0.07 0 0.03 0 0 0 0 0 0 BPT 0.11 0.05 0.06 0.08 0.11 0.05 0.08 0 0 0.05 0.04 0.06 0 0.09 0.06 0.05 0.03 0 0.03 0.14 0 0.13 0 0.08 0.06 BSA 0.04 0.05 0.1 0.08 0.11 0 0.04 0 0 0 0 0 0 0.05 0 0 0.07 0 0 0 0 0 0 0 0.11 BST 0.07 0.1 0.1 0.08 0 0.1 0.08 0.09 0 0 0.04 0 0.13 0.09 0 0 0.1 0 0.06 0 0.09 0 0 0 0.03 CPA 0.07 0.05 0.03 0.04 0.05 0.05 0.13 0 0 0.05 0.04 0.11 0 0.09 0.11 0.05 0.03 0 0.03 0 0 0 0 0.08 0.08 CPT 0.04 0 0.16 0 0 0.1 0.13 0 0.22 0.1 0.04 0.11 0 0.05 0.11 0.1 0.07 0.25 0.18 0 0.09 0 0 0 0.14 CSA 0 0 0.03 0.04 0.05 0.05 0 0 0 0.05 0.04 0.06 0 0.14 0 0 0 0 0 0 0 0 0 0 0 CST 0.04 0.05 0.03 0.12 0.05 0.05 0 0 0 0.05 0 0 0 0 0 0 0.03 0.13 0.03 0.14 0.09 0 0 0 0.08 DPA 0.04 0.05 0 0 0.05 0.05 0.04 0 0 0 0 0.06 0 0.05 0.06 0.05 0.03 0 0.03 0 0.09 0 0 0.08 0.03 DPT 0.15 0.25 0.13 0.12 0.16 0.1 0.13 0.27 0.33 0.24 0.13 0.06 0.13 0.09 0.11 0.1 0.21 0.25 0.18 0.29 0.18 0.25 0.16 0.08 0.08 DSA 0 0.05 0.03 0.04 0.05 0.05 0 0 0 0 0.04 0 0 0.05 0 0 0 0 0 0 0 0 0 0.08 0 DST 0.07 0.1 0.06 0.08 0.05 0.14 0.04 0 0 0.05 0.09 0.11 0 0.05 0.06 0.1 0.03 0 0.09 0.14 0.09 0 0 0.15 0.08 EPT 0.22 0.2 0.23 0.2 0.26 0.19 0.17 0.45 0.33 0.29 0.39 0.22 0.63 0.09 0.17 0.25 0.24 0.25 0.26 0.29 0.36 0.5 0.79 0.38 0.22 EST 0.11 0.05 0.03 0.12 0.05 0.05 0.13 0.18 0.11 0.1 0.09 0.11 0.13 0.05 0.11 0.1 0.07 0.13 0.09 0 0 0.13 0.05 0.08 0.08 the DFA (Table 4), again highlighting this characteristic of open habitats. Closed habitats were differentiated from open and mixed habitats by large representation in three variables (Fig. 2; Table 3). All of these were arboreal categories (small and medium primary consumers, and medium secondary consumers). Smaller proportions of large terrestrial primary and secondary consumers differentiated between closed and mixed/ open habitats along PC 2. Of the ecological variables helping to separate closed habitats, only medium arboreal secondary consumers were selected by the DFA in the classiﬁcation of habitats (Table 4). Mixed habitats were separated in the PCA by large proportions of four ecological variables: small terrestrial primary consumers, small secondary consumers, both arboreal and terrestrial, and medium terrestrial secondary consumers (Fig. 2; Table 3). Of these only the proportion of small and medium terrestrial secondary consumers was selected by the DFA in differentiating between habitats (Table 4). We note that no large mammals were necessary in differentiating mixed habitats. The bias in most SE Asian fossil sites towards larger and more iconic species may therefore provide a potential bias against a mixed habitat classiﬁcation. 6 On the basis of these results, some expectations of changes in the proportions of the ecological variables can be determined across suspected ecological changes. In particular, the presence of a central savanna corridor running through the middle of Sundaland, as proposed by a number or researchers (e.g. Heaney, 1991; Meijaard, 2003; Bird et al., 2005), should be traceable through an ecological analysis of fossil faunas. A change from a closed (e.g. rain forest) habitat to either a mixed or an open (e.g. open forest or grassland) habitat should be reﬂected by a signiﬁcant decrease in the proportion of medium arboreal secondary consumers. Mammals from Pleistocene sites described by this variable include medium-sized cats (e.g. Catopuma and Prionailurus species), the pangolin (Manis), the clouded leopard (Neofelis) and the binturong (Arctictis). Concomitant with these are relative increases in the proportions of large and very large terrestrial primary consumers, and small and medium terrestrial secondary consumers. Mammals from Pleistocene sites described by these variables include the dhole (Cuon), Asian spotted hyena (Crocuta), as well as cervids, bovids, rhinos and elephants. Differentiating between open and mixed habitats is more difﬁcult than differentiating between closed and open habitats. Although component 2 in the PCA separates closed habitats, Journal of Biogeography ª 2010 Blackwell Publishing Ltd Palaeoecology of Southeast Asian megafauna-bearing sites Table 3 Principal components analysis of modern medium- and large-bodied mammalian communities found in select Asian national parks and reserves, with communities broken down according to ecological guilds. Eigenvalues, percentage variance and principal components (PC) loading for the ﬁrst three principal components are listed. B, small; C, medium; D, large; E, very large; A, arboreal; T, terrestrial; P, primary consumer; S, secondary consumer. Ecological category Eigenvalue Percentage variance accounted for BPA BPT BSA BST CPA CPT CSA CST DPA DPT DSA DST EPT EST PC 1 0.02315 42.775 )0.3263 0.2528 0.3011 0.4228 )0.1462 )0.2869 )0.0203 0.4047 3.99 · 10)24 )0.4847 )0.0221 0.06011 )0.2212 0.0663 PC 2 0.010194 18.835 0.3207 0.1088 0.1408 )0.3239 0.3201 0.04314 0.4429 )0.1861 6.97 · 10)21 )0.6055 )0.02196 )0.1915 )0.1248 0.07741 PC3 0.007198 13.299 )0.0397 )0.1243 )0.3479 0.3873 0.07771 0.649 )0.06991 )0.05982 7.84 · 10)19 )0.1614 0.1508 )0.05152 )0.4704 0.06015 Table 4 Discriminant functions analysis of modern mediumand large-bodied mammalian communities found in select Asian national parks and reserves, with communities broken down according to ecological guilds and values standardized. Canonical discriminant function coefﬁcients are listed. B, small; C, medium; D, large; E, very large; A, arboreal; T, terrestrial; P, primary consumer; S, secondary consumer. Function 1 BST CSA CST DPT EPT 1.029 0.799 1.207 0.731 )1.183 2 )0.260 )0.309 0.530 1.213 )0.399 Figure 2 Principal components analysis (PCA) and biplot of modern reserves and parks in Asia using ecological characteristics of their mammal faunas. (a) PC 1 versus PC 2; (b) PC1 versus PC 3; (c) PC 2 versus PC 3. B, small; C, medium; D, large; E, very large; A, arboreal; T, terrestrial; P, primary consumer; S, secondary consumer. Journal of Biogeography ª 2010 Blackwell Publishing Ltd it does not separate mixed from open habitats. Nevertheless, signiﬁcant relative changes between the proportions of large terrestrial primary consumers and small and medium terrestrial secondary consumers should indicate a transition from mixed to open habitat. The results of the DFA of modern faunas are shown in Fig. 3 and Table 4. Five variables were selected by the DFA as useful in classiﬁcation, these being BST (small terrestrial secondary consumers), CSA (medium arboreal secondary consumers), DPT (large terrestrial primary consumers), CST (medium terrestrial secondary consumers) and EPT (very large terrestrial primary consumers). The ﬁrst function calculated accounted 7 J. Louys and E. Meijaard Figure 3 Discriminant functions analysis showing the classiﬁcation of select Southeast Asian Pleistocene sites (circles). Colour scheme: dark grey circles represent closed; light grey circles represent mixed; black circles represent open habitat classiﬁcations. Dark grey diamonds represent closed; light grey squares represent mixed; and black crosses represent open modern parks and reserves. Black solid squares represent the group centroids for each habitat type. for 62.1% of the variance observed (0.953 canonical correlation), and the second accounted for 37.9% of the variance (0.926 canonical correlation). All reserves were correctly reclassiﬁed in the leave-one-out cross-validation analysis. Each fossil site was subsequently allocated to a habitat on the basis of these functions (Table 5). Palaeoenvironmental classiﬁcation The sites of Thum Phra Khai Phet and Kao Pah Nam (Thailand) are classiﬁed as open habitats. Although these sites have sufﬁcient numbers of species to differentiate between open and closed habitats, they do not have enough to allow differentiation between open and mixed environments. On the basis of their discriminant scores, however, they are more likely to represent open than mixed habitats. Trinil and Hang Hum II are also classiﬁed by the DFA as open, although an examination of their position relative to the habitat group centroids shows that they fall much closer to mixed habitats than the aforementioned