CHAPTER VI

THE GEOGRAPHICAL DISTRIBUTION OF MAMMALS

To every one who has paid the slightest attention to the subject, it is a familiar fact that different parts of the earth have different animals; school-children learn from their geographies that kangaroos are found in Australia, the Hippopotamus in Africa, the Tiger in southern Asia, armadillos and llamas in South America. These examples are all taken from distant lands, yet the zoölogical difference between two given land-areas is by no means proportional to the distance between them. An Englishman landing in Japan finds himself surrounded by animals and plants very like and often identical with those which he left at home, while the narrow Strait of Lombok, east of Java, separates two profoundly different regions. In crossing Mexico from east to west, the traveller meets very different animals in closely adjacent areas; and, at first sight, the arrangement of animals appears to be so capricious as to admit of no formulation in general laws.

In pre-Darwinian times, when it was the almost universal belief that each species had been separately created and was exactly fitted to the region which it inhabits, no explanation of the geographical arrangement of animals was possible, but the acceptance of the theory of evolution demanded that such an explanation should be found. A failure to devise any rational and satisfactory account of the geography of animal life would be a fatal weakness in the evolutionary theory, hence the facts of distribution were subjected to a renewed and searching analysis as one of the best means of critically testing the new doctrine. Not that the subject had received no attention before the publication of Darwin’s book; on the contrary, it had attracted much interest as a study of facts, and this study was one of the principal avenues by which Darwin approached his great generalization. In his autobiographical fragment he tells us: “I had been deeply impressed by discovering in the Pampean formation great fossil animals covered with armour like that on the existing armadillos; secondly, by the manner in which closely allied animals replace one another in proceeding southward over the Continent; and third, by the South American character of most of the productions of the Galapagos archipelago and more especially by the manner in which they differ slightly in each island of the group.”

Obviously, before attempting to explain the facts of the geographical distribution of mammals, we must first ascertain what those facts are. The following brief sketch of the terms used in describing geographical arrangement is summarized from Mr. Wallace’s “Island Life.”

Though with fluctuating boundaries and subject to slow secular changes, a mammalian species is limited to a fairly definite area, which may be of immense or very restricted extent, and throughout which it may be found in greater or less abundance. Many species, however, are not distributed continuously over the areas which they inhabit, but occur only in suitable _stations_ adapted to their habits and mode of life. Thus, some will be found only where there are trees, others in the neighbourhood of water, others only on open plains, etc. A _specific area_ is then the whole extent of country within which the species may be found, while the _stations_ are the limited districts contained in the area which are exactly suited to the habits of the species in question; these stations may be hundreds of miles apart, as in the case of mountain-tops, or they may be close together. A marsh-living species, for example, will occur in all the marshes throughout its area, whether these be many or few, near together or widely scattered; for such a species marshes are its stations.

_Generic areas_ differ in character according as the genus is large, that is, comprising many species, or small and having but few species, or, it may be, a single one. The species, as a rule, occupy each its own area, and the areas may be entirely distinct, or they may be contiguous and more or less extensively overlapping, though it seldom happens that two or more species of the same genus inhabit exactly the same area. Often some physical feature, such as a range of high mountains, a great river, the edge of a forest, plain or desert, exactly defines the limits of species of the same genus. The Amazon, for example, acts as such a boundary to many species. It was to this change of related species from one area to another that Darwin referred in the passage quoted above, saying that he had been deeply impressed “by the manner in which closely allied animals replace one another in proceeding southward over the Continent [_i.e._ South America].” On the other hand, the overlapping of areas may be very extensive, and one species of great range may cover the whole area of another and much more besides.

