McDonald 313 PALEOECOLOGY OF EXTINCT XENARTHRANS AND THE GREAT AMERICAN BIOTIC INTERCHANGE H. Gregory McDonald1 The Xenarthra were the most successful South American mammals to participate in the Great American Biotic Interchange (GABI) and representatives of each family dispersed to at least the middle latitudes in North America. The ability of xenarthrans to insinuate themselves into the North American fauna suggests their ecology was sufficiently different from endemic North American taxa to avoid direct competition and sufficiently different amongst themselves to avoid competition with each other. Despite the diversity of xenarthrans in South America at the time of the interchange only one member of each major xenarthran group (roughly equivalent to a family) reached temperate North America. The morphological diversity displayed by ground sloths, glyptodonts, pampatheres, armadillos and anteaters, indicates that no single ecological explanation can account for their successful dispersal into North America as each represents a distinct ecological adaptation, including a variety of feeding strategies such as omnivore/insectivore, specialized myrmecophage, and herbivory including low browser, high browser, aquatic grazer, intermediate feeder on coarse vegetation/grazer, and large grazer/intermediate feeder. Despite this ecological diversity there do seem to be some trends common to the group. Ecologically the dispersing member was more of a generalist than other members of the family and dispersing members tended to be small with a subsequent increase in size following their entry into North America. Key Words: Great American Biotic Interchange; Xenarthra; paleoecology; biogeography INTRODUCTION In recent years there has been a renewed interest in the Great American Biotic Interchange (GABI) (Stehli & Webb 1985 and papers within). The dispersal of some North American mammals into South America and the reciprocal dispersal of some South American forms north- ward has been looked at from a variety of angles, par- ticularly the resulting impact of the appearance of these “exotic” species on the respective native fauna of each continent (Webb 1991). Usually the emphasis has been on the greater impact of the northern forms on the South American native fauna than vice versa. There is no question that in terms of number of taxa, there was dis- equilibrium with regard to the dispersers, with the num- ber of taxa of North American origin going south ex- ceeding that of the southern invaders going north. How- ever, of all the groups of South American origin with members dispersing northward, none were as success- ful as the xenarthrans. Not only did more members of the Xenarthra reach temperate North America than any other South American disperser, but once present, most became an integral part of the North American fauna and were common enough to be present in numerous individual local faunas. The xenarthrans were able not only to compete ecologically with the native North Ameri- can forms, in terms of food resources, but were able to survive the native predators, the group whose North American members have been identified as having had the greatest impact on the native fauna after entering South America (Marshall 1981). Their success contin- ued until the end of the Pleistocene when along with the natives they also became extinct. This seems to be a rather unexpected history given our generally biased view of this metabolically challenged group. As noted by Marshall and Hecht (1978), the trophic role assumed by an animal in a community is best un- derstood when the structure of the community is viewed within a historical perspective. A critical component of this historical perspective for the entrance of xenarthrans into North America includes understanding what facili- tated the dispersal of certain taxa through the Panama- Bull. Fla. Mus. Nat. Hist. (2005) 45(4): 313-333 1Museum Management Program, National Park Service, 1201 Oakridge Drive, Suite 150, Fort Collins, CO 80525; 314 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb nian land bridge. Was it their physical or behavioral attributes or merely the opportunity to exploit available adaptive zones or niches present in both the tropical and temperate parts of North America? In order to under- stand the success of xenarthrans in establishing them- selves within the North American fauna it is equally im- portant to understand what permitted them to success- fully disperse. GENERALIZATIONS ABOUT XENARTHRAN GENERALISTS In addition to a distinctive suite of skeletal features that unite the xenarthrans, a physiological trait common to all members is a low basal metabolism. This low basal metabolism is an important factor that determines many of their ecological attributes, such as long gestation peri- ods, low growth rates, extended periods of parental care, and small litter size (McNab 1985). The result is that xenarthrans are K-selected with an overall low rate of reproduction. The one possible exception is Dasypus, which may be relatively more r-selected than other xenarthrans given its unusual approach to reproduction, especially in the subgenus Dasypus. Any consideration of the ecology of xenarthrans, the ability of some not only to disperse into North America but also their long- term success until the Pleistocene extinction event must include a careful consideration of their metabolism, whether considered as a detriment or an advantage. Traditionally the low basal metabolism of xenarthrans has been considered as indicative of their primitive status as the earliest offshoot from other euth- erians (Englemann 1985). However, McNab (1978) has argued that for many of the living xenarthrans a low basal metabolism is directly linked to the nutritive levels of their food sources – sloths as arboreal folivores, and anteaters as termite and ant specialists. Currently there is only one study that indicates that a fossil xenarthran, the sloth Nothrotheriops shastensis, had like its extant relatives, a low basal metabolism and consequently a lower core body temperature than expected for its body size (Ho 1967). An estimated core body temperature of 35.3 + 0.5° C, based on the hydroxyproline content of bone collagen, was calculated for Nothrotheriops shastensis. This lower metabolism in herbivorous xenarthrans such as sloths and glyptodonts may have permitted them to eat either nutritionally poorer plants than could be used by mammals with a higher metabo- lism or to consume plants containing higher levels of secondary metabolites. There are two broad categories of herbivores based on digestive strategy and the part of the gut that is en- larged in order to facilitate microbial fermentation of veg- etation. Extant mammals with an enlarged foregut to provide the main site of microbial activity include rumi- nants, camelids, suiforms, leaf-eating monkeys, sloths, and macropod marsupials. Those mammals with an enlarged hindgut and in which gastric and intestinal di- gestion occurs before microbial fermentation include perissodactyls, lagomorphs, capybaras, hyrax and pro- boscideans. In all of these latter herbivores there is an enlarged caecum and colon (Parra 1978). The signifi- cance of differences between fore- and hindgut ferment- ers in ungulates and their relative evolutionary success with regard to the utilization of available vegetation has been discussed by Janis (1976). Previously the digestive physiology of the extinct ground sloths has been considered to be similar to that of monogastric hindgut fermenters or caecalids such as proboscideans and perissodactyls (Guthrie 1984), so it has been with these animals that they were considered to have competed as megaherbivores. However, the stomach in the living tree sloths is large and complex; being composed of four chambers (Goffart 1971; Bauchop 1978) and composes 20 to 30% of the body weight of living sloths (Britton 1941). Its overall similar- ity to a rumen has been noted (Grassé 1955). Bauchop (1978) considered its general development in modern sloths to be broadly similar to that of a camel and the presence of numerous pillars also gives it a more rumi- nant-like structure. Despite the overall general similari- ties, the construction of the sloth stomach is not as com- plex as in ruminants and lacks the equivalent of the ru- minant omasum. While there is no direct evidence for the structure of the stomach in any of the extinct xenarthrans, since both of the living genera, Bradypus and Choloepus, which are placed in separate families, have similar stomach morphology, it is not unreasonable to extrapolate that the stomach in the extinct ground sloths was also chambered. Modern armadillos have a simple sac-like stomach, simpler in construction than sloth’s, reflecting food habits that tend to be insectivo- rous. Their extinct herbivorous relatives, pampatheres and glyptodonts, probably also had simple stomachs, which may account for the greater complexity of their teeth and associated masticatory apparatus for the me- chanical processing of vegetation, particularly in the glyptodonts. It should also be noted that all living xenarthrans, including both extant tree sloths, lack a caecum (Goffart 315 1971) and this is most likely also true for their extinct relatives. In living sloths the large intestine is short (Bauchop 1978) and therefore also departs from the di- gestive anatomy of proboscideans and perissodactyls and other hindgut fermenters. The feeding strategy of large caecalid grazers (hind- gut digesters) is to consume the entire grass plant, re- sulting in a dilution of the