by Karl Patterson Schmidt (1946)
The bridge-building school of zoogeographers has found these relations [[p. 145]] puzzling, for it is certainly difficult to account for them by supposing a former connection of Europe and California (for example) by a trans-oceanic land-bridge that leaves out eastern North America. This very hypothesis, however, has been proposed by Scharff (1912), who extends the land connection from the Mediterranean region to Central America and Mexico (and so to western North America) via the West Indies. He is not very clear about the connection with Asia, which is to explain the fauna of the southeastern United States, but follows Ortmann in allowing for a broad connection at Bering Strait in the Cretaceous.
A certain amount of new and interesting data has become available for the study of these problems, even from the purely faunistic viewpoint. The vital blow dealt to the bridge-building school by Matthew's masterly résumé of the paleontological evidence (1915) will not be touched upon here, except insofar as I shall try to bring a little of the evidence from the existing fauna into line with Matthew's general hypothesis of dispersal.
Turning first to the faunal relations between western North America and Europe, especially western Europe, I may summarize briefly the principal groups of animals concerned, drawing heavily for this purpose from Scharff, whose diligent assembling of evidence redeems for me an otherwise futile work. In spite of divergence of opinion as to their classification, a number of snails and slugs seem to exemplify this relationship, and the family of slugs Arionidae is plainly confined to the western coast of the United States and nearly confined to western Europe, ranging, however, sparingly into Siberia and China. Among insects, the general aspect of the Californian butterfly fauna is strikingly European and strikingly unlike that of southeastern North America, and in this group nine species are considered identical in the two areas. The Eleodinae, a subfamily of the tenebrionid beetles, is characteristic of California and the Mediterranean region. Certain genera of dung-beetles have a similar distribution. The remarkable honey-ants are confined to southwestern North America, the Mediterranean region, and Central Asia. Other ants have a similar distribution, and the common Formica cinerea and F. rufibarbis are considered to be identical in the two areas in question. A species of centipede, Theotops erythrocephalus, is found in California and Oregon and in southwest and central Europe. The common European cray-fish (Potamobius) is represented by species of the genus in western North America, while eastern North America and eastern Asia are inhabited by other genera (Cambarus and Cambaroides).
Among vertebrates, by far the most striking case is that of the plethodont salamanders of the genus Hydromantes, with a species in the high Sierras of California and two species in Europe, confined to Sardinia and to northern Italy and France. The poverty of the fresh-water turtle fauna in the two areas is surprising, and each has only a single Hyla.
I have long been of the opinion that zoogeographic hypotheses should be based upon geological facts, and not geological hypotheses upon zoogeographic facts. In this respect, I feel bound to reject the 'land-bridge' speculations, which have had a considerable vogue among students of animal distribution, and to range myself with Wallace and Matthew in regarding the [[p. 146]] permanence of the continents (in their larger outlines) as the only scientific basis for zoogeographic speculation. In casting about for geological factors to explain the faunal relations above outlined, I examined the maps of the Pleistocene glaciation in the current geological manuals (see especially Chamberlain and Salisbury, 1907, 3: 423, fig. 529). The distribution of continental glaciers is represented as highly eccentric. North America and Europe are shown to have been heavily glaciated, while northern Asia was almost entirely free from ice. A variety of explanations suggest themselves for this surprising fact, of which the simplest would be to ascribe the want of indicated glaciation to insufficient knowledge. The Russian geologists, however, must have an intimate acquaintance with glacial phenomena at home, and could scarcely have overlooked them in Siberia. Dr. Frederick K. Morris of the geological staff of the American Museum of Natural History's "Third Asiatic Expedition" assures me, in conversation, that there is no evidence of glaciation in Mongolia, which lies between latitudes 42° and 54° N., while the southern limit of the continental glacier in North America was at about 38° and in Europe 50°.
Leaving the explanation of this world pattern of glaciation to the geologists, the facts clearly have an important bearing on certain problems of historical zoogeography. Its importance does not seem to have been realized in current discussions of animal distribution. I venture to suggest that the freedom from ice of northeastern Asia may afford a direct explanation of the faunal relations between Europe and western North America.
During the advance of the continental glaciers in Europe, the animal population retreated to the south, and the glacial relict fauna of the Alps clearly illustrates this southward advance of the northern fauna. A second and perhaps even more important refuge for the preglacial European animals was available in unglaciated northern and central Asia, and the eastward flow of the refugees must have been accentuated by the east-west trends of the Eurasian mountain systems. On the other hemisphere, the North American fauna, as far west as the Rocky Mountains, must have retreated rather completely to the southern half of the continent. The detailed evidence that retreat and advance of the fauna accompanied advance and retreat of the ice is presented by Chamberlain and Salisbury.