A remarkable example of the widely separated areas of species belonging to the same genus is that of the tapirs. Of this genus there are two or three species in Central and South America and one inhabiting the Malay Peninsula and Borneo, almost as wide a separation as the size of the earth permits. Discontinuous distribution of this character can be explained in terms of the evolutionary theory only in one of two ways. Either (1) the American and Asiatic species developed independently of one another from different ancestors, or (2) the regions intervening between these widely separated areas once formed a continuous land, occupied by species of the genus which have become extinct. From all that we know concerning the operation of the evolutionary process, the first alternative may be set aside as altogether improbable, and, even had we no information concerning the history of the tapirs and their former distribution, the second explanation would be chosen as incomparably the more likely. As a matter of fact, we have definite knowledge that tapirs once ranged all over Europe and North America and doubtless over northern Asia, as well, and, further, that North America was joined to Asia by a land occupying the place of the shallow Bering Sea, at a time when the tapirs were able to take advantage of this means of passing from one continent to the other. Such appears to be the invariable explanation of discontinuous distribution, though we may not always be able to give so clear a proof of it.

The genera of a family are distributed in much the same fashion as the species of a genus, but, as a rule, much more widely. While no genus of terrestrial mammals is cosmopolitan (_i.e._ universally distributed), at least as genera are defined and limited by most modern systematists, certain families are represented in every continent. If the extremely peculiar and isolated Australian continent be excepted, the number of such cosmopolitan families is considerable and wide separation between the genera is frequent. Of the camel family, for instance, one genus, that of the true Camel (_Camelus_), is confined to the northern hemisphere and the Old World, the other (_Lama_), comprising the Llama, Guanaco, etc., is found only in the southern hemisphere and the New World. Less extreme instances of the discontinuous distribution of a family are common enough.

The principles of distribution are the same when applied to families and orders. Most of the mammalian orders are very widely distributed and many are cosmopolitan, except for Australia, though some are confined to one or two continents. The monotremes are limited to Australia and Tasmania, the marsupials to Australia and the Americas, the edentates to the latter, the elephants and hyracoids to Africa and Asia. Carnivores and rodents, on the contrary, are found in every continent, even Australia.

We have next to inquire what is the nature of the obstacles or barriers that prevent the indefinite spread of terrestrial mammals, so that the mammalian fauna of the whole earth, and even of a single continent, is not uniform, but highly variegated. The rate of multiplication of animals is so rapid that, under normal conditions, the animal population is always pressing hard upon the means of subsistence and every species that is increasing in numbers must constantly extend its range in search of food. Every species would increase indefinitely, if there were no countervailing checks. Were all the young to survive and breed in their turn, “even large and slow-breeding mammals, which only have one at a birth, but continue to breed from eight to ten successive years, may increase from a single pair to 10,000,000 in forty years” (Wallace). Evidently, a species must spread from its place of origin until stopped by insuperable obstacles, the most obvious of which are wide seas. A few land mammals are not only excellent swimmers, but will cross straits without hesitation, as the Guanaco has been seen to swim the Straits of Magellan; for the great majority, however, a very few miles of sea form an impassable barrier. As was shown above, a broad or deep river is sufficient to limit many species, as the Santa Cruz River in Patagonia marks the southern boundary of the armadillos.

Important geographical changes, such as the joining of lands that before were separate, or the dividing of continuous lands by transgressions and incursions of the sea, must necessarily have a profound effect upon the distribution of land mammals. Separated land-areas, however similar may have been their faunas at the time of separation, will, through the operation of the divergent evolutionary process, grow more unlike in proportion to the length of time that the separation continues. Regions which have been severed within a short time (in the geological sense of a short time) are zoölogically very similar or even identical, while those that have long been isolated are correspondingly peculiar. Attention has already been called, in another connection, to the contrasted cases of such great continental islands as Great Britain, Java, Sumatra, etc., on the one hand, and Australia, on the other. The continental islands, which have but lately been detached from the neighbouring main lands, are hardly more peculiar zoölogically than equal areas of the adjoining continents, while the long-continued isolation of Australia has made it the most peculiar region of the earth. Climatic changes, which, as we saw in Chapter I, have indubitably taken place many times, have also had a great effect in shifting the distribution of mammals, which in its present form is the outcome of a very long series of geographical and climatic changes, on the one hand, and of adaptive changes in the animals themselves, on the other.