leaf nutrients. This feeding strategy results in the consumption of large volumes of plant material of low to modest nutrient value and also allows these animals to utilize plant parts not available to other herbivores including those parts high in silica and fiber. Based on the plant parts preserved in the dung of Nothrotheriops, a browser, and Mylodon, a grazer (Moore 1978), sloths were more selective in their feed- ing habits than caecalids. Given the large body size of Eremotherium it has been assumed that it would have been in competition with the other megaherbivores present in North America. This would have been initially with both gomphotheres such as Cuvieronius and the mastodon Mammut during the Blancan, with the addition of mammoths (Mammuthus) during the Irvingtonian and Rancholabrean. In South America, Eremotherium and the gomphothere Haplomastodon are closely associ- ated in their distribution and the presence of one in a fauna is often indicative of the presence of the other. The ability of these large bodied herbivores to coexist and share habitat has a more reasonable explanation if we consider that Eremotherium did not have a digestive physiology similar to that of proboscideans, but instead was more similar to non-ruminant artiodactyls. This would have permitted these animals to better partition the habitat they shared with regard to either the types of plants they consumed or at least the plant parts. Janzen and Martin (1982) have postulated that frugivory by ex- tinct horses, gomphotheres, and ground sloths may ex- plain the reproductive traits of some plants in the Cen- tral American lowland forests and that these animals may have been important dispersal agents that influenced the distribution of these plant species. Recent stable isotope studies by MacFadden et al. (1994) for the gomphothere Cuvieronius resulted in a wide range of values for 13C (-8.5 to -0.1‰) suggesting that it fed on a wide range of plant foots and was not restricted to fruits. Pending the discovery of more direct evidence on the diet of these animals, I suggest that if both gomphotheres and Eremotherium were in competition foraging in trees, then their respective roles may have been analogous to modern African elephants and giraffes with Eremotherium filling the “giraffe niche” and feeding on selected leaves and twigs, thus functioning as a high browser, in the New World ecosystems. In examining the success of herbivorous xenarthrans in North America, it is more likely that the best possible candidates for direct competition were non- ruminant artiodactyls such as peccaries and camelids. As will be discussed below there are functional similari- ties in the premaxillae between sloths and those artio- dactyls that lack upper incisors. Besides Eremotherium, the smaller browsing sloths Megalonyx and Nothrotheriops may have been selec- tive feeders and were capable of utilizing the more nu- trient rich parts of plants along with some ingestion of twigs and branches of mature parts of the plant. Lacking direct evidence of diet, such as dung, in most fossil species we often have to rely on distinctive anatomical features to serve as proxies that permit some reasonable inferences as to the animal’s feeding strate- gies. Often this is accomplished by looking for a corre- lation between a specific anatomical structure and the diet in extant forms and then looking for similar struc- tures in extinct forms. One approach has been to ex- amine the relationship of the shape of the premaxillae and the proportions of the anterior part of the skull to infer the general feeding strategies of ruminant artio- dactyls and their sister group the Tylopoda (Solounias et al. 1988; Janis & Ehrhardt 1988; Dompierre & Churcher 1996). Given the general similarities in the edentulous premaxillae of artiodactyls and sloths, the relationship between its shape and diet seen in this one set of herbi- vores seems applicable to the other and most likely per- formed similar functions with regard to feeding. Using this ungulate model it is possible to make some generali- zations with regard to the feeding strategies of those sloths that dispersed northward in comparison to those taxa restricted to South America. It appears that the dispersing sloth taxa all had more primitive or generalized premaxillae in comparison to other members of their family present at the time of the dispersal event. The premaxillae of Eremotherium are triangular and more primitive in comparison to those of Megatherium (Fig. 1), which are highly modified, and suggest a more specialized feeding function, perhaps lim- iting it to a smaller selection of plants or plant parts upon which it fed. Likewise, the premaxillae of Glossotherium/ Paramylodon are intermediate in structure between scelidotheres or even the more closely related genus MCDONALD: Xenarthrans and the Great American Biotic Interchange 316 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb Mylodon in which the premaxillae and facial region are long and narrow, and that of either Oreomylodon or Lestodon in which the anterior part of the muzzle is expanded along with a related widening of the premax- illae (Fig. 2). While we do not have specific details on the relationship of premaxillae and muzzle shape or pro- portions and feeding habits in either scelidotheres or lestodonts, they do represent extremes in skull morphol- ogy within the family Mylodontidae and presumably, like Megatherium, reflect some type of feeding specializa- tion. The shape of the premaxillae and muzzle in the Glossotherium/Paramylodon lineage represents an in- termediate and relatively unspecialized morphology and most likely more generalized and less specialized feed- ing habits. Of all the North American xenarthrans, the one taxon that we have the most dietary information for is Nothrotheriops shastense, thanks to the preservation of its dung in numerous dry caves in the southwestern United States. Numerous workers have analyzed the dung of Nothrotheriops (Hansen 1978; Laudermilk & Munz 1934, 1938; Martin et al. 1961; Thompson et al. 1980) and these studies have provided detailed informa- tion on the variety of vegetation in the diet of this extinct species. Hansen (1978) identified 72 species of plants in the dung of Nothrotheriops from Rampart Cave, the vast majority of which are low growing bushes and shrubs. While most of the species consumed are xerically adapted desert vegetation, indicating that Nothrotheriops was capable of surviving in relatively dry environments, other taxa such as the reed Phragmites indicates the animal fed in a variety of habitats. Mylodont sloths have generally been thought of as grazers although Naples (1989) suggested that based on her analysis of the skull morphology of Paramylodon harlani from Rancho La Brea, this species may have Figure 1. Comparison of the shape of the premaxillae of Eremotherium and Megatherium. 317 Figure 2. Comparison of the shape of the premaxillae of Paramylodon with a scelidothere, Proscelidodon and a lestodont, Lestodon, showing its intermediate morphology. been an intermediate in its habits and included browse as well as graze. Coltrain et al. (2004) analyzed the 13C and 15N in Paramylodon from Rancholabrean and ob- tained mean values of -20.99 ‰ for carbon and 7.93 ‰ for nitrogen based on 10 specimens. The value for ni- trogen was more positive than for equids or true rumi- nants from Rancho La Brea and placed the sloth be- tween nonruminants and ruminants isotopically. McDonald (1995) examined the relationship between jaw depth and alveolar length of the tooth row in Paramylodon and noted a relatively greater increase in jaw depth suggesting a greater degree of hypsodonty in later forms. Despite the presence of ever-growing teeth, the lack of enamel on the teeth in xenarthrans can result in a rapid rate of wear if they consume highly abrasive food such as grasses. An increased depth of the jaw permitted an overall increase in the height of the tooth and thus compensated for the more rapid rate of wear. This suggests that later members of the lineage were more adapted to grazing or at least consuming more abrasive vegetation than their predecessors. If this was the case, then the ancestors of this lineage participating in the faunal interchange may have been more general- ists in their feeding and less dependent on grasses, thus facilitating their dispersal. Dasypus novemcinctus is morphologically derived compared to other armadillos and shows specializations for myrmecophagy, although behaviorally it is one of the most generalized of the armadillos with regard to its food habits (Smith & Redford 1990) and amongst all the ar- madillos is one of the most omnivorous (Redford 1985). Since we have no direct evidence for the diet of D. bellus it may be unreasonable to only consider the diet of D. novemcinctus to interpret the food habits of D. bellus and an examination of the diet in other species of Dasypus can provide a somewhat broader perspective. MCDONALD: Xenarthrans and the Great American Biotic Interchange 318 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb In examining the diet of Dasypus Redford (1986) found that in the tropics the diet of D. novemcinctus is similar to that of D. sabanicola consisting predominately of ants and termites. In contrast, the stomach contents of Dasypus kappleri had only 29% ants and termites and in Dasypus hybridus it was 58%. The overall impres- sion is that all species of Dasypus are somewhat labile in their diet and what is consumed is dependent on the local environment and what food is available. Smith and Redford (1990) interpreted the food habits of D. novemcinctus as a dietary generalist that along with its unique reproductive habits of monozygotic polyembry- ony permitted its success in expanding into new habitats and environments. Consequently it has the widest dis- tribution of all armadillos ranging from South to North America. While we have no direct evidence of the food hab- its of Dasypus bellus, there is direct evidence that like the other species of the genus, D. bellus produced mul- tiple young. Usually four young are produced in D. novemcinctus and D. sabanicola and 4 to as many as 12 young have been documented in D. hydridus (Wetzel & Mondolfi 1979). The remains of multiple young were found associated with the skeleton of an adult D. bellus preserved in a cave in Florida (Auffenberg 1957). Based on a functional analysis of the skull of Holmesina and comparisons with some extant species of armadillo, Vizcaino et al. (1998) concluded that it fed on a variety of vegetable matter and although not nec- essarily strict grazers, Holmesina probably fed on coarse vegetation. In their analysis they compared feeding strat- egies in armadillos and African antelopes and noted a relationship between size and feeding strategies. Small armadillos, like small antelopes, tend to be browsers and that with an increase in size there is a shift in the degree of grazing, with glyptodonts being primarily grazers and pampatheres occupying an intermediate feeding strat- egy. With the size increase in the North American lin- eage of Holmesina (see below) it is possible that H. floridanus was more of a browser while the larger H. septentrionalis included a larger proportion of graze in its diet. While Webb (1978) included the glyptodonts as one of nine genera of savanna grazers, Gillette and Ray (1981) suggested that they were likely “aquatic graz- ers” along with the capybaras. Recently Fariña (1995) questioned Gillette and Ray’s interpretation of Glyptotherium having a preference for marshy, low- land habitats based on a biomechanical analysis of dif- ferent South American genera of glyptodonts. His analy- sis was based on the size relationship between glyptodont feet and estimates of body weight. He concluded that the feet were too small in proportion to the body size to support the animal in this type of terrain. He also noted the absence of glyptodont tracks at Bahia Blanca in Argentina, which had a soft substrate at the time the tracks were formed. However, Aramayo and Manera de Bianco (1996) later reported a single track of a glyptodont from the locality. While not as well repre- sented as other ichnotaxa reported from the fauna, it does indicate that at least one individual of glyptodont did wander onto this soft substrate. The jaw and denti- tion of Glyptotherium as in other glyptodonts is well adapted for processing grasses. These features include a deep jaw with extremely hypsodont, ever-growing teeth, the complexity of the teeth with an arrangement of osteodentine to compensate for the lack of enamel, and the structure of the zygomatic arch to orient the masseter muscles to grind abrasive vegetation such as grasses (Ferigolo 1985; Fariña 1985, 1988) At first glance it would seem that Myrmecophaga is a major contradiction to the argument that those xenarthrans that dispersed northward were generalists. The giant anteater is an obligate myrmecophage with a diet consisting almost entirely of ants and termites. Redford (1985) noted that it is not specialized on either ants or termites per se but rather small arthropods found in dense aggregations. It can however, in terms of diet, be considered the most highly specialized xenarthran to participate in the faunal interchange. In this regard it provides an interesting contrast to Dasypus novemcinctus and by extrapolation Dasypus bellus which although morphologically could be considered a myrmecophage probably had a diet consisting of a broad range of invertebrates, vertebrates, and plant material. While its diet is specialized, Myrmecophaga inhabits a large variety of habitats from tropical rainforest to the xeric scrub of the Chaco (Eisenberg & Redford 1999). Since all of the living xenarthran anteaters are highly specialized for feeding on ants and termites, within this context, the giant anteater, Myrmecophaga, can be con- sidered the most generalized in terms of inhabiting a greater variety of habitats. The smallest form, Cyclo- pes, the most highly adapted, is strictly arboreal in hab- its, while the intermediate sized Tamandua is semi-ar- boreal with a distribution restricted to forest with closed canopy, so both are restricted to forests with a closed canopy. Since Myrmecophaga favors open country 319 savanna, it is not surprising that it was a least a minor participant in the GABI, since this type of habitat is con- sidered to have provided the major access between North and South America (Webb 1978). Its dispersal north- ward was probably not limited by available habitat as much as by an available food source in the form of colo- nial ants and termites. With its distribution controlled pri- marily by the distribution of its food source and ant and termite diversity is greatest in the tropics and decreases rapidly with increasing latitude (Redford 1986), it is not surprising that the northern limit of Myrmecophaga in North America at El Golfo, Mexico at about 30 degrees, is similar to its southernmost latitude of 28 degrees in Argentina. The record of the anteater in northern Mexico is not only limited geographically but also chronologi- cally, as it is based on a single record from the Irvingtonian El Golfo fauna (Shaw & McDonald 1987). Further study of Irvingtonian faunas in Mexico should provide addi- tional records and enhance our understanding of the en- vironmental and climatic conditions that permitted its short foray northward. THE FEW, THE PROUD, THE CHOSEN The corridor through the Panamanian isthmus connect- ing North and South America acted as a filter and pre- vented a wholesale exchange of taxa between the two continents (Webb 1978). At each stage of the inter- change, with the first xenarthrans dispersing northward in the Huayquerian to the last in the Lujanian there was a diverse xenarthran fauna in South America from which to derive the dispersers. Yet despite the large number of potential candidates, whether pilosans (Fig. 3) or cingulates (Fig. 4), at each phase of the interchange only one taxon from each major group passed through the filter and reached temperate North America. Rather than thinking of the Panamanian corridor as a filter, per- haps a more accurate model with regard to the xenarthrans is that of a nested sieve. The different size holes in each layer of the sieve represent different eco- logical parameters that allowed different taxa to disperse varying distances northward but at the same time held back other taxa. This sieving essentially resulted in only one member of each xenarthran family at any one time during the interchange to disperse as far northward as temperate North America while others were restricted to the tropics (see discussion below). As previously dis- cussed, each xenarthran participating in the interchange seems to have been a generalist, at least in comparison to other members of its family. As such, each xenarthran that participated in the interchange was capable of not only avoiding potential ecological competition with the endemic North American mammals but with the more specialized members of their own respective family and with other members of its order. Each represents a distinct morphological and presumably ecological type within the broader categories of omnivory, insectivory, and herbivory (both browsers and grazers). This is re- flected not only in each animal’s anatomy but in size as well. The earliest representatives of each group in North America fall into well separated size classes and this distinction tends to be maintained even as each lineage increases in size (Fig. 5). The only exception seems to be in Megalonyx, a browser, which from the Blancan to the end of the Pleistocene has members comparable in size to Paramylodon, a grazer or at least a mixed intermediate feeder. While the premaxillae of Nothrotheriops and Eremotherium are somewhat similar in morphology (Fig. 6) and both have been interpreted as browsers, the dif- ference in size between the two genera would certainly indicate different feeding strategies. Nothrotheriops can be considered a low browser while Eremotherium would have been a high browser. The distribution of both genera overlapped in the eastern United States during the Irvingtonian but by the Rancholabrean there was no overlap in their distribution suggesting even more pronounced differences in their preferred habitat reflect- ing significance differences in their ecology and pre- sumably diet. Much closer in size and presumably with greater potential for ecological competition would have been Nothrotheriops and Megalonyx, both interpreted as browsers. Each has its own distinctive skull and dental morphology, which reflect different feeding strategies. Both genera co-occur in faunas in North America dur- ing the Rancholabrean but in inverse proportions so when one is common the other is rare. McDonald (1996) ex- amined the distribution and association of ground sloths in parts of the western United States where the distribu- tion of the two genera overlapped and found that it was more likely that either of the two genera would be found with the grazer Paramylodon in a fauna than with each other. Within the cingulates, there is an omnivore/insecti- vore specialist for ants and termites in the form of Dasy- pus, the larger herbivorous pampathere Holmesina, and the even larger grazing glyptodont Glyptotherium. To this suite of cingulates in the Irvingtonian we can add MCDONALD: Xenarthrans and the Great American Biotic Interchange 3 2 0 C E N O Z O IC V E R T E B R A T E S : P apers to H onor S . D avid W ebb Figure 3. Chronologic distribution of sloths in North and South America during the Great American Biotic Interchange. 3 2 1Figure 4. Chronologic distribution of cingulates in North and South America during the Great American Biotic Interchange. M C D O N A LD : X enarthrans and the G reat A m erican B iotic Interchange 322 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb Figure 5. Change in body weight through time for North American xenarthrans. the strange form, Pachyarmatherium, which unfortu- nately is not well represented in the fossil record thus preventing any real inferences about its paleoecology. For the other better known cingulates there does not seem to be any overlap in their ecology and as armored mammals, definitely nothing comparable existed in the North American fauna. Webb (1976) identified a number of features of the interchange that need to be considered with regard to trying to determine the ecological attributes of the xenarthrans involved. One consideration is topographic. South American vertebrates dispersing northward may have utilized the mountain corridor connecting the Co- lombian Andes with the Central Highlands of Panama (Webb’s high road) or alternatively followed a lowland coastal route as they moved northward (Webb’s low road). Secondly, the types of habitat present in each corridor might aid in identifying more specifically what animals could or could not move northward through the region. Webb (1976) suggested that at the time of the interchange there was a broad spectrum of habitats rep- resented in the crucial tropical regions, including mesic forests, an array of low gradient streams, and major throughways of scrub and savanna. Recent studies of pollen from the Panamanian isthmus by Graham (1992) indicates that by the middle Pliocene grass goes from virtually absent to a maximum of 7.5% indicating the first appearance of some version of an open tropical dry forest. Prior to this time what is now the Panamanian isthmus consisted of low-lying islands with maximum elevations of 1200 to 1400 meters and habitat type was limited to those of the lowland tropics. These are the conditions that would have existed at the time of the dispersal of the two sloths Pliometanastes and Thinobadistes, but prevented the other xenarthrans as well as other South American mammals from dispersing northward. The presence of moderate habitat diversity with temperate environments at higher elevations, tropi- cal dry forests with a significant grass component, and sufficient topographic relief to produce different rainfall regimes on the Atlantic and Pacific sides of the isthmus all developed after the middle Pliocene and it was at this time that the greatest variety of xenarthrans moved north- ward. This diverse set of habitats was eventually re- placed by closed tropical forest resulting in the last wave of xenarthran dispersers adapted to this closed forest habitat but also unable to disperse any farther north than its northernmost extent. To quote Graham (1992:126) “The earlier stepping stones were highly selective in terms of habitat requirements for potential migrants, and the present bridge is a very recent pathway for dispersal …” While his discussion centered on plants, the de- scription is equally applicable to xenarthran dispersal via the isthmus. During the time of the interchange, the 323 Figure 6. Shape of the premaxillae in four North American ground sloths. Nothrotheriops, Eremotherium and Megalonyx are interpreted as browsers, Paramylodon as a grazer or intermediate mixed feeder. MCDONALD: Xenarthrans and the Great American Biotic Interchange 324 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb creation of a diversity of habitats provided the opportu- nities for a member of each of the different xenarthran groups to disperse northward. Given the wide range of feeding strategies and habitat preference represented by the xenarthrans that participated in the interchange, the xenarthran diversity in each habitat was probably low with different habitat preferences for each species. For example, Williams (1960) described the Tertiary to- pography of the interchange corridor in Guatemala as consisting of meandering rivers on broad floodplains and locally shallow lakes, a description that matches closely the interpretation of Gillette and Ray (1981) that Glyptotherium lived in lake and stream communities near permanent bodies of water. However, this type of habitat may not have been equally suitable for the sloth Nothrotheriops, which seems to have utilized drier habi- tats. It is very likely that many of the habitat prefer- ences of xenarthrans eventually seen in faunas in North America may not have existed at the time of the inter- change and are secondary adaptations. The entrance of different xenarthrans into North America at different times, Glyptotherium with its preference for wetter habitat in the Blancan and the more arid adapted Nothrotheriops in the Irvingtonian reflects the dynam- ics of changing habitat in the Panamanian isthmus over time which contributed to it acting as a sieve. STARTING SMALL, GROWING LARGE Except for Dasypus bellus all of the late Pleistocene xenarthrans in North America are considered part of the megafauna, i.e. > 100 kg body mass and our view of these animals is generally one of large size. However as shown in Figure 5, the founding representative of each lineage in North America was smaller than later mem- bers. Previous discussions of trends for an increase in size includes Dasypus bellus (Klippel & Parmalee 1984), Holmesina floridanus to H. septentrionalis (Edmund 1987; Hulbert & Morgan 1993), the Megalonyx series of chronospecies (McDonald 1977), Nothrotheriops texanum to N. shastensis (McDonald 1995), “ Glossotherium” chapadmalense to Paramylodon harlani (McDonald 1995) and Thinobadistes segnis to T. wetzelli (Webb 1989). In order to calculate an estimate of the body mass for each species of North American xenarthran the length of the femur was used. The femur was selected since it is preserved for all taxa. Fariña and Vizcaino (1997) have shown that relationship of the femur length to body mass in the armadillos scales closely to that of general- ized mammals, while other bones such as the humerus scale differently because of its functional tie to digging. Choice of a single element reduces the error that might be generated by utilizing different skeletal elements in different taxa and at least makes the data more useful for comparison and identifying trends in size changes in the various lineages through time. Choice of this for- mula, (Formula F1 of Fariña et al. 1998 and Scott 1990) using the length of the femur between proximal and dis- tal articular surfaces to calculate body weight, also has the added advantage as it generates a body weight esti- mate that can be compared with those calculated by these workers for various South American late Pleis- tocene xenarthran taxa. The formula utilized is: log 10 mass = 3.485 x log 10 (femur length in cm) – 2.9112. The femora lengths used in calculating body weights are de- rived from personal measurements of specimens and from the literature. Whenever possible averages of speci- mens were used but in some cases only a single speci- men is available for some taxa for a particular land mam- mal age. The calculated weights of the various taxa listed in Table 1 provide a reasonable approximation for each species at different points in time. There have been few calculations of body weight for any of the North American xenarthrans although Jerison (1973) reported an estimated weight for Paramylodon harlani of 1100 kg, which is close to the value of 1392 kg derived from the formula used here. All of the xenarthrans show a roughly similar in- crease in size: Dasypus bellus 1.8, Holmesina 4.9, Pliometanastes to Megalonyx 5.9, Nothrotheriops 3, and Glossotherium-Paramylodon 4.5. The smallest increase is in Eremotherium with 1.5, a not unexpected value given the already large size of E. eomigrans. The only exception to this trend is Glyptotherium. While there is a 3.4 increase in size from the Blancan to Irvingtonian, the Rancholabrean species is smaller, just 0.71 times the size of its Irvingtonian predecessor. A decrease in size has been documented in other Pleis- tocene lineages of both herbivores such as Bison (Wil- son 1992) and carnivores such as jaguars (Seymour 1993) and it appears that perhaps Glyptotherium in a similar way was responding to the dramatic environmental changes taking place in North America during the Rancholabrean. Hulbert and Morgan (1993) in their study of the evolution of Holmesina in Florida noted that the rate of change in this lineage was variable, with rates being more 325 rapid, by a factor of 3 to 10 times, during some intervals than others. They observed that over a 2.3 million year interval the evolutionary rate of this lineage was two to three times greater than the “normal” evolutionary rate of 0.06 to 0.07 d calculated by Gingerich (1983). While similar studies have not been performed on any other North American xenarthrans and would be instructive, the study by Hulbert and Morgan does indicate that fol- lowing its appearance in North America Holmesina un- derwent a rapid evolutionary response, as indicated by the increase in body size. As they noted, what was unusual about the evolution of Holmesina was not the rate of size increase, but rather that it was sustained almost continuously for such a long interval. This rapid evolutionary rate suggests that there were few ecologi- cal constraints on this lineage after entering North America and there existed only limited competition, if any, with any of the endemic mammals. That the lin- eage was able to evolve rapidly in size suggests the ab- sence of competition that previously existed with related species in South America. Marshall and Hecht (1978) described the North American ungulates as insinuating themselves into the South American biota, but this term seems equally applicable to all of the xenarthrans that entered North America and became successfully inte- grated into the fauna. As a result of this increase in body size over time the various species of endemic mammals with which each xenarthran lineage would have ecologically com- peted would have changed. This would have been par- ticularly true of the herbivorous sloths, pampatheres, and glyptodonts. One consequence resulting from a change in body size of the herbivorous xenarthrans would have been their feeding strategies. Smaller herbivores tend Taxon North American Length of Calculated Body Land Mammal Age Femur in cm Weight in kg Dasypus bellus Blancan 13.72 10.07 Dasypus bellus Rancholabrean 15.6 17.68 Holmesina floridanus Blancan 20.3 44.27 Holmesina septentrionalis Irvingtonian 25 91.48 Holmesina septentrionalis Rancholabrean 33.2 216.27 Glyptotherium texanum Blancan 32.6 230.74 Glyptotherium arizonae Irvingtonian 46.4 789.68 Glyptotherium floridanum Rancholabrean 42.1 562.65 Pachyarmatherium leiseyi Irvingtonian 15.0 15.42 Pliometanastes protistus Hemphillian 30.6 185.05 Megalonyx curvidens Hemphillian 30.6 185.05 Megalonyx leptostomus Blancan 39.3 442.64 Megalonyx wheatleyi Irvingtonian 42.8 595.93 Megalonyx jeffersonii Rancholabrean 50.9 1090.35 Nothrotheriops texanum Irvingtonian 35.9 322.91 Nothrotheriops shastensis Rancholabrean 39.8 462.59 Eremotherium eomigrans Blancan 65.2 2584.40 Eremotherium laurillardi Irvingtonian 69 3148.54 Eremotherium laurillardi Rancholabrean 73.7 3961.48 Thinobadistes segnis Hemphillian 43.8 645.89 “Glossotherium”chapadmalense Blancan 35.5 310.54 Paramylodon harlani Irvingtonian 48.4 914.81 Paramylodon harlani Rancholabrean 54.6 1392.47 Table 1. North American xenarthrans from each North American Land Mammal Age and their calculated body weights. MCDONALD: Xenarthrans and the Great American Biotic Interchange 326 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb to be more selective in their feeding especially with re- gard to consuming those parts of the plant with higher nutritive value and least defended, while large herbivores tend to consume a wider selection of plant parts with varying nutritive value (Guthrie 1984). The ability of these relatively small herbivorous xenarthrans to disperse may have been facilitated by their ability to utilize the parts of plants with the highest nutritional value present in the more tropical parts of the corridor. As they in- creased in size, their ability to compete with the native herbivores and success in becoming integrated into the North American fauna may have been facilitated by being able to utilize a wider variety of plant parts with rela- tively lower nutritional value which would have been facilitated by their low basal metabolism. Their success in temperate North America with increased seasonality and a restricted seasonal flush of nutrients during a more limited growing season may have been a result of the combination of their inherent digestive capability result- ing from an extended digestive period and better ab- sorption of nutrients which would have increased along with their increase in body size which would have in- creased the transit time in the intestines. INTEGRATION BUT NOT RADIATION As noted above, generally only one representative of each major xenarthran group (essentially equivalent to the family) dispersed into North America. While most of the xenarthrans that dispersed into North America can be considered “successful” in the sense that many survived until the Pleistocene extinctions, this is not true for all. There are some taxa that following their dis- persal northward did not become established and are present in North America for a very limited time. The first mylodont to enter North America in the Hemphillian, Thinobadistes, appears to have become extinct by the end of that land mammal age (Webb 1989). The strange cingulate Pachyarmatherium is known from the late Blancan through the middle Irvingtonian (Downing & White 1995) and the only record of the giant anteater, Myrmecophaga tridactyla, is known from a single Irvingtonian locality (Shaw & McDonald 1987). While those xenarthrans that did survive filled a niche that per- mitted them to become integrated into the North Ameri- can mammal fauna and survive until the Pleistocene extinction event, in none of these groups is there any type of radiation or diversification similar to what oc- curred in South America with many of the northern in- vaders. While there are evolutionary changes in each of the lineages after their appearance in North America there are few that show any major modifications of the basic bauplan. Each lineage consists of a series of chronospecies distinguished by various morphological characters or a difference in size but none greatly de- part from the ancestral morphology. The modification of the caniniform teeth from Pliometanastes to Megalonyx reflects one of the more radical morpho- logical changes in any of the North American xenarthrans. The occlusal surfaces of the caniniforms in Pliometanastes form at an oblique angle to the axis of the tooth so that occlusion produced a shearing of vegetation. This pattern of occlusion is present in the earliest megalonychids and is retained to some degree in most of the later members of the family. Megalonyx departs from this pattern in that the occlusal surface of the caniniforms is perpendicular to the axis of the tooth so that they functioned to crush or hold rather than shear. As the caniniforms in Megalonyx become greatly en- larged there is a concurrent reduction in the size of the mandibular spout (Fig. 7). While this morphology in Megalonyx is specialized relative to other megalonychids, it did not develop until after the lineage was present in North America. The morphology of the caniniform in Pliometanastes, the ancestor of Megalonyx, is more similar to that of the basic mor- phology seen in other megalonychids and presumably more generalized ecological specializations compared to Megalonyx. In the Glossotherium-Paramylodon lineage one striking evolutionary change is the reduction in the size of the anteriormost upper teeth or caniniforms (McDonald 1995). There are two morphologies to the occlusal surface of the tooth in all members of the lin- eage from the Blancan to the Rancholabrean and in the earlier members there is also a distinct size difference between the two morphs. In Rancholabrean P. harlani the two morphs are still present but there is no longer any distinction in size and in many individuals the caniniforms were never present. The two tooth morphs are associated with two skull morphs and the inference is that this represents sexual dimorphism in the species. It would appear that there is a decrease in sexual dimor- phism through time in this mylodont lineage. Like the other North American xenarthrans, the mylodonts dis- played evolutionary changes within the lineage but no morphological innovation that allowed them to diversify. Within the Holmesina lineage numerous skeletal and dental modifications have been described that per- mit the earliest and latest species to be readily distin- guished (Edmund 1987; Hulbert & Morgan 1993), yet 327 as in other xenarthrans there are no new modifications or real taxonomic diversity, merely a series of chronospecies. All in all each lineage of North American xenarthrans remains rather conservative in its morphol- ogy. As discussed above the one trait they all have in common is a general increase in body size. DISTRIBUTION – OPPORTUNITIES AND LIMITS The only extant xenarthran in temperate North America is the armadillo, Dasypus novemcinctus, whose pres- ence is essentially a Holocene event. While this species is the most widespread of all living armadillos, extending as far south as southern Brazil, the northern edge of its range in the United States is limited and essentially con- fined to the Gulf Coast and southeastern United States, although on the Great Plains the species has dispersed as far north as Nebraska. There is a strong similarity in its distribution and that of two of its extinct cingulate relatives in North America, Dasypus bellus and Holmesina spp., and the sloth, Eremotherium. This list can be somewhat expanded if Glyptotherium is included. While the Blancan records of Glyptotherium extend Figure 7. Comparison of the caniniform and shape of the mandibular spout in Pliometanastes and Megalonyx. MCDONALD: Xenarthrans and the Great American Biotic Interchange 328 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb westward to southern Arizona, there is a progressive contraction of the range of the genus to the Gulf Coast and southeastern United States and by the Rancholabrean its distribution roughly coincides with that of Dasypus novemcinctus today. These patterns of distribution gen- erally correspond to North American patterns of pre- cipitation in which seasonally the greatest amount of rainfall in the United States is in the southeastern part of the continent (World Meteorological Organization [F. Steinhauser, Tech. Super] 1979). Gillette and Ray (1981) described the habitat requirements of Glyptotherium as restrictive, requiring proximity to standing water in a region with extensive lowland terrain, lush tropical or subtropical vegetation, a warm climate without exces- sive extremes of temperature, and relatively high con- stant moisture. The last factor, which is dependent on the annual precipitation, was needed to maintain the per- manent bodies of water and abundant vegetation. While these factors may have limited the northward expansion of Glyptotherium in temperate North America, none of these would have necessarily been a limiting factor to the northward dispersal through the tropics of the earli- est members of this lineage. A long-term drying trend from the Blancan to Rancholabrean resulting in increas- ing aridity in the southwest (Thompson 1991) may ac- count for the contraction of the range of Glyptotherium eastward and the concurrent westward contraction of the range of Nothrotheriops. In contrast to Glyptotherium, the ground sloth Nothrotheriops, contracts its range in the opposite di- rection. Initially widespread in the Irvingtonian from coast to coast by the Rancholabrean the genus is re- stricted to the western United States with the largest number of occurrences in the southwest (Akersten & McDonald 1991). The single record of giant anteater from El Golfo, Mexico also indicates a maximum northward expansion of the range of this species in the Irvingtonian (Shaw & McDonald 1987) followed by a retraction to its present northern limit in southern Guatemala. As with the changes in distribution of Glyptotherium and Nothrotheriops this probably reflects the environmen- tal change in North America that took place from the Irvingtonian to the Rancholabrean. The impact on the distribution of Myrmecophaga due to climatic change may not have been direct, but rather was a secondary effect, the primary impact being on the animal’s food source, colonial termites and ants. Not all of the North America xenarthrans were as restricted in their distribution. By the Rancholabrean, the oldest lineage of sloths in North America, Megalonyx, is also the most widespread and ranged above the Arc- tic Circle (McDonald et al. 2000) to both coasts and into southern Mexico (McDonald 2002). While not quite as widespread as Megalonyx, the mylodont Paramylodon is known from as far north as Washington and Montana, from coast to coast and south into Mexico. The earliest records of both lineages, presumably recording their appearance shortly after dispersal into North America are limited to the southern United States and Mexico but seem to be followed by a fairly constant expansion of their respective ranges across the continent during the Irvingtonian and Rancholabrean. THE FORGOTTEN DISPERSERS – XENARTHRANS IN CENTRAL AMERICA Webb (1985) identified two principle phases in the GABI. The first phase includes those animals that entered ei- ther continent prior to the establishment of the land bridge. The earliest arrivals from South America, two sloths in the Hemphillian of North America were termed “her- alds of the south”; this was followed by a large set of bridge-crossers, which he called “legions”. Both of these groups dispersed northward at a time when the Pana- manian land bridge contained a mosaic of habitats that preceded the establishment of the tropical lowland for- est that exists today. There is however a third set of dispersers including a number of xenarthrans present in Central America today that probably entered the region after the establishment of the tropical lowland forest, a habitat to which they are restricted. This stage is iden- tified in Webb (1985:fig. 1) as the final tropical mingling. In keeping with Webb’s terminology of “heralds” and “legions” I propose that this last group be called “camp followers”. Unlike the earlier xenarthran dispersers, which utilized a variety of habitats in the interchange corridor, most members of this last suite are closely as- sociated with a specific habitat, tropical rainforest. While the focus tends to be on those xenarthrans (and other dispersers of South American origin) that were able to disperse to temperate latitudes in North America, there is a subset of dispersing xenarthrans and other South American mammals restricted to the tropi- cal portions of North and Central America. Some are extinct, such as the toxodont, Mixotoxodon. However, with regard to the xenarthrans, unlike their temperate counterparts that became extinct, many are still present in Central America and comprise a significant part of the fauna. These include the anteaters, Myrmecophaga tridactyla, Tamandua mexicana, and Cyclopes didactylus, the tree sloths Bradypus variegatus and 329 Choloepus hoffmanni and the armadillos, Cabassous centralis and Dasypus novemcinctus. Except for D. novemcinctus, which as a generalist has the widest dis- tribution of any living xenarthran, these xenarthrans are restricted to a warmer tropical habitat, which has lim- ited their northward expansion. Unfortunately the fossil record of Central America is poorly documented and none of these taxa including D. novemcinctus is known from fossils, so we have no idea how long any of them have been in the region. While it is assumed that they are relatively new arrivals and most likely did not move northward until after the formation of the tropical for- ests, it is possible they may have a long established his- tory in the northern Neotropics. However, what their modern presence does indicate is that the dispersal of xenarthrans out of South America is richer than the tra- ditional view of only looking at those forms that made it into the North American temperate zones. Just as there are modern xenarthrans unable to disperse farther north than the tropics so the northern part of their range is restricted to Central America, such seems to have also been the case for some extinct taxa as well. The fossil record of xenarthrans in Central America is limited and most records are from the wide- spread Eremotherium laurillardi (Cartelle & DeIuliis 1995). One good example of the potential richness of the diversity of the xenarthran dispersal limited to tropi- cal North and Central America is the endemic megalonychid sloth genus, Meizonyx from El Salvador (Webb & Perrigo 1985). This large species is known from a single specimen but clearly indicates that along with those taxa that were able to disperse farther north, there are taxa that participated in the interchange that are restricted to the tropics. It is very likely that contin- ued fieldwork in Central America will result in the dis- covery of other extinct xenarthra that were part of the GABI but because of their ecology did not disperse north of the tropics. SUMMARY While topography and ecology have been two major fac- tors that have been discussed with regard to the GABI, another factor that needs to be examined in more detail is climate at the time of the interchange. Thompson (1991) reviewed evidence that indicates that during the early and middle Pliocene (prior to ~ 2.4 Ma, the time of appearance of Webb’s legion from the south) the cli- mate in the western United States was less seasonable (more equable) and generally more humid and the ter- restrial record parallels that seen in oxygen isotope records in the North Pacific. While these factors may not have been quite as critical to other mammals, this may be a much more critical factor to xenarthrans with their overall lower basal metabolism (McNab 1985). Climate during the interchange was more equable (Thompson 1991) and upon their initial appearance in temperate North America many xenarthran taxa became widely distributed. The increase in seasonality during the Pleistocene resulted in major shifts in the range, of- ten a reduction for some taxa such as Nothrotheriops and Glyptotherium. Their sensitivity to temperature reflects their low basal rates of metabolism and a high conductance resulting in a low body temperatures and a high lower limit of thermoneutrality at given body mass (McNab 1985). This thermal sensitivity may also be an important factor with regard to the trend of increase in overall body size in all of the North American xenarthrans. An increase in overall size would have provided greater thermal inertia. This was probably an important factor particularly for the cingulates with their general lack of hair as insulation. It would seem that this group would be most susceptible to climate, particu- larly seasonal temperature extremes and this is reflected in their more limited distribution in North America com- pared to sloths. Temperature (and possibly humidity) also seems to be a primary factor regarding the distribu- tion of two sloths, Nothrotheriops and Eremotherium, and the anteater, Myrmecophaga. As the largest of the sloths, it would seem that Eremotherium would pos- sess the greatest amount of thermal inertia and thus have a greater tolerance to colder temperatures. Two other sloths, Megalonyx and Paramylodon, do not seem to have been affected by this environmental change and seemed to have been widespread until the time of their extinction. The range of both genera extended farther north than any of the other xenarthrans, with Megalonyx crossing the Arctic Circle (McDonald et al. 2000). This may have been a consequence of their larger size and subsequently better thermal inertia and by the end of the Pleistocene both genera are exceeded only by Eremotherium in size. It is possible both may have de- veloped better insulation or even had higher basal me- tabolisms but we have no direct evidence for either of these adaptations. If temperature was not a limiting factor in the continued northward dispersal in Eremotherium then the availability of food may have been the limiting factor. While Eremotherium laurillardi is widely distributed (Cartelle & De Iuliis 1995), its dis- tribution is centered in tropical South America and only extends to the subtropics in South America and as a MCDONALD: Xenarthrans and the Great American Biotic Interchange 330 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb general rule to the subtropics of North America. The distribution of Eremotherium may have therefore been limited secondarily by preferred vegetation found in tropi- cal to semi-tropical rainforests and not by temperature. While this paper attempts to review what is known of the paleoecology of fossil xenarthrans as a means to better understand their success as members of the GABI, much more data is needed. The data currently available ranges from very comprehensive such as for the sloth Nothrotheriops, for which we have dung and detailed knowledge of its diet to the small cingulate, Pachyarmatherium, which consists of a few meager specimens in the fossil record. Despite this spotty data- base there do seem to be a few common features shared by those xenarthrans that participated in the GABI and established themselves as members of the North Ameri- can mammalian fauna. 1. Despite the potential variety of xenarthrans in South America only one member of each major xenarthran group (roughly equivalent to a family) participated in the faunal interchange and dis- persed northward into temperate North America at any particular phase of the GABI. 2. Each of these dispersers represent an ecological adaptation distinct from all other xenarthran dis- persers so there does not appear to have been any ecological competition between any members of a particular xenarthran cohort participating at any phase of the interchange. This includes a variety of feeding strategies including omnivore/ insectivore – Dasypus bellus; specialized myrmecophage – Myrmecophaga tridactyla; a variety of herbivores including two low browsers – Nothrotheriops and the Pliometanastes/ Megalonyx lineage, each with different skull morphologies indicating different feeding strate- gies; a high browser – Eremotherium; an aquatic grazer – Glyptotherium; an intermediate feeder on coarse vegetation/ grazer – Holmesina and a large grazer/ intermediate feeder – Glossotherium/Paramylodon lineage. There is insufficient data on the small cingulate, Pachyarmatherium, to make any inferences as to its feeding strategy or ecology. Given the vari- ety of habitats present in the corridor used by these dispersers, each was utilizing different habitats with probably limited overlap. Another aspect of the xenarthrans participating in the GABI is dis- tinct body sizes with very little overlap unless they have different feeding strategies i.e. by the late Pleistocene, Megalonyx, a browser is similar in body size to Paramylodon, a grazer or interme- diate mixed feeder. 3. Ecologically the dispersing member was more of a generalist than other members of the family. This may include morphological features that lack some of the extreme specializations seen in other members of the family or by having a more gen- eralized ecology or habitat preference. Any spe- cialization seems to have evolved after the lin- eage entered North America. 4. Dispersing members tended to be small with a subsequent increase in size following their entry into North America. This size increase may indi- cate they lacked any ecological competitors among the native North American mammals and thus filled an empty niche or alternatively may merely indicate a response to cooling temperatures and greater seasonality in a temperate environment and provided thermal inertial to compensate for a low basal metabolism and high thermal conduc- tance. 5. While evolutionary change did take place in the North American xenarthrans after they became integrated into the North American mammalian fauna, they did not diversify or undergo any adap- tive radiation. They seem to have avoided eco- logical competition with endemic members of the North American fauna but the ecological niche they filled did not permit any diversification. ACKNOWLEDGEMENTS At the time I went to Florida to study with Dave Webb for my masters thesis, my interest was confined to ground sloths. Thanks to both his mentoring and through dis- cussions and his publications, my horizons were broad- ened to consider the broader implications of the role of xenarthrans within the scope of the Great American Biotic Interchange. The development of the ideas pre- sented here has not been one of punctuated equilibrium but rather gradualistic and have evolved over many years. An initial version of this paper was first presented in a talk entitled “Gallavanting Gravigrades and Sedentary Sloths: Ecological Adaptations of Xenarthrans and the Great American Faunal Interchange” presented at the 331 Fifth International Theriological Congress in Rome, Italy in August, 1989 at a symposium organized by Dave Webb and Rosendo Pascual entitled Late Cenozoic Mammals: Dispersal Between Americas. During the intervening years many of the initial ideas have been refined and it is with great pleasure that I contribute this paper on a topic that has been of major interest to my professor, Dave Webb. For better or for worse, the ideas devel- oped here are a result of his influence. I would like to extend my appreciation to the two anonymous review- ers for their insightful comments. LITERATURE CITED Akersten, W. A., & H. G. McDonald. 1991. Nothrotheriops (Xenarthra) from the Pleistocene of Oklahoma and pa- leogeography of the genus. The Southwestern Natural- ist, 36(2):178-185. Aramayo, S. A., & T. Manera de Bianco. 1996. Edad y nuevos hallazgos de icnita de mamiferos y aves en el yacimiento paleoicnologico de Pehuen-Co (Pleistoceno tardio), Provincia de Buenos Aires, Argentina. Asociación Paleontólogica Argentina, Publicación Especial 4:47-57. Auffenberg, W. 1957. A note on an unusually complete skel- eton of Dasypus bellus (Simpson) from Florida. Quar- terly Journal of the Florida Academy of Science, 20(4):233-237. Bauchop, T. 1978. Digestion of leaves in vertebrate arboreal folivores. Pp. 193-204 in G. G. Montgomery, ed. The Ecology of Arboreal Folivores. Smithsonian Institution Press, Washington, D.C. Britton, W. S. 1941. Form and function in the sloth. Quarterly Review of Biology, 16:190-207. Cartelle, C., & G. DeIuliis, G. 1995. Eremotherium laurillardi: the panamerican late Pleistocene megatheriid sloth. Jour- nal of Vertebrate Paleontology, 15(4):830-841. Coltrain, J. B., J. M. Harris, T. E. Cerling, J. R. Ehleringer, M.-D. Dearing, J. Ward, & J. Allen. 2004. Rancho La Brea stable isotope biogeochemistry and its implications for the palaeoecology of late Pleistocene coastal southern California. Palaeogeography, Palaeoclimatology, Palaeoecology, 205(3-4):199-219. Dompierre, H., & C. S. Churcher. 1996. Premaxillary shape as an indicator of the diet of seven extinct late Cenozoic New World camels. Journal of Vertebrate Paleontology, 16(1):141-148. Downing, K. F., & R. White. 1995. The cingulates (Xenarthra) of Leisey Shell Pit 1A (Irvingtonian), Hillsborough County, Florida. Bulletin of the Florida Museum of Natu- ral History, 39(12):375-396. Edmund, A. G. 1987. Evolution of the genus Holmesina (Pampatheriidae, Mammalia) in Florida, with remarks on taxonomy and distribution. Pearce-Sellards Series, Texas Memorial Museum, 45:1-20. Eisenberg, J. F., & K. H. Redford. 1999. Mammals of the Neotropics, The Central Neotropics, Vol. 3 Ecuador, Peru, Bolivia, Brazil. The University of Chicago Press, Chi- cago, 624 p. Englemann, G.F. 1985. The phylogeny of the Xenarthra. Pp. 51-64 in G. G. Montgomery, ed. The Evolution and Ecol- ogy of Armadillos, Sloths, and Vermilinguas. Smithsonian Institution Press, Washington, D.C. Fariña, R. A. 1985. Some functional aspects of mastication in Glyptodontidae. Fortschritte der Zoologie, 30:277-280. Fariña, R. A. 1988. Observaciones adicionales sobre la biomecánica masticatoria en Glyptodontidae. Boletín de la Sociedad Zoológica del Uruguay, 4:5-9. Fariña, R. A. 1995. Limb bone strength and habits in large glyptodonts. Lethaia, 28(3):189-196. Fariña, R. A., & S. F. Vizcaino. 1997. Allometry of the bones of living and extinct armadillos (Xenarthra, Dasypoda). Zeitschrift für Saugetierekunde, 62:65-70. Fariña, R. A., S. F. Vizcaino, & M. S. Bargo. 1998. Body mass estimations in Lujanian (Late Pleistocene-Early Holocene of South America) mammal megafauna. Mastozoologia Neotropical, 5(2):87-108. Ferigolo, J. 1985. Evolutionary trends of the histological pattern in the teeth of Edentata (Xenarthra). Archives of Oral Biology, 30:71-82. Gillette, D. D., & C. E. Ray. 1981. Glyptodonts of North America. Smithsonian Contributions to Paleobiology, 40:1-255. Gingerich, P. D. 1983. Rates of evolution: effects of time and temporal scaling. Science, 222:159-161. Goffart, M. 1971. Function and form in the sloth. Pergamon Press Inc. New York, 225 p. Graham, A. 1992. Utilization of the isthmian land bridge dur- ing the Cenozoic –paleobotanical evidence for timing, and the selective influence of altitudes and climate. Re- view of Paleobotany and Palynology, 72(1-2):119-128. Grassé, P.