Thai sites. They preserve, however, sufﬁcient numbers of species to differentiate between open and mixed habitats, and as such should be classiﬁed as open. Only two sites could conﬁdently be allocated to a mixed habitat: Tam Hang (south), Laos and Ban Fa Suai, Thailand. Their positions are, however, close to those of Trinil and Hang Hum II, both classiﬁed as open sites. These four sites therefore in all likelihood preserve a gradation from more open habitats (Trinil and Hang Hum II) to slightly more wooded habitats [Tam Hang (south) and Ban Fa Suai], on the basis of the mammals present. All other sites, classiﬁed as either mixed or closed habitats, do not preserve sufﬁcient numbers of species for them to be conﬁdently attributed to either habitat. What can be determined, however, is that these sites, with the exception of Ci Saat, Mogok Caves and Kinta Valley, are unlikely to represent 8 open habitats. These last three sites, although preserving sufﬁcient numbers of species to differentiate between closed and open habitats, do not preserve sufﬁcient numbers to differentiate between mixed and open and mixed and closed habitats, therefore making unambiguous classiﬁcations difﬁcult to establish. DISCUSSION Sundaland Java The climate of Java during much of the Miocene is thought to have been wet and warm, and land would have been covered mostly by rain forest and peat-swamps (Morley, 1991, 2000; Whitten et al., 1996). During the Late Miocene and Pliocene, cooler and drier conditions probably prevailed (Morley, 1999; Morley et al., 2000), with climatic drying and concomitant expansion of savanna in the Central Java region suggested at the Pliocene–Pleistocene boundary (Rahardjo, 1999). This change in vegetation is not reﬂected in our data, which suggest that both the Early Pleistocene Ci Saat (c. 1200 ka; van den Bergh et al., 2001) and Kali Glagah faunas in Central Java are classiﬁed as representing closed habitats (although Ci Saat may represent mixed or open habitats and Kali Glagah mixed habitats). Ci Saat has previously been interpreted to represent a more open habitat on the basis of pollen evidence (de Vos et al., 1994), and a similar palaeoenvironment was suggested for the Early Pleistocene Mojokerto region, where pollen indicates extensive grassland with hardly any trees (R. Morley, Palynova, Indonesia, pers. comm., 12 April 2003). However, an examination of Ci Saat’s position relative to the group centroids for the habitats (Fig. 3) shows that its position Journal of Biogeography ª 2010 Blackwell Publishing Ltd Table 5 Habitat allocations of select Pleistocene fossil sites found in Southeast Asia based on a discriminate functions analysis of modern medium- and large-bodied mammalian communities found in Asian national parks and reserves, with communities broken down according to ecological guilds and values standardized. Other habitat possibilities are listed for sites where the number of taxa in that site does not reach the minimum number required to differentiate conﬁdently between habitat types. Possible oxygen isotope stages (OIS) corresponding to each site are also indicated. Other DF habitat possibilities (more likely/less likely) Age (ka) Chronological reference Journal of Biogeography ª 2010 Blackwell Publishing Ltd Site Primary DF classiﬁcation Glacial/Interglacial (OIS) Irrawaddy beds Ci Saat Kali Glagah 1200 van den Bergh et al. (2001) Mixed/Open Closed Mixed Mixed Mixed/Open Mixed Closed Mixed 690 169 169 900 700–800 van den Bergh et al. (2001) van den Bergh et al. (2001) ?Intergacial (OIS23) Largely Glacial (OIS17-20) Mogok Caves Phnom Loang Trinil Kedung Brubus Tam Hang (south) Tam Hang (north) Kinta Valley Kao Pah Nam Thum Wimam Nakin Thum Phra Khai Phet Ban Fa Suai Tham Om Tham Khuyen Closed Closed Closed Closed Mixed Mixed Mixed Closed Closed Mixed 475 39–45 128 128 128 80–140 80–140 128 20–30 140–250 475 Tham Hai Niah Punung Lida Ajer Sibrambang Hang Hum I Hang Hum II Lang Trang Keo Leng Closed Mixed Open Mixed Closed Mixed Closed Closed Mixed Mixed Mixed Mixed Closed Mixed Open Closed Mixed Closed Closed Open Mixed Open Closed Closed Closed Mixed Mixed/Open ?Glacial (OIS36) Pope et al. (1981) Esposito et al. (2002) Based on correlation with Thum Wiman Nakin, as per Tougard (2001) Olsen & Ciochon (1990) Ciochon et al. (1996) Based on correlation with Tham Khuyen, as per Olsen & Ciochon (1990) ?Interglacial (OIS17) Glacial (OIS6) Glacial (OIS6) Interglacial (OIS3) Interglacial (OIS5) Interglacial (OIS5) Interglacial (OIS5) ?Interglacial (OIS5)/?Glacial(OIS4,6) ?Interglacial (OIS5)/?Glacial(OIS4,6) Interglacial (OIS5) ?Glacial (OIS2)/?Interglacial (OIS3) Barker et al. (2007) Westaway et al. (2007) Based on correlation with Punung, as per de Vos (1983) Based on correlation with Punung, as per de Vos (1983) Olsen & Ciochon (1990) Olsen & Ciochon (1990) Based on correlation with Punung, as per de Vos (1983) Olsen & Ciochon (1990) Period Country Early Pleistocene Early Pleistocene Early Pleistocene Myanmar Java Java Middle Middle Middle Middle Middle Middle Middle Middle Middle Middle Pleistocene Pleistocene Pleistocene Pleistocene Pleistocene Pleistocene Pleistocene Pleistocene Pleistocene Pleistocene Myanmar Cambodia Java Java Laos Laos Malaysia Thailand Thailand Thailand Middle Pleistocene Middle Pleistocene Middle Pleistocene Thailand Vietnam Vietnam Middle Pleistocene Vietnam ?Glacial (OIS6,8)/?Interglacial (OIS7) Boundary between glacial (OIS12) and interglacial (OIS13) Boundary between glacial (OIS12) and interglacial (OIS13) Late Pleistocene Late Pleistocene Late Pleistocene Borneo Java Sumatra Late Pleistocene Sumatra Late Pleistocene Late Pleistocene Late Pleistocene Vietnam Vietnam Vietnam Palaeoecology of Southeast Asian megafauna-bearing sites Late Pleistocene Vietnam 9 J. Louys and E. Meijaard approaches that of open habitats. Mixed habitats, with some forest but also signiﬁcant open (i.e. grassland) areas, are more probable for these sites. These habitats can occur in areas with long dry seasons, where abrupt forest to open grassland ecotones can arise, possibly maintained by natural ﬁre. This interpretation is supported by palaeosol studies of the Solo Basin, central Java (Bettis et al., 2009), which indicate an increased duration of the dry season during the early part of the Pleistocene. Extensive forest areas remain in wetlands, around rivers, and on mountain slopes, with drier areas kept open by ﬁre and grazing. This seems to be conﬁrmed by the presence of species such as Chiropodomys gliroides, which indicate a forested environment at the time of the early Pleistocene Satir fauna, whereas wide-toothed species of mouse (Mus sp.) suggest the presence of grasslands (van der Meulen & Musser, 1999). The different vegetation zones would have shifted with the increasingly severe glacial cycles that characterize much of the Pleistocene and which would have contracted and expanded forests and mixed areas inversely with more open ones (Meijaard, 2004). Trinil’s (c. 900 ka; van den Bergh et al., 2001) classiﬁcation in our analysis largely agrees with previous palaeoenvironmental reconstructions. The almost intermediate position of Trinil between open and mixed habitats in the DFA suggests primarily grassland conditions, but does not speciﬁcally exclude the presence of some trees. On the basis of the large numbers of bovids present at the site, the predominant vegetation was previously classiﬁed as open woodland (de Vos et al., 1994; van den Bergh et al., 2001; Storm, 2001). Finds of Leptoptilos cf. dubius (an adjutant stork) and Ephippiorhynchus cf. asiaticus (black-necked stork of mainland Asia – but not Sundaland – and Australia) suggest the presence of grassy areas, mud banks, mangroves, swamps, or open and lightly wooded areas, which also seems to ﬁt the picture for Pavo muticus (green peafowl) habitat (MacKinnon & Phillipps, 1993; Simpson & Day, 1996). The open environment reconstructed in this analysis is conﬁrmed by the presence of the wide-toothed species of rat (Rattus sp. A), from Trinil (van der Meulen & Musser, 