Of almost equal importance as a barrier is climate and especially temperature. Not that similar climates can produce similar forms in separate areas. Regions of almost exactly similar climate in Australia, Africa and South America have totally different faunas, but, _within continuous land-areas_, the most effective of barriers is temperature. This acts differently in the case of limiting the northward spread of southern forms and the southward spread of northern species. Dr. Merriam’s long study of this problem has led him to the conclusion that southern species are bounded on the north by the temperature of the breeding season, in which the total quantity of heat must reach a certain minimum, while “animals and plants are restricted in southward distribution by the mean temperature of a brief period covering the hottest part of the year.” On the Pacific coast there is a remarkable mingling in the same areas of species which, east of the high mountains, are distributed in sharply separated zones. This is explained by the mild and equable climate of the coastal belt, where the hottest season of the year does not reach the limiting maximum for the northern species, while the total quantity of heat in the breeding season is sufficient to enable southern species to thrive and maintain themselves.

Dr. Merriam thus sums up the effects of climatic factors upon distribution: “Humidity and other secondary causes determine the presence or absence of particular species in particular localities within their appropriate zones, but temperature pre-determines the possibilities of distribution; it fixes the limits beyond which species cannot pass.” “Concurrently with these changes in vegetation from the south northward occur equally marked differences in the mammals, birds, reptiles, and insects. Among mammals the tapirs, monkeys, armadillos, nasuas, peccaries, and opossums of Central America and Mexico are replaced to the northward by wood-rats, marmots, chipmunks, foxes, rabbits, short-tailed field-mice of several genera, shrews, wild-cats, lynxes, short-tailed porcupines, elk, moose, reindeer, sables, fishers, wolverines, lemmings, musk-oxen, and polar bears.”

Dr. J. A. Allen has reached closely similar conclusions. “Of strictly climatic influences, temperature is by far the most important, although moisture plays an influential part. Where a low temperature prevails life, both animal and vegetable, is represented by comparatively few forms; under a high temperature it is characterized by great diversity and luxuriance. Within the Arctic Circle the species of both animals and plants are not only few, but they are widely distributed, being for the most part everywhere the same. Under the tropics they are a hundred fold more numerous and of comparatively restricted distribution.” “The influence of temperature is perhaps most strikingly displayed in the distribution of life upon the slopes of a high mountain, especially if situated near the tropics. While its base may be clothed with palms and luxuriant tropical vegetation, its summit may be snow-capped and barren.... The animal life becomes likewise correspondingly changed, tropical forms of mammals, birds, and insects of the lower slopes gradually giving place to such as are characteristic of arctic latitudes.” “The effect of humidity upon plant life is thus obvious, but it is equally potent, though less evident, upon animal life. Many animals ... are so fitted for a forest life, as regards both food and shelter, that their very existence depends upon such surroundings.... Thus moisture alone may determine the character of life over extensive regions.”

While climate is thus the most important of the barriers which determine distribution in continuous land-areas, the absence of any particular species from a given region is no proof that the climate is unsuitable to that species. This is sufficiently shown by the manner in which animals introduced into a new country often run wild and multiply to an incredible extent, as the rabbits have done in Australia, the Mongoose in Jamaica, horses on our western plains, horses and cattle on the Pampas of Argentina, etc.

Topographical features, such as great mountain-ranges and plateaus, also limit many species, not only by the difficulty of crossing them, but also by the effect which they have upon temperature and moisture. For this reason long ranges of mountains and table-lands may carry a northern fauna very far to the south of its ordinary range, as do the mountain-systems of North America in a very conspicuous manner. The great Mexican plateau is zoölogically a part of North America, while the low coastal lands as far as southeastern Texas have Central American affinities.

