by George Gaylord Simpson (1943)
The living animals of South America are so distinctive that the students of biological geography and faunal zones set this continent apart as the Neotropical Region. Besides South America, this region is usually taken to include Central America and tropical Mexico, and the West Indies are also frequently included in it, so that its boundaries nearly coincide with those of Latin America. Disregarding the marginal areas, for the moment, South America is known to have been a great center of evolution where many types of animals have originated and [[p. 414]] it has also served as a refuge where animals have survived after they became extinct elsewhere. The origin of the mixture that has thus come to make up the unique Neotropical fauna presents a multitude of interesting and baffling problems that have attracted popular curiosity and scientific study since the days of the conquistadores.
The composition of a fauna is the result of a long sequence of historical events, a history extending over many millions of years. New groups of animals evolve and they tend to spread as widely as geographic, climatic, and other factors will permit. In the meantime old types disappear, either because they have evolved into something different or because they did not successfully meet changing conditions and became extinct. Understanding the origin of any fauna involves following a remarkably complex succession of changes not only within the fauna itself but also in all the surrounding circumstances and especially in those factors of geography and climate that either prevent or promote the migrations of animals. The peculiarity of the Neotropical fauna, specifically, is due, more than to anything else, to a series of unusual geographical conditions during the last 50,000,000 to 75,000,000 years affecting animals that normally live on land or in fresh water.
These conditions have influenced all the life of Latin-America, and the Neotropical fauna is recognizably distinctive among many different sorts of animals, such as worms, insects, fresh-water and land shells, lake and river fishes, amphibians, and the non-marine reptiles and mammals. In general the peculiarity of the Neotropical animals is inversely proportional to the facility with which they spread and especially with their ability to cross marine waters, being least among animals that swim and can survive immersion in salt water or that can fly for considerable distances, and greatest among animals that live on dry land and that only rarely and accidentally cross bodies of water. Since there are no animals that cannot somehow cross water, this distinction is only relative, but it is often well marked.
Complete understanding of the fauna will require the study and interpretation of all of its constituent animals and also of all the evidence of botany, geology, geography, climatology, and some other sciences. Complete attainment of this goal is still far in the future, but remarkable progress is being made. Such historical study is particularly difficult and inconclusive [[p. 415]] for groups of animals that are not preserved or have not been found as fossils, for instance for worms, which are very rarely preserved as identifiable fossils, or for birds, the fossil bones of which are readily identified when found but of which far too few have as yet been found in South America to provide good evidence on the origin of the Neotropical bird fauna. Historical study must be devoted first of all to groups of which a considerable number of South American fossils are known, and then other groups must be interpreted in the light of this evidence.
For the study of the Neotropical fauna this means attention must be primarily focused on mammals, because more fossil mammals have been found in South America than all other non-marine fossil animals put together. As a background for the present study, the briefest possible summary of conclusions concerning the origin of the mammalian fauna will be given.1 This paper is devoted to a very different group of animals, one that is much less well known historically but regarding which some fairly positive conclusions are nevertheless now possible: the Chelonia, an order of reptiles comprising the turtles and tortoises. They will be considered in the light of what is already known from other sources and of the evidence more directly concerning them. Some of the most important items of this evidence are so new that they had not appeared in print when this paper was written. It will be seen that the probable interpretation of the Neotropical chelonians not only agrees well with the outlines of mammalian history but also supplements parts of that history in an interesting way.
PALEOGRAPHY AND NEOTROPICAL MAMMALS.
The most important fact about the ancient geography of South America is that it was an island continent during most of the Tertiary period, the Age of Mammals. At about the beginning of that period or toward the end of the preceding period, the Cretaceous, there must have been a land connection with other parts of the world, but this was certainly broken, sunk beneath the sea, early in the Tertiary and remained so until near the end of the period, when connection with North America was established. These facts are in no doubt, but there is dispute regarding three important questions of detail: (1) where [[p. 416]] was the early or pre-Tertiary land connection, (2) did any land animals enter South America while it was isolated and if so, how, and (3) just when was the North American connection completely established, or re-established?
No attempt will be made here to discuss these questions, which are not absolutely vital for the present topic, and I shall only give my own answers to them, with the understanding that these answers are not conclusive and that excellent authority can be found for rejecting them. (1) It is most likely that the early connection, like the later connection now in existence, was with North America.2 (2) It is fairly certain that a few, random groups of land animals did enter South America while it was completely surrounded by the sea; these may have had more than one origin, but some, at least, probably came from North America across the Central American strait then existing, facilitated by islands in that flooded region, making it a sweepstakes route3 before it was a land bridge. (3) It is probable that the first land migrants from North America followed that sweepstakes route and that complete land communication did not exist until almost the end of the Tertiary, in the latter part of its last epoch, the Pliocene.
In correlation with this geographic history (largely but not solely deduced from their evidence) the living and extinct South American mammals can be divided into four principal groups:
1. Extinct forms, for the most part unlike any living mammals or any mammals formerly native to other parts of the world, that developed in South America while it was isolated and whose ancestors were the very primitive stocks that reached there around the beginning of the Tertiary. Some of these, e.g. pyrotheres and astrapotheres, never reached any other continent, as far as known. Others, e.g. glyptodonts and ground-sloths, spread to North America late in their history.
2. Surviving mammals that evolved in South America, like the mammals in the first category, but that did not become [[p. 417]] extinct. These are the Neotropical mammals in the fullest, most characteristic sense, although they are now a small minority in the Neotropical fauna. Examples: armadillos, tree sloths.
3. Waifs that entered South America by some sweepstakes route while the continent was isolated. Although some of their evolutionary branches died out, the major groups that belong in this category survived in each case and their members are now among the most typical Neotropical mammals. Those that entered early in the Tertiary, e.g. the monkeys and the porcupine-like rodents, had time to develop varied and distinctive families in South American isolation. Later migrants of this sort, e.g. the coatis and their allies, became generically differentiated but still belong to North American families.
4. North American mammals that reached South America at or since the end of the Tertiary over the Central American land bridge. Some of these later became extinct on both continents, e.g. horses and mastodons. Others survived only in the Neotropical Realm and so are now considered as characteristic Neotropical animals although they are historically Nearctic, that is, North American, for example tapirs, llamas. Others still survive in both the Neotropical and Nearctic regions, e.g. pumas, deer, mouse-like rodents. The period of time involved was considerable. Many of the earlier immigrants from North America have been in South America long enough to develop distinctive genera (but not families) there and even the later migrants among mammals are, as a rule, specifically distinct in the Neotropical Region.
REVIEW OF NEOTROPICAL CHELONIA.
Turning to the Chelonia, what is known of past and present distribution of the Neotropical tortoises and fresh-water turtles will be summarized family by family before reviewing the geographic relationships of the South and Central American Chelonia as a whole.
Chelydridae: The snapping turtles are typically North American but the single species, Chelydra serpentina, ranges as far as Ecuador. A very isolated genus, Devisia, occurs in New Guinea. No fossils of this family are known from South America, but they occur in the Miocene of North America and Europe. These facts, together with the very close relationship of North and South American forms, make it highly probable [[p. 418]] that the Chelydridae are old, Holarctic4 forms and that the South American representatives are late invaders.
Kinosternidae: This family, the musk turtles and mud turtles, is confined to the New World. It is most differentiated in southern North America but ranges throughout Central America and as far as Bolivia and Brazil. The fossil history is practically unknown, the oldest fossil known to me being from the early Pleistocene of Arizona. The distribution suggests origin in southern North America or in the ancient part of Central America, in what is now but was not then part of the Neotropical Realm, but no fully positive conclusion can be reached until more fossils are found.
Dermatemydidae: This small family of fresh-water turtles is now restricted to Central America. It is in some respects intermediate between the two preceding families and may have arisen from the same ancestry, or may represent that ancestry in a general way. Members of this group were common in North America during the Cretaceous and early Tertiary and in the European early Tertiary and no sure trace of them has ever been found south of Guatemala and British Honduras. This strongly suggests that they evolved in the north and that in the increasingly cool later Tertiary they became concentrated in the warmest part of the northern continent and were then incorporated in the Neotropical Region, more or less accidentally from the point of view of their own history, when this faunal region spread north over the isthmus. The stronger evidence for this family reinforces the possibility that the related Kinosternidae had a similar history, although they now have more northern representatives.
Testudinidae: The family of the true tortoises and their allies is sometimes divided into two, and these will here be considered subfamilies and discussed separately.
The Emydinae are more aquatic than the Testudininae. They range over almost the whole world except in Australia and in the colder regions, but they are particularly characteristic of North America and southern Asia and adjacent islands, that is, of the Nearctic and Oriental regions. The collective [[p. 419]] genus Chrysemys, now often split into three or four genera or subgenera, includes many of the so-called terrapins of the United States, the painted tortoise, "cooter," and others. The genus or subgenus Trachemys of this group ranges from the United States to the West Indies, Central America, and as far south as southern Brazil and Argentina. The genus itself is not known before the Pleistocene, but it has allies throughout the Tertiary in North America and it seems probable that this was its place of origin and that it was a latest Tertiary or Pleistocene invader in South America. Like Kinosternum and some others, it is probable, but as yet hypothetical, that the genus was already at home in part of Central America before that area was incorporated in the Neotropical Region and that this promoted its subsequent inclusion in the Neotropical fauna.
The other Neotropical emydine, Geoemyda, has a very peculiar distribution, since it occurs in Central America, Venezuela, Brazil, and adjacent countries and also throughout the Oriental Region but not in Nearctic North America. It has no known fossil history in the Americas, but a close ally occurs in the Eocene of Japan.5 The insufficient evidence suggests that the genus was formerly common to North America and Asia and that it survived the Pleistocene cold only in the warm est parts of both continents. Its incorporation into the Neotropical fauna would then be latest Tertiary. A possible but less probable hypothesis is that the genus already existed at the beginning of the Tertiary, then became isolated in South America, and moved into Central America at the end of the Tertiary. Only the discovery of fossils in North or South America can determine this point with certainty.
The problem has been complicated rather than solved by the discovery of a possible emydine in Patagonia, supposed by its describer, Staesche, to be Cretaceous in age but probably from the Paleocene.6 The remains do not suffice for exact identification, but Staesche thought that they belonged to a new genus allied to Gyremys of the North American Cretaceous. The affinities of Gyremys are also uncertain, but it may belong in the Emydinae. The discovery could mean that the Emydinae are old natives in South America, although this would not prove [[p. 420]] that the living genera arose there. The identification is, however, too uncertain to warrant much speculation.
The Testudininae, true tortoises, are also distributed throughout the tropical and warm temperate regions of the globe except Australia. Of the several genera of this subfamily only the commonest and most widespread, Testudo, occurs in the Neotropical fauna. The South American species, including those so highly differentiated in the Galápagos Islands, may represent the derivatives of a single geographic stock. Several fossil forms have been found in the late Tertiary of Argentina7 and these could be interpreted either as earlier members of a South American group or as invaders accompanying the mammals then beginning to immigrate from North America. More light is cast on this problem by the discovery of a true Testudo in the Miocene of Patagonia, in beds older than the beginning of the mammalian invasion from North America.8 This new species is of a type suitable to be ancestral to the later South American and Galápagos Islands tortoises. From the evidence of the abundant fossil tortoises of North America and elsewhere, it seems probable that the genus Testudo, proper, did not evolve until after South America was isolated at the beginning of the Tertiary. Thus the new discovery makes it probable that Testudo reached South America during the first half of the Tertiary, while that continent was isolated. In that case it crossed a sea barrier, which is a reasonable possibility because living tortoises of this genus float and can survive long periods in salt water. This also contributes to the problem of the origin of the Galápagos Islands tortoises. They almost certainly reached the islands overseas from South America and not by any land connection. Such an event is improbable in a short period of time, but now that we know that appropriate ancestors were in South America at least as early as the Miocene, ample time is available to make this spread probable. This also allows time for the specific differentiation of the tortoises within the islands.
Pelomedusidae: This family of snake-necked, river turtles occurs in South America, Madagascar, and Africa and the one South American genus, Podocnemis, has a species in [[p. 421]] Madagascar. This peculiar distribution has long attracted attention and it is one of the standard arguments for great land-bridges formerly directly connecting the southern continents. Recent fossil discoveries support a very different explanation.
Podocnemis has been continuously in South America since the late Cretaceous. It is known in the Cretaceous of Brazil9 and the Eocene of Peru.10 In the supposed Cretaceous, probably Paleocene, of Patagonia were found some fragments referred by Staesche11 to Naiadochelys but barely different from Podocnemis and perhaps to be placed in that genus. Podocnemis has now also been found in the mid-Tertiary of Venezuela12 and of Brazil.13 Although not numerous, these discoveries suffice to prove that the present South American distribution of the genus is of long standing. Now the genus has also been found in the Cretaceous of North America.14 Its associations there are purely marine and prove that this genus, now exclusively fresh water, could and did enter marine waters. The genus is also known from the Paleocene, Eocene, and Miocene of Africa, the Eocene, Oligocene, and Miocene of Europe, and the Eocene of India. The evidence as a whole proves that Podocnemis was once nearly world-wide, that the present limited distribution indicates relict survival in part of its former area and not any direct connection between the regions now occupied, that migration occurred through and along the northern continents, and that in any case no continuous land bridges were necessary to allow the early spread of the genus. It is not now determinable whether the late Cretaceous Podocnemis originated in South America and migrated thence to North America and the rest of the world or whether it originated elsewhere and migrated to South America. Possible Mesozoic ancestors are known in North America and origin there may be somewhat more probable, but this is uncertain.
Chelyidae: This is the most characteristic chelonian element [[p. 422]] in the Neotropical fauna, which includes six genera of the family, all confined to that realm. The best known but most aberrant form is the matamata, Chelys fimbriata, and all the common, smaller side-neck or snake-necked, river turtles of South America are chelyids. It is interesting and significant that chelyids do not occur in Central America or west of the Orinoco-Amazon basin. The family includes four genera in Australia and New Guinea and this fact has been adduced as further evidence of former union of the southern continents.
There can be little doubt that this is an old group in South America, but its fossil record there is poor. It is known in Argentina from beds possibly Miocene and surely Pliocene and Pleistocene in age.15 Ameghino also reported chelyids in the Guaranitic, a series late Cretaceous and early Tertiary in age, but he gave no details. This early occurrence is probable and some description of the specimens is urgently needed. In any case, the South American genera appear to have arisen there and they must represent the differentiation of an old stock in that continent. The most reasonable explanation of the Australian genera is that they represent a similar but independent differentiation from the same very remote ancestry, which probably reached Australia from Asia and not more directly from South America. This probability is greatly enhanced, first, by the occurrence of a probable chelyid in the Eocene of India and, second, by the fact that forms near or in the ancestry of Pelomedusidae and Chelyidae occurred in the Cretaceous of North America.16
Meiolaniidae: This extinct family is mentioned because it includes the only known South American fossil turtles that are not closely allied to living forms and because it has been supposed to give evidence of continental connections with Australia. It was formerly stated, and the statement is still sometimes reported by uncritical or ill-informed students, that the extinct horned turtle Meiolania occurs both in Australia and in South [[p. 423]] America and that it was allied to the Chelyidae, which now occur on those two continents. I have, however, recently pointed out (a) that the Australian Pleistocene Meiolania and the South American Cretaceous or Eocene Niolamia and Eocene Crossochelys are not identical genera but are very distinct, (b) that they probably had a remote common ancestry but are only doubtfully referable to the same family, (c) that they are not related to the Chelyidae or the suborder Pleurodira but are specialized survivors of an older, essentially Mesozoic group that was widespread in the Northern Hemisphere, and (d) that the known facts about these genera do not support and cast no light on the postulated South American-Australian land connection.17 Niolamia and Crossochelys cannot be ancestral to any living chelonians and they have no direct bearing on the origin of the recent Neotropical fauna.
Trionychidae: Among the negative features of the recent Neotropical chelonian fauna, the complete absence of the soft-shelled river turtles, trionychoids, is most striking. Ameghino reported a "Trinnyx argentina" in beds that would now be considered late Cretaceous or early Tertiary.18 As far as I know, the specimens have never been clearly described and this is still a nomen nudum. It was omitted from Ameghino's later work, including an expansion of this same study. Without further data, the record cannot be accepted. Trionychoids were already common in North America before the end of the Cretaceous. There is no impossibility but there is some improbability in their early South American occurrence. Chelyids are very probably of equal antiquity in that continent, and chelyids and trionychids are now geographically exclusive and perhaps ecologically incompatible. Yet their ancestors may have lived together. If so, it is now impossible to say whether (a) the groups expanded together and subsequently had a lethal competition with success for the chelyids in the south and for trionychids in the north or (b) whether the chelyids arose first and were not able to hold their own against later, more progressive trionychids except with the protection of geographic and climatic barriers against the latter.
[[p. 424]] SUMMARY OF GEOGRAPHIC RELATIONSHIPS.
From a paleographic point of view, the Chelonia of the Neotropical fauna fall into the same four categories as the mammals, with greater or lesser certainty in different cases:
1. Extinct, specialized forms strongly aberrant from the point of view of the fossil Holarctic or recent Neotropical faunas: Niolamia and Crossochelys. It is somewhat odd that no such chelonians are known from the mid-Tertiary, when the peculiarity of the South American mammalian fauna reached its apogee, but this may be due to lack of discovery.
2. Surviving ancient endemic forms which, whatever their ultimate origin, had ancestors in South America by about the end of the Cretaceous and become differentiated there during Tertiary isolation: Pelomedusidae and Chelyidae (the Suborder Pleuorida). The Neotropical differentiation of the Pelomedusidae was specific only, within the formerly more widespread genus Podocnemis, and that of the Chelyidae was generic. Extinct aberrant species are known, but none of the fossils surely belongs to extinct genera.
3. Waifs that reached South America during its marine isolation: Testudininae, Testudo only. The recent species (including those of the Galápagos Islands) probably represent the survivors of a late Tertiary specific radiation, some of the lines of which became extinct.
4. Members of the Tertiary North American fauna that became incorporated in the Neotropical fauna since the rise of a land bridge in the late Pliocene: Kinosternidae, Dermatemydidae, Chelydridae, and in the Emydinae Trachemys and probably Geoemyda.
Like the mammals, the chelonians of group 4 can be divided into those that survive only in the Neotropical fauna, the Dermatemydidae and Geoemyda, and those that also survive in the Nearctic fauna, the Kinosternidae, Chelydridae, and Trachemys.
In one important and interesting respect the chelonians do not merely agree with the faunal history already and better known for mammals but also supplement this history by a factor less clear for mammals but also applicable to them. Even for the chelonians the evidence on this point is indirect, but it is highly suggestive.
That part of the present Neotropical Region that lies north [[p. 425]] and west of Colombia was not part of Tertiary South America, where the fully endemic Neotropical forms arose. Variable portions of this region were either under water, constituted impermanent islands, or were united to North America. Their fauna was then perhaps wholly and certainly predominantly of immediately Nearctic origin and affinities. There is abundant indirect evidence of this and some direct evidence that is almost conclusive in itself.19 Climatic zones must have existed in North America during the Tertiary. They may have been broader and less sharply differentiated in the early and middle Tertiary, but they became more marked toward the later Tertiary as shown, among other things, by the southward movement of the margins of subtropical and warm temperate forests. In the glacial stages of the Pleistocene the climatic extremes and gradient must have been greater than now. Under such conditions the most southern areas, to some extent southeastern United States and to greater extent Mexico and Central America, had a differentiated warm climate fauna zonally different from that of the rest of the continent although of the same origin. Here species and in some cases genera evolved that could not tolerate the colder conditions of the middle latitudes. Others that had occurred in the middle latitudes became confined to this warm zone. Some of these spread again to the north in inter-glacial and post-glacial times but some continued as southern zonal forms.
Thus Central America and Mexico had a fauna that was distinctive but that was historically Nearctic. The North American invaders into South America necessarily came from this region and not from the middle latitudes which are paleontologically much better known. Some of these "warm Nearctic" forms stayed in middle America and neither migrated to South America nor back into North America, proper. Since a great part of South America was climatically similar to Central America, many of the surviving endemics of that continent also occupied this region when continuity was established and the present nominally Neotropical fauna of Central [[p. 426]] America is a mixture of warm climate ancient Neotropical elements, really new to the region, and warm climate Nearctic elements, really native to the region.
Reptiles and amphibians are more sensitive to climatic factors, especially temperature and precipitation, than are mammals and the effects of this history are more clearly seen in them than in the latter. Among the chelonians, the Dermatemydidae may almost certainly be placed as "Neotropical" only nominally or in the special sense of being historically warm Nearctic rather than temperate or boreal. Geoemyda probably belongs here also and so in all probability do the Kinosternidae, which may be middle American invaders in the more northern North American as well as in the South American faunas. It is less clear whether Chelydra and Trachemys should be so classed or whether, as I think a little more probable, they are fully Nearctic forms recently intrusive in the Neotropical fauna.
The same sort of history would explain most of the apparent discontinuity between the Pleistocene and recent South American mammals of North American origin and their presumed ancestors and allies in North America. For instance the striking differentiation of the smaller cats in the Neotropical fauna has no apparent parallel or roots in the middle latitudes of North America, through which the ancestors must nevertheless have come. This differentiation may be assumed with considerable probability to have occurred in "warm Nearctica."
Up to this point no mention has been made of the paper in which Dunn20 covered the same field, and more, over a decade ago, because it was desired to present an independent investigation carried out in more detail as regards the Chelonia and with some crucial evidence that was not available to Dunn. This evidence inevitably results in a few minor modifications, but the general conclusions here reached are in essential agreement with Dunn's in every respect. His short and somewhat neglected paper should become a classic of zoögeography and the present contribution may be considered a tribute to his insight and a corroboration of his judgment.
[[p. 427]] GENERAL COMMENT ON AMERICAN FAUNAL REGIONS.
The foregoing summary and the evident difficulty of expressing it in the conventional zoögeographic terms emphasize the unsuitability of the classic nomenclature in its application to historical studies. This unsuitability was also felt by Dunn, who abandoned the use of "Nearctic" and "Neotropical" and divided the American (or Western Hemisphere) herpetological fauna into the following elements:
1. Holarctic--Modern, circumpolar element. (Not to be confused with the much broader use of Holarctic to include also north African and temperate, not only circumpolar, European and Asiatic elements).
2. Old Northern--This element now characterizes the southern United States and Central America.
3. South American--The old element in South America, proper, elaborated there during the isolation of the continent.
It is to be stressed that these terms apply to faunal elements, not to zoögeographic areas. From a more directly zoögeographic viewpoint, present evidence suggests the convenience of dividing the continental Americas into five regions or zones, as follows:
1. Boreal North America--The northern zone that has been periodically continuous with Boreal Asia and essentially part of the same evolutionary center. Especially characterized by Holarctic elements in the limited sense of Dunn. This is now essentially what is called the Boreal Region of Nearctica by neomammalogists.
2. Middle North America--The great zone comprising most of what is now the United States. This has been a major evolutionary center in itself and new "Holarctic" stocks, subsequently differentiated here, have periodically come into it not from the Old World but from Boreal North America. It has also contributed to the "Holarctic" fauna through spread of selected elements of its fauna into Boreal North America. This now approximately corresponds with Austral Nearctica21 of many neomammalogists or with C. H. Merriam's and [[p. 428]] Lydekker's "Sonoran Subregion," broader than the "Sonoran Zone" of most recent authors.
3. Southern North America--Historically the warmer part of North America which constituted a secondary but important evolutionary center for the differentiation and radiation of stocks received from Middle North America. Besides these (essentially the "Old Northern" elements of Dunn), the region is now characterized by more recent immigrants from Middle North America and from Equatorial South America, neither group of which has yet undergone important evolution within this area. Essentially the Central American and Mexican parts (or the Mexican Subregion) of the Neotropical Region of neomammalogists, but the West Indian Islands are excluded. The boundary against the "Nearctic" is broad and transitional and there are lingering effects of a time when the Southern zone extended well into the United States, so that considerable parts of the so-called Austro-Riparian and, especially, lower Sonoran zones may as well be assigned to my Southern as to my Middle North American Zone.
4. Equatorial South America--The more northern and a considerable portion of the central part of South America. There is indirect evidence that this area was a secondary evolutionary center in Cenozoic South America and that, with considerable over-simplification, the faunal history of that continent was one of interchange between broadly northern and southern regions with local differentiation or radiation in each.22 More recently this zone has been an important secondary center for stocks from the Southern North American Zone and its present fauna is a thorough admixture of ultimately Old Northern and South American elements in the sense of Dunn. This is now approximately the "Subregión Guayano-brasileña" of Cabrera and Yepes23, the "Guiano-Brazilian Subregion" of Sclater.24
[[p. 429]] 5. Austral South America--The cooler part of the continent, in general the southern part beginning at varying distances below the Tropic, encroaching northward along the Andes as that region rose. This was the other secondary Cenozoic center in South America and it has now also received various offshoots of the differentiation of Southern North American stocks in Equatorial South America. These do not as yet appear to have undergone particularly pronounced further differentiation in the Austral South American Zone. This is now approximately the "Subregión Patagónica" of Cabrera and Yepes and "Patagonian Sub-region" of Sclater, which includes a great deal more than Patagonia.
These zones cannot be bounded by lines on a map. Not only do they interdigitate but they also intergrade very broadly and must be thought of somewhat differently for different groups of animals. In addition to this, they have not been stationary historically but have expanded and contracted and moved north and south--locally even east and west--while retaining a certain vague sort of entity.
I would not include the islands (except Trinidad and some others that are obviously parts of the continents from a faunal viewpoint) in any of these zones, nor yet in the classic Nearctic and Neotropical Regions, either historically or on the basis of recent zoögeography. The West Indian faunas are almost entirely derived from the Southern North American and the Equatorial South American Zones, but they are not themselves to be allocated with either zone.
The names here applied to these five zones are admittedly clumsy and it would be preferable eventually to coin short and simple designations. I feel, however, that it would be unwise to do so in this paper. Much about them is frankly hypothetical and their usefulness and validity in zoögeographical discussion should be tested more carefully before adding to the already confusing array of special technical terms. It may also be difficult to find designations brief, unambiguous, and suitable. In the meantime longer and almost self-explanatory descriptive phrases may serve.
2. This does not exclude the possibility of other connections at dates still earlier or even at the same time, but such additional connections are not, in my opinion, either adequate or necessary to explain the recent faunal relationships. [[on p. 416]]
3. Simpson, G. G.: 1940, Mammals and land bridges. Jour. Washington Acad. Sci., XXX, pp. 137-163. On this and on the possible location of earlier land routes see also: Simpson, G. G.: (1939) 1940, Antarctica as a faunal migration route. Proc. 6th Pacific Sci. Cong., pp. 755-768. [[on p. 416]]
4. There are various different usages, but the Holarctic is commonly taken to include Europe, northern Asia, and most of North America. Within this area, the North American part is the Nearctic Region. Southern Asia and the many adjacent islands form the Oriental Region, which will also be mentioned in the course of this study. [[on p. 418]]
12. Near Zaraza in beds of Miocene age, found by me while working under the auspices of the Ministerio de Fomento of Venezuela and the American Museum of Natural History. A description is in press. [[on p. 421]]
15. Wieland, G. R.: 1923, A new Paraná Pleurodiran. Amer. Jour. Sci. (5) V, pp. 1-14. Ameghino, F.: 1913-1936, Obras. (Collected reprinting). 24 volumes, La Plata. See references to Platemys in index, vol. XXIV. Wieland did not note that occurrence of chelyids in the Paraná had been repeatedly mentioned long before his work, by Bravard, Burmeister, Ameghino, and others and he made no comparisons with the several previously named forms, to one of which his supposedly new species probably belongs. [[on p. 422]]
18. Ameghino, F.: 1900-1903, L'age des formations sédimentaires de Patagonie. An. Soc. Ci. Argentina, L, pp. 109-130, 145-165, 209-229; LI, pp. 20-39, 65-91; LII, pp. 189-197, 244-250; LIV, pp. 161-180, 220-249, 283-342. [[on p. 423]]
19. Notably: Olson, E. C., and McGrew, P. O.: 1941, Mammalian fauna from the Pliocene of Honduras. Bull. Geol. Soc. Amer., LII, pp. 1219-1244. It is "most likely . . . during the lower Pliocene . . . that the Honduras fauna was continuous with that of North America . . . No mammals of southern origin were encountered." [[on p. 425]]
20. Dunn, E. R.: 1931, The herpetological fauna of the Americas. Copeia, 1931, No. 3, pp. 106-119. I am indebted to Doctor Dunn for constructively critical comments on the present paper. [[on p. 426]]
21. Nomenclature the unsuitability of which is particularly noticeable. This is the middle, now mainly warm temperate part of North America, not its southern or Austral part, and it has nothing directly to do with the Arctic, as "Nearctic" inevitably suggests. Similarly a large and important part of the so-called Neotropical Realm is far from being tropical and part of truly tropical America is historically "Nearctic." [[on p. 427]]
22. But I do not see any good evidence that northern and southern South America were separated by a marine barrier during part of the Tertiary and still less that the northern part was associated with Central America and the West Indies more closely than with the southern part, theses advanced by Ihering and somewhat uncritically accepted by a few other students: Ihering, H. von: 1927, Die Geschichte des Atlantischen Ozeans. Jena. [[on p. 428]]