Human influence on distribution and extinctions of the late Pleistocene Eurasian megafauna

Extinction in the Anthropocene.

Taphonomic and paleoecological aspects of large mammals from Sudety Mts (Silesia, SW Poland), with particular interest to the carnivores

Etude de la mobilite des groupes aurignaciens du bassin de la Vezere a partir de l'economie de la matiere premiere lithique et de la saisonnalite animale. Ces elements fournissent des donnees nouvelles sur l'extension du territoire exploite, sur le degre de sedentarisation et de saisonnalite humaine

GC: Your professional life is closely tied to New York City, but the Second City [Chicago] was the first chronologically. What about your life in Chicago was significant to your subsequent career in anthropology? SS: The Chicago time of my life had everything to do with the way I've functioned since. I was born in the shadow of the 1933 World's Fair, in the middle of the Depression, into an Orthodox Jewish family, the youngest of seven children. My parents had both immigrated from Lithuania as young adults. My father was a rabbi, a scholar, a teacher. To support his family he became a ritual slaughterer of poultry. He came from a distinguished line of rabbis and scholars. My mother came from a simpler background and only went through the third grade. While she was raising all her children and also supporting the family during the war years, she went back to school to get her 8th-grade diploma, and her ambition was to get her high school diploma. She was an extremely intelligent woman who could do everything: she was a dressmaker, a professional cook, had a talent for business; she helped all her brothers and sisters get started in business. She gave me a sense that everything was possible for me.

Eight anatomically and taxonomically different finds are presented in this paper, and they belong to four taxa: woolly mammoth (Mammuthus primigenius), giant deer (Megaloceros giganteus), red deer (Cervus elaphus), and dog (Canis familiaris). All specimens represent allochthonous Late Pleistocene and Holocene animal remains, and all were dredged during the gravel exploitation at the Sekuline site near Molve (Podravina region, SW Pannonian basin, NE Croatia). Mammoth remains (bone and tusk fragments) were radiocarbon dated, and these are the first absolute dates on mammoths in Croatia. One upper last left deciduous premolar (dP4 sin.) also belongs to the same species. Ascribed to a dog is one well-preserved skull with a peculiar abscess scar on the maxillary bone as the result of an inflammatory process on the carnassial (P4) premolar. The Late Pleistocene cervid remains are giant deer, while the other cervid finds were determined to be red deer of the Holocene age. Morphometrical and taphonomical data are presented for each specimen. Such fossil and recent bone/tooth aggregates are characteristic of fluvial deposits and selective collecting. Although lacking stratigraphic provenance, these finds help to fulfil the gaps in palaeoenvironmental, palaeoecological, and palaeoclimate reconstructions of Podravina and its neighbouring areas.

Bone and tooth microstructure in extinct and extant mammals and implications for growth and life history evolution, with an emphasis on cervids as a case study

Recent studies connecting the decline of large predators and consumers with the disintegration of ecosystems often overlook that this natural experiment already occurred. As recently as 14 ka, tens of millions of large‐bodied mammals were widespread across the American continents. Within 1000 yr of the arrival of humans, ∼ 80% were extinct including all > 600 kg. While the cause of the late Pleistocene (LP) extinction remains contentious, largely overlooked are the ecological consequences of the loss of millions of large‐bodied animals. Here, we examine the influence of the LP extinction on a local mammal community. Our study site is Hall's Cave in the Great Plains of Texas, which has unparalleled fine‐grained temporal resolution over the past 20 ka, allowing characterization of the community before and after the extinction. In step with continental patterns, this community lost 80% of large‐bodied herbivores and 20% of apex predators at the LP extinction. Using tightly constrained temporal windows spanning full glacial to modern time periods and comprehensive faunal lists, we reconstruct mammal associations and body size distributions over time. We find changes in alpha and beta diversity, and in the statistical moments associated with periods of climate change as well as with the LP extinction event. Additionally, there is a fundamental change in the composition of herbivores, with grazers being replaced by frugivores/granivores starting about 15 ka; the only large‐bodied grazer remaining today is the bison Bison bison . Moreover, the null model program PAIRS reveals interesting temporal patterns in the disassociation or co‐occurrence of species through the terminal Pleistocene and Holocene. Extinct species formed more significant associations than modern ones, and formed more aggregated pairs than do modern species. Further, negative species associations were about three times stronger than positive ones, suggesting that competitive interactions or environmental filtering are a strong force in community structure.