A different kind of obstacle to the spread of a species into a new area may be the pre-occupation of that area by another species. The pre-occupier may be one that plays so similar a part in the economy of nature as to leave no opportunity for the newcomer to establish itself. On the other hand, the obstructing form may be an active enemy and of a totally different character from the intruder, as in the case of the Tse-tse Fly in parts of Africa. The bite of the fly is fatal to horses and oxen, so that these mammals are unable to enter the fly-infested regions. Many times in the course of the Tertiary period various mammals reached North America from the south or from the Old World, which were unable to gain a permanent foothold and speedily died out. At this distance of time it is seldom, if ever, possible to explain why a species which succeeded in reaching this continent could not maintain itself, though the most probable assumption is that the forms already in possession of the land were an insuperable obstacle to the intruders.

The rate of dispersal of a species into new areas may be fast or slow, according as the conditions are more or less favourable. Newly introduced insect-pests, like the Gypsy and the Brown-tailed Moths in New England, often spread with portentous rapidity; and introduced mammals have frequently taken possession of vast areas in a surprisingly short time. One of the most remarkable of these cases is cited by Darwin. “In the time of Sarmiento (1580) these Indians had bows and arrows, now long since disused; they then also possessed some horses. This is a very curious fact, showing the extraordinarily rapid multiplication of horses in South America. The horse was first landed at Buenos Ayres in 1537 and the colony being then for a time deserted, the horse ran wild; in 1580, only forty-three years afterwards, we hear of them at the Strait of Magellan!” (“Voyage of a Naturalist,” pp. 232-233.) In this example, something must be allowed for human agency, but even so, it is very surprising.

In the case of lands newly raised above the sea and connecting formerly separated areas, it is necessary that they should first be taken possession of by vegetation, before they can become passable by animals, for the migration of mammals from continent to continent is an entirely distinct phenomenon from the annual migration of birds. The latter, though a fact familiar to every one, is an unexplained mystery, and it is somewhat unfortunate that the same term should be used for the completely different process of the spread of mammals into newly opened land. This spread is purely unconscious and is due to the pressure of increasing numbers upon the means of subsistence, each new generation ranging farther and farther from the original home of the species and continuing so to extend until some insuperable obstacle is encountered. When a sea-barrier is removed by upheaval and the newly formed land rendered habitable for mammals through the invasion of plants, the interrupted process is resumed and an interchange of species between the areas thus connected is brought about. The interchange is, however, always an incomplete one, certain forms not being able so to extend their range, because of climatic differences, pre-occupation or some such barrier.

It is customary to give a graphic expression to the facts of animal distribution by dividing the land surface of the earth into districts which are characterized by their faunas. It is not possible to construct a geographical scheme which will be equally satisfactory for all classes of animals, because the geological date of most rapid development and diffusion was so different in the various classes. The geographical and climatic conditions which favoured a particular geographical arrangement of one class had been so completely altered that the class coming in later could not attain a similar distribution. For this reason, land mammals are chosen as affording the best criteria; their adaptability is such that they are found all over the earth, their dispersal is primarily dependent upon the arrangement and connections of the continental land-masses, modified by the topographical and climatic conditions, and they, with the birds, are the latest of the vertebrate classes to assume a dominating importance. Their history is the most fully known and falls within the best understood portion of the earth’s history, making it possible to follow their migrations with a precision which is seldom feasible for the other classes of animals, and thus to correlate the successive physical and organic changes. A particularly great advantage which mammals possess for this purpose is that the mutual relationships of the various kinds are better understood than in the case of most other groups of animals. It is true that we shall find a great many unsolved problems, upon which the most divergent opinions are held, but the main outlines of the scheme are quite generally agreed upon.

Many plans for the zoölogical division of the continental areas have been proposed by various writers on the subject, some differing very radically from others. It would be useless and tedious to review even the more important of the many proposals and suggestions which have been made in the last half-century; and we may, with advantage, adopt an eclectic scheme which has been slowly reached by successive approximations to a satisfactory arrangement.

