the great american biotic interchange: patterns and...

The Great American Biotic Interchange: Patterns and ProcessesAuthor(s): S. David WebbSource: Annals of the Missouri Botanical Garden, Vol. 93, No. 2 (Aug., 2006), pp. 245-257Published by: Missouri Botanical Garden PressStable URL: http://www.jstor.org/stable/40035724 .

Accessed: 08/04/2014 23:14

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

Missouri Botanical Garden Press is collaborating with JSTOR to digitize, preserve and extend access toAnnals of the Missouri Botanical Garden.

http://www.jstor.org

This content downloaded from 137.111.226.20 on Tue, 8 Apr 2014 23:14:05 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=mobot

http://www.jstor.org/stable/40035724?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


THE GREAT AMERICAN BIOTIC INTERCHANGE: PATTERNS AND PROCESSES1

S. David Webb2

Abstract

When the Panamanian land bridge was emplaced about 2.7 Ma, it triggered the Great American Biotic Interchange (GABI), a major mingling of land mammal faunas between North and South America. Four families of northern immigrants (Procyonidae, Felidae, Tayassuidae, and Camelidae) diversified at moderate rates, while four others, Canidae, Mustelidae, Cervidae, and especially Muridae, evolved explosively. As a consequence, half of living South American genera are descendants of northern immigrants. The other major consequence of the interchange was the conquest of tropical North America by immigrants from Amazonia, an episode that justifies the term Neotropical Realm.

Key words: biostratigraphy, evolutionary rates, isthmian land bridge, mammals, Neotropical Realm.

Resumen

Cuando el corredor terrestre panameno se ubico hace cerca de 2.7 Ma, desencadeno el Gran Intercambio Biotico Americano (GABI), una mezcla importante de las faunas de mamiferos terrestres entre America del Norte y del Sur. Cuatro familias de inmigrantes del norte (Procyonidae, Felidae, Tayassuidae, and Camelidae) se diversificaron a un ritmo moderado, mientras que otras cuatro, Canidae, Mustelidae, Cervidae y especialmente Muridae, evolucionaron en forma explosiva. Por consiguiente, la mitad de generos suramericanos actuales son descendientes de inmigrantes del norte. La otra principal consecuencia del intercambio fue la conquista de Norteamerica tropical por inmigrantes amazonicos, un episodio que justifica el termino de Reino Neotropical.

The roots of the Latin American biota can be traced down through the epochs of earth history to several formative chapters. During most of the Mesozoic Era, the South American biota was an integral part of the Gondwanan biota. Through most of the Cenozoic Era, the terrestrial and freshwater biotas of South America were isolated from those of North America and all other continents by deep ocean troughs. The most recent chapter begins with the Great American Biotic

Interchange (GABI). This occurred approximately three million years ago during the late Pliocene, when the powerful forces of plate tectonics raised the isthmian land bridge in a final phase of uplifting the northern and southern cordilleras to connect them as the backbone of the Americas. Such dramatic physical changes inevitably triggered major new biological interactions between the two previously separate American continental biotas, and that is the focus of this paper.

In the following pages, I consider the major evolutionary features of the GABI as registered by land mammals. The mammals offer the richest, most

nearly coherent record of faunal changes in both American continents during the late Cenozoic. The

living faunas, along with their fossil predecessors, record both the rapidly changing distributions and the dynamic evolutionary diversifications that illuminate the origins of the Latin American biota.

My own introduction to the nature of the GABI began in the 1960s when the Florida Museum of Natural History was actively filling chronological gaps in that state's rich late Cenozoic record of land vertebrates. One fossil site in particular opened a critical window into the late Pliocene, which is known in the system of North American land mammal ages as the Blancan. Unlike any preceding stage, that interval introduced to North America a large cohort of land mammals previously restricted to South America.

Our site, Inglis 1A (Fig. 1), yielded the richest sample of late Pliocene vertebrates in eastern North America. Our team excavated and screenwashed about 100 tons of fossiliferous sediment, which yielded about 120 species of land and freshwater animals. Early in our studies, I realized that ten genera of new immigrant taxa from South America accounted for about 20% of the entire rich fauna (Webb, 1976). There were three families of ground sloths; three families of shelled edentates, ranging

1 1 thank Peter Raven and Alan Graham for inviting me to participate in the 51st Annual Systematics Symposium at the Missouri Botanical Garden. I am also grateful to many colleagues who have expanded my perspectives in both biological and geological sciences, in particular R. H. Tedford, M. C. McKenna, B. J. MacFadden, the late J. F. Eisenberg, and the participants in our symposium. This paper is Contribution No. 575 from the Florida Museum of Natural History. 2 Florida Museum of Natural History, University of Florida, Gainesville, Florida 32611, U.S.A. [email protected]

Ann. Missouri Bot. Gard. 93: 245-257. Published on 23 August 2006.



246 Annals of the Missouri Botanical Garden

Figure 1. Paleontological excavations by the Florida Museum of Natural History at Inglis 1A (late Pliocene) at cross-state barge canal, west-central Florida. Photo by author.

from tiny Dasypus to gigantic "\Glyptotherium3; as well as two families of rodents, namely Erethizontidae and

Hydrochoeridae. The truly surprising forms were a vampire bat, Desmodus Wied, and a giant predaceous bird named "\Titanis Brodkorb. Additionally, nearly 40% of the fauna at Inglis 1A represented North American genera, representatives of which had dispersed to South America during roughly the same time in the late Pliocene. There were relatives of such living Neotropical taxa as llamas, peccaries, tapirs, jaguars, raccoons, foxes, and spectacled bears (cf. Table 1 for taxa). Thus, a majority of the land mammal fauna recovered from Inglis 1A, and living in Florida nearly three million years ago, were participants, in one direction or the other, of the GABI. I was also astonished to realize how similar the mingled elements of the Inglis 1A local fauna were to contemporaneous interchange faunas through the American tropics and all the way into Argentina and the "southern cone" of the continent. Thus, in Florida we had stumbled fortuitously into the acme of the GABI (Webb, 1976).

The interamerican evolution of land mammals represents a dramatic experiment in rapid cladogen-

3 f represents extinct taxa.

esis on an intercontinental scale. We may ask how the modern South American fauna differ from the preceding fauna living in isolation. Likewise, we may ask how the present Central American fauna differ from the preceding fauna that had lived in the southern

portion of North America. Clearly, the land bridge catalyzed a rapid remodeling of the Latin American fauna. Such geological changes represented the

underlying cause that triggered major biological effects.

We begin by indicating some preliminary land mammal dispersals that preceded the GABI. I refer to these as herald taxa. Thereafter, we will consider the much larger cohorts that burst across a fully accessible Panamanian land bridge, which I refer to as

legions taxa. We also briefly review current geochronological evidence regarding the timing of the GABI. Finally, we will consider evolutionary and

ecological factors that influenced the origin of the South and Central American land mammal faunas.

Herald Taxa

A very small number of interamerican land mammal dispersals have been recorded earlier than the latest Pliocene in North and South America. These



Volume 93, Number 2 Webb 247 2006 Great American Biotic Interchange

Table 1. Families of land mammals in the GABI.

Family Common name

Northern legion taxa (17 families) Soricidae shrews

Leporidae rabbits

Heteromyidae pocket mice

Geomyidae pocket gophers Sciuridae squirrels Muridae field mice Felidae cats Mustelidae skunks and otters Canidae foxes and wolves

Procyonidae raccoons Ursidae bears

Gomphotheriidae elephantoids Tapiridae tapirs Equidae horses

Tayassuidae peccaries Camelidae llamas Cervidae deer

Southern legion taxa (2 1 families) Didelphidae opossums Dasypodidae armadillos

Chlamytheriidae giant armadillos

Glyptodontidae tank-like edentates

Megalonychidae bear-sized ground sloths

Mylodontidae middle-sized ground sloths

Megatheriidae elephantine ground sloths

Bradypodidae three-toed tree sloths

Myrmecophagidae anteaters Desmodontidae vampire bats Callithricidae marmosets Cebidae New World monkeys Hydrochoeridae capybaras Erethizontidae porcupines Caviidae guinea pigs Agoutidae pacas Dasyproctidae agoutis Echimyidae spiny rats Toxodontidae rhinoceros-like ungulates Trichechidae sea cows Phororhachidae giant predaceous birds

herald taxa appear during the middle Miocene some six or seven million years earlier than the legion taxa. Both Mexico and Central America have increasingly strong records of late Cenozoic land mammals, yet in those areas no South American immigrants other than the herald taxa have been recognized before the latest Pliocene (Webb, 1997; Ferrusquia-Villfranca, 2003). Some native North American groups (e.g., the

slingshot-horned Protoceratidae) that had become

increasingly rare at temperate latitudes evidently thrived in the more tropical landscapes of Mesoamer- ica (Webb et al., 2003). As stated by Ferrusquia- Villfranca (2003: 321), all known mammal faunas in

the Miocene of Mesoamerica "show strict North American affinities."

According to present records, the herald taxa consist of only two distinct genera of ground sloths in North America and one genus of large raccoon in South America. As indicated by Tedford et al. (2004), "\Thinobadistes Hay, a mylodontid sloth, and "fPlio- metanastes Hirshfeld & Webb, a megalonychid sloth, arrived in the early Hemphilllian about 9 million years ago (Ma). One record of "fPliometanastes from California yields the best age estimate for these early immigrant sloth genera. There it occurs in sediments 4 meters below the Mehrten Tuff, which is radiometri-

cally dated at 8.2 Ma (Hirschfeld & Webb, 1968). "\Arctonasua Baskin is a large procyonid, well

known in North America during the Clarendonian and

Hemphillian (Baskin, 2003). "fCyonasua Ameghino, a closely related genus, appears in the Chiquimil fauna of northwestern Argentina in the Huayquerian land mammal age, middle to late Miocene (Marshall, 1985; Flynn & Swisher, 1995). By Chapadmalalan time, "\Cyonasua had been replaced by its very large descendant, "\Chapalmalania Ameghino.

Flynn et al. (2005) suggested that two additional

genera of South American ground sloths separately invaded North America. This suggestion is based on late Hemphillian collections from near Guanajuato in central Mexico. It may be more parsimonious, however, to view the two genera, identified as

"\Megalonyx Leidy and ^Glossotherium Harlan, as North American descendants of the closely related sloth taxa already present in the early Hemphillian. The question may be resolved by more detailed

analysis of the Hemphillian sloth material from

Guanajuato. For three reasons, the appearances of these herald

taxa have been interpreted as waif dispersals, not as

crossings of a continuous land bridge. Firstly, the hiatus of some six million years, during which no other

crossings have been documented, suggests a fortuitous mode of island hopping. Secondly, living sloths and raccoons are particularly adept at floating and

swimming, so that they are among the most likely groups of land mammals to make water crossings (Webb, 1985). Thirdly, the very limited number of other concurrent arrivals in either direction provides strong negative evidence that no broad highway had

opened between 9 and 3 Ma. The strength of this

interpretation depends on one's optimism about the

completeness of relevant parts of the fossil record. Suffice it to state that the late Cenozoic record of land mammals in North and South America is exceedingly rich and continues to improve each year.

Occasionally, one reads that the murid rodents also

jumped the isthmian gap and entered South America




before the latest Pliocene. One early paper by Marshall (1979) encouraged the idea of some mid- Tertiary murid immigration, even though the word "model" in the title made it quite clear that his discussion was hypothetical. In their discussion of the timing of the GABI at about 2.7 Ma, Woodburne and Swisher (1995) reiterate the possibility that two murid

genera had reached South America during the Montehermosan stage, generally correlated with the latest Miocene and earliest Pliocene (ca. 7 to 4 Ma). The supposed records that they cite (by Reig, cited in Marshall, 1985), however, were not explicit and have not been verified by subsequent faunal and strati-

graphic studies. Indeed, in his review of fossil murid records in the Pampean region of Argentina, Pardinas (1995: 229) finds the stratigraphic provenance of

Reig's two type specimens to be "dubious." An additional strong inference comes from paleon-

tological studies of murids in Miocene-Pliocene sites in North America. There the presence of appropriate plesiomorphic sister groups such as 'fProsigmodon Jacobs & Lindsay and "\Bensonomys Gazin, just prior to the time of the interchange, strengthens the view that these lineages did not jump the gap prematurely. Jacobs and Lindsay (1984), for example, make a strong case for Miocene-Pliocene diversification of murids in North America followed by dispersal of multiple branches during, but not before, the Blancan GABI.

Legions Taxa

Table 1 records the families of land mammals that moved in each direction across the land bridge. In addition to including land and freshwater mammals, I have included the f Phororhachidae, a family of large, predaceous birds that followed some of their favorite mammalian prey northward across the land bridge. The record is based on "\Titanis from Inglis 1A and other Blancan sites in Florida. These data indicate a balanced reciprocal faunal interchange, as noted by Webb (1976), Webb and Marshall (1982), and Marshall et al. (1982b). The body of paleontological evidence also indicates that the mingled fauna on each American continent supported a greater diversity than before the GABI. This increase in apparent diversity lasted about a half a million years before it decreased to its prior level (Webb, 1976; Marshall et al., 1982b). Wilson (1992: 120) explains this bio-

geographic pattern as follows: "When biologists see a number go up following a disturbance and then fall back to the original level, whether body temperature, density of bacteria in a flask, or biological diversity on a continent, they suspect an equilibrium." The

interchange numbers neatly fit a large-scale example of the species equilibrium hypothesis. After the initial

mutual enrichment of both continental faunas, they encounter limits to diversity, a kind of carrying capacity that limits the number of broadly similar mammalian groups that one continent can support.

Chronology of the Interchange

The reciprocal arrival of a cohort of terrestrial mammal immigrants deep within each of the two American continents marks the inception of the interamerican land bridge and a major faunal interchange. It is important to date the GABI as precisely as possible, because it represents the datum by which ensuing evolutionary and biogeographic changes can be calibrated.

The best method to date emplacement of the Panamanian land bridge comes, perhaps surprisingly, not from any geologic data in Panama but from excellent stratigraphic sequences with rich fossil faunas in temperate latitudes of North and South America. Only a few examples of this paleontological record need be cited here to establish the quality of the record and the basic chronological framework.

The Anza- Borrego Desert lies in southern Califor- nia, just north of the Salton Sea Basin and within the San Andreas Fault zone. Sediments in that region accumulated very rapidly during the Pliocene and Pleistocene, first entombing marine fossils within the Imperial Formation, an arm of the Gulf of California, and later, non-marine vertebrates and their trackways within the Vallecito Creek Formation. The latter formation provides a wealth of land mammals, including early appearances of some South American immigrants, notably "\Erethizon stirtoni (White) Fra- zier, in a vertical sequence of some 4000 meters. Furthermore, these fossil beds include a radiometric date of 2.3 Ma and are directly intercalated with a long, detailed paleomagnetic profile (Opdyke et al., 1977). Figure 2 exemplifies the outstanding exposures of long fossiliferous sections in the Anza-Borrego Desert.

An equivalent fossiliferous section in North America, in Arizona, known as the 111 Ranch Sequence, produces the Tusker Fauna, which is associated both with an ash dated at 2.33 Ma and an excellent magnetic profile (Galusha et al., 1984). Similarly, in southwestern New Mexico, local faunas from the lowermost strata in the Mesilla Basin (Mesilla A in the Camp Rice Formation) and Pearson Mesa clearly tie South American immigrants into magnetostratigraphically controlled sections (Morgan & Lucas, 2003).

In each of these southwestern examples, the new

immigrants occur within an interval of normally oriented magnetic samples, constrained above by the




Figure 2. Vallecito Creek Formation Pliocene-Pleistocene exposures in Anza Borrego State Park, California. Dr. John A. White indicating collecting areas of Los Angeles County Museum. Photo by author.

radiometric date and the Gauss/Matuyama magnetic reversal boundary (fixed at 2.58 Ma) and below by the Kaena Subchron (at 3.04 Ma). On this basis, mammalian stratigraphers often cite the time of first appearance for Neotropical immigrants into North America as 2.6 or 2.7 Ma (Woodburne & Swisher, 1995; Bell et al, 2004).

The foregoing faunas from the southwestern United States that most concisely date the first appearances fall within the late Blancan land mammal stage. This same biostratigraphic rubric includes several Florida faunas, most notably the Inglis IA site, which records the greatest wealth of Neotropical immigrants from any one locality. The unfortunate absence of radiometric control on Florida faunas weakens our ability to specify how concurrently the entire cohort of Blancan immigrants from South America arrived in temperate North America. It is reasonable to suggest that they had all come within about a two-million-year period, for that is the approximate span of the Blancan interval (Bell et al., 2004).

Some evidence of Blancan immigrants comes from lower latitudes in Mesoamerica. One unique genus is "\Meizonyx Webb & Perrigo from El Salvador. This megalonychid ground sloth is larger than any Blancan

species within its temperate sister genus, "\Megalonyx, yet it possesses triangular (plesiomorphic) caniniform teeth (Webb & Perrigo, 1985).

Intensive geochronological and biostratigraphic studies of sedimentary sequences near Guanajuato in central Mexico record several immigrant genera from South America earlier in the Blancan than the well-dated southwestern faunas cited above (Miller & Carranza-Castaneda, 2001; Flynn et al., 2005; Montellano-Ballesteros & Jimenez-Hidalgo, 2006). 'fPlaina, a small sister genus of "\Holmesina Simpson, is recorded at about 4.6 or 4.7 Ma, in the very early Blancan. ~\Glyptotherium and "\Neochoerus appear in the interval between 3.9 and 3.1 Ma (Flynn et al., 2005). This may suggest, as proposed by Flynn et al.

(2005), that central Mexico represented a different mammalian province than southwestern United States. It is possible that some Florida Blancan faunas, notably Haile 15A and Santa Fe River 1, producing these and other immigrant land mammals from South America, may also pertain to the early Blancan. These faunas are not well enough constrained by geochronology, and it is notoriously difficult to distin-

guish early from late Blancan biostratigraphically (Bell et al., 2004). The new discoveries of earlier




Figure 3. Barranca de Los Lobos, Buenos Aires Province, Argentina, Pliocene-Pleistocene exposures. Photo by author.

Blancan immigrants in Guanajuato indicate that the legion taxa may have reached temperate North America on a more haphazard schedule than previously suggested.

Critical evidence from southern, temperate parts of South America produces results broadly similar to those from the southwestern United States. Long sequences of fairly continuous fossiliferous sediments provide an essential framework for integrating biostratigraphic evidence with both radiometric dates and paleomagnetic reversal chronology (Marshall et al., 1982a; Flynn & Swisher, 1995).

A central part of Barranca de Los Lobos near Mar del Plata in Argentina (Fig. 3), along with other typical Uquian sites in the province of Buenos Aires, records a number of first appearances of North American taxa, including equids, mustelids, tayas- suids, camelids, and murid rodents (Pascual et al., 1985; Alberdi et al., 1995). Paleomagnetic work, integrated with land-mammal biostratigraphy, places these northern immigrants just below the Gauss/ Matuyama boundary, justifying a first appearance datum of about 2.5 Ma (Marshall et al., 1982a), later revised to about 2.7 Ma (Woodburne & Swisher, 1995).

A curious claim of a much earlier immigration of northern land mammals into South America has been advanced by Campbell et al., (2000: 33A) in a brief

abstract. Based on presumed correlations to sediments underlying an ash date of 9.01 Ma in the Madre de Dios Formation in eastern Peru, these authors

recognize "numerous Amazonian paleofaunas" that include "at least two species each of proboscideans, peccaries, tapirs and camelids." The discrepancy between their evidence from unspecified Amazonian faunas versus the wealth of previously documented, late Miocene land mammal faunas in South America is

profound- The classical sequences at Chiquimil and Corral Quemado in Catamarca Province, northwestern

Argentina, offer a wealth of excellent fossil mammals

typifying the Huayquerian land mammal stage. Additionally, rich faunas from Micana in the eastern

altiplano of Bolivia also span relevant, late Miocene intervals (see recent references in Flynn & Swisher, 1995). Yet the only hint of any North American element in any of these well-studied faunas is

"\Cyonasua, discussed earlier from the Chiquimil fauna from Argentina (Marshall, 1985; Flynn & Swisher, 1995). Even the next younger Chapadmalalan and/or Montehermosan mammal ages give no hint of the mysterious immigration proposed by Campbell et al. (2000). In apparent anticipation of such skepticism, these authors claim that "ecological restrictions

undoubtedly hindered the movement of the earliest

dispersers out of lowland tropical forests" (2000: 33A). This ad hoc defense of such a novel hypothesis




underlines its improbability. It would be quite astonishing if eight or more species of large, mostly herding herbivore taxa had remained sequestered in southeastern Peru for six or more million years.

Recently, the assertion of Campbell et al. (2000) received minor encouragement from Ferrusquia-Vil- lafranca (2003) on the grounds that three of the four family groups in question ("fGomphotheriidae, Tayas- suidae, and Camelidae) are known in the middle Miocene of Mesoamerica. This might suggest that these three family groups could have been available for their supposed early dispersal into one corner of the Amazon Basin. Such a charitable view does not relieve those who would hypothesize a much earlier dispersal by a novel, geographic route from the need to provide more substantial evidence.

Later Features of the Interchange

After the early events of the GABI, with balanced dispersals and augmented familial diversity in both continental faunas, the evolutionary pathways of immigrant land mammals in the two American continents diverge quite dramatically. Evidently, the GABI conformed to the balanced mutual enrichment, predicted by the species equilibrium hypothesis, during the first million years, within an ecological time frame. Thereafter, however, within a more extended evolutionary time frame, the northern and southern legions embarked on radically different evolutionary pathways.

In South America, the most remarkable result of the GABI is the rapid diversification of many of the immigrants from the north. In each successive stage, the fossil record throughout the continent reflects a continual increase in the numbers of taxa derived from the first wave of immigrants. Webb and Marshall (1982: 50) quantified this trend by a census of the modern mammalian fauna at the generic level. They found that more than half (about 53%) of the present South American land mammal fauna consisted of taxa that had entered the continent by crossing the isthmian land bridge less than three million years ago.

This result becomes even more notable when one turns to the North American record and follows the very different evolutionary fate there of immigrants from South America. Little or no diversification takes place. Some genera may be considered quite success- ful by some measures. For example, "fMegalonyx, a bear-sized ground sloth, evolved fairly rapidly and extended its range very widely, even reaching Alaska. However, none of the South American immigrants diversified above the species level. In the recent fauna of temperate North America, only four genera of

southern immigrants survive, namely Didelphis L., Dasypus L., Erethizon Cuvier, and Trichechus L.

Why Did Northern Taxa Succeed in South America?

This striking asymmetry in the results of the GABI requires explanation. In particular, the immense success of northern immigrants in South America represents one of the greatest natural experiments in rapid evolution on a large scale. The most cogent general hypothesis may be that the northern groups that spread into South America had a long, wide- ranging history not only in North America, but also before that in Eurasia. This cosmopolitan background strongly contrasts with the long Cenozoic isolation experienced by the pre-GABI South American fauna.

A particularly intriguing part of this inequity is the presence among the northern legions of efficient mammalian carnivores. In South America, by contrast, the only mammalian carnivores were primitively modified members of the marsupial family, Borhyae- nidae. The group had declined markedly before the interchange. It seems evident that the few remaining taxa could not compete with progressive carnivorans that rapidly spread into their native continent. Furthermore, the large and small herbivores of South American descent, lacking any previous experience with advanced mammalian carnivores, were over- whelmed by the rapid onslaught of wolves, jaguars, sabercats, and many smaller kinds of carnivores in the families Felidae, Canidae, and Mustelidae.

To place the GABI in a broader context, let us turn briefly to the earlier and more diffuse intercontinental interchange between Asia and North America. Table 2 lists the times and generic numbers of immigrant cohorts that entered North America from Asia. These data summarize recent tabulations for Miocene land mammals by Tedford et al. (2004). Of particular interest in this context are the several sets of immigrants that reached the New World in the middle to late Miocene. The greatest of these episodes occurred about 19 Ma (middle Miocene), when 17 Eurasian genera are recorded as new arrivals in this continent. In Woodburne and Swisher's (1995) similar codification of Cenozoic immigration events in North America, this is Interchange Event ̂ 6. Two million years later, another large wave of immigrations appeared in Interchange Event #7 (Woodburne & Swisher, 1995). A particularly important member of that cohort was "\Copemys Wood, a probable descendant of a Eurasian murid such as "fDemocriceto- don Fahlbusch. The future impact of this North American immigration event with respect to the GABI surely could not have been anticipated. Nonetheless, one epoch later, when a few inconspicuous descen-




Table 2. Miocene immigrants to North America from Asia.*

Age (Ma) No. of genera Selected genera**

6 7 Lutra (Mustelidae), "\Eocoileus (Cervidae), "fMegantereon (Felidae), "fTrigonictis (Mustelidae)

7 8 Felis (Felidae) 8 9 -fPlionarctos (Ursidae) 9 6 12 2 15 9 "\Gomphotherium

(Gomphotheriidae) 16 2 17 11 "fCopemys (Muridae) 19 17 20 4 23 6

* Figures compiled from Tedford et al. (2004). ** See text. t Includes both fossil and living genera.

dants of these murid rodents colonized South America, they produced by far the greatest monofamilial

diversity of land mammal genera in that continent. The next middle Miocene immigration episode

introduces a taxon that is extremely far removed from murid rodents, namely, the first proboscidean in the New World. Out of Africa by way of Asia came the

elephantoid genus "fGomphotherium Burmeister. New World descendants of this basal member of the family Gomphotheriidae diversified greatly in Central Amer- ica and, following the GABI, South America.

Following a mid-Miocene lull, the late Miocene in North America witnessed another series of strong immigration episodes from Asia. New immigrant genera included the first New World deer, Eocoileus

Webb; otters of the modern genus Lutra Bliinnich; "\Trigonictis Hibbard, a large extinct mustelid more or less closely related to the living tropical genus Eira

(Hamilton) Smith (the tayra); and "\Plionarctos Frick, the oldest known tremarctine (spectacled) bear, presumably an early sister-genus of the living Neotropical Tremarctos Gervais. Finally, two immi-

gration events occurred involving the large cats: Felis L. and "fMegantereon Croizet & Jobert, which is the

probable antecedent of "fSmilodon Lund, the great sabercat of the Pliocene-Pleistocene. All eight of these Miocene immigrants contributed key elements to the North American cohort during the GABI.

Evidently, these Miocene immigrants extended their ranges through the polar latitudes of the Bering Strait before climatic deterioration produced extreme-

ly cold, dry conditions. Virtually all of the omnivores, frugivores, and herbivores involved, based on reason-

able interpretations of their dental adaptations and on the habits of their living relatives, required relatively mesic habitats with year-round provisions of nutritious

browse, fruit, or seeds.

By the beginning of the Pliocene, the ecological valences of immigrants from Asia began to shift toward forms adapted to cold steppes. Such grazers as

'fMammuthus Brookes, f Bison H. Smith, and diverse microtine rodents spread widely through most of North America but were not subsequently involved in the GABI. Evidently, they were too late to avail themselves of an open-country route through the isthmian region. There is good evidence that "\Mam- muthus, during the early Pleistocene, and f Bison, in the latest Pleistocene, extended their ranges southward along the relatively dry Pacific terrain into Honduras and Nicaragua. These areas today support extensive pine savannas and thorn scrub, especially along the seasonally drier Pacific slopes. Beyond that, however, these grazers did not penetrate, presumably because they were blockaded by the prevalent rainforest (Webb & Perrigo, 1984; Webb, 1997).

The purpose of this digression into Holarctic faunal movements is to show that the Bering Strait had an

important effect on the GABI. Clearly, many groups of land mammals that joined the legion taxa from North America had previous evolutionary records in the World Continent. We suggest that this background may help explain their success in South America. More than two million years after the GABI, murid

rodents, cervid ruminants, and other northern families with similar backgrounds burgeoned into more than half of the genera of modern South American mammals (Webb & Marshall, 1982). Wilson (1992: 130) offered the following eloquent explanation: "The World Continent . . . has tested more evolutionary lines, built tougher competitors, and perfected more defenses against predators and disease. This advan-

tage has allowed its species to win by confrontation.

They have also won by insinuation: many were able to

penetrate sparsely occupied niches more decisively, radiating and filling them quickly. With both confrontation and insinuation, the World Continent mammals gained the edge."

Evolutionary Patterns of Northern Interchange Taxa

To develop a full understanding of how the legions of land mammals from the north realized their full

evolutionary potential in South America remains a great challenge. More than a century of dedicated

paleontological spadework on both American continents has provided a reliable framework in two

respects. Firstly, it has identified the set of 17 families of land mammals (Table 1). Secondly, it has de-




termined a fairly well-dated beginning point for reciprocal immigrations of legion taxa between about 3.1 and 2.7 Ma (Webb, 1985; Flynn et al., 2005).

Yet, from an evolutionary perspective, this is just the beginning of the story. The other end of the story has also been well wrought by several generations of mammalogists working out the modern systematics, ecology, and distribution of Neotropical mammal groups. Taken together, these basic facts imply some of the most dynamic examples of mammalian evolution on any continent within the entire Age of Mammals. However, most of the real phylogenetic history of these groups still remains to be told.

Presumably, the opening of a new continent after

long isolation, accompanied by wholesale introduction of new kinds of mammals, provides a perfect theater for rapid, innovative evolution. The diverse nature of topographic and ecologic opportunities in South America was already evident in the mid Tertiary, especially after the uplift of the Andes (Flynn & Swisher, 1995; Burnham & Graham, 1999). This

grand panorama of opportunities may have expanded even more broadly after the GABI as a consequence of increased climatic and sea level fluctuations during the Pleistocene.

The basic fossil mammal data from the late Pliocene also permit an ecological generalization about the availability of an open country route. At the time of the GABI, the isthmian region was broadly accessed by herds of grazers and browsers (Webb, 1991). The legions of the north were primarily populated by such large herbivores as horses, llamas, and deer. An especially significant role may be attributed to gomphotheres; presumably they modified the landscape by clearing trees, especially during dry seasons, much in the manner of African elephants today.

Likewise, a majority of the southern immigrant taxa reflect the same set of open landscapes. The analogy is weakened somewhat because most of the key families of large herbivores lack living relatives. Nonetheless, one can reasonably impute to toxodonts the role of rhinocerotids and confirm it with carbon

isotopic data. Three or four distinct families, broadly brigaded as ground sloths, included grazers, mixed feeders, and browsers. The tank-like glyptodonts are

generally perceived as herding grazers of savanna

settings. In the southern set, there were also some

aquatic herbivores, namely the Hydrochoeridae (capybaras) and the Trichechidae (seacows); both groups were grazers, the former arrayed along lowland waterways and the latter surely moving along the coast. In general, as emphasized by Webb (1978; 1997), the GABI data imply a broad highway of savanna and open woodland that extended from the

Great Plains of North America to the Pampas of Argentina during the late Pliocene.

The most controversial part of this supposed savanna highway is the isthmian portion. In their comprehensive review of the complex late Cenozoic history of Neotropical vegetation, Burnham and Graham (1999: 560) grant that "fossil plant data are generally consistent with estimates of habitats (savannas) and climate (temperate) based on mammalian fossil faunas, but with some differences." It may be necessary to argue for the GABI savanna by calling for relatively short-lived rain shadows during the late Pliocene. GABI data do not conflict with younger evidence indicating that broad belts of rainforest probably dominated the region during most, if not all, of the Pleistocene (Colinvaux, 1997).

In South America, the remarkable radiations of northern immigrants, briefly outlined below, moved along similar ecogeographic lines. Marquet and Cofre (1999) conducted detailed analyses of South Ameri- can mammal diversity partitioned into ten biomes. Their results find northern immigrant taxa dominant in the following three biomes: temperate grassland, cold winter desert, and mountain systems.

Continuing systematic studies of each of the South American phylogenetic lines promise to fill in the detailed patterns of their deployment across the continent during the Pleistocene. Molecular studies provide a particularly important set of independent evidence. Based on the degree of its generic diversification, each immigrant family may be assigned a bradytelic, horotelic, or tachytelic rate of phylogenetic evolution (Table 3). The nine bradytelic families are Soricidae, Sciuridae, Heteromyidae, Geomyidae, Leporidae, Ursidae, Gomphotheriidae, Tapiridae, and Equidae. Five of these families still consist only of the same genus that entered the continent. The four families exemplifying intermediate (horotelic) rates are Procyonidae, Felidae, Tayassuidae, and Camelidae. With nine or ten genera each, Canidae, Mustelidae, and Cervidae exemplify rapid (tachytelic) taxonomic evolution following the GABI. The fourth tachytelic family, the Muridae, exhibits truly explosive taxonomic evolution, producing dozens of genera and several tribes uniquely differentiated in South America.

The dazzling diversity of Neotropical Muridae and their interrelationships continue to challenge sys- tematists. These field mice experienced multiple diversifications in southern deserts, in tropical low- lands, and in Andean uplands. The sense of their succcess everywhere is captured nicely by Pearson (1982: 280) in his work on small mammals of the altiplano and desert regions of Peru, wherein he cites the Muridae as "a tribute to the evolutionary ferment in South America."




Table 3. Generic diversification among northern land mammals in South America.

Family No. of genera

Bradytelic Soricidae 1 Sciuridae 3 Heteromyidae 1 Geomyidae 1 Leporidae 1 Ursidae 2 tGomphotheriidae 3 Tapiridae 1 Equidae 3

Horotelic

Procyonidae 6 Felidae 4 Tayassuidae 4 Camelidae 5

Tachytelic Canidae 10 Mustelidae 9 Cervidae 9 Muridae 46-60

f Includes both fossil and living genera.

An important consideration that will increasingly concern students of phylogenetic differentiation of northern mammals in South America is the question of how much diversification had already been accom-

plished in tropical North America before the GABI. As noted above, the paleontological record documents that some Murid diversity (e.g., "fProsigmodon and

"fBensonomys) was staged before the GABI in low latitudes of North America during the Hemphillian and Blancan (Jacobs & Lindsay, 1984). This means that molecular clock estimates of divergence times

among murid rodents may date events that preceded the actual entry of those branches into South America.

Baskin (2003) recently described a surprising new

example of pre-GABI differentiation: the new genus and

species "fParapotos tedfordi Baskin representing the

procyonid tribe Potosinae in the middle Miocene

(Barstovian) of Texas. This taxon is the only extinct member and the first recorded fossil of the Potosinae. Without this ancient kinkajou record, it would seem reasonable to assume that modern forms developed their

arboreal, nocturnal habits and their marvelous pre- hensile tails as post-GABI evolutionary products of their

present range in multistratal rainforests of the Amazon and lowland Central America (Eisenberg, 1989).

A recent phylogenetic analysis of New World deer

(Webb, 2000) documents the geographic complexity of another radiation that had been thought to be purely South American (Hershkovitz, 1982). The Andean

endemics, Hipppocamelus Leuckart and Pudu Gray, are shown to be members of the tribe Rangiferini (as defined by Webb, 2000) with an extinct sister taxon, f Navahoceros Kurten, in the southwestern parts of

temperate North America. This tribe of mountain deer had already diverged in North America from Mazama

Rafinesque, Ozotoceros Ameghino, and Blastocerus

Gray, members of the tribe Odocoileini (as defined by Webb, 2000) before their separate southward dis-

persals. It seems likely that there were three or more

separate entries of what are presently South American deer across the Panamanian land bridge. Thus, the issue of Central American diversification and staging of northern groups must be considered in phylogenetic analyses of South American immigrant groups.

A Southern Success Story within the American Tropics

The great success of the northern immigrants in South America has been well documented by paleontologists working in Andean localities, Argen- tina, and other parts of temperate South America. Those efforts, however, largely miss the contrary success story in which Central America was con-

quered by the tropical fauna of South America. A few sites in tropical parts of South and Central America must be more intensely developed before the full

paleontological perspective can elaborate on this

tropical counterdispersal. I like to call this episode The Central American

Paradox. During the Miocene, Central America had been a southern adjunct of North America, widely and

deeply separated from northern South America by the Bolivar Trough. Vertebrate fossil sites in Mexico, Honduras (Gracias Formation), El Salvador, and Panama (Cucaracha Formation) corroborate the pa- leogeographic data by producing no evidence of any land mammals from South America (Webb & Perrigo, 1984; Ferrusquia-Villafranca, 2003). Yet today, Central America and southern Mexico are faunisti-

cally and floristically brigaded within the Neotropical Realm. This was obvious to A. R. Wallace in his great work on biogeography (Wallace, 1876). In that classic

synthesis, he extended the Neotropical Realm from its

major center in Amazonia up to about the Tropic of Cancer. The predominant groups of mammals in Central America today had their origins in Amazonia before the establishment of the land bridge. These were stalwart members of the ancient isolated fauna, including a diversity of New World monkeys; edentates, such as armadillos, anteaters, and sloths; and hystricognath rodents.

Unfortunately, the critical, transitional phases of this Neotropical success story are nearly opaque to

paleontologists. In essence, one can study only the




Figure 4. Paleontological excavations by Universidade Federal do Acre party, led by Dr. Alceu Rancy, on Rio do Acre in late Cenozoic sediments. Photo by author.

Miocene, which is too early, or the modern fauna, which, in some sense, is too late. It is a problem familiar to any paleontologist who has attempted to overcome the refractory nature of the fossil record in tropical settings. The pressing need in this case is to redouble our efforts in primary field work in Pliocene- Pleistocene exposures throughout tropical America. Ultimately, this challenge will be answered and this mystery will yield to such ongoing endeavors as those of Laurito et al. (1993) in Costa Rica and Rancy in the western Amazon (Webb & Rancy, 1996). Figure 4 illustrates one of the valuable late Cenozoic localities on the Rio do Acre, Brazil accessible during the dry season.

Conclusions

The causes of the GABI can be traced readily and powerfully to the forces of plate tectonics in lower Central America. Hundreds of cubic kilometers of ignimbrites and other massive volcanic productions had already greatly augmented the volume of nuclear Central America during the middle Miocene. The gradual shallowing of seaways in the isthmian region, driven by persistent subduction of oceanic crust by Pacific plates, had major consequences well before the final emplacement of a continuous land bridge (Cronin & Dowsett, 1996). Jones and Hasson (1985) reviewed

the many studies that demonstrate the separation and divergence of marine faunas on both sides of the isthmus. In general, deepwater species undergo geminate speciation earlier than shallow-water forms, and such patterns can be demonstrated in many organisms both across a broad range of taxonomic levels and also by rapidly diverging molecular systems, notably by mitochndrial DNA (Collins, 1996). Rapid divergence between Atlantic and Pacific descendants of planktonic foraminifera, such as the globigerinid "\Globorotalia multicamerata, indicates that the seaway was breaking up between about 3.6 to 3.1 Ma. The

importance of this process in reorganizing oceano-

graphic currents and intensifying the Gulf Stream was indicated by Keigwin (1982) and Cronin and Dowsett (1996), who also noted its probable role in generating northern hemisphere ice sheets beginning about 3.2 Ma and again around 2 Ma.

The final emplacement of a continuous isthmian land bridge is faithfully recorded by the appearance of

legions of land vertebrates reciprocally dispersing between North and South America. Substantial evidence from long sections in temperate latitudes of both continents places the GABI events between about 2.7 and 3.1 Ma (Webb, 1985; Flynn et al., 2005).

During the first million years or so of the GABI, the contribution of each continent to the faunal in-




terchange was roughly equal at the family level, and each continental fauna was taxonomically enriched. Subsequently, by early Pleistocene time, the interchange, as recorded by new immigrants, was essentially completed and family-group diversities had dropped back to previous levels, in accord with predictions of species equilibrium theory.

The northern legion taxa experienced astonishing evolutionary success as measured in the modern South American fauna, capturing more than half of the present generic diversity. When that broad record is partitioned, it is seen that nine families simply entered the continent with little or no taxonomic diversification above the species level (cf. Table 3, bradytelic families). Four other families of northern land mammals evolved taxonomically at moderate (horotelic) rates: these include Procyonidae, Felidae, Tayassuidae, and Camelidae. The four families that really seized the new opportunities in South America, evolving at tachytelic rates, were Canidae, Musteli- dae, Cervidae, and especially Muridae. It is worthy of note that their success focused ecologically on the very biomes that most nearly resembled temperate biomes from whence they had come. According to Marquet and Cofre's (1999) detailed analyses of mammalian diversity in ten biomes, northern immigrant taxa are dominant in temperate grassland, cold winter desert, and mountain systems.

A countervailing conquest of Central America by the Amazonian biota, while nearly opaque to paleontologists, has been fully appreciated by most biologists since Alfred Russell Wallace (1876) proposed the

Neotropical Realm. There are hints in the relatively weak tropical portions of the vertebrate fossil record that this phase of the GABI came after the Pliocene and that during the Pleistocene the Panamanian land bridge was more fully occupied by rainforest (Webb, 1991; Webb & Rancy, 1996). Paleobotanical evidence makes a strong case for this (Colinvaux, 1997; Burnham & Graham, 1999).

Among land mammals, the prominent protagonists in this second phase, into Central America, were old southern taxa including marsupials, primates, xenar- thrans (both cingulates and pilosans), and hystricognath rodents. It is not surprising that these same old southern taxa outnumber the northern newcomers in the two South American biomes designated by Marquet and Cofre (1999) as temperate rain forest and tropical humid forest.

Literature Cited

M. T. Alberdi, G. Leone & E. P. Tonni (editors). 1995. Evolucion Biologica y Climatica de la Region Pampeana durante los Ultimos Cinco Millones de Anos. Monogr. Mus. Nac. Ci. Nat., Madrid 12.

Baskin, J. 2003. New Procyonines from the Hemingfordian and Barstovian of the Gulf Coast and Nevada, including the first fossil record of the Potosini. Pp. 125-146 in L. Flynn (editor), Vertebrate Fossils and Their Context: Contributions in Honor of Richard H. Tedford. Bull. Amer. Mus. Nat. Hist.

Bell, C. J. et al. (+ 8 authors). 2004. The Blancan, Irvingtonian, and Rancholabrean Mammal Ages. Pp. 232- 314 in M. 0. Woodburne (editor), Late Cretaceous and Cenozoic Mammals of North America: Biostratigraphy and Geochronology. Columbia Univ. Press, New York.

Burnham, R. J. & A. Graham. 1999. The history of Neotropical vegetation: New developments and status. Ann. Missouri Bot. Gard. 86: 546-589.

Campbell, K. et al. (+ 5 authors). 2000. Late Miocene dynamics of the Great American Interchange: Waifs are out. J. Vertebrate Paleontol. (abstract supplement) 20: 33A.

Colinvaux, P. 1997. The history of forests on the isthmus from the Ice Age to the Present. Pp. 123-136 in A. G. Coates (editor), Central America: A Natural and Cultural History. Yale Univ. Press, New Haven.

Collins, T. 1996. Molecular comparisons of transisthmian species pairs: Rates and patterns of evolution. Pp. 303- 334 in J. B. C. Jackson, A. F. Budd. & A. G. Coates (editors), Evolution and Environment in Tropical America. Univ. Chicago Press, Chicago.

Cronin, T. M. & H. J. Dowsett. 1996. Biotic and oceanographic response to the Pliocene closing of the Central American isthmus. Pp. 76-104 in J. B. C. Jackson, A. F. Budd. & A. G. Coates (editors), Evolution and Environment in Tropical America. Univ. Chicago Press, Chicago.

Eisenberg, J. F. 1989. Mammals of the Neotropics: The Northern Neotropics, vol. 1. Univ. Chicago Press, Chicago.

Ferrusquia-Villafranca, I. 2003. Mexico's Middle Miocene mammalian assemblages: An overview. Pp. 321-347 in L. Flynn (editor), Vertebrate Fossils and Their Context: Contributions in Honor of Richard H. Tedford. Bull. Amer. Mus. Nat. Hist.

Flynn, J. J. & C. C. Swisher III. 1995. Cenozoic South American land mammal ages: Correlation to global geochronologies. Pp. 317-333 in W. A. Berggren (editor), Geochronology, Time Scales and Global Stratigraphic Correlation. SEPM Special Publication No. 54.

et al. (+ 6 authors). 2005. Geochronology of Hemphillian-Blancan aged strata, Guanajuato, Mexico, and implications for timing of the Great American Biotic Interchange. J. Geol. 113(3): 287-307.

Galusha, T., N. M. Johnson, E. H. Lindsay, N. D. Opdyke & R. H. Tedford. 1984. Biostratigraphy and magnetostrati- graphy, late Pliocene rocks, 111 Ranch, Arizona. Geol. Soc. Amer. Bull. 95: 714-722.

Hershkovitz, P. 1982. Neotropical deer (Cervidae), Part 1. Pudus, Pudu Gray. Fieldiana, Zool. 11: 1-86.

Hirschfeld, S. E. & S. D. Webb. 1968. Plio-Pleistocene Megalonychid sloths of North America. Bull. Florida State Mus., Biol. Sci. 12: 213-296.

Jacobs, L. L. & E. H. Lindsay. 1984. Holarctic radiation of Neogene muroid rodents and the origin of the South American Cricetids. J. Vertebrate Paleontol. 4: 265-272.

Jones, D. S. & P. F. Hasson. 1985. History and development of the marine invertebrate faunas separated by the Central American isthmus. Pp. 325-355 in F. G. Stehli & S. D. Webb (editors), The Great American Biotic Interchange, Topics in Geobiology, vol. 4. Plenum Press, New York.




Keigwin, L. 1982. Isotopic paleoceanography of the Carib- bean and east Pacific: Role of Panama uplift in Neogene time. Science 217: 350-353.

Laurito, C. A., W. Valerio & E. Vega. 1993. Nuevos hallazgos paleovertebradologicos en la Peninsula de Nicoya: Implicancias paleoambientales y culturales de la fauna de Nacaome. Revista Geol. Amer. Centr. 16: 113-115.

Marquet, P. A. & H. Cofre. 1999. Large temporal and spatial scales in the structure of mammalian assemblages in South America: A macroecological approach. Oikos 85: 299-309.

Marshall, L. G. 1979. A model for paleobiogeography of South American cricetine rodents. Paleobiology 5: 125-132.

. 1985. Geochronology and land-mammal biochronology of the transamerican faunal interchange. Pp. 49-85 in F. G. Stehli & S. D. Webb (editors), The Great American Biotic Interchange, Topics in Geobiology, vol. 4. Plenum Press, New York.

, R. Butler, R. E. Drake & R. E. Curtis. 1982a. Geochronology of type Uquian (Late Cenozoic) land mammal age, Argentina. Science 216: 986-989.

, S. D. Webb, J. J. Sepkoski & D. M. Raup. 1982b. Mammalian evolution and the Great American Inter- change. Science 215: 1351-1357.

Miller, W. E. & 0. Carranza-Castaneda. 2001. Late Cenozoic mammals from the basins of central Mexico. Boll. Soc. Paleontol. Ital. 40: 235-242.

Montellano-Ballesteros, M. & E. Jimenez-Hidalgo, (editors). Mexican fossil mammals: Who, when, where? Pp. 249-273 in F. Vega et al. (editors), Studies on Mexican Paleontol- ogy. Springer, Dordrecht.

Morgan, G. & S. G. Lucas. 2003. Mammalian biochronology of Blancan and Irvingtonian (Pliocene and early Pleisto- cene) faunas from New Mexico. Pp. 269-320 in L. Flynn (editor), Vertebrate Fossils and Their Context: Contribu- tions in Honor of Richard H. Tedford. Bull. Amer. Mus. Nat. Hist.

Opdyke, N. D., E. H. Lindsay, N. M. Johnson & T. Downs. 1977. The paleomagnetism and magnetic polarity stratig- raphy of the mammal-bearing section of Anza-Borrego State Park, California. Quatern. Res. 7: 316-329.

Pardinas, U. F. J. 1995. Los roedores Cricetidos. Pp. 229- 253 in M. T. Alberdi, G. Leone & E. P. Tonni (editors), Evolucion Biologica y Climatica de la Region Pampeana durante los Ultimos Cinco Millones de Anos. Monogr. Mus. Nac. Ci. Nat., Madrid 12.

Pascual, R., M. G. Vucetich, G. J. Scillato-Yane & M. Bond. 1985. Main pathways of mammalian diversification in South America. Pp. 219-247 in F. G. Stehli & S. D. Webb (editors), The Great American Biotic Interchange, Topics in Geobiology, vol. 4. Plenum Press, New York.

Pearson, 0. 1982. Distribucion de pequenos mamiferos en el altiplano y los desiertos del Peru. Pp. 263-284 in P. Salinas (editor), Actas del VIII Congreso Latinoamericano de Zoologia, Merida.

Tedford, R. H. et al. (+ 7 authors). 2004. Mammalian biochronology of the Arikareean through Hemphillian Interval (Late Oligocene through Early Pliocene epochs). Pp. 169-231 in M. 0. Woodburne (editor), Late Cretaceous and Cenozoic Mammals of North America: Biostratigraphy and Geochronology. Columbia Univ. Press, New York.

Wallace, A. R. 1876. The Geographical Distribution of Animals, vols. 1 and 2. Harper & Brothers, New York.

Webb, S. D. 1976. Mammalian faunal dynamics of the Great American Interchange. Paleobiology 2: 216-234.

. 1978. A History of savanna vertebrates in the New World, Part II: South America and the Great Interchange. Ann. Rev. Ecol. Syst. 9: 393^26.

. 1985. Late Cenozoic mammal dispersals between the Americas. Pp. 357-386 in F. G. Stehli & S. D. Webb (editors), The Great American Biotic Interchange, Topics in Geobiology, vol. 4. Plenum Press, New York.

. 1991. Ecogeography and the Great American Interchange. Paleobiology 17: 266-280.

. 1997. The Great American Faunal Interchange. Pp. 97-122 in A. G. Coates (editor), Central Ameri- ca: A Natural and Cultural History. Yale Univ. Press, New Haven.

. 2000. Evolutionary History of New World Cervidae. Pp. 38-64 in E. S. Vrba & G. B. Schaller (editors), Antelopes, Deer and Relatives: Fossil Record, Behavioral Ecology, Systematics and Conservation. Yale Univ. Press, New Haven.

& L. G. Marshall. 1982. Historical biogeography of recent South American land mammals. Pp. 39-54 in M. A. Mares & H. H. Genoways (editors), Evolution of Neotropical Mammals, vol. 6. Pymatuning Laboratory of Ecology, Univ. Pittsburgh.

& S. C. Perrigo. 1984. Late Cenozoic vertebrates from Honduras and El Salvador. J. Vertebrate Paleontol. 4: 237-254.

& . 1985. New Megalonychid sloths from El Salvador. Pp. 113-120 in G. G. Montgomery (editor), The Evolution and Ecology of Armadillos, Sloths and Vermi- linguas. Smithsonian Institution Press, Washington, D.C.

& A. Rancy. 1996. Late Cenozoic evolution of the Neotropical mammal fauna. Pp. 335-358 in J. B. C. Jackson, A. F. Budd & A. G. Coates (editors), Evolution and Environment in Tropical America. Univ. Chicago Press.

, B. L. Beatty & G. Poinar Jr. 2003. New evidence of Miocene Protoceratidae, including a new species from Chiapas, Mexico. Pp. 348-367 in L. Flynn (editor), Vertebrate Fossils and Their Context: Contributions in Honor of Richard H. Tedford. Bull. Amer. Mus. Nat. Hist.

Wilson, E. 0. 1992. The Diversity of Life. Harvard Univ. Press, Cambridge.

Woodburne, M. 0. & C. C. Swisher III 1995. Land-mammal high-resolution geochronology, intercontinental overland dispersals, sea level, climate and vicariance. Pp. 335-364 in W. A. Berggren (editor), Geochronology, Time Scales and Global Stratigraphic Correlation. SEPM Special Publication No. 54.



the great american biotic interchange: patterns and...

Documents