The final retreat of the glaciers opened up vast territories on both continents for repopulation by more southern forms of animals and plants. The fossils from interglacial periods exhibit the extent of the earlier northward migrations, and during a part of the Pleistocene, Bering Strait is known to have been closed, and an Asiatic-North American connection was thus available for that part of the fauna which could exist under the Alaskan climatic conditions. The expansion of a part of the previously crowded Siberian fauna into Alaska and thence into western North America, with a contemporaneous and more complete westward migration into Europe, affords an extremely satisfactory explanation of the remarkable similarities between the west American and European faunae. Scharff has adequately emphasized the westward migration of the Siberian element in the European fauna, but has not recognized the similar element in western North America as a correlative and
World pattern of glaciation in the northern hemisphere and central Asia being without a continental glacier. After Chamberlain and Salisbury.
[[p. 148]] probably contemporary migration from the same center, a part of the general postglacial expansion of the Siberian fauna. It is not impossible that the "Völkerwanderungen" represent a late phase of this phenomenon, and in historic times the spread of the black rat and the brown rat suggest that it still continues. The resurge northward of the semi-tropical Chinese fauna with modification of the postglacial climate would add to the "faunal pressure" in eastern Asia. The Rocky Mountains, presenting a highway for the southward migration of northern forms, offer an effective barrier to the westward spread of the preglacial fauna of the southeastern United States. A further avenue of southward dispersal may have been a hypothetical corridor between the mountain glaciers of the Rockies and the western border of the great northeastern continental glacier.
The circumpolar distribution of the more boreal forms is an allied phenomenon, but it seems likely that it depends on a number of other factors, and that ice floes may have played a part in the dispersal of life in this region. The circumpolar species of fresh water fishes (Esox lucius, the northern pike, Coregonus, the whitefish, Perca sp., the yellow perch, and the ling, Lota) contribute an item of evidence on the relatively recent existence of a closure of Bering Strait. The individuals of these forms in Europe, Siberia, and North America are either conspecific or so closely allied as to be scarcely distinguishable. They may have spread through coastwise marshes, or their dispersal may have been furthered by a freshening of the surface waters of the Arctic Ocean with the melting ice in spring, since fresh water tends to float on the oceanic salt water.
In addition to the fairly clear cut evidence of an alliance between the west European and west American faunae, there are a large number of related forms that fall in line with the general argument, but appear to represent earlier phases of this wave of dispersal and, in some cases, perhaps, a parallel wave of much earlier date. The present distribution of the bisons may be instanced. The two living species are confined to western North America and Europe (the Carpathians and Caucasus), but there was a Pleistocene development in North America of rather numerous species, suggesting the expansion of a group in a newly occupied territory. The Rocky Mountain goat is perhaps related to the chamois, but the amount of difference between them is too great for any evidence of recent migration. The wild sheep of the Rockies and of the mountains of Asia do appear to be a part of a Pleistocene and recent dispersal. The Pacific Coast skink of western North America (Eumeces skiltonianus) agrees well enough with the hypothesis of a northern Asiatic origin, but its closest relatives are in southern China instead of in Europe. Emys, with a species in Central Europe and one in the mid-western United States presents a curious parallel with the bisons.
The resemblance between the west European and west North American faunae thus seems to be essentially a relation based on a late Tertiary, Pleistocene, and Recent dispersal. I wish to discuss the second relation, between the animals of eastern North America and eastern Asia, in this connection because it presents a strong contrast with the first, as a much earlier series of migrations, probably in late Mesozoic and early Tertiary times, interchanging [[p. 149]] via the same Bering Strait bridge. For the consideration of this second problem the herpetological evidence is especially instructive.
The faunae of eastern Asia and eastern North America present certain general resemblances in their composition. Both are rich in salamanders and in freshwater turtles of the family Emydidae. The still more aquatic soft-shelled turtles of the genus Trionyx are well represented. Of the groups of amphibians and reptiles now confined to these two regions (and found in both) the most important are the cryptobranchid salamanders, the crocodilian genus Alligator, and the spoon-billed sturgeons, Psephurus in China and Polyodon in the Mississippi. The two areas do not have any species of amphibians or reptiles in common, though it was long a commonplace of zoogeographic compilations to state that the common ground skink of the southeastern United States (Leiolopisma laterale) is indistinguishable from a similar form in China. Re-examination of this question with the ample material collected by Mr. Clifford H. Pope, of the Third Asiatic Expedition of the American Museum of Natural History, shows that the Chinese skinks that have been referred to Leiolopisma laterale fall into four or five distinct forms, and all of them are distinguishable from the American species, though the Yangtze valley species (L. modestum) and the northernmost form (L. septentrionale) approach it rather closely. The difficulty had been with the lack of ample series and the want of conspicuous characters on one hand, and the general conservatism of the scincid lizards on the other.