Just as in political geography it is found necessary to recognize divisions of different rank and scope, like nation, state, county, township, the facts of zoölogical geography require divisions of different orders of importance. Thus, in descending order, the terms _realm_, _region_, _subregion_, _province_, etc. are commonly employed, but unfortunately they are often used loosely and even interchangeably; yet it is desirable to attach a more or less precise significance to each and more terms are needed for an accurate expression of the many complex facts.

The extreme zoölogical peculiarity of Australia is recognized by making that continent and its adjoining islands one of the great primary divisions, of which the other includes all the rest of the world; the former is characterized by its almost exclusively marsupial fauna, while the other continents are inhabited by the Monodelphia or placental mammals. Aside from Australia, by far the most isolated and peculiar region of the earth is South America, and this fact is expressed by constituting it into a _realm_, or division of the second order, and to this realm is given the name _Neogæa_. The remaining continents, North America, Europe, Asia and Africa, make up the other realm, _Arctogæa_, in which there is an unmistakable general likeness among the mammals. The three continents of the Old World form a vast, connected land-mass, and the final separation of North America from this great complex is an event of geologically recent date. For reasons that will be made clear in the course of the history, the junction of the two Americas has had comparatively little effect upon the zoölogy of the northern continent, except in its tropical portion. It is obvious from a glance at the map, that the great zoölogical divisions are of very unequal size, but the arrangement is made on the basis of degrees of difference in the mammalian faunas. These degrees of difference are, in turn, an expression of length of separation or of the difficulty of communication between connected lands.

The following table gives the major divisions of the earth apart from Australia:

I. NEOGÆIC REALM. _Neotropical Region._—South and Central America, lowlands of Mexico, the West Indies.

II. ARCTOGÆIC REALM. { 1. _Malagasy Region._—Madagascar. { 2. _Ethiopian Region._—Africa south of the { Sahara Desert. { 3. _Oriental Region._—Southern peninsulas of { Asia, Malay Archipelago. { 4. _Holarctic Region._—N. Africa, Europe, Asia, { (except southern part), boreal N. America. { 5. _Sonoran Region._—Remainder of N. America { (except lowlands of Mexico).

North America, as is expressed by this scheme, is zoölogically composite; the northern half, including nearly all of Canada, belongs to the vast Holarctic Region, which also comprises Europe, Africa north of the Sahara and Asia north of the Himalaya Mountains. The remainder of the continent, exclusive of the Mexican coastal lowlands, is set off as the Sonoran Region. Inasmuch as we have here to do with broadly continuous land-areas, not demarcated by great physical features, and as the genera and species of mammals differ greatly in regard to their ability to withstand a wide range of climatic variations, it is not to be expected that the boundaries between the regions which make up North America should be very sharply drawn. It is not surprising, therefore, to find a transition zone, extending all across the continent, in which the Holarctic and Sonoran faunas mingle, or that Central America should, in considerable measure, be transitional to South America, though zoölogically a part of the latter.

[Illustration: FIG. 53.—Zoölogical Divisions of North America. (After Merriam.)]

Dr. Merriam’s arrangement, which deals only with North America without reference to the Old World, divides the land into a series of transcontinental zones, which he calls the Arctic, Boreal, Upper and Lower Sonoran and Tropical. These zones have very irregular and sinuous boundaries, which follow lines of equal temperature (isothermal lines) during the breeding season, May, June and July, the tortuous boundaries being conditioned by topographical features, which deflect the isothermal lines.

[Illustration: FIG. 54.—Polar Bear (_Thalarctus maritimus_).—By permission of the N.Y. Zoölog. Soc.]

The Arctic zone is part of a circumpolar area, which is very much the same in North America, Asia and Europe; and in any of these continents the fauna differs much more from that of the contiguous zone to the south than from the Arctic fauna of another continent. There are some local differences, but the characteristic mammals of this Arctic zone are the Polar Bear, Arctic Fox, Musk Ox, Barren-ground Caribou, Lemming, Arctic Hare, and a marmot. Most, if not all, of these forms are of Old World origin.