Land bridges once assured transcontinental connectivity, but climate-induced habitat loss resulted in the extinc- tion of numerous North American large mammals. Using GPS technology on the formerly widespread but now endan- gered saiga in Mongolia, we identified a fine-scaled 5-km wide critical corridor, whose protection is critical for maintain- ing migration and meta-population structure. The world's great overland migrations are disappearing, truncating fundamental processes that have contributed to ecosystem functioning for millennia. With more people reli- ant on lands that were once remote, intact habitats are now replaced by livestock, fences, and people (1). Nevertheless, expansive grasslands and deserts in China, Mongolia, Ka- zakhstan, and Russia still sustain extraordinary movements between winter and summer ranges, including those of chiru (Pantholops hodgsonii), white-(Procapra gutturosa) and black-tailed (Gazella subgutturosa) gazelles, khulan (Equus hemionus), saiga (Saiga tatarica), and Bactrian camels (Camelus bactrianus) (2 - 4). The latter three species - all recognized as endangered by the IUCN - along with mam- moths, Panthera lions, and wild horses (Equus spp.) once moved between Asia and North America using the Bering Land Bridge. That conduit for inter-continental movement collapsed during the Late Pleistocene when climate changed, ocean levels rose, and the arid Mammoth Steppe of Beringea disappeared (5). Although saiga, camels, and wild horses were dependent upon these cold-adapted xeric grasslands, such remnant Pleistocene habitats and their associated fauna currently persist in situ only in Central Asia (5). While the Being Land Bridge once enabled connectivity across broad landscapes, today's conservation challenges are appropriately more linked to the maintenance of connectivity among population subunits, protecting corridors at a fine scale, and understanding the direct impacts of humans under an umbrella of climate change. Long-term conservation of fragmented populations requires maintenance of meta- population structure. With the above little-known relicts re- stricted to Central Asia, a critical challenge has been the ac- quisition of knowledge, not only about migration routes, if any, but the protection of land(s) critical to sustaining movements among disjunct population segments (6).

The faunal succession of Japanese Quaternary mammals is described within the stratigraphic framework provided mainly by KAMEI, KAWAMURA and TARUNO (1988). Descriptions are given separately for Hokkaido, Honshu-Shikoku-Kyushu and the Ryukyu Islands.In Hokkaido, Pleistocene mammalian remains are too scarce to provide a detailed faunal succession, but abundant remains of Holocene age suggest that the fauna was almost identical to that of the present day. Large mammals recorded for the Late Pleistocene are therefore considered to have been extinct by the Holocene.In Honshu-Shikoku-Kyushu, the Early Pleistocene fauna is of temperate forest type, and related to those of north China. Almost all the components are, however, assigned to extinct endemic species. The Middle Pleistocene fauna is characterized by the presence of extant species. In fact, they exceed half of the components in the middle Middle Pleistocene fauna, and are still more common in the later faunas. This fauna is also dominated by temperate forest elements and endemic species. Immigration from south China in the middle Middle Pleistocene is more limited than previously thought, and only a few forms migrated from north and northeast China in the late Middle Pleistocene. The Late Pleistocene fauna is basically identical with that of the Middle Pleistocene except for the absence of several extinct species and several exotic species which still survive in other regions. Although the fauna seems to have been isolated from those of the adjacent continent in the early Late Pleistocene, immigration of large herbivores from the northern part of the continent was recognized in the late Late Pleistocene. Most of the extinct and exotic species were eliminated from the fauna between 20, 000 and 10, 000 years BP, and thus the fauna became almost identical with that of the present day by the early Holocene.In the Ryukyu Islands, Early and Middle Pleistocene faunas are almost unknown, while Late Pleistocene and Holocene ones are relatively well recorded. The Late Pleistocene fauna is of insular type, and includes several species endemic to the islands. Some of them are extinct species. From the end of the Pleistocene to the Holocene, insularity of the fauna was enhanced by the extinction of major species and by extreme reduction in habitat areas of the survivors.