The related groups in the two areas may be examined more in detail. The Cryptobranchidae are represented by Cryptobranchus in eastern North America and by Megalobatrachus in China and Japan. The differences between the two genera are not great, but neither are they widely distinct from a fossil species (Andrias scheuchzeri) in the Oligocene of Europe. It is evident that the group is an ancient one and that the close resemblance between the three forms is due to conservatism rather than to a recent origin of the group. In general structure, and in the fact that fertilization is external, the group is distinctly primitive. The newt fauna (Triturus) of the two areas is poor when compared with that of Europe, where the newts have undergone an expansion comparable with that of the Hynobiidae in eastern Asia and the Ambystomidae in North America. The latter, and the still more specialized American family Plethodontidae are unrepresented in Asia, and the extraordinary development of the American salamander fauna is strikingly shown by the presence of two highly peculiar types, the Sirenidae and the Amphiumidae. The family Proteidae being represented by the widespread Necturus, the North American fauna, in the southeastern United States, exhibits seven of the eight families of salamanders. The Hynobiidae, which are both primitive and generalized, scarcely reach Europe, and have developed five genera and twenty-two species in eastern Asia.
The frogs and toads common to the two areas belong to the genera Bufo, Hyla, and Rana. None of these genera are confined to the two areas nor, indeed, to the Holarctic Region, and thus they do not contribute evidence of special relationship between the faunae. The wood-frog group, with a circumpolar distribution, belongs rather with European and west American faunae than with the Chinese and east American.
[[p. 150]] One of the most striking resemblances between the two faunae lies in the development of the freshwater turtles. The more aquatic Testudinidae, perhaps best distinguished as the family Emydididae, have a most striking distribution. Entirely absent from Australia, from Africa, with the exception of a single species, and with only a pair of outlying forms in South America, the group is primarily Holarctic. The inequality of their distribution in both Eurasia and North America has already been mentioned. Three species, Clemmys leprosa and caspica and Emmys orbicularis, are found in Europe, and a single Clemmys (C. marmorata) on the Pacific slope of North America. In contrast with this sparse population, eastern Asia has five genera and a dozen species (including four Clemmys), and eastern North America eight genera and twenty-three species (Clemmys with three species). Clemmys and Emys appear to be the only genera that appear in both hemispheres, and the family accordingly presents the aspect of an old and well differentiated group. The parallel between the Old World and the New World faunae is continued by the fact that the emydine forms overflow into southeastern Asia and into Central America. The soft-shelled turtles (Trionyx) range into the Oriental and African regions, and do not belong to the restricted faunal element under consideration.
The allied forms of lizards involve some of the most striking and abundant forms. The genus Leiolopisma has already been mentioned. The genus Eumeces is differentiated into numerous species in the two areas, and in some measure parallels the development of the emydid turtles. It presents a close parallel to the interesting distribution of the limbless Ophisaurus, the "glass snakes," whose five species are found in north Africa, central and southeastern Europe, the eastern Himalayas, southern China, and southeastern North America. Both Eumeces and Ophisaurus are absent from western Europe, and Eumeces is only sparingly represented in western North America, the glass snake not at all. Their distribution is a curiously peripheral one with relation to the more strictly Holarctic fauna.
The genera of snakes in common are Natrix, Opheodrys, Elaphe, and Agkistrodon. Of these Opheodrys is strictly confined to the two areas. Both Elaphe and Natrix are sparingly represented in Europe but absent from Pacific America. Both, however, have their chief centers of speciation in China and southeast North America, and thus strikingly parallel the distribution of the Emydidae. Agkistrodon reaches south Mexico on the American Continent, and Java in the Orient, and thus requires mention again in connection with the more shadowy faunal resemblance between the East Indies and Central America.
The only remaining reptiles that effect the present problem are the two species of Alligator, one widespread in the southeastern United States, the other with an extremely limited range on the lower Yangtze. This genus is well represented by fossil species, in the Tertiary with an especially instructive species from the middle-Miocene of Wyoming (A. thompsoni Mook). The living alligators are evidently relicts of a widespread late Mesozoic and early Tertiary Holarctic range. All of the group for which there is fossil evidence (spoonbill, sturgeons, alligators, cryptobranchids) combine to [[p. 151]] suggest that the existing resemblances between eastern Asia and eastern America are all that remain of a wide-spread Holarctic fauna of an earlier period, and one in numerous respects peripheral to the Holarctic fauna of the present day.
The elements of the two principal faunal resemblances discussed in this paper are rather strikingly contrasted:
Western . . . . . Eastern
It is an essential part of Matthew's theory of distribution that there have been successive migration waves spreading outward from the northern land masses. The east American-east Asiatic common element appears to represent one of these great faunal waves, in accordance with this theory, and one which largely antedates the Tertiary mammalian evolution. The European-west American relation depends upon dispersals conditioned by the relatively recent Ice Age.
Traces of still earlier faunal waves or, perhaps, in part, of the earliest phase of the east Asiatic-east American wave in question, are to be seen in the faunal element common to Central America and the East Indies. The Central American fauna presents certain obvious relations with eastern North America (Chelydra, Pseudemys, Ophisaurus), which are paralleled in the Orient by an east Asiatic element in the fauna of the East Indies (emydid turtles, "Ophisaurus"). The persistence of the remarkable lizards of the family Helodermatidae with Heloderma in Mexico and the southwestern United States and Lanthanotus in Borneo forms a striking parallel to the distribution of the tapirs, but is probably to be referred to a much earlier date. The snapping turtles with Chelydra in Central America and Devisia in New Guinea fall into line with this element. The Oriental and the Central American species of the genus Geoemyda (Emydidae) afford a clear example of this peripheral wave. The species of pit-vipers of the genus Trimeresurus in the Orient and in tropical America are also in agreement, but their relations with the more primitive but more northern Agkistrodon complicate the question of the age of their dispersal. It is evident that the peripheral distribution of primitive forms postulated by Matthew may be obscured by the fact that the retreat of a group before its more advanced competitors may be an ecological instead of a geographic one. Primitive forms are often left behind by more advanced ones, a part of whose superiority may rest on a greater adaptability to climatic conditions, with capacity for rapid spread.
The parallel between the southward extension of the two faunae is a striking one when tabulated:
[[p. 152]] The distribution of the crocodilians falls satisfactorily into successive waves of dispersal. The earliest is represented by the South American caimans, whose broad snouts and heavy dermal armor seem to make them the most primitive living members of the order. The true crocodiles have two primitive genera in equatorial Africa and primitive short-snouted forms in India and Central America, and there is some paleontological evidence to show that these forms are an early and distinct phase of the dispersal of the more modern, longer-snouted species, all of them belonging to the genus Crocodylus, which has two African, four Oriental, three Australian, and four tropical American species. The genus is completely absent from eastern and central South America, where the caimans hold their own. The widespread American C. acutus just reaches Florida, while the still more widely ranging C. porosus reaches south China. The alligators fall into line as a fourth wave of dispersal but not, I believe, as the last, in spite of their more northern range. They are plainly "hanging relicts"3 of the Mesozoic dispersal waves, obviously later than the caimans but earlier than the crocodiles.
There is a general agreement of the South American, Australian, and African faunae in certain primitive elements, among which may be mentioned lung fishes; leptodactylid frogs; pleurodiran turtles; the more primitive groups of snakes and lizards; and the marsupials (absent from Africa). The list might be greatly extended among invertebrate groups; it includes forms about whose ancient or primitive nature there is no question; and where there are paleontological records, these point to an early Mesozoic or at least pre-Cretaceous dispersal. These primitive faunae are probably the accumulated remnants of repeated dispersals in the late Paleozoic and early Mesozoic, i.e. from the "Holarctic" fauna of those early ages.
CHAMBERLAIN, T. C., and R. D. SALISBURY
1At the instance of my colleague
in zoogeographic interests, Mr. M. Graham Netting, I have slightly modified,
corrected, and amplified the essay of the same title written in 1928 and
published in the Lingnan Science Journal, Vol. 10, in 1931. The essay
had been prepared for the 1928 Science Conference at Lingnan University
but was mislaid at Lingnan after being somewhat delayed in transmission,
and there was no opportunity to correct MS. or proof, even after the four-year
lapse between preparation and publication. It is felt that the interpretation
of post-glacial faunal resurgence, in the light of the eccentricity of
the world glaciation, justifies the emphasis of republication. [[on