1999), which was presumably a grassland species. The Trinil conditions changed during the deposition of the Kedung Brubus sediments, when a wave of Asiatic mammals entered the region, probably during a glacial maximum (Musser, 1982). The Kedung Brubus fauna, which has been dated at 0.8–0.7 or 1.3–0.8 Ma (see van den Bergh et al., 2001; Larick et al., 2001), brought in several new medium- to largesized mammals [Rhinoceros unicornis (Indian rhinoceros), Tapirus indicus (Malayan tapir), Manis palaeojavanica (giant tapir), Pachycrocuta brevirostris (giant hyena), Lutrogale palaeoleptonix (otter)] (Aimi & Aziz, 1985; Bandet et al., 1989; van den Bergh et al., 2001). Although allocated to a closed habitat, we cannot discount, owing to the low number of taxa preserved, that Kedung Brubus may represent a mixed environment such as open woodland. This latter interpretation would be more congruent with previous qualitative analyses that describe the Middle Pleistocene environments of Java as 10 very diverse with tall evergreen forests, shrubby or woodland vegetation, and extensive grasslands, reminiscent of a mosaic of forest, scrub and grassland, such as seen today in the wetter, hillier parts of East Africa (e.g. western Uganda) or India (Medway, 1972; de Vos et al., 1994; van den Bergh et al., 2001), as well as with the open/mixed forest reconstructions of the Middle Pleistocene deposists of the Sangiran dome (Bouteaux, 2005). Climatic conditions at the beginning of the Late Pleistocene are considered to have been drier than today. Around 135 ka, annual precipitation on Java, inferred from palaeosol development, is estimated to have been 750–1000 mm (van der Kaars & Dam, 1995, 1997), whereas today annual rainfall amounts to between 1500 and 2000 mm (Whitten et al., 1996). However, between 126 and 81 ka, very warm and humid conditions prevailed, with up to 2000 mm of annual precipitation. The site of Punung, dated at between 128 ± 15 and 118 ± 3 ka (Westaway et al., 2007), preserves for the ﬁrst time characteristic rain forest taxa [e.g. Symphalangus (siamang), Pongo (orangutan), Catopuma (golden cat), Helarctos (Malayan sun bear) and Tapirus (Malay tapir)] (Storm et al., 2005). The palaeoenvironmental reconstruction for this fauna by the DFA is unsurprisingly that of a closed forest. At around 81 ka the vegetation in the Bandung plain changed from closed freshwater swamp forest to open herbaceous swamp, dominated by grasses and sedges, suggesting a reduction in annual precipitation to c. 1000 mm (Smit-Sibinga, 1947; van der Kaars & Dam, 1995, 1997; van der Kaars, 1998 cited in Sun et al., 2000). However, between 74 and 47 ka, slightly warmer and wetter conditions existed. Sumatra Late Pleistocene environments of Sumatra were warm and wet, similar to those at present, if perhaps a little drier (Maloney & McCormac, 1996). The sites of Lida Ajer and Simbrambang, Sumatra, from the Late Pleistocene, share the same classiﬁcation as Punung, which is not surprising because they share a high level of taxonomic similarity (de Vos, 1983). Furthermore, the faunas from these sites are essentially modern – neither Lida Ajer nor Simbrambang records any extinctions (Louys, 2007) – and it is expected that these faunas would correspond to Sumatra’s present-day predominant vegetation types, namely closed forests. Borneo Early Miocene coals in the Maliau Basin in Sabah, Borneo, suggest warm everwet climates (Tjia et al., 1990). The conifers Phyllocladus and Podocarpus ﬁrst appear in the Late Pliocene of Borneo, with the former arriving at the Pliocene–Pleistocene transition, indicating either cooler climates and/or substantially higher mountains in Borneo than at present (Muller, 1971). As elsewhere in the region, the increasingly severe glacials of the Pliocene–Pleistocene resulted in occasionally more open vegetation types in an evergreen matrix. Journal of Biogeography ª 2010 Blackwell Publishing Ltd Palaeoecology of Southeast Asian megafauna-bearing sites These patterns are especially clear in Late Pleistocene Borneo, for which several sources suggest considerably colder temperatures and reduced rainfall (Thomas, 1987; Thorp et al., 1990). These conditions, for example those in the penultimate glacial period, are characterized by the expansion of montane vegetation, the contraction of lowland cover and a reduction of mangrove vegetation (Jirin, 1993). In east and south Borneo, dated charcoal ﬁnds suggest that the coastal area was more continental and drier than at present; forest formations were more seasonal and probably had a temporary ﬁre-climax character (Goldammer & Seibert, 1989). Grasslands or wooded savannas also appear to have been present (Caratini & Tissot, 1988). This would be congruent with the faunal evidence from Niah Cave (39–45 ka, Barker et al., 2007). In our analysis, Niah Cave is classiﬁed as a mixed habitat, although not enough species are preserved at this site to enable it to be conﬁdently distinguished from a closed habitat. The fauna from Niah Caves occupies a unique position in the DFA, and is strongly indicative of a mixed habitat, most probably including woodlands and grasslands. This classiﬁcation agrees with previous qualitative reconstructions on the basis of fauna (e.g. Medway, 1977; Harrison, 1996) and ﬂora (e.g. Barker et al., 2007). By contrast, Majid (1982) suggested that, at the height of the last glacial maximum, the Niah area in Sarawak was covered by deciduous monsoon forest, a drier forest type but not as open as suggested by other authors. The data from the literature are, however, equivocal about the predominant vegetation characteristics of Late Pleistocene Borneo. Visser et al. (2004) found evidence based on lignin phenol ratios in coastal sediments that indicated that Late Pleistocene vegetation in eastern Borneo did not signiﬁcantly change from tropical rain forest during the last two glacial periods, relative to subsequent interglacial periods. Furthermore, pollen records suggest that during the Last Glacial Maximum (LGM) the upper Kapuas area in west Borneo was probably covered in rain forest, very similar to that found during the mid-Holocene (Anshari et al., 2001; Kershaw et al., 2001). Moreover, pollen data from peat swamps in the Sebangau area in south Borneo (Page et al., 1999) indicate that the climate remained relatively warm and wet during the LGM. These mixed signals from various data sets may suggest that there was signiﬁcant spatial and temporal variation in the palaeonvironments of Borneo during the Late Pleistocene. This would not be unlike the present-day situation, in which rainfall patterns and concomitant vegetation types vary widely, longitudinally as well as elevationally. and to a lesser extent by gymnosperms (Hing & Leong, 1990). Based on palaeobotanical data, Morley (1999) suggested that Pinus savanna was probably widespread on the Malay Peninsula at 660, 480, 200 and 22 ka. All of these, apart from 200 ka, are periods of low sea-level and presumably drier, slightly cooler conditions (see Meijaard, 2004 for an overview). During the interstadials, the climate in the lowlands of the Malay Peninsula was probably similar to that prevailing today, as suggested by pollen found at interglacial deposits dated at c. 80 and 55 ka (Kamaludin & Azmi, 1997). Elsewhere, Ayob (1970) provided carbon-dated peat and wood samples from deposits containing pollen indicating perhumid vegetation; these samples were dated at c. 36.4 and 41.2 ka, indicating that before the LGM an evergreen vegetation type existed in this area close to present-day Kuala Lumpur. Price et al. (1997) suggested that bauxite formation on the Malay Peninsula took place continuously from 115 ka to the present, which would suggest mostly warm and wet conditions. It therefore appears that vegetation zones along the peninsula shifted from open woodlands to closed deciduous and evergreen rain forest with the changing climatic patterns of the glacial–interglacial ﬂuctuations. The fauna from the Kinta Valley represents one of the few discoveries of Pleistocene fossils on the