[Illustration: FIG. 55.—Musk Ox (_Ovibos wardi_) female; the males have much larger horns.—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 56.—Arctic Fox (_Vulpes lagopus_) in winter dress.—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 57.—Arctic Fox in summer dress.—By permission of the N.Y. Zoölog. Soc.]

The American portion of the great Holarctic region is called by Mr. Lydekker, who uses Wallace’s term, the “Canadian subregion,” and by Dr. Merriam the “Boreal region.” Not that there is any difference of principle involved in this varying nomenclature, for Dr. Merriam says: “It so happens that the Boreal element in America resembles that of Eurasia so closely that in the judgment of many eminent authorities the two constitute a single primary region—a view in which I heartily concur.” The Canadian or Boreal subregion of the Holarctic is the great belt of coniferous forest, which extends obliquely across North America from Alaska to New England; its frontier with the Arctic zone is the northern limit of trees and it is divided from the Transition zone approximately by the line of latitude 45° N., though with a sinuous course, and it is carried far to the south by the wooded heights of the Appalachian, Rocky and Sierra Nevada Mountains, and along the Pacific coast, the mixed character of which has already been explained; it extends almost to San Francisco. The subregion is further divisible into northern and southern belts, called the Hudsonian and Canadian faunas, the limit between them approximately following the isothermal line of 57° F. The mammals of this subregion are largely of Old World origin, many of them coming in with the great immigrations of the Pliocene and Pleistocene epochs; but there are also native American elements and even one genus of South American origin, the Short-tailed or Canada Porcupine (_Erethizon_).

[Illustration: FIG. 58.—Canada Porcupine (_Erethizon dorsatus_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 59.—Woodchuck or Marmot (_Marmota monax_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 60.—Mink (_Lutreola vison_).—By permission of the N.Y. Zoölog. Soc.]

In considering the mammals of this subregion, it should be remembered that they are not uniformly distributed throughout even one subdivision, but in a scattering way and in accordance with their habits and stations, and also in accordance with a gradual change to the south, following the changing temperature. The Muskrat will not be found far from water or the Porcupine from woods. Especially characteristic of the Canadian subregion are the Old World types of deer, none of which range farther south than the Transition zone. The Wapiti, erroneously called the Elk (_Cervus canadensis_), is very closely allied to the European Stag (_C. elaphus_) and still more closely to the Stag of the Thian Shan in Central Asia (_C. eustephanus_). So great is the resemblance, that some naturalists would refer all three forms to a single species. The Moose (_Alce americanus_), which should be called the Elk, is so near to the Scandinavian Elk (_A. machlis_) that it is hardly distinguishable as a separate species, and the Woodland Caribou (_Rangifer caribou_) is the American representative of the Lapland Reindeer (_R. tarandus_). The so-called Rocky Mountain Goat (_Oreamnos montanus_), a peculiar and aberrant form of the Chamois subfamily of the Antelopes, is confined to the subregion. The Mountain Sheep (_Ovis montana_, _O. dalli_) are represented by three or four species, one of which extends into the Sonoran region, as does also the Bison, wrongly called Buffalo (_Bison bison_), which is nearly allied to the European _B. bonasus_. In Cæsar’s time the European Bison (German, Wisent) ranged through Germany and is described in his account of the Hercynian Forest; but the advance of civilization has almost exterminated it, only a few small herds being maintained by the most rigid protection in Russia and in the Carpathian Mountains. Of the Carnivora, the weasels, martens, Fisher, Mink and Ermine are Boreal, as are the Wolverene (_Gulo_) and the Grey Wolf (_Canis_), the three last-named extending also into the Arctic zone. Essentially Boreal, though reaching and entering the Sonoran, are the bears (_Ursus_), the red foxes (_Vulpes_), the otters (_Lutra_) and the Old World shