<<

Sigma Xi, The Scientific Research Society

Land and the Great American Interchange Author(s): Larry G. Marshall Reviewed work(s): Source: American Scientist, Vol. 76, No. 4 (July-August 1988), pp. 380-388 Published by: Sigma Xi, The Scientific Research Society Stable URL: http://www.jstor.org/stable/27855307 . Accessed: 16/05/2012 18:07

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 of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Sigma Xi, The Scientific Research Society is collaborating with JSTOR to digitize, preserve and extend access to American Scientist.

http://www.jstor.org Land Mammals and Great American the Interchange

LarryG. Marshall

that Xhe of , , , 1982). It is only during the last decade, moreover, in the sediments , and were once joined in a large land greater precision dating containing taxa has a firm time frame for mass in the southern hemisphere called . interchange provided various the event. It is now to assess About 100 million ago (mya) South America began aspects of possible in and to the and to separate fromAfrica, moving in a primarily westward the interchange detail, analyze tempo and the rates of and direction. There is no convincing geological evidence to mode of dispersal origina a indicate that South America had a continuous land tion in successive faunas through time. As result, the connection with any other _ _Great American Interchange repre until about 3 mya, when the Bolivar sents the best-documented example in the record of the intermin Trough marine barrier disappeared The the and the were united the emergenceof of two continen by Panamanian land gling long-separated emergence of the Panamanian tal faunas. the time of the bridge. The long-isolated continental threemillion By interchange, biotas of North and South America years ago the land faunas of North and South America had distinct his were brought into contact, resulting permittedthe mingling tories that their character and in an intermingling that has come to the shaped be known as the Great American of long-separated taxonomic composition. During the about 66 mya to the Interchange (Webb 1976; Fig. 1). The North and (from faunas of was con site of the former Bolivar Trough is present), South America nected at one time with and thus the gateway for this event, de Europe on occasions with via noting the historical boundary be multiple in recurrent additions of Old World tween two biotic provinces (Fig. 2). Although many , resulting in the taxa Russell and Zhai different groups of plants and took part (McKenna 1975; Repenning 1980; on are When the Great American the interchange, I will focus land mammals, which 1987). Interchange began, North American land mammal fauna was of the vast the most thoroughly studied of the participants. part was families and occurred The Great American Interchange first recog Holarctic realm, and many genera another hun in North and nized by Wallace (1876), but it has taken simultaneously America, Asia, Europe, Africa. The taxa inNorth America at the start of dred years of intense paleontological study by Amegh present the were thus the survivors of earlier ino, Matthew, Scott, Patterson, Simpson, Webb, and interchange many tested and at others to clarify patterns of dispersal (see, for example, exchanges, repeatedly by immigrations Marshall 1981, 1985;Webb 1985;Webb and Marshall tempted immigrations from the Old World.

380 American Scientist, Volume 76 By contrast, during most of Cenozoic time South nent. The important point here is that the stratum 1 an were America was island continent, like Australia today. groups the first to radiate to fill land mammal As a result of this isolation, South American land niches and adaptive zones in South America in early mammals evolved in a world of their own; genera, Cenozoic time. families, and most orders were autochthonous and Stratum 2 groups include caviomorph ?for endemic to the continent, being found there and no example, and ?and monkeys, where else. When the Great American Interchange be both of which are first recorded in rocks of middle gan, the land mammals of South America were thus Oligoc?ne to early age. The oldest caviomorph a brought into contact with a major influx of potential rodents are known from level dated about 34 mya in mya competitors and predators for the first time in their ; monkeys first appear about 26 in history. In addition, South American taxa that dispersed (Marshall 1985; MacFadden et al. 1985). These groups toNorth America were first-time immigrants, entering a arrived either fromAfrica or (more probably) fromNorth fauna that had known numerous earlier invasions. America sometime during the late time or earlier, These differences in the histories of the North and South traveling by the process of waif dispersal across the American faunas signal the fact that aspects of the water barrier which then isolated South America from interchange will be different on each continent. the other continents (Patterson and Wood 1982). Waif dispersal is thought to occur during times of flooding South American land mammals and high water levels, when rafts of may break away from the banks of swollen and be sea. A brief look at the history of South American land carried to Some of these rafts, it is speculated, may mammals permits the identification of some direct con contain animals that can subsist on thematerials provid sequences of the interchange, as opposed to changes ed by the raft itself. These miniature "Noah's arks" were that inevitable or related to trends begun earlier. (McKenna 1973) may be carried by prevailing winds and The constraints imposed on patterns of dispersal by currents to distant shores; upon successful docking, the South America's status as an island continent allow us to voyagers disembark to colonize new lands. A prerequi a distinguish threemain strata of land mammals (Simpson site for successful waif dispersal is the survival of 1980;Fig. 3). pregnant female, a female with young, or a male-female a Stratum 1 consists of groups present in South pair able to perpetuate the in new land. The are America at or just before the beginning of the Cenozoic. chances of a successful crossing clearly low, and Included are Marsupialia, Proteutheria, , 1. The of the Panamanian land set in Condylarthra, and , which are first record Figure appearance bridge motion the event now known as the Great American ed in rocks of late age, and , Interchange, which resulted in a of the , and major restructuring widely differing , , , biotas of North and South America. The land mammals shown on which are first found in rocks of middle and late the and at the left are of the 38 South in and and de representative age Argentina (Marshall across American genera that walked north the land bridge; at the Muizon 1988; Marshall 1985). There ismuch debate as to right are representatives of 47 dispersants from North America that whether some or all of these evolved in South groups arrived in SouthAmerica byway of the land bridge. Unlike South or America from long-established stock wheth American taxa, which showed little diversification after their er they arrived there from elsewhere?Africa, North immigration, North American land mammals experienced an America, or Australia via Antarctica?just before or explosive diversification following their arrival on the South simultaneous with their first appearance on that conti American continent.

Cervidae

Ursidae

Tayassuidae

1988 July-August 381 rocks of late to age. Thinoba distes is firstknown from the early Hemphillian of and occurs with in faunas of middle Hem phillianage inFlorida and (Webb 1985). North American immigrants of the family are first found in South America in Argen tinian rocks of age dating from about 7.5 mya (Butler et al. 1984). The earliest procyonid , Cyonasua, was about the size of a large modern raccoon; by Ensenadan time this genus gave rise to the slightly

Miniature "Noah's arks"may be carried byprevailing winds and currentsto distant shores

larger Brachynasua, and later to the -size Chapalma lania(Marshall 1985). These waif dispersants had little im pact on the overall diversity of the faunas they joined. However, they did become firmly established and them selves show a low level of diversification. The members a of all three families appear to have been adapted to wide range of ; Thinobadistes and Pliometanastes were large generalized , whereas Cyonasua was the of Figure 2. The Panamanian land bridge formed by uplift was a large omnivorous (Webb 1985). This the earth's crust in the of the Bolivar marine barrier on region Trough adaptiveness would have been advantageous rafts, which once connected the Sea and the (colored area), where food was limited, because these voyagers could across what is now southern and have eaten available to them. Judg northwestern . Colored dots indicate sites at which virtually everything ing from relatives, Cyonasua may also have had of late Miocene to age have living interchange early abilities thatwould have been useful had the been found in Argentina, the southern and southwestern United swimming raft sunk before States, and . Virtually nothing is known about the early part docking. The second of was created the of the interchange in and northern South America. group immigrants by emergence of the Panamanian land bridge, which result ed from a combination of tectonic changes and decreases many voyagers that survived were no doubt unsuccess in the sea level related to ice-cap formation, the separate a are to Savin ful in establishing foothold. Nevertheless, many such effects ofwhich difficult isolate (Cronin 1981; American Noah's arks probably existed, and some containing and Douglas 1985). The first record of South across rodents and monkeys managed to reach South America. animals that walked north the newly emerged date Stratum 3 includes participants in the Great Ameri land bridge occurs in rocks of late age that are on mya. mammals and can Interchange. Two groups recognized, based from about 2.5 Seven genera of land were one almost in the time and mode of dispersal: taxa that waif large ground appear simultaneously in New dispersants in the lateMiocene, before the emergence of faunas of this age Florida, Texas, Mexico, across of the Panamanian land bridge, and taxa thatwalked , and . These immigrants consisted mya. two a arma the land bridge after its final emergence about 3 ( and Kraglievichia), giant dillo-likeglyptodont (), two ground a ( and ), (Erethi Waifs and walkers a zon), a large (), and phororhacoid EXtring lateMiocene time, a limited interchange of land ground bird (Titanis). mammals occurred between the Americas either by The most interesting of these early dispersants is or to have reached a of rafting across the Bolivar Trough by island-hopping Titanis, which is believed height im over m Phororhacoids were through the Antilles archipeligoes. South American 3 (Brodkorb 1963; Fig. 4). carnivorous showed marked migrants of the ground families flightless, ground that are inmiddle and late fau (Pliometanastes) and (Thinobadistes) first running specializations Tertiary nas recorded inNorth America in rocks of early Hemphillian in South America (Patterson and Kraglievich 1960; were age, i.e., by 8 mya (forNorth and South American ages, Marshall 1978). They the only large terrestrial was a on that continent when the land see Fig. 5). Pliometanastes, which the size of bridge modern black bear, makes its first appearance in local appeared, and Titanis was the only large South American faunas of this age in central Florida, , and carnivore to disperse to North America, where it is a central California. This family gave rise toMegalonyx, recorded in faunas of late Blancan and early occurs larger and more specialized that in age in Florida (Marshall 1977;Webb 1985). Phororha

382 American Scientist, Volume 76 coids have one MYA Period Stratum Land mammal distant living relative in Paraquay, Car groups a iama, long-legged, long-necked bird about 0.7 m tall, a capable of running at speed of 25 miles an hour; it resorts to spurts of short-distance flight only when necessary. A second major contingent of South American taxa that crossed land the bridge appears in rocks of early Irvingtonian age dating from about 1.9 mya in north western Mexico and numerous localities across the southern . These taxa include a giant = a ( Pampatherium), ground sloth a (=), giant (Myrmeco a phaga), and capybara (). An opposum () appears in the late Irvingtonian of Florida, another ground sloth (Meizonyx) in the Irvingtonian of , and a rhino-like toxodont (Mixotoxodori) in the Rancholabrean of southern Central America. Of these genera, the rhino-like Mixotoxodori and the are ground sloth Meizonyx known only from southern Central America, and the anteater Myrmecophaga only a from single site in northwestern Mexico (Webb and Perrigo 1984; Shaw and McDonald 1987). Records of the two latter have been established only within the last four years, demonstrating that some aspects of the North American are part of the interchange still poorly docu mented. We now know many of the South American dispersants that reached the southern part of the United States but little of what happened along the way. , Meizonyx, and Myrmecophaga did not go to the end of the highway north, and further research in 3. The fact that South America was once an Central America will no doubt reveal that other Figure isolated island disper continent makes it an ideal for the of sants also failed to do so. The number of laboratory study dispersal dispersants patterns. The first stratum of South American land mammals should decrease as one goes from Panama to the United consists of early groups, either indigenous or of unknown origins, States, but due to the of the fossil record and a vagarities that radiated to fill niches and adaptive zones within the continent. lack of of what in Central Ameri knowledge happened By contrast, stratum 2 consists of waif dispersants that arrived the reverse now true. across ca, is water barriers in the Oligoc?ne. Stratum 3 shows the impact The first unequivocal record of the presence of of the land bridge, with the appearance of 16 new families. Some North American land mammals that walked south Cricetidae and Procyonidae were included among these "walkers/7 across other and other Cricetidae as well, the land bridge occurs in Argentinian rocks of although Procyonidae, perhaps certainly arrived earlier water. Chapadmalalan age, dating from 2.8 to 2.5 mya. Two by are taxa represented: a skunk (Conepatus) and a (). A (Hippidiori) appears at about the same time in faunas of early Uquian age (2.5 mya) in 54 living genera of cricetids in South America. This northwestern Argentina. group is not discussed above with the waifs or the Evidence of themain of North American contingent walking dispersants because there is a great deal of dispersants begins to appear in rocks of late Uquian age debate about when and how it arrived in South America. in Sixteen nine (2 mya) Argentina. genera representing Webb (1985) has admirably summarized two possible families have been found: dogs (Dusicyon, Protocyon), scenarios. The first attempts to account for the remark and saber-tooths skunks {, ), (, able present-day diversity of cricetids by hypothesizing Stipanicicia), (Arctodus), -like gomphoth that they arrived as waif dispersants earlier than the first eres (), (Onohippidium), (Tapirus), documented appearance in time and (, , ), and (Hershkovitz 1966; Reig 1978). Reig observes that known and a (Blastocerus,Morenelaphus, Ozotoceros). (Sylvi fossils strongly suggest much earlier arrival, probably and in the at lagus) squirrels (Sciurus) appear ; the beginning of theMiocene, with themain episodes shrews and (Cryptotis), pocket (Orthogeomys), of development taking place in the northern and central are known in () only living , where fossil deposits are still poorly known. A faunas. slightly modified version of this theory holds that crice Field found in North America as as 9 mice, early tids arrived during the period of world-wide low sea mya, first in South America in faunas in appear Argentinian levels the late Miocene, when the Bolivar Trough of late Montehermosan from 3 to 2.8 mya. was narrower age, dating marine barrier and conditions fordispersal Two of these cricetids are genera represented there were optimal (Marshall 1979). and and four (Auliscomys Bolomys), additional genera A second scenario accepts the fossil record at face and in on (Akodon, Dankomys, Graomys, Reithrodori) appear value, building the assumption that the first appear Chapadmalalan faunas (Reig 1978). There are now about ance of cricetids in the late Montehermosan marks or

1988 July-August 383 approximates their time of arrival (Patterson and Pascual barrier to true steppe biota throughout its existence, a are 1972; Jacobs and Lindsay 1984). According to this scenar because no species adapted to an as io, cricetids either experienced yet undocumented involved in any phase of the interchange. The shifting radiation in southern Central America, and were thus distribution of subhumid on the or taxonomically diverse at the time of their arrival, they land bridge thus influenced the dispersal of the taxa an underwent explosive adaptive radiation following living in them during the time of the interchange. their arrival in South America. If this scenario is correct, cricetids were the firstNorth American group to walk south across the Panamanian land newly emerged Man has beencredited with the bridge. being in The total generic diversity of dispersant families sole or primaryagent in a processof North and South America through time is summarized was an overkill in Figure 5. In South America there exponential called "blitzkrieg" increase of genera in families that arrived from North America, whereas in North America the increase of on South American immigrants was significantly lower. The The history of savanna habitats the land bridge possible reasons for this difference will be explored has been studied in some detail (Raven and Axelrod below. 1975; Webb 1977, 1978). During glacial advances in temperate regions and at high tropical elevations, low Filter effects of the land areas in equatorial latitudes became cool and dry, result bridge ing in the shrinking of wet tropical habitats to savanna All the "walkers" preserved as fossils represent taxa that island-like refuges and the expansion of dryer were apparently tolerant of or specifically adapted to habitats (Haffer 1974; Van der Hammen 1974). The savanna a ecosystems. This indicates the presence of reverse occurred during times of glacial retreat. Several area continuous corridor or, at the very least, a shifting marine regressions occurred in the Caribbean dur mosaic of open-country habitats through the American ing times of glacial advance, providing optimal ecological savanna tropics (Webb 1978, 1985).Webb points out that the windows for the reciprocal dispersal of biotas a southern part of the land bridge probably served as between the Americas (Cronin 1981). One such regres sion is documented at about 3 mya (? 0.2), another about 2 mya, and a third about 1.4 mya. These times approximate those of the major episodes of reciprocal dispersal as recorded in the fossil record: 2.8 to 2.5 mya, 2 to 1.9 mya, and 1.4 mya. These "sister" dispersal events record the existence of savanna corridors, and explain the pulsations of interchange shown in Figure 5. as During times of glacial retreat, such today, the distribution of savanna consists of disjunct habitats (Fig. 6). In times of glacial advance, however, these habitats would have been united by a corridor along the eastern or side of the Andes?Webb's so-called "high road" a "Andean route" (1978). This corridor provided north savanna south route that permitted the dispersal of biotas within South America; more important, it contin ued across the Panamanian land bridge into the south ern United States, extending eastward into Florida. In were addition to creating this corridor, glacial advances m sea accompanied by drops of as much as 50 in level, resulting in a widening of the land bridge. During times of glacial advance the savanna habitats of the southern United States and southern South America were thus mutually accessible. The principal obstacles to a com plete intermingling of their biotas were distance and competitive exclusion. Many species in South America today have popula a tions restricted to disjunct savanna habitats, and similar situation is found in Central America and north ern South America, where some are restricted to 4. The bird Titaniswas species Figure flightlessphororhacoid ground savanna habitats now separated by 1,700 km of wet probably an important predator of small to medium-sized land as forest 1966, 1972;Webb mammals such capybaras. Titanis was the only large terrestrial tropical (Hershkovitz 1985). These to the existence of the carnivore from South America to take part in the interchange, and is disjunct populations testify last savanna corridor between the which was found in North America only in Florida. It was comparable in size Americas, to a m. in to Megatherium, attaining height of over 3 A figure is operation 12,000 10,000 years ago (Bradbury 1982; shown at the left for scale. Markgraf and Bradbury 1982).

384 American Scientist, Volume 76 Epoch North American Age South American Age

i i 9L_-1_L_I _L_I?_ 4020 20 0 40 80100 60 Generic diversity over the 9 million shows dramatic Figure 5. A comparison of the generic diversity of North and South American land mammals last years the there was an increase effect of the appearance of the land bridge 3 MYA (color). In addition, it is apparent that in South America exponential was in genera that arrived from North America, whereas in North America the increase of South American immigrants much lower. Four with of when the of a major dispersal events marked by sharp increases in diversity appear to coincide periods glacial advance, presence savanna corridor or a mosiac of open-country habitats may have facilitated immigration.

is the of new a During the time of the interchange, recurrent glacial until it disrupted by appearance taxa, a in or a combination of the events thus produced filtering effect that determined change physical environment, which types of animals could disperse and when. The two. taxa in the show expansion of the savanna during glacial advances has its Do the involved interchange 7 antithesis in the expansion of the tropics in times of evidence of turnover induced by immigration? Figure stratum 1 families in South glacial retreat such as the present. Taxa in wet tropical shows that the diversity of seldom if ever leave a fossil record, and we America began to decrease steadily in time and more in late therefore know virtually nothing about the dynamics of slightly sharply time, stratum families to there in the interchange during such periods. Many wet tropical when 3 began appear large in stratum 1 families between the forest taxa are present today in both southern North numbers. The decrease America and northern South America, but we have no middle Miocene and late Pliocene is offset by an increase one was record of which taxa went north and which went south in stratum 2 families, suggesting that being before the The during the last threemillion years. Our understanding of replaced by the other interchange began. 3 have accelerated the de the Great American Interchange is thus biased by the appearance of stratum may crease but continued decline of fact that the evidence comes only from times when the of stratum 1 families, the savanna habitat was at itsmaximum. stratum 1 families during late Cenozoic time clearly a It is represents trend begun before the interchange. some or most of the Faunal thus possible to speculate that dynamics decrease in the diversity of stratum 1 families during the to the of The success or failure of the dispersants can be investi time of the interchange is unrelated appearance stratum 3 decreased in gated by analyzing various aspects of taxonomic evolu groups. Among genera, diversity strata 1 and 2 in the late Pliocene and Pleistocene but tion?that is, by measuring changes in the total number same of taxa or in the number of taxa within over time increased markedly in stratum 3 during the time. (Marshall et al. 1982; Webb 1984). MacArthur and Wil The appearance of stratum 3 groups thus resulted in over a a in total but a son's equilibrium theory (1967) predicts that time only minor increase family diversity in region such as a continent will become saturated with significant increase generic diversity. that the taxa of taxa, reaching a level of diversity where rates of turnover It has been suggested -like are stratum and stochastically constant. Equilibrium will then persist 1?litopterns notoungulates?were actively

1988 July-August 385 birds, whose extinction in South America coincides with the arrival of this group (Marshall 1977). Among mam mals, the only potential example of competitive replace ment is the extinction of the saber- , which coincided with the arrival of the true saber-tooth placental, Smilodon. The concept of "prey naivete" has broad implica tions for the mixing of faunas (Diamond 1984). The r introduction of new predators in historic times has repeatedly shown that native species may be naive about new predators in general and predators in particular. The arrival of stratum 3 carnivores, "the likes of which southern had never before experienced," as Webb points out (1976,p. 225),may explain the factthat a decrease in stratum 1 and 2 ungulate-like forms coincided with an increase in stratum 3 Figure 6. Recurring glacial advances and retreats appear to have immigrant It is that the of played an important part in the rhythm of dispersal. In times of ungulates. possible replacement large stratum 1 and 2 herbivores stratum 3 was glacial retreat such as the present, savanna habitats contract into by ungulates rather than due to active disjunct areas, as shown at the left; in periods of glacial advance, passive, being competition. The however, these habitats were united by a corridor along the eastern immigrant carnivores may simply have killed off some of the a route that side of Andes (right), providing north-south the native prey, making room available for immigrant extended into the southern United States and east to Florida. prey. Arrows indicate route and avenues within South this of dispersal At the end of the last 12,000 to 11,000 America. glaciation, years ago, migrated out of Asia by way of the Bering land bridge, passing southward through North America and into South America. Within possibly 1,000 man or replaced by immigrating stratum 3 ungulates, which years, apparently occupied most all of the New included horses, camels, and deer (Webb 1976). Howev World (Martin 1973). Humans were the last stratum 3 er, much of the extinction of native ungulate-like taxa in dispersants to South America, and their arrival coincides South America took place before the interchange, and with the extinction of most large-bodied taxa in both was thus unrelated to the arrival of stratum 3 groups North and South America (Marshall et al. 1984). Man has or (Patterson and Pascual 1972). The replacements that been credited with being the sole primary agent in this a occurred during the interchange involved an interplay phenomenon through process of overkill called "blitz in among the native ungulates, some ungulate-like cavio krieg," which sudden followed the initial a mass morph rodents, ground sloths, , and the colonization of land inhabited by animals that were new immigrant North American taxa (Marshall 1981). There especially vulnerable to the human predator. no The is clear evidence of replacement of any ungulate disappearance of these large-bodied species in the at group during interchange time due to rampant competi Americas the end of the Pleistocene is certainly a a tion. consistent with such process, although model based on a Among the large carnivorous groups, the doglike climatic change remains viable alternative (Markgraf borhyaenid were extinct before the time of 1985). ex the interchange. If any competitive interaction with the Whatever the cause, the consequences of this invading carnivores of stratum 3 occurred, itwould have tinction event in South America were dramatic. Of the 37 been with the large flightless phororhacoid ground land mammal families recorded in Lujanian faunas, 8

Stratum Stratum Stratum Total MYA 2 1 3 diversity Stratum i Stratum 2 Stratum 3 Total diversity

10

20

30

40 I

50 110 families

60

70

Figure 7. Spindle diagrams of family and generic diversity among South American mammals help to distinguish trends begun before the a interchange from the direct effects of the land bridge. The decline of stratum 1 families visible at the far left clearly represents trend predating the appearance of the land bridge 3 MYA, suggesting that this decrease in diversity is unrelated to the arrival of stratum 3 families. 1 For genera, however, as shown at the right, diversity simultaneously decreased in strata and 2 and increased significantly in stratum 3, demonstrating some replacement between native and immigrant taxa. (After Marshall and Cifelli, in press.)

386 American Scientist, Volume 76 (21%), were extinct by time, even though all 8 North America is unique and asymmetrical. Several families included more than one genus. Of the 153 land theories have been proposed to explain this difference. was mammal genera documented in Lujanian faunas, 56 The classic view during the firsthalf of this century were now (37%) extinct, including 35 belonging to those that the taxa of North America were competitively 8 families. Of the extinct genera, 54 (96%) were of large superior to those of South America. This view was body size, suggesting that this event was selective; the founded on the belief that the taxa existing in North taxa were a were that became extinct clearly not random America at the beginning of the interchange the sample of Lujanian land mammal fauna (Marshall and survivors of numerous earlier invasions and thus tested Cifelli, in press). and "worldly wise/' possessing such varied advantages narrower as more rapid reproductive rates, niche selec new The success ofNorth American taxa tion, and more rapid evolutionary responses to opportunities(Webb 1985). or more sees In evaluating the relative success failure of North and A recent view, compatible with this, the South American interchange species, two factors must North American taxa as "insinuators" able to exploit be kept inmind. First, it is necessary to distinguish true niches and adaptive zones not occupied by native South are or a dispersants, which represented by the same American taxa (Patterson and Pascual 1972; Hershkovitz on sister their native continent, from pseudodis 1972). Thus the North American immigrants would not are persants, which derived from true dispersants. Just have competed directly with South American natives, a a were because genus belongs to family that dispersed from but able simply to radiate and fill unoccupied another continent, it does not follow that it itself dis space. Another theory holds that ongoing geological persed. For example, the elephant-like a Cuvieronius, true dispersant to South America, is known in North America. from pre-interchange faunas the and After reaching South America it gave rise toHaplomasto Nearly halfof families genera don and , which are pseudodispersants; now on the South American continent although they belong to a family that dispersed from in and are to that North America, they themselves evolved belong groups emigratedfrom endemic to South America (Webb 1985). Such pseudo NorthAmerica the last three dispersants must be identified and factored out when during analyzing dispersal events. Second, the number of po million years tential dispersants is directly related to the size of the source faunas. A larger geographic area will predictably more a have taxa and hence more dispersants than activity created new habitats and changed the old ones, some smaller area. This, too, must be taken into account. resulting in the extinction of native taxa before and Of the members of South American families that during the time of the interchange. The opening of walked toNorth America, all 38 may be regarded as true niches and adaptive zones through the disappearance of dispersants, indicating that little or no diversification prey-naive natives would have allowed theNorth Amer vacuums. occurred after their immigration. By contrast, of the ican immigrants to disperse into ecological can North American walkers (including cricetids) only 47 These features would have facilitated both the arrival of be regarded as true dispersants, whereas 72, from 8 true dispersants and their radiation into pseudodisper families, represent pseudodispersants. Thus about 60% sants. This theory implies that the changes which took of the North American genera in South America appar place in the composition of the South American land ently evolved in situ on that continent, demonstrating mammal fauna were due primarily to passive replace were that these immigrants experienced considerable diversi ment, and that the North American groups simply fication after their arrival. timely invaders. The total surface area of North America and Central A fourth theory combines aspects of all these views, America (24 million km2) is greater than that of South suggesting that the two principal evolutionary theaters America (18 million km2). This fact, along with the for the interchange were not South and North America known fossil record, explains why North America had but South America on the one hand and North America an average of 60% greater generic diversity and hence and on the other (Webb 1985). This theory area more potential dispersants than South America during stresses the imbalance between the available to the time of the interchange (Marshall et al. 1982). When northern taxa adapted to temperate conditions and that the differences in the size of the source faunas are taken available to the southern taxa, and holds that the pres are a area into account and only true dispersants considered, ence of vast Holarctic staging rather than any the interchange can be seen to be balanced, with the number of true dispersants proportional to the size of the source faunas in both continents. More taxa dis Larry G. Marshall is Senior Research Scientist at the Institute ofHuman He Northern Arizona and to South America than toNorth America Origins, Berkeley. graduated from University, persed simply was awarded hisM.Sc. Monash (Australia) in 1974 and his because there were more inNorth from University potential dispersants Ph.D. in 1976 from theUniversity of California, Berkeley. Since 1975 he has America. on been working the evolutionary history of South American land mammals, This of the is aspect interchange predicted by equi and has participated in 17 expeditions toArgentina, Bolivia, Brazil, , librium the diversifica theory. However, later explosive Colombia, Mexico, , and . Address: Institute ofHuman as to 2453 CA 94709. tion of true dispersants in South America opposed Origins, Ridge Road, Berkeley,

1988 July-August 387 transamerican faunal In The Great American Biotic Inter inherent biological superiority explains the greater suc interchange. change, ed. F. G. Stehli and S. D. Webb, pp. 49-81. Plenum. cess of northern taxa after the interchange. Marshall, L. G., and R. L. Cifelli. In press. of Cenozoic Although the reasons for the success of the North Restructuring land mammal faunas of South America. In Long-term Restructuring in American remain debatable, the on immigrants impact Late Cenozoic Terrestrial Ecosystems, ed. K. Luchterhand. Evanston: South America is undisputed. When the interchange Wm Caxton, Ltd. about 7.5 60% American began mya, about of the South Marshall, L. G., and C. de Muizon, 1988. The dawn of the Age of families derived from stratum 1 and 40% from stratum 2; Mammals in South America. Nat. Geog. Res. 4:23-55. L. 70% of the genera came from stratum 1 and 30% from Marshall, G., S. D. Webb, J. J. Sepkoski, and D. M. Raup. 1982. stratum 2. the of South American Mammalian evolution and the Great American Interchange. Science Today composition 215:1351-57. land mammal fauna is quite different: 19% of the families Marshall, L. G., et al. 1984. of the continental mammal come from stratum 1, 37% from stratum 2, and 44% from Geochronology bearing of South America. Palaeovertebrata, Mem. Extr. stratum 3; 17% of the genera derive from stratum 1, 29% 1984:1-76. from stratum and 54% from stratum 3. These data 2, Martin, P. S. 1973. The discovery of America. Science 179:969-74. demonstrate that the Great American firmly Interchange McKenna, M. C. 1973. Sweepstakes, filters, corridors, Noah's arks, a and in resulted in major restructuring. Nearly half of the beached Viking funeral ships paleogeography. In Implications Continental to the Earth ed. D. H. and S. K. families and genera now on the South American conti of Drift Sciences, Tarling nent to that fromNorth Ameri Runcorn, pp. 21-46. Academic Press. belong groups emigrated -. ca 1975. Fossil mammals and early Eocene North land during the last 3 million years. continuity. Ann. Miss. Bot. Garden 62:335-53. Patterson, B., and J. L. Kraglievich. 1960. Sistematica y nomenclatura aves de las forrorracoideas del Plioceno argentino. Pub. Mus. Munie. Cienc. Nat., Mar del Plata 1:1-49. References Patterson, B., and R. Pascual. 1972. The fossil mammal fauna of South Bradbury, J. P. 1982. Holocene chronostratigraphy of Mexico and America. In Evolution, Mammals, and Southern Continents, ed. A. Central America. In Chronostratigraphic Subdivision of theHolocene, ed. Keast, F. C. Erk, and B. Glass, pp. 247-309. State Univ. of New York J.Mangerud, H. J. B. Birks, and K. D. Jager. Striae 16:46-48. Press. Brodkorb, P. 1963. A giant from the Pleistocene of Patterson, B., and A. E. Wood. 1982. Rodents from the Florida. Auk 8:111-15. Oligoc?ne of Bolivia and the relationships of the . Bull. Butler, R. F., L. G. Marshall, R. E. Drake, and G. H. Curtis. 1984. Mus. Comp. Zool. 149:371-543. and 40K-40Ar of lateMiocene Magnetic polarity stratigraphy dating Raven, P. H., and D. I. Axelrod. 1975. History of the flora and fauna of and early Pliocene continental deposits, Catamarca Province, NW Latin America. Am. Sei. 63:420-29. /. Geol 92:623-36. Argentina. Reig, O. A. 1978. Roedores cricetidos del plioceno superior de la T. M. 1981. Rates and causes neotectonic vertical Cronin, possible of Provincia Buenos Aires (Argentina). Pub. Mus. Munie. Cienc. Nat., crustal movements of the emerged southeastern United States Mar del Plata 2:164-90. Atlantic Coastal Plain. Geol. Soc. Am. Bull. 92:812-33. A. Repenning, C. 1980. Faunal exchange between and North M. 1984. Historic extinctions: A rosetta stone for under Diamond, J. America. Can. J. Anthropol. 1:37-44. extinctions. In Extinctions: A Prehis standing prehistoric Quaternary Russell, D. E., and R.-J. Zhai. 1987. The of Asia: Mammals toricRevolution, ed. P. S. Martin and R. G. Klein, pp. 824-62. Univ. of and Stratigraphy. Mem. Mus. Nat. Hist. Natur., Sei. de la Terre 52:1 Arizona Press. 488. Haffer, J. 1974. Avian in South America. Pub. Nuttall speciation tropical Savin, S. M., and R. G. Douglas. 1985. Sea level, climate, and the Ornith. Club 14:1-390. Central American land bridge. In The Great American Biotic Inter P. 1966. land and Latin America faunal Hershkovitz, Mice, bridges change, ed. F. G. Stehli and S. D. Webb, pp. 303-24. Plenum. In Parasites Panama, ed. R. L. Wenzel and V. J. interchange. of Shaw, C. A., and H. G. McDonald. 1987. First record of 725-47. Field Museum of Natural Tipton, pp. : History. (Xenarthra, Myromecophagidae) in North America. Science 236:186 -. 1972. The recent mammals of the Neotropical Region. In 88. Evolution, Mammals and Southern Continents, ed. A. F. Keast, F. C. Simpson, G. G. 1980. Splendid Isolation: The Curious History of South Erk, and B. Glass, pp. 311-431. State Univ. of New York Press. American Mammals. Yale Univ. Press. Jacobs, L. L., and E. H. 1984. Holarctic radiation of Lindsay. Van der Hammen, T. 1974. The Pleistocene changes of vegetation and Muroid rodents and the of South American cricetids. /. Vert. origin climate in tropical South America. /. Biogeog. 1:3-26. Paleont. 4:265-72. Wallace, A. R. 1876. The Geographical Distribution ofAnimals. Macmillan. MacArthur, P. H., and E. O. Wilson. 1967. The Theory of Island Webb, S.D. 1976. Mammalian faunal dynamics of the Great American . Princeton Univ. Press. Interchange. Paleobiology 2:216-34. MacFadden, B. J., et al. 1985. and Magnetic polarity stratigraphy -. 1977. A of savanna in the New World. Part mammalian of the Deseadan history biostratigraphy (Middle I: North America. Ann. Rev. Ecol. Syst. 8:355-80. Early Miocene) Salla Beds of northern Bolivia. /. Geol. 93:223-50. -. 1978. A of savanna vertebrates in the New World. Part V. 1985. faunal extinctions in southern history Markgraf, II: South America and the Great Ann. Rev. Ecol. Science 228:1110-12. Interchange. Syst. . 9:393-426. and P. 1982. Holocene climatic of Markgraf, V., J. Bradbury. history -. 1984. On two kinds of rapid faunal turnover. In Catastrophies South America. In Chronostratigraphic Subdivision of theHolocene, ed. J. and Earth History, ed. W. A. and J.A. Van Couvering, pp. H. B. Birks, and K. D. Striae 16:40-45. Berggren Mangerud, J. Jager. 417-36. Princeton Univ. Press. Marshall, L. G. 1977. Evolution of the carnivorous zone in adaptive -. 1985. Late Cenozoic mammal between the Ameri South America. In Patterns in ed. M. K. dispersals Major Evolution, cas. In The Great American Biotic Interchange, ed. F. G. Stehli and S. D. Hecht, P. C. Goody, and B. M. Hecht, pp. 709-21. Plenum. Webb, pp. 357-86. Plenum. -. 1978. The terror bird. Field Mus. Nat. Hist. Bull. 49:6-15. Webb, S. D., and L. G. Marshall. 1982. Historical of -. biogeography 1979. A model for of South American Recent paleobiogeography South American land mammals. In Mammalian Biology in cricetine rodents. 5:126-32. South Paleobiology America, ed. M. A. Mares and H. H. Genoways, pp. 39-52. -. 1981. The Great American An invasion induced Interchange: Special Pub. 6, Pymatuning Laboratory of Ecology, Univ. of crisis for South American In Biotic Crises in and mammals. Ecological Pittsburg. Time, ed. M. Niticki, pp. 133-229. Academic Press. Evolutionary Webb, S. D., and S. C. Perrigo. 1984. Late Cenozoic vertebrates from -. 1985. Geochronology and land-mammal biochronology of the and El Salvador. /. Vert. Paleont. 4:237-54.

388 American Scientist, Volume 76