Australia's recently established predators restore complexity to food webs simplified by extinction

Comments on Steadman, D.W., Martin, P.S., MacPhee, R.D.E., Jull, A.J.T., McDonald, H.G., Woods, C.A., Iturralde-Vinent, M. & Hodgins, G.W.L. (2005) Asynchronous extinction of late Quaternary sloths on continents and islands. Proceedings of the National Academy of Sciences USA, 102, 11763–11768. The debate over the causes of the Pleistocene megafaunal extinction dates back to the early 19th century (Grayson, 1984), and continues to generate considerable controversy (e.g. Grayson & Meltzer, 2003; Araujo et al., 2004; De Vivo & Carmignotto, 2004; Fiedel & Haynes, 2004; Burney & Flannery, 2005; Wroe et al., 2006). Typically, protagonists in this debate can be classified into two groups. One group argues that Late Pleistocene megafaunal extinctions were primarily caused by direct and indirect human action through hunting, habitat modification or introduction of new predators (Burney & Flannery, 2005, 2006; Barnosky et al., 2004; Fiedel & Haynes, 2004). The other interpretation is that humans had at most a minor role in the megafaunal extinction, and that the loss was attributable principally to a climatic cause (Ficcarelli et al., 2003; Grayson & Meltzer, 2003, 2004; Barnosky et al., 2004; De Vivo & Carmignotto, 2004; Boeskorov, 2006; Guthrie, 2006; Wroe et al., 2006; Wroe & Field, 2006). Here we contest the position of Steadman et al. (2005), who favour the overkill hypothesis to explain the ground sloth extinction in the Americas. Although making an important contribution to the debate on extinction of the New World megafauna, Steadman et al. (2005) make some important assumptions in their analysis. Steadman et al. (2005) argue that the extinction of ground sloths in the New World was concomitant with, and a consequence of, the human occupation of the Americas. Their argument is two-fold. First, the radiocarbon dates (14C) accepted by them for the last appearance dates (LADs) of these animals roughly correspond to megafaunal extinction dates in South and North America and the West Indies. These dates coincide with the human colonization of these regions and they argue that this supports the thesis that human arrival caused extinction of the ground sloth. Second, according to Steadman et al., extinctions caused by climatic fluctuation would result in concomitant LADs across the entire continent and associated islands, as they viewed these fluctuations as being widespread and uniform, whilst they found that the LADs for the West Indies, around 4400 14C yr bp [c. 4800–5050 calibrated years before present (cal. bp); dates calibrated with calib 5.0, Stuiver et al., 2005], are much younger than those found in the continent (c. 11,000 14C yr bp; c. 12,880–12,950 cal. bp for North America and c. 10,500 14C yr bp; c. 12,390–12,640 cal. bp for South America). We contend that the chronological data presented by Steadman et al. (2005) are incomplete, especially when considering South America. While Steadman et al. (2005) suggest that there are no acceptable Holocene LADs for ground sloths, a large number of Holocene dates generated through direct dating of bone and dung remains are indeed available in the literature. Barnosky et al. (2004; supporting material) revised the radiocarbon dates available for megafaunal remains throughout the world. In South America, they listed four articles with remains of megafauna dated within the Holocene, based both on direct and indirect dates. Even when considering only the results based on direct dates of bone remains, sufficient evidence still supports Holocene LADs for subequatorial ground sloths. For instance, from Argentina, Borrero et al. (1998) presented a total of seven 14C dates consistent with a Holocene survival of megafauna, albeit two of these ages are potentially unreliable, and four were obtained from one single specimen (indeed, one of the unreliable dates comes from this specimen; Table 1). Other reports not included in Barnosky et al. (2004) provide two direct radiocarbon ages of megafaunal bone remains from central Brazil at the Pleistocene/Holocene boundary (Table 1; Neves & Piló, 2003; Araujo et al., 2004). Politis et al. (2004; also not included in Barnosky et al., 2004) presented two additional Holocene direct radiocarbon ages of Megatherium americanum (Blumenbach) specimens (Table 1) and a third one from the Holocene/Pleistocene boundary (10,190 ± 120 14C yr bp; c. 11,820–12,020 cal. bp; Table 1), all in Argentina; and Marshall et al. (1984; also not included in Barnosky et al., 2004) reported a single Holocene age of 8910 ± 200 14C yr bp (c. 9780–10,150 cal. bp; GIF-4116) of a Scelidodon chiliensis (Lydekker) in Peru (Marshall et al., 1984;Pujos & Salas, 2004). Four of the sites where these dates were obtained are located in Argentina, while two are located in central Brazil and the last in Peru (Fig. 1). All the Argentinean sites (Arroyo Seco 2, La Moderna, Campo Laborde and Paso Otero 5) are open-air archaeological sites, i.e. the megafaunal remains are associated with prehistoric human occupations (see Borrero et al., 1998; Politis et al., 2004 for detailed descriptions). Arroyo Seco 2 is interpreted as a base camp where a large variety of activities were undertaken (Politis et al., 2004), including the exploitation of ground sloths and other megafauna by humans. However, Borrero et al. (1998) and Politis et al. (2004) do not state clearly if the two specimens (M. americanum and Equus neogeus Lund) that dated to the Holocene (Table 1) showed marks of human manipulation. The remaining open-air sites are believed to be sites used for specific activities (Politis et al., 2004): La Moderna is interpreted as an occasional megafaunal processing site, where the remains of a single glyptodont (Doedicurus clavicaudatus Owen) dated to the Holocene (Table 1) were recovered; Campo Laborde presents evidence that it was used as a hunting and processing site for ground sloths (M. americanum; Table 1); and Paso Otero 5, was also identified as a hunting and processing site for local megafauna. Archaeological and palaeontological sites in South America presenting direct Late Pleistocene/Early Holocene radiocarbon (14C) dates for megafaunal remains. Circles represent sites with no evidence of human exploitation of the megafaunal remains, whereas triangles represent sites with evidence of human exploitation of megafauna. 1, Gruta Cuvieri; 2, Escrivânia 5; 3, Gruta del Indio; 4, La Moderna; 5, Campo Laborde; 6, Arroyo Seco 2; 7, Paso Otero 5; 8, Pampa de los Fósiles. The two Brazilian sites, in contrast, are exclusively palaeontological, i.e. they are not associated with human occupations, and are located in limestone caves in the karstic region of Lagoa Santa. Gruta Cuvieri is a cave where three vertical chambers functioned as natural traps for the now extinct megafauna and other animals. The only megafauna species found so far is Catonyx cuvieri (Lund), a medium-sized ground sloth. The Holocene date presented in Table 1 was obtained from one of these ground sloths, found at the surface of one of the chambers. The other Brazilian site, Escrivânia 5, is part of a complex of caves, generically referred to as Escrivânia, representing one of the richest palaeontological limestone outcrops known at Lagoa Santa. Together with tons of animal fossil bones, in one of the chambers (Escrivânia 3) an almost complete human skeleton was also recently recovered, dated to 7650 ± 80 14C yr bp (c. 8370–8420 cal. bp; Beta 174734). The Peruvian site, Pampa de los Fósiles, is also a palaeontological site located in the Cupisnique Desert. Several archaeological sites in the region have revealed no evidence of human interaction with the megafauna in the region (Pujos & Salas, 2004). In addition to these reported dates, Steadman et al. (2005; supporting material) disqualified two other Holocene dates as unreliable (they also rejected a third date, but it has a very large margin of error). These were the only Holocene dates found in their bibliographical revision and they ‘have means that are up to 1000 years younger than means of any [of the accepted LADs] [Supplementary online material]’. As 10 reliable Holocene direct radiocarbon dates for megafauna are described here, there is no further reason to reject the dates of 8990 ± 90 14C yr bp (c. 9920–10,190 cal. bp; LP-925; Garcia, 2003) and 9560 ± 90 14C yr bp (c. 10,680–10,860 cal. bp; GrN-5772; Long et al., 1998) as unacceptable outliers. These two dates are from an Argentinean site, Gruta del Indio (Fig. 1; see Long et al., 1998; Garcia, 2003 for detailed descriptions). This site is a rockshelter, and although it presents chronological information placing humans together with megafauna in time, there is no evidence of humans exploiting the local megafauna (Long et al., 1998; Garcia, 2003). As presented in Table 1, from the 14 existing Holocene dates we found for megafaunal remains in South America eight are derived from ground sloths, which severely weakens the position of Steadman et al. (2005), that there are no acceptable Holocene LADs for ground sloths in the Americas. Assuming that human groups already inhabited South America around 12,500 14C yr bp (c. 14,300–14,950 cal. bp; Dillehay, 2000), the argument that the ground sloth LADs were concomitant with the human arrival in the New World can no longer be accepted, at least not as an immediate phenomenon. The second argument presented by Steadman et al. (2005) is that the apparent delay observed in the LADs of Central America islands, when compared with the continental ones, favours the overkill hypothesis. Delayed LADs in insular regions have been found in other parts of the world, independent of human presence (Guthrie, 2004; Boeskorov, 2006). Boeskorov (2006) showed that in northern Eurasian islands, megafauna survived into the Holocene, e.g. the mammoths of Wrangel Island. Nonetheless, the extinction of megafauna in Eurasia as a whole is believed to be primarily due to climatic changes (Barnosky et al., 2004; Boeskorov, 2006), particularly because no human presence is found in the Wrangel Islands until well after the extinction of the megafauna (Boeskorov, 2006). Although these data do not peremptorily disqualify Steadman’s argument, they do bring into question whether the overkill hypothesis is the most parsimonious explanation for megafaunal extinctions. Finally, it must be emphasized that there is a general lack of evidence of sloth remains in archaeological contexts in the Americas as a whole (but see Politis et al., 2004 for an exception), which also speaks against the overkill hypothesis. Specifically, in Lagoa Santa, despite the excavation of dozens of archaeological sites dated to the Pleistocene/Holocene transition (showing human evidence as old as 11,000–11,500 14C yr bp; c. 12,880–13,400 cal. bp; Neves et al., 1999), evidence is lacking of megafaunal use by humans, either as a source of food or raw material (Kipnis, 1998; Prous & Fogaça, 1999). In North America, a similar situation is observed. According to Grayson & Meltzer (2003), there are only two genera of megafauna (Mammuthus Burnett, 1830 and Mammut Blumenbach, 1799) known to have been hunted by humans during the Clovis period (Grayson & Meltzer, 2003). This scenario is accepted even by Fiedel & Haynes (2004), strong defenders of the overkill hypothesis. Thus, at least in South America (and most probably in North and Central America as well), the idea that ground sloths went extinct due to overkill lacks archaeological support. In conclusion, the ground sloth overkill hypothesis, as defended by Steadman et al. (2005), is not sufficiently supported in the empirical world. As we have briefly pointed out: (1) a considerable number of reliable Holocene dates for megafaunal specimens in South America already exist, including for ground sloths; (2) the existence of late megafaunal LADs in Central America islands can be equally well explained through overkilling or environmental changes; and (3) the general lack of megafaunal killing sites and megafaunal remains in archaeological contexts is inconsistent with the overkill hypothesis. Nonetheless, it is important to emphasize that the amount of information regarding the presence of megafauna in archaeological sites is still too small to be considered as strong evidence against human predation of megafauna, and thus this piece of information must be interpreted as complementary to the others. Collectively, the data presented here are more consistent with a model explaining megafaunal extinction through climatic fluctuations, although in South America the poor chronological contextualization of the megafaunal decline does not yet allow for a percentage estimate of megafaunal genera that survived until human arrival. In North America (Grayson & Meltzer, 2002, 2003) and in Australia (Wroe et al., 2006; Wroe & Field, 2006), this percentage seems to have been small, suggesting that the megafaunal extinction was a protracted process, beginning much earlier than the human settlement of these continents. Such a decline may have been the case in South America, as only a few megafaunal genera apparently survived until the Holocene. While a human presence could have accelerated the process of extinction of the remaining megafaunal genera, climatic fluctuations could also have been responsible. Araujo et al. (2005) suggested a period of drought during the mid-Holocene in central Brazil, based on a general abandonment of the region by humans and also on palaeoenvironmental data. At least for central Brazil, megafaunal extinction could thus be also explained by the dry period that started between 8500 and 7500 14C yr bp (c. 9520–8190 cal. bp). Furthermore, according to Araujo et al. (2005) several authors recognize the existence of dry climatic periods during the early and mid-Holocene in South America. Bush et al. (2005) also found evidence suggesting the existence of this drier period in the Andes region (between 0° and 24°), although in this case it was not a single or synchronous event. Even if asynchronous, the important point here is that this dry period seems to have been a widespread phenomenon in South America. Thus, we concur with Borrero et al. (1998, p. 197) who propose that ‘people played at most a secondary role in the mega mammal extinctions, perhaps accelerating a process already underway before human arrival in South America’. We would like to thanks Rodolfo Salas for his kindness in assisting us in determining the Holocene date in Peru. Our long-term research in Lagoa Santa is funded by FAPESP (grant 04/01321-6) and by scholarships given to AH (FAPESP 04/11485-6), MH (FAPESP 04/01253-0) and to WAN (CNPQ 305918/85-0). Alex Hubbe is a graduate student at the Laboratory for Human Evolutionary Studies, Instituto de Biociências, Universidade de São Paulo. His main interests are the palaeoecology and extinction of the South America megafauna. Mark Hubbe is an investigator at the Instituto de Investigaciones Arqueológicas y Museo, Universidad Católica del Norte, Chile. His main research interest is the origin and dispersion of the First Americans. Walter Neves is the coordinator of the Laboratory for Human Evolutionary Studies, Instituto de Biociências, Universidade de São Paulo. His main research interest is the origins and adaptations of the First Americans. Editor: Mark Bush