Malay Peninsula. It is classiﬁed by the DFA as representing a closed habitat, although a mixed habitat is also likely. However, the dating of this site is not well constrained: although assigned to the Middle Pleistocene based on the presence of Palaeoloxodon (e.g. Hooijer, 1963; Medway, 1972), on the basis of geological evidence the site could well be Late Pleistocene in age (Kamaludin et al., 1993; Thorp & Thomas, 1993; Kamaludin & Azmi, 1997; Teeuw et al., 1999). Muhammad et al. (2002) report an age of 228 ka for a Kinta Valley vertebrate fossil, but it is unclear how that particular fossil relates to the fauna described by Hooijer (1963). It is unfortunate that neither the age nor the habitat type of this site can be constrained on the basis of current data, as the Kinta Valley site lies on the proposed savanna corridor of Heaney (1991). Mainland SE Asia Mammalian fossil assemblages in Yunnan Province, southern China, indicate that between the Early and Middle Pliocene a mixed forest and open bush–grassland vegetation existed, with several carnivores typical of forested areas, and ungulate species suggesting more open, grassy vegetation (Pan, 1993). Slightly wetter conditions seem to have prevailed further east and south-east of Yunnan. Reconstructions of the climate and vegetation of this area in the Middle Pliocene (3 Ma) indicate that there was a considerable expansion of evergreen forest in southern China, whereas mainland SE Asia was covered mostly by rain forest, possibly with patches of deciduous forest in the area of present-day Myanmar (Dowsett et al., 1994). The Middle–Late Pleistocene in the eastern parts of SE Asia saw some considerable changes in vegetation distribution. For instance, Zheng & Lei’s (1999) palaeoenvironmental data for the 11 Malay Peninsula The coal composition of Pliocene–Early Pleistocene deposits in the Kepong and Kluang-Niyor Basins (Stauffer, 1973) and the Merapoh Basin (Hing & Leong, 1990) indicates that these deposits were derived mainly from woody and herbaceous plants as well as from spore-bearing plants, as would have been found in a tropical forest dominated by angiosperms Journal of Biogeography ª 2010 Blackwell Publishing Ltd J. Louys and E. Meijaard Leizhou Peninsula, north of Hainan Island, suggest signiﬁcant changes over the last 400 kyr. Starting with slightly drier and cooler conditions than today around 400 ka, there was a change to a wetter and warmer climate with predominant evergreen oak forest between 340 and 280 ka. When mean annual temperatures were more than 4°C lower than today, between 280 and 240 ka, a substantial increase in typical montane conifer forest elements occurred. The Middle to Late Pleistocene saw ﬂuctuations in vegetation with the waxing and waning of glacials and interglacials, with open vegetation peaking around the LGM, when mean temperatures dropped between 5 and 6°C and dense forests were transformed to grassland or savanna (Zheng & Lei, 1999). This scenario ﬁts with data from Ferguson (1993), who reported an extremely cold and arid phase at the end of the Pleistocene (20–15 ka) for China. In Jiangxi Province, central–south-east China, the Late Pleistocene climate (between 12.8 and 10.5 ka) was also drier and cooler than today’s, and the subtropical, mixed deciduous–evergreen broad-leaved forest, which is now in the area, was reduced and herbaceous cover expanded (Jiang & Piperno, 1999). Overall, mainland SE Asia appears to have been cooler and more seasonal than today during much of the Pleistocene. The dominant vegetation type for mainland SE Asia during this period appears to have been evergreen, semi-evergreen and coniferous forests, with varying amounts of grasses and other herbaceous vegetation. The results of our mammal fauna analysis are in general agreement with these overall patterns, although mostly indeﬁnite about exactly which vegetation types the faunas correspond to. be more consistent with the interpretations discussed above. Furthermore, the climatic ﬂuctuations of the Pleistocene would have resulted in expanding and contracting forests within a more open vegetation matrix. As the age of the Pleistocene faunas remains unresolved, it might be that they relate to the warmer and more humid phases of the Pleistocene, which would have coincided with more forested environments. Thailand Thick laterites of Pliocene–Early Pleistocene age in the Lower Central Plain of Thailand (Thiramongkol, 1986) indicate seasonality in a generally humid tropical climate (Whittow, 1984). Based on rodent fossil distributions, Chaimanee (1998) reconstructed Late Pliocene–Early Pleistocene palaeoenvironments for several locations in Thailand. The site of Khao Samngam (99°42¢ E, 13°27¢ N) probably had an environment with some mixed vegetation, with grasslands in the ﬂoodplain and forests on the surrounding limestone hills. Of the SE Asian Middle Pleistocene sites, only Kao Pah Nam (c. 690 ka, Pope et al., 1981) and Thum Phra Khai Phet (c. 169 ka, see Table 5) are classiﬁed as open habitats. It should be noted, however, that both these sites are represented by relatively small numbers of species (although they are not the only sites to exhibit such low species numbers). The palaeoenvironment of Kao Pah Nam was interpreted by Pope et al. (1981, p. 161) as ‘relatively open, dry dipterocarp woodland’. However, this description would be more congruent with our mixed habitat description, as opposed to the open classiﬁcation obtained. Thum Phra Khai Phet has been considered, on the basis of taxonomic similarity, to be contemporaneous with Thum Wimam Nakin (Tougard, 1998, 2001). In our analysis, this latter site (Thum Wimam Nakin) is classiﬁed as representing a mixed habitat, which agrees with previous palaeoenvironmental reconstructions suggesting that it represents ‘a slightly open forested environment’ (Tougard & Montuire, 2006). Thus, the classiﬁcation of Thum Phra Khai Phet as open is also likely to be a result of the inadequate number of large-bodied taxa preserved at the site (Table 2), which prohibits differentiation between open and mixed habitats. Ban Fa Suai, the only Middle Pleistocene site from Thailand to preserve more than 32 species, is conﬁdently assigned as mixed. Myanmar Palaeobotanical data from India and from Myanmar indicate that a rich tropical to subtropical vegetation covered the region c. 5 Ma under a prevailing warm humid climate (several references in Poole & Davies, 2001). During the Pliocene, however, an increasingly arid climate, engendered by the rising Himalayas, caused a change in the vegetation of this region (Poole & Davies, 2001). The vegetation changes are also reﬂected in faunal turnovers in southern Asia: browsers are replaced by grazers, and rodents show considerable turnover (Barry et al., 1985; Quade et al., 1989). These palaeoenvironmental changes are also reﬂected in northern Thailand, where coals in Middle Miocene sediments indicate a peat swamp, ﬂuvial or lacustrine environment, whereas no coal has been found in Late Miocene, Pliocene and Quaternary sediments (Morley et al., 2000), probably indicating a climate that was either too cool or too dry for coal formation. Evidence for grasslands is very limited through most of the Pliocene, but subsequently shows a marked increase during the Early Pleistocene, indicating the expansion of savanna vegetation, which, together with evidence of charred grass cuticle, suggests it was subject to burning (Morley, 1999). Our classiﬁcations of Mogok Caves and the Irrawaddy faunas as representing closed habitats are probably a result of the limited number of taxa from these sites, and the classiﬁcation of mixed or open would 12 Vietnam and Cambodia Vietnam and Cambodia have received little previous palaeoecological attention. The three Middle Pleistocene Vietnamese sites examined in this analysis [Tham Khuyen (475 ka, Ciochon et al., 1996), Tham Hai (475 ka, see Table 5), Tham Om (140–250 ka, Olsen & Ciochon, 1990)] are each classiﬁed in our analysis as mixed, but they do not preserve sufﬁcient species to enable us to distinguish conﬁdently between mixed and closed habitats. Of these three sites, Tham Om is the most southern, and with 31 species is actually quite likely to be correctly classiﬁed. With respect to the DFA, it is closest to Journal of Biogeography ª 2010 Blackwell Publishing Ltd Palaeoecology of Southeast Asian megafauna-bearing sites that of the mixed habitat group centroid. In terms of age, at 140–250 ka (Olsen & Ciochon, 1990), it is closest to that of Tham Wimam Nakin, Thailand, which is also considered a mixed habitat. The site of Phnom Loang, Cambodia, is also classiﬁed as mixed habitat, but again too few species are preserved to rule out the possibility that it may represent a closed habitat. The discriminant functions calculated for this site are almost identical to those of Tham Hai, Vietnam, suggesting that these two sites sample the same habitat. Middle Pleistocene megafaunal communities across Indochina are therefore reconstructed by the DFA as most probably representing mixed habitats, that is, heterogeneous habitats consisting of patches of grassland and forest, most probably of the dry evergreen type. These include sites in Cambodia (Phnom Loang), Thailand (Thum Wimam Nakin, Ban Fa Suai) and Vietnam (Tham Om, Tham Khuyen, Tham Hai). The presentday vegetation in this area, prior to human disturbance, is dense evergreen forest. Bacon et al. (2008b) suggested a humid forested area with some open areas for the Late Pleistocene site of Duoi U’Oi (c. 66 ka), whereas the site of Ma U’Oi (c. 49 ka) in northern Vietnam showed drier and more open conditions (Bacon et al., 2006). Between 62 and 28 ka and between 28 and 19 ka, the reworking and redeposition of aeolian sands along the south-eastern Vietnam coast points to reduced vegetation cover and to landscape instability in this area (Murray-Wallace et al., 2002). In our analysis, Late Pleistocene sites from Vietnam represent a variety of habitat types. Interestingly, the only site to preserve sufﬁcient numbers of species to enable a habitat type to be conﬁdently assigned is Hang Hum II (80–140 ka, Olsen & Ciochon, 1990), which is classiﬁed as open. An examination of the discriminant functions shows that Hang Hum II is the furthest ‘open’ site from the open habitat group centroid. On the basis of its discriminant functions, it is quite close to Tam Hang (south), a site conﬁdently assigned as mixed. The faunas preserved in these two sites therefore most likely represent an ecotone between mixed and open habitats. Other sites from Vietnam are classiﬁed as either mixed (Hang Hum I, Lang Trang) or closed [Keo Leng (20–30 ka, Olsen & Ciochon, 1990)]; however, the numbers of species preserved at these sites mean that they can be excluded conﬁdently only from being open. small number of species preserved at this site makes this classiﬁcation uncertain, and it may also in fact represent a mixed habitat. CONCLUSIONS Our results indicate that, throughout much of the Pleistocene, SE Asia was characterized by a heterogenous vegetation complex, with habitats comprising a mix of both woodlands and grasslands. The Sundaic subregion was dominated by mixed habitats during most of the Early and Middle Pleistocene, certainly for Java; and evidence for closed habitats during this period is seen only in the Malay Peninsula. Both the faunal evidence and other palaeoenvironmental studies demonstrate changes of vegetation structure with glacial and interglacial cycles during the Late Pleistocene, with glacial periods associated with drier and more mixed habitats, and interglacials with more humid, closed habitats. Evidence for environmental changes in the Indochinese subregion is more limited, but consistently indicates the dominance of mixed habitats through most of the Pleistocene; support for extensive closed habitats in this subregion is found only for the Pliocene and the end of the Late Pleistocene. Our study of habitat types of 25 Pleistocene sites in SE Asia through synecological methods is in general agreement with the overall palaeoenvironmental information obtained from other sources and through different methods. This indicates that our approach can provide useful additional information on palaeoenvironments where only mammal fossils are available. Even when insufﬁcient numbers of fauna are present at a site, the position of that site with regard to the centroids of closed, mixed and open vegetation provides an indication of the overall vegetation type present. Although our analysis does not directly predict certain environmental factors (for example relative precipitation), when our results are placed within a regional palaeoenvironmental framework this method can provide signiﬁcant new insights into palaeoenvironmental changes over time. Not only is such information of considerable use in the study of the evolution of faunas in relation to environmental changes, but with the presently rapid changes in global climate, an understanding of palaeoenvironments in relation to palaeoclimates can help us to predict the impacts that such changes might have on the distribution of present-day species. ACKNOWLEDGEMENTS J.L. acknowledges the support of the Leverhulme Trust (FC00754C). We thank L.R. Heaney and an anonymous referee for their constructive comments. REFERENCES Aimi, M. & Aziz, F. (1985) Vertebrate fossils from the Sangiran Dome, Mojokerto, Trinil and Sambungmacan areas. Quaternary geology of the hominid fossil bearing formations in Java. Report of the Indonesia–Japan Joint Research Project 13 Laos The palaeoenvironmental reconstructions for Laos represent some of the ﬁrst for this country. Tam Hang (south) is one of two sites conﬁdently assigned, on the basis of both DFA and species numbers, as a mixed habitat. However, its position relative to the DFA group centroids also reveals that in character it approaches some open sites. Therefore, it is herein interpreted as a mixed habitat with signiﬁcant open areas. This reconstruction is at odds with the palaeoenvironmental interpretations of Bacon et al. (2008a), who suggest that this site is more indicative of closed forests. Interestingly, Tam Hang (north) is classiﬁed as a closed habitat; however, the Journal of Biogeography ª 2010 Blackwell Publishing Ltd J. Louys and E. Meijaard CTA-41, 1976-1979 (ed. by N. Watanabe and D. Kadar), pp. 155–198. Geological Research and Development Centre, Bandung, Indonesia. Andrews, P. & Humphrey, L. (1999) African Miocene environments and the transition to early hominines. African biogeography, climate change, and human evolution (ed. by T.G. Bromage and F. Schrenk), pp. 282–300. Oxford University Press, Oxford. Andrews, P., Lord, X. & Evans, E.M.N. (1979) Patterns of ecological diversity in fossil and modern mammalian faunas. Biological Journal of the Linnean Society, 11, 177–205. Anshari, G., Kershaw, A.P. & van der Kaars, S. (2001) A Late Pleistocene and Holocene pollen and charcoal record from peat swamp forest, Lake Sentarum Wildlife Reserve, West Kalimantan, Indonesia. Palaeogeography, Palaeoclimatology, Palaeoecology, 171, 213–228. Arambourg, C. & Fromaget, J. (1938) Le gisement quaternaire ˆ de Tam Nang (Chaıne Annamitique septentrionale). ´ Sa stratigraphie et ses faunes. Comptes Rendus de l’Academie des Sciences, 207, 793–795. Ayob, M. (1970) Quaternary sediments at Sungei Besi, West Malaysia. Bulletin of the Geological Society of Malaysia, 3, 53–61. Bacon, A.-M., Demeter, F., Schuster, M., Long, V.T., Thuy, N.K., Antoine, P.-O., Sen, S., Nga, H.H. & Huong, N.M. (2004) The Pleistocene Ma U’Oi cave, northern Vietnam: palaeontology, sedimentology and palaeoenvironments. Geobios, 37, 305–314. Bacon, A.-M., Demeter, F., Rousse, S., Long, V.T., Duringer, P., Antoine, P.-O., Thuy, N.K., Mai, B.T., Huong, N.T.M., Dodo, Y., Matsumura, H., Schuster, M. & Anezaki, T. (2006) New palaeontological assemblage, sedimentological and chronological data from the Pleistocene Ma U’Oi cave (northern Vietnam). Palaeogeography, Palaeoclimatology, Palaeoecology, 230, 280–298. Bacon, A.-M., Demeter, F., Tougard, C., de Vos, J., Sayavongkhamdy, T., Antoine, P.-O., Bouasisengpaseuth, B. & ´ Sichanthongtip, P. (2008a) Redecouverte d’une faune ´ ` pleistocene dans les remplissages karstiques de Tam Hang au ´ Laos: Premiers resultats. Comptes Rendus Palevol, 7, 277–288. Bacon, A.-M., Demeter, F., Duringer, P., Helm, C., Bano, M., Long, V.T., Thuy, N.T.K., Antoine, P.-O., Mai, B.T., Huong, N.T.M., Dodo, Y., Chabaux, F. & Rihs, S. (2008b) The Late Pleistocene Duoi U’Oi in northern Vietnam: palaeontology, sedimentology, taphonomy and palaeoenvironments. Quaternary Science Reviews, 27, 1627–1654. Bandet, Y., Semah, F., Sartono, S. & Djubiantono, T. (1989) ` ´ Premier peuplement par les mammiferes d’une region de ` ` ˆ Java Est, a la ﬁn de Pliocene: age de la faune du Gunung ` ´ Butak, pres de Kedungbrubus (Indonesie). Comptes Rendus ´ ´ ´ de l’Academie des Sciences Paris, serie II. Mecanique, Physique, Chimie, Sciences de l’univers, Sciences de la Terre, 308, 867–870. Barker, G., Barton, H., Bird, M. et al. (2007) The ‘human revolution’ in lowland tropical Southeast Asia: the antiquity 14 and behaviour of anatomically modern humans at Niah Cave (Sarawak, Borneo). Journal of Human Evolution, 52, 243–261. Barry, J.C., Johnson, N.M., Raza, S.M. & Jacobs, L.L. (1985) Neogene mammalian faunal change in southern Asia: correlations with climatic, tectonic, and eustatic events. Geology, 13, 637–640. ´ ´ ´ Beden, M. & Guerin, C. (1973) Le gisement de vertebres du Phnom Loang (province du Kampot, Cambodge). Faune ´ ` pleistocene moyen terminal. Travaux et Documents de l’O.R.S.T.O.M, 27, 1–97. van den Bergh, G.D., de Vos, J., Sondaar, P.Y. & Aziz, F. (1996) Pleistocene zoogeographic evolution of Java (Indonesia) and glacio-eustatic sea level ﬂuctuations: a background for the presence of Homo. Bulletin of the Indo-Paciﬁc Prehistory Association, 14, 7–21. van den Bergh, G.D., de Vos, J. & Sondaar, P. (2001) The Late Quaternary palaeogeography of mammal evolution in the Indonesian Archipelago. Palaeogeography, Palaeoclimatology, Palaeoecology, 171, 385–408. Bettis, E.A., III, Milius, A.K., Carpenter, S.J., Larick, R., Zaim, Y., Rizal, Y., Ciochon, R.L., Tassier-Surine, S., Murray, D., Suminto & Bronto, S. (2009) Way out of Africa: Early Pleistocene paleoenvironments inhabited by Homo erectus in Sangiran, Java. Journal of Human Evolution, 56, 11–24. Bird, M.I., Taylor, D. & Hunt, C. (2005) Palaeoenvironments of insular Southeast Asia during the Last Glacial Period: a savanna corridor in Sundaland? Quaternary Science Reviews, 24, 2228–2242. ´ ´ ´ Bouteaux, A. (2005) Paleontologie, paleoecologie et taphonomie ` ´ ` ` des mammiferes du Pleistocene moyen ancien du site a ´ ´ hominides de Sangiran (Java central, Indonesie). PhD Thesis, ´ Museum National d’Histoire Naturelle de Paris, Paris. Caratini, C. & Tissot, C. (1988) Paleogeographical evolution of the Mahakam delta in Kalimantan, Indonesia during the Quaternary and late Pliocene. Review of Palaeobotany and Palynology, 55, 217–228. Chaimanee, Y. (1998) Plio-Pleistocene rodents of Thailand. Thai Studies in Biodiversity, 3, 1–303. ´ ¨ Ciochon, R.L., Long, V.T., Larick, R., Gonzalez, L., Grun, R., de Vos, J., Yonge, C., Taylor, L., Yoshida, H. & Reagan, M. (1996) Dated co-occurrence of Homo erectus and Gigantopithecus from Tham Khuyen Cave, Vietnam. Proceedings of the National Academy of Sciences USA, 93, 3016–3020. Colbert, E.H. (1943) Pleistocene vertebrates collected in Burma by the American Southeast Asiatic expedition. Transactions of the American Philosophical Society, 32, 395–429. Dowsett, H., Thompson, R., Barron, J., Cronin, T., Fleming, F., Ishman, S., Poore, R., Willard, D. & Holtz, T.J. (1994) Joint investigations of the Middle Pliocene climate I: PRISM palaeoenvironmental reconstructions. Global and Planetary Change, 9, 169–195. Dubois, E. (1908) Das geologische Alter der Kendeng- oder Trinil-fauna. Tijdschrift van het Koninklijk Nederlandsch Aardrijkskundig Genootschap, 2, 1235–1270. Journal of Biogeography ª 2010 Blackwell Publishing Ltd Palaeoecology of Southeast Asian megafauna-bearing sites Esposito, M., Reyss, J.-L., Chaimanee, Y. & Jaeger, J.-J. (2002) U-series dating of fossil teeth and carbonates from Snake Cave, Thailand. Journal of Archaeological Science, 29, 341–349. Ferguson, D.K. (1993) The impact of Late Cenozoic environmental changes in East Asia on the distribution of terrestrial plants and animals. Evolving Landscapes and Evolving Biotas of East Asia since the Mid-Tertiary: proceedings of the Third Conference on the Evolution of the East Asian Environment (ed. by N.G. Jablonski and S. Chak-Lam), pp. 145–196. Centre of Asian Studies, Hong Kong. ´ Fromaget, J. (1936) Sur la stratigraphie des formations recentes de la Chaine annamitique septentrionale et sur l’existence de ´ l’Homme dans le Quaternaire inferieur de cette partie de ´ l’Indochine. Comptes Rendus de l’Academie des Sciences, 203, 738–741. Gagnon, M. (1997) Ecological diversity and community ecology in the Fayum sequence (Egypt). Journal of Human Evolution, 32, 133–160. Goldammer, J.G. & Seibert, B. (1989) Natural rain forest ﬁres in Eastern Borneo during the Pleistocene and Holocene. Naturwissenschaften, 76, 518–520. Hammer, Ø., Harper, D.A.T. & Ryan, P.D. (2001) PAST: Palaeontological Statistics software package for education and data analysis. Palaeontologia Electronica, 4, 9, http:// palaeo-electronica.org/2001_1/past/issue1_01.htm. Harrison, T. (1996) The palaeoecological context at Niah Cave, Sarawak: evidence from the primate fauna. Bulletin of the Indo-Paciﬁc Prehistory Association, 14, 90–100. Heaney, L.R. (1984) Mammalian species richness on islands on the Sunda Shelf, Southeast Asia. Oecologia, 61, 11–17. Heaney, L.R. (1986) Biogeography of mammals in SE Asia: estimates of rates of colonization, extinction and speciation. Biological Journal of the Linnean Society, 28, 127–165. Heaney, L.R. (1991) A synopsis of climatic and vegetational change in Southeast Asia. Climatic Change, 19, 53–61. Hing, T.T. & Leong, L.K. (1990) Environmental placer gold deposits in northern Pahang. Bulletin of the Geological Society of Malaysia, 26, 1–11. Holloway, J.D. & Hall, R. (1998) SE Asian geology and biogeography: an introduction. Biogeography and geological evolution of SE Asia (ed. by R. Hall and J.D. Holloway), pp. 1–23. Backhuys Publishers, Leiden. Hooijer, D.A. (1963) Report upon a collection of Pleistocene mammals from tin-bearing deposits in a limestone cave near Ipoh, Kinta Valley, Perak. Federated Museum Journal, 7, 1–5. Hooijer, D.A. (1975) Quaternary mammals west and east of Wallace’s line. Netherlands Journal of Zoology, 25, 46–56. Jiang, Q. & Piperno, D.R. (1999) Environmental and archaeological implications of a Late Quaternary palynological sequence, Poyang Lake, southern China. Quaternary Research, 52, 250–258. Jirin, S. (1993) Palynology of Late Quaternary sediments of piston-core KL-139 from Sabah Trough, East Malaysia (abstract). Warta Geologi, 19, 117. Journal of Biogeography ª 2010 Blackwell Publishing Ltd van der Kaars, W.A. & Dam, M.A.C. (1995) A 135,000-year record of vegetational and climatic change from the Bandung area, West-Java, Indonesia. Palaeogeography, Palaeoclimatology, Palaeoecology, 117, 55–72. van der Kaars, W.A. & Dam, M.A.C. (1997) Vegetation and climate change in West-Java, Indonesia during the last 135,000 years. Quaternary International, 37, 67–71. Kamaludin, H. & Azmi, H.Y. (1997) Interstadial records of the last glacial period at Pantai Remis, Malaysia. Journal of Quaternary Science, 12, 419–434. Kamaludin, H., Nakamura, T., Price, D.M., Woodroffe, C.D. & Fuji, S. (1993) Radiocarbon and thermoluminescence dating of Old Alluvium from a coastal site in Perak, Malaysia. Sedimentary Geology, 83, 199–210. Kershaw, A.P., Penny, D., van der Kaars, S., Anshari, G. & Thamotherampillai, A. (2001) Vegetation and climate in lowland southeast Asia at the Last Glacial Maximum. Faunal and ﬂoral migration and evolution in SE Asia–Australasia (ed. by I. Metcalfe, J.M.B. Smith, M. Morwood and I. Davidson), pp. 227–236. A.A. Balkema Publishers, Lisse, The