-P. 1955. Ordre des Edentes (Edentata Cuvier, 1798; Edentati Vicq D’Azyr, 1792; Bruta Linne, 1758), formes actuelles. Pp. 1182-1246 in P.-P. Grassé ed. Traite de Zoologie: anatomie, systematique, biologie mammiferes, Tome 17, fascicule II, Masson et Cie, Paris Guthrie, R. D. 1984. Mosaics, allochemics and nutrients an ecological theory of Late Pleistocene megafaunal extinc- tions. Pp. 259-298 in P. S. Martin & R. G. Klein, eds. Quaternary Extinctions: a Prehistoric Revolution. The University of Arizona Press, Tucson. Hansen, R. M. 1978. Shasta ground sloth food habits, Ram- part Cave, Arizona. Paleobiology, 4(3):302-319. Hirschfeld, S. E., & S. D. Webb. 1968. Plio-Pleistocene megalonychid sloths of North America. Bulletin of the Florida State Museum Biological Sciences, 12(5):213-296. Ho, T-Y. 1967. Relationship between amino acid contents of mammalian bone collagen and body temperature as a basis for estimation of body temperature of prehistoric mammals. Comparative Biochemistry and Physiology, 22:113-119. MCDONALD: Xenarthrans and the Great American Biotic Interchange 332 CENOZOIC VERTEBRATES: Papers to Honor S. David Webb Hulbert, R. C., Jr., & G. S. Morgan. 1993. Quantitative and qualitative evolution in the giant armadillo Holmesina (Edentata: Pampatheriidae) in Florida. Pp. 134-177 in R. A. Martin & A. D. Barnosky, eds. Morphological Change in Quaternary Mammals of North America. Cambridge University Press, Cambridge. Janis, C. 1976. The evolutionary strategy of the Equidae and the origin of rumen and caecal digestion. Evolution, 30(4):757-774. Janis, C., & D. Ehrhardt. 1988. Correlation of relative muzzle width and relative incisor width with dietary preference in ungulates. Zoological Journal of the Linnean Society, 92(3):267-284. Janzen, D. H., & P. S. Martin. 1982. Neotropical anachro- nisms: the fruits the gomphotheres ate. Science, 215:19- 27. Jerison, H. J. 1973. The evolution of the brain and intelli- gence. Academic Press, New York, 482 p. Klippel, W., & P. W. Parmalee. 1984. Armadillos in North American late Pleistocene contexts. Carnegie Museum of Natural History Special Publication, 8:149-160. Laudermilk, J. D., & P. A. Munz. 1934. Plants in the dung of Nothrotherium from Gypsum Cave, Nevada. Carnegie Institution of Washington Publication, 453:29-37. Laudermilk, J. D., & P. A. Munz. 1938. Plants in the dung of Nothrotherium from Rampart and Muav Caves, Arizona. Carnegie Institution of Washington Publication, 487:271- 281. MacFadden, B. J., Y. Wang, T. E. Cerling, & F. Anaya. 1994. South American fossil mammals and carbon isotopes: a 25 million-year sequence from the Bolivian Andes. Palaeogeography, Palaeoclimatology, Palaeoecology, 107(3-4):257-268. Marshall, L. D. 1981. The Great American Interchange – an invasion induced crisis for South American mammals. Pp. 133-229 in M. Nitecki, ed. Biotic Crises in Ecological and Evolutionary Time. Academic Press, New York. Marshall, L. D., & M. K. Hecht. 1978. Mammalian faunal dynamics of the Great American Interchange: an alterna- tive interpretation. Paleobiology, 4(2):203-209. Martin, P. S., B. E. Sabels, & D. Shutler Jr. 1961. Rampart Cave coprolite and ecology of the Shasta ground sloth. Ameri- can Journal of Science, 259:102-127. McDonald, H.G. 1977. Description of the osteology of the extinct gravigrade edentate, Megalonyx, with observa- tions on its ontogeny, phylogeny and functional anatomy. M.S. Thesis, University of Florida, Gainesville, 328 p. McDonald, H.G. 1989. Gallivanting gravigrades and seden- tary sloths: ecological adaptations of Xenarthrans and the Great American Faunal Interchange. Fifth Interna- tional Theriological Congress 1:273-274. McDonald, H. G. 1995. Gravigrade xenarthrans from the middle Pleistocene Leisey Shell Pit 1A, Hillsborough County, Florida. Bulletin of the Florida Museum of Natural His- tory Biological Sciences, 37(11):345-373. McDonald, H. G. 1996. Biogeography and paleoecology of ground sloths in California, Arizona and Nevada. San Bernardino County Museum Association Quarterly, 43(1):61-65. McDonald, H. G. 2002. Fossil Xenarthra of Mexico: a review. Pp. 227-248 in M. Montellano Ballesteros and J. Arroyo Cabrales, eds. Avances en los estudios paleomastozoológicos en México. Serie Arqueología, Instituto Nacional de Antropología e Historia. McDonald, H. G., C. R. Harington, & G. De Iuliis. 2000. The ground sloth, Megalonyx, from Pleistocene deposits of the Old Crow Basin, Yukon, Canada. Arctic, 53(3):213- 220. McNab, B. K. 1978. Energetics of arboreal folivores: physi- ological problems and ecological consequences of feed- ing on an ubiquitous food supply. Pp. 153-162 in G. G. Montgomery, ed. The Ecology of Arboreal Folivores. Smithsonian Institution Press, Washington, D.C. McNab, B. K. 1985. Energetics, population biology, and dis- tribution of xenarthrans, living and extinct. Pp. 219-232 in G. G. Montgomery, ed. The Evolution and Ecology of Armadillos, Sloths and Vermilinguas. Smithsonian In- stitution Press, Washington, D.C. Moore, D. M. 1978. Post-glacial vegetation in the south Patagonian territory of the giant ground sloth, Mylodon. Botanical Journal of the Linnean Society, 77(3):177-202. Murie, J. 1872. On the habits, structure, and relationships of the Three-banded Armadillo (Tolypeutes conurus, Is. Geoff.) Transactions of the Linnean Society, 30:71-132. Naples, V. L. 1989. The feeding mechanism in the Pleistocene ground sloth, Glossotherium. Natural History Museum of Los Angeles County, Contributions in Science, 415:1- 23. Parra, R. 1978. Comparison of foregut and hindgut fermenta- tion in herbivores. Pp. 205-229 in G. G. Montgomery, ed. The Ecology of Arboreal Folivores. Smithsonian Insti- tution Press, Washington, D.C. Redford, K. H. 1985. Food habits of armadillos (Xenarthra: Dasypodidae). Pp. 429-437 in G. G. Montgomery, ed. The Evolution and Ecology of Armadillos, Sloths and Vermilinguas. Smithsonian Institution Press. Washing- ton, D.C. Redford, K. H. 1986. Dietary specialization and variation in two mammalian myrmecophages (variation in mamma- lian myrmecophagy). Revista Chilena de Historia Natu- ral, 59:201-208. Scott, K. 1990. Postcranial dimensions of ungulates as pre- dictors of body mass. Pp. 301-355 in J. Damuth and B. J. MacFadden, eds. Body Size in Mammalian Paleobiology: Estimation and Biological Implications. Cambridge Uni- versity Press, Cambridge. Seymour, K. L. 1993. Size change in North American Quater- nary jaguars. Pp. 343-372 in R. A. Martin and A. D. Barnosky, eds. Morphological Change in Quaternary 333 Mammals of North America. Cambridge University Press, Cambridge. Shaw, C. A., & H. G. McDonald. 1987. First record of giant anteater (Xenarthra, Myrmecophagidae) in North America. Science, 236(4798):186-188. Smith, K. K., & K. H. Redford. 1990. The anatomy and func- tion of the feeding apparatus in two armadillos (Dasypoda): anatomy is not destiny. Journal of Zool- ogy, London, 222:27-47. Solounias, N., M. Teaford, & A. Walker. 1988. Interpreting the diet of extinct ruminants: the case of the non-browsing giraffid. Paleobiology, 14(3):287-300. Stehil, F. G., & S. D. Webb, eds. 1985. The Great American Biotic Interchange. Topics in Geobiology Vol. 4. Plenum Press, New York, 532 p. Thompson, R. S. 1991. Pliocene environments and climates in the western United States. Quaternary Science Reviews, 10:115-132. Thompson, R. S., T. R. Van Devender, P. S. Martin, T. Foppe, & A. Long. 1980. Shasta ground sloth (Nothrotheriops shastense Hoffstetter) at Shelter Cave, New Mexico: en- vironment, diet, and extinction. Quaternary Research, 14(3):360-376. Vizcaino, S. F., G. De Iuliis, & M. S. Bargo. 1998. Skull shape, masticatory apparatus, and diet of Vassallia and Holmesina (Mammalia: Xenarthra: Pampatheriidae): When anatomy constrains destiny. Journal of Mamma- lian Evolution, 5(4):291-322. Webb, S. D. 1976. Mammalian faunal dynamics of the great American interchange. Paleobiology, 2(3):220-234. Webb, S. D. 1978. A history of savanna vertebrates in the New World. Part II: South America and the Great Inter- change. Annual Review of Ecology and Systematics, 9:393-426. Webb, S. D. 1985. Late Cenozoic mammal dispersals between the Americas. Pp. 357-386 in F. G. Stehli and S. D. Webb, eds. The Great American Biotic Interchange. Topics in Geobiology Vol. 4. Plenum Press, New York. Webb, S. D. 1989. Osteology and relationships of Thinobadistes segnis, the first mylodont sloth in North America. Pp. 469-532 in K. H. Redford and J. F. Eisenberg, eds. Advances in Neotropical Mammalogy. Sandhill Crane Press, Gainesville. Webb, S. D. 1991. Ecogeography and the Great American Interchange. Paleobiology, 17(3):266-280. Webb, S. D., & S. Perrigo. 1985. New megalonychid sloths from El Salvador. Pp. 113-120 in G. G. Montgomery, ed. The Evolution and Ecology of Armadillos, Sloths, and Vermilinguas. Smithsonian Institution Press, Washing- ton, D.C. Wetzel, R. M., & E. Mondolfi. 1979. The subgenera and species of long-nosed armadillos, genus Dasypus L. Pp. 43-63 in J. F. Eisenberg, ed. Vertebrate Ecology in the Northern Neotropics. Smithsonian Institution Press, Washington, D.C. Williams, H. 1960. Volcanic history of the Guatemalan high- lands. University of California Publications, Geological Sciences, 38:1-86. Wilson, M. C. 1992. Bison in Alberta. Pp. 1-17 in J. Foster, D. Harrison & I. S. MacLauren, eds. Buffalo. University of Alberta Press, Edmonton. World Meteorological Organization [F. Steinhauser, Tech. Super]. 1979. Climatic Atlas of North and Central America – Maps of Mean Temperature and Precipitation. World Meteorological Organization, Geneva. MCDONALD: Xenarthrans and the Great American Biotic Interchange