Holocene survival of Late Pleistocene megafauna in China: a critical review of the evidence

The woolly rhinoceros (Coelodonta antiquitatis) was a cold-adapted herbivore, widely distributed from western Europe to north-east Siberia during the Late Pleistocene. Previous studies have associated the extinction of the species ∼14,000 calendar years before present to climatic and vegetational changes, suggesting the later survival of populations in north-east Siberia may have related to the later persistence of open vegetation in the region. Here, we analyzed carbon (δ13C) and nitrogen (δ15N) stable isotopes and mitochondrial DNA sequences to elucidate the evolutionary ecology of the species. Our dataset comprised 286 woolly rhinoceros isotopic records, including 192 unpublished records, from across the species range, dating from >58,600 to 12,135 14C years before present (equivalent to 14,040 calendar years ago). Crucially, we present the first 71 isotopic records available to date of the 15,000 years preceding woolly rhinoceros extinction. The data revealed ecological flexibility and geographic variation in woolly rhinoceros stable isotope compositions across time. In north-east Siberia, we detected stability in δ15N through time, which could reflect long-term environmental stability, and may have enabled the later survival of the species in the region. To further investigate the paleoecology of woolly rhinoceroses, we compared their isotopic compositions with other contemporary herbivores. Our findings suggested isotopic similarities between woolly rhinoceros and both musk ox (Ovibos moschatus) and saiga (Saiga tatarica), albeit at varying points in time, and possible niche partitioning between woolly rhinoceros and both horse (Equus spp.) and woolly mammoth (Mammuthus primigenius). To provide phylogeographic context to the isotopic data, we compiled and analyzed the 61 published mitochondrial control region sequences. The genetic data showed a lack of geographic structuring; we found three haplogroups with overlapping distributions, all of which showed a signal of expansion during the Last Glacial Maximum. Furthermore, our genetic findings support the notion that environmental stability in Siberia influenced the paleoecology of woolly rhinoceroses in the region. Our study highlights the utility of combining stable isotopic records with ancient DNA to advance our knowledge of the evolutionary ecology of past populations and extinct species.