Netherlands. Larick, R., Ciochon, R.L., Zaim, Y., Sudijono, Suminto, Riza, Y., Aziz, F., Reagan, M. & Heizler, M. (2001) Early Pleistocene 40Ar/39Ar ages for Bapang Formation hominins, Central Java, Indonesia. Proceedings of the National Academy of Sciences USA, 98, 4866–4871. Legendre, S. (1986) Analysis of mammalian communtities from the late Eocene and Oligocene of southern France. Palaeovertebrata, 16, 191–212. ´ ` Legendre, S. (1989) Les communautes de mammiferes du ´ ` ` ` Paleogene (Eocene et Oligocene) d’Europe occidentale: ´ ¨ structures, milieu et evolution. Munchner Geowissenschaftliche Abhandlungen, 16, 1–110. Lekagul, B. & McNeely, J.A. (1988) Mammals of Thailand. Association for the Conservation of Wildlife, Bangkok, Thailand. Louys, J. (2007) Limited effect of the Quaternary’s largest super-eruption (Toba) on land mammals from Southeast Asia. Quaternary Science Reviews, 26, 3108–3117. Louys, J. (2008) Quaternary extinctions in Southeast Asia. Mass extinction (ed. by A.M.T. Elewa), pp. 159–189. Springer-Verlag, Heidelberg. Louys, J., Curnoe, D. & Tong, H. (2007) Characteristics of Pleistocene megafauna extinctions in Southeast Asia. Palaeogeography, Palaeoclimatology, Palaeoecology, 243, 152–173. Louys, J., Travouillon, K.J., Bassarova, M. & Tong, H. (2009) The use of natural protected areas in palaeoecological analyses: assumptions, limitations and application. Journal of Archaeological Science, 36, 2274–2288. MacKinnon, J. & Phillipps, K. (1993) A ﬁeld guide to the birds of Borneo, Sumatra, Java, and Bali. Oxford University Press, Oxford. Majid, Z. (1982) The West Mouth, Niah in the prehistory of Southeast Asia. Special Monograph No. 3. The Sarawak Museum Journal, 31, 1–200. 15 J. Louys and E. Meijaard Maloney, B.K. & McCormac, F.G. (1996) Palaeoenvironments of North Sumatra: a 30,000 year old pollen record from Pea Bullok. Bulletin of the Indo-Paciﬁc Prehistory Association, 14, 73–82. Medway, Lord (1960) The Malay Tapir in late Quaternary Borneo. Sarawak Museum Journal, 9, 356–360. Medway, Lord (1972) The Quaternary mammals of Malesia: a review. Transactions of the Second Aberdeen–Hull Symposium on Malesian Ecology, Hull (ed. by P. Ashton and M. Ashton), pp. 63–98. University of Hull Department of Geography Miscellaneous Series 13, Hull, UK. Medway, Lord (1977) The Niah excavations and an assessment of the impacts of early man on mammals in Borneo. Asian Perspectives, 20, 51–69. Meijaard, E. (2003) Mammals of South-East Asian islands and their Late Pleistocene environments. Journal of Biogeography, 30, 1245–1257. Meijaard, E. (2004) Solving mammalian riddles. A reconstruction of the Tertiary and Quaternary distribution of mammals and their palaeoenvironments in island South-East Asia. PhD Thesis, School of Archaeology and Anthropology, The Australian National University, Canberra. Meiri, S., Meijaard, E., Wich, S.A., Groves, C.P. & Helgen, K.M. (2008) Mammals of Borneo – small size on a large island. Journal of Biogeography, 35, 1087–1094. Mendoza, M., Janis, C.M. & Palmqvist, P. (2005) Ecological patterns in the trophic-size structure of large mammal communities: a ‘taxon-free’ characterization. Evolutionary Ecology Research, 7, 505–530. van der Meulen, A.J. & Musser, G.G. (1999) New paleontological data from the continental Plio-Pleistocene of Java. Elephants have a snorkel! (ed. by J.W.F. Reumer and J. de Vos), pp. 361–368. Deinsea (Jaarbericht van het Natuurmuseum Rotterdam), Vol. 7, Rotterdam. Morley, R.J. (1991) Tertiary stratigraphic palynology in Southeast Asia: current status and new directions. Bulletin of the Geological Society of Malaysia, 28, 1–36. Morley, R.J. (1999) Palynological evidence for Tertiary plant dispersals in the SE Asian region in relation to plate tectonics and climate. Biogeography and geological evolution of SE Asia (ed. by R. Hall and J.D. Holloway), pp. 211–234. Backhuys Publishers, Leiden. Morley, R.J. (2000) Origin and evolution of tropical rain forests. John Wiley & Sons, Ltd, Chichester, UK. Morley, R.J., Lelono, E.B., Nugrahaningsih, L. & Hasjim, N. (2000) LEMIGAS Tertiary Palynology Project: aims, progress and preliminary results from the Middle Eocene to Late Pliocene of Sumatra and Java. Geological Research and Development Centre, Paleontology Series, 10, 27–47. Muhammad, R.F., Yoshida, D., Tani, A. & Smart, P.L. (2002) Implications of electron spin resonance and uranium-series dating techniques on speleotherms in the Kinta and Lenggong Valleys, West Malaysia. Advances in ESR Applications, 18, 19–26. Muller, J. (1971) Palynological evidence for change in geomorphology, climate, and vegetation in the Mio-Pliocene of 16 Malesia. Transactions of the Second Aberdeen–Hull Symposium on Malesian Ecology, Hull (ed. by P. Ashton and M. Ashton), pp. 6–32. University of Hull Department of Geography Miscellaneous Series 13, Hull, UK. Murray-Wallace, C.V., Jones, B.G., Nghi, T., Price, D.M., Vinh, V.V., Tinh, T.N. & Nanson, G.C. (2002) Thermoluminescence ages for a reworked coastal barrier, southeastern Vietnam: a preliminary report. Journal of Asian Earth Sciences, 20, 535–548. Musser, G.G. (1982) The Trinil rats. Modern Quaternary Research SE Asia, 7, 65–85. Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B. & Kent, J. (2000) Biodiversity hotspots for conservation priorities. Nature, 403, 853–858. Nowak, R.M. (1999) Walker’s mammals of the world. The Johns Hopkins University Press, London. Olsen, J.W. & Ciochon, R.L. (1990) A review of evidence for postulated Middle Pleistocene occupations in Viet Nam. Journal of Human Evolution, 19, 761–788. Page, S.E., Rieley, J.O., Shotyk, Ø.W. & Weiss, D. (1999) Interdepence of peat and vegetation in a tropical peat swamp forest. Philosophical Transactions of the Royal Society B: Biological Sciences, 354, 1885–1897. Pan, Y. (1993) The relationships between Late Cenozoic mammalian faunas from Yuanmou and their environment. Evolving landscapes and evolving biotas of East Asia since the Mid-Tertiary (ed. by N.G. Jablonski and S. Chak-Lam), pp. 207–220. Centre of Asian Studies, Hong Kong. Poole, I. & Davies, C. (2001) Glutoxylon Chowdhury (Anacardiaceae): the ﬁrst record of fossil wood from Bangladesh. Review of Palaeobotany and Palynology, 113, 261–272. Pope, G.G., Frayer, D.W., Liangchareon, M., Kulasing, P. & Nakabanlang, S. (1981) Palaeoanthropological investigations of the Thai–American expeditions in northern Thailand (1978–1980): an interim report. Asian Perspectives, 21, 147–163. Price, G.D., Valdes, P.J. & Sellwood, B.W. (1997) Prediction of modern bauxite occurrence: implications for climate reconstructions. Palaeogeography, Palaeoclimatology, Palaeoecology, 131, 1–13. Quade, J., Cerling, T.E. & Bowman, J.R. (1989) Development of Asian monsoon revealed by marked ecological shift during the latest Miocene in northern Pakistan. Nature, 402, 163–166. Rahardjo, A.T. (1999) Perubahan iklim dan batas umur Pliosen – Plistosen berdasarkan analisis forminifera dan palinologi di daerah Mojoroto, Mojokerto – Jawa Timur. Buletin Geologi, 31, 1–14. Reed, K.E. (1997) Early hominid evolution and ecological change through the African Plio-Pleistocene. Journal of Human Evolution, 32, 289–322. Reed, K.E. (1998) Using large mammal communities to examine ecological and taxonomic structure and predict vegetation in extant and extinct assemblages. Paleobiology, 24, 384–408. Journal of Biogeography ª 2010 Blackwell Publishing Ltd Palaeoecology of Southeast Asian megafauna-bearing sites Ruedi, M. (1996) Phylogenetic evolution and biogeography of Southeast Asian shrews (genus Crocidura: Soricidae). Biological Journal of the Linnean Society, 58, 197–219. Sandrock, O., Kullmer, O., Schrenk, F., Juwayeyi, Y.M. & Bromage, T.G. (2007) Fauna, taphonomy, and ecology of the Plio-Pleistocene Chiwondo Beds, Northern Malawi. Hominin environments in the East African Pliocene: an assessment of the faunal evidence (ed. by R. Bobe, Z. Alemseged and A.K. Behrensmeyer), pp. 315–332. Springer, Dordrecht. Simpson, K. & Day, N. (1996) Field guide to the birds of Australia. Penguin Books Australia Ltd, Ringwood, Australia. Smit-Sibinga, G.L. (1947) De jongste bodembewegingen in en om de Java Zee en de eustatische beweging van de zeespiegel. Tijdschrift Nederlandsch Aardrijkskundig Genootschap, 64, 455–465. Soligo, C. & Andrews, P. (2005) Taphonomic bias, taxonomic bias and historical non-equivalence of faunal structure in early hominid localities. Journal of Human Evolution, 49, 206–229. Stauffer, P.H. (1973) Cenozoic. Geology of the Malay Peninsula (West Malaysia and Singapore) (ed. by D.J. Gobbett and C.S. Hutchison), pp. 143–176. John Wiley and Sons, New York. Storm, P. (2001) The evolution of humans in Australasia from an environmental perspective. Palaeogeography, Palaeoclimatology, Palaeoecology, 171, 363–383. Storm, P. & de Vos, J. (2006) Rediscovery of the Late Pleistocene Punung hominin sites and the discovery of a new site Gunung Dawung in East Java. Palaeobiodiversity and Palaeoenvironments, 86, 271–281. Storm, P., Aziz, F., de Vos, J., Kosasih, D., Baskoro, S., Ngaliman & van den Hoek Ostende, L.W. (2005) Late Pleistocene Homo sapiens in a tropical rain forest fauna in East Java. Journal of Human Evolution, 49, 536–545. Sun, X., Li, X., Luo, Y. & Chen, X. (2000) The vegetation and climate at the last glaciation on the emerged continental shelf of the South China Sea. Palaeogeography, Palaeoclimatology, Palaeoecology, 160, 301–316. Teeuw, R.M., Rhodes, E.J. & Perkins, N.K. (1999) Dating of Quaternary sediments from Western Borneo, using optically stimulated luminescence. Singapore Journal of Tropical Geography, 20, 181–192. Thiramongkol, N. (1986) Neotectonism and rate of uplift in the eastern margin of the lower central plain of Thailand. Proceedings of the Workshop on Economic Geology, Tectonics, Sedimentary Processes and Environment of the Quarternary in Southeast Asia. Haay Yai, 3–7 February 1986 (ed. by N. Thiramongkol), pp. 35–44. Department of Geology, Chulalongkorn University, Bangkok, Thailand. Thomas, M.F. (1987) Quaternary sedimentation in W. Kalimantan (Indonesian Borneo). Tropical Geomorphology Newsletter, 4, 23. Thorp, M.B. & Thomas, M.F. (1993) Discussion: Late Pleistocene sedimentation and landform development in western Kalimantan (Indonesian Borneo). Reply by the authors. Geologie en Mijnbouw, 71, 363–368. Journal of Biogeography ª 2010 Blackwell Publishing Ltd Thorp, M.B., Thomas, M.F., Martin, T. & Whalley, B. (1990) Late Pleistocene sedimentation and landform development in western Kalimantan (Indonesian Borneo). Geologie en Mijnbouw, 69, 133–150. Tjia, H.D., Komoo, I., Lim, P.S. & Surat, T. (1990) The Maliau basin, Sabah: Geology and tectonic setting. Bulletin of the Geological Society of Malaysia, 27, 261–292. ` Tougard, C. (1998) Les faunes de grands mammiferes du ´ ` ¨ Pleistocene moyen terminal de Thaılande dans leur cadre ´ ´ ´ ´ phylogenetique, paleoecologique et biochronologique. PhD ´ Thesis, Universite Montpellier II, Montpellier. Tougard, C. (2001) Biogeography and migration routes of large mammal faunas in South-East Asia during the Late Middle Pleistocene: focus on the fossil and extant faunas from Thailand. Palaeogeography, Palaeoclimatology, Palaeoecology, 168, 337–358. Tougard, C. & Montuire, S. (2006) Pleistocene paleoenvironmental reconstructions and mammalian evolution in SouthEast Asia: focus on fossil faunas from Thailand. Quaternary Science Reviews, 25, 126–141. Travouillon, K.J. & Legendre, S. (2009) Using cenograms to investigate gaps in mammalian body mass distributions in Australian mammals. Palaeogeography, Palaeoclimatology, Palaeoecology, 272, 69–84. Visser, K., Thunell, R. & Goni, M.A. (2004) Glacial–interglacial organic carbon record from the Makassar Strait, Indonesia: implications for regional changes in continental vegetation. Quaternary Science Reviews, 23, 17–27. de Vos, J. (1983) The Pongo faunas from Java and Sumatra and their signiﬁcance for biostratigraphical and paleo-ecological interpretations. Proceedings of the Koninklijke Nederlandse Akadademie van Wetenschappen, 86, 417–425. de Vos, J., Sondaar, P.Y., van den Bergh, G.D. & Aziz, F. (1994) The Homo bearing deposits of Java and its ecological context. Courier Forschung-Institut Senckenberg, 171, 129–140. Westaway, K.E., Morwood, M.J., Roberts, R.G., Rokus, A.D., Zhao, J.-x., Storm, P., Aziz, F., van den Bergh, G., Hadi, P., Jatmiko & de Vos, J. (2007) Age and biostratigraphic signiﬁcance of the Punung Rain forest Fauna, East Java, Indonesia, and implications for Pongo and Homo. Journal of Human Evolution, 53, 709–717. Westaway, K.E., Roberts, R.G., Sutikna, T., Morwood, M.J., Drysdale, R., Zhao, J.-X. & Chivas, A.R. (2009a) The evolving landscape and climate of western Flores: an environmental context for the archaeological site of Liang Bua. Journal of Human Evolution, 57, 450–464. Westaway, K.E., Sutikna, T., Saptomo, W.E., Jatmiko, Morwood, M.J., Roberts, R.G. & Hobbs, D.R. (2009b) Reconstructing the geomorphic history of Liang Bua, Flores, Indonesia: a stratigraphic interpretation of the occupational environment. Journal of Human Evolution, 57, 465–483. Whitmore, T.C. (1987) Biogeographical evolution of the Malay Archipelago. Clarendon Press, Oxford. Whitten, T., Soeriaatmadja, R.E. & Aﬁff, S.A. (1996) The ecology of Java and Bali. Periplus Editions, Singapore. 17 J. Louys and E. Meijaard Whittow, J. (1984) Dictionary of physical geography. Penguin Books, London. Woodruff, D.S. & Turner, L.M. (2009) The Indochinese– Sundaic zoogeographic transition: a description and analysis of terrestrial mammal species distributions. Journal of Biogeography, 36, 803–821. Zeitoun, V., Seveau, A., Forestier, H., Thomas, H., Lenoble, A., Laudet, F., Antoine, P.-O., Debruyne, R., Ginsberg, L., Mein, P., Winayalai, C., Chumdee, N., Doyasa, T., Kijngam, A. & ´ Nakbunlung, S. (2005) Decouverte d’un assemblage ` faunique a Stegodon–Ailuropoda dans une grotte du Nord de ¨ la Thaılande (Ban Fa Suai, Chiang Dao). Comptes Rendus Palevol, 4, 255–264. Zheng, Z. & Lei, Z.Q. (1999) A 400,000 year record of vegetational and climatic changes from a volcanic basin, Leizhou Peninsula, southern China. Palaeogeography, Palaeoclimatology, Palaeoecology, 145, 339–362. SUPPORTING INFORMATION Additional supporting information may be found in the online version of this article: Appendix S1 Species and ecovariables from natural protected areas used in the analyses. Appendix S2 Species and ecovariables from fossil sites used in the analyses. Appendix S3 Description and examples of ecological categories used in the analyses. As a service to our authors and readers, this journal provides supporting information supplied by the authors. Such materials are peer-reviewed and may be re-organized for online delivery, but are not copy-edited or typeset. Technical support issues arising from supporting information (other than missing ﬁles) should be addressed to the authors. BIOSKETCHES Julien Louys is a postdoctoral research assistant at the Research Centre for Evolutionary Anthropology and Palaeoecology at Liverpool John Moores University. His current research interests include the development of new approaches in palaeocommunity and vertebrate palaeoecology analyses, as well as palaeoenvironmental reconstructions and the biogeographical history of the African and Asian Plio-Pleistocene. Erik Meijaard has worked in Indonesia as a conservation scientist since 1992. He has worked for various conservation NGOs and research organizations. His research interests are broad and vary from conservation economics and management studies to more fundamental research on mammal evolution, biogeography and taxonomy. He presently works as forest director for People and Nature Consulting International, a conservation consultancy based in Bali, but his real ambition is to become a farmer in the Scottish Highlands. Author contributions: J.L. and E.M. contributed equally to this research. Editor: Mark Carine 18 Journal of Biogeography ª 2010 Blackwell Publishing Ltd

Palaeoecology of Southeast Asian megafauna-bearing sites from the Pleistocene and a review of environmental changes in the region more

Log In

Reset Password