Mammals as a rule have seven cervical vertebrae, a number which remains remarkably conserved. Occasional deviations of this number are usually due to the presence of cervical ribs on the seventh vertebra, indicating a homeotic transformation from a cervical rib-less vertebra into a thoracic rib-bearing vertebra. These transformations are often associated with major congenital abnormalities or pediatric cancers (pleiotropic effects) that are, at least in humans, strongly selected against. Based on data from Late Pleistocene mammoths (Mammuthus primigenius) and woolly rhinoceroses (Coelodonta antiquitatis) from the North Sea, we hypothesized that high incidences of cervical ribs in declining populations are due to inbreeding and/or adverse conditions impacting early pregnancies. In this study, we investigated the incidence of cervical ribs in an extinct Late Pleistocene megaherbivore, giant deer (Megaloceros giganteus) from Ireland and in the extant highly inbred Père David deer (Elaphurus davidianus) and in twenty other extant species. We show that the incidence of cervical ribs is exceptionally high in both the Irish giant deer and the Père David deer and much higher than in extant outbred deer. Our data support the hypothesis that inbreeding and genetic drift increase the frequencies of maladaptive alleles in populations at risk of extinction. The high incidence of cervical ribs indicates a vulnerable condition, which may have contributed to the extinction of megaherbivore species in the Late Pleistocene. We argue that cervical rib frequency may be a good proxy for extinction risk in inbred populations.

The timing and causes of extinctions of West Indian land mammals during three time intervals covering the last 20000 years (late Pleistocene and early Holocene, Amerindian, and post-Columbian) are discussed in detail. Late Pleistocene extinctions are attributed to climatic change and the post-glacial rise in sea level, whereas most late Holocene extinctions are probably human caused, resulting from predation, habitat destruction and introduction of exotic species. Extinctions have dramatically altered the composition of the non-volant mammal fauna, but have had a lesser impact on bats. Of the 76 recognized species of living and extinct non-volant mammals in the West Indies, 67 species (88%) have gone extinct since the late Pleistocene, whereas only eight of the 59 species of bats (14%) have disappeared during this same time interval. A larger percentage of Antillean bat species (24%) have suffered localized extinction on certain islands, particularly obligate cave-dwelling forms. These local extinctions occurred primarily on small islands, and probably resulted from changes in cave microclimates and flooding of low-lying caves by rising sea levels. The majority of West Indian bats and all of the edentates, primates and rodents are Neotropical in origin. The South American fossil record indicates that most West Indian terrestrial mammals did not evolve until the early Miocene or thereafter. The Caribbean islands had assumed essentially their modern position and configuration by the Miocene, thus leaving overwater dispersal as the primary mechanism by which these endemic South American mammal groups reached the islands. The primitive insectivores, Solenodon and Nesophontes, are derived from Early Tertiary forms in North America that may have reached the islands through vicariance by way of a proto-Antillean archipelago. Many of the bats are either conspecific or congeneric with mainland taxa, suggesting that most species reached the islands by overwater dispersal during the Late Cenozoic, primarily from Central and South America. Two hypothetical immigration rates are calculated for West Indian land mammals, one assuming the earliest colonization in the late Eocene and the other based on an early Miocene origin. The known Late Quaternary and living Antillean land mammal fauna was derived from approximately 50 separate colonization events (13 for non-volant mammals and 37 for bats) giving immigration rates of one species per 800000 years since the late Eocene, or one species per 400000 years since the early Miocene. Immigration rates for bats are approximately three times greater than those for non-volant mammals throughout the Tertiary and eight times greater in the Pleistocene, presumably reflecting their greater dispersal abilities. These immigration rates should be considered rough values, owing to deficiencies in the fossil record, especially the absence of pre-Pleistocene fossils. Extinction rates calculated for the last 20 000 years demonstrate that an average of one species of mammal went extinct every 267 years during that time period. Since the arrival of man in the West Indies some 4500 years ago, 37 species of non-volant mammals have disappeared giving the rapid extinction rate of one species every 122 years. Island area-species diversity curves are plotted for both the current and late Pleistocene mammal faunas. All Caribbean islands with a reasonably complete fossil record have more species in the late Pleistocene and Holocene than in the living fauna. The living non-volant mammals of the West Indies do not constitute a natural fauna, but are an impoverished subset of species that managed to escape the extinctions that decimated the remainder of the fauna. Historical or theoretical biogeographic analyses of Antillean mammals that fail to incorporate extinct forms will be unlikely to elicit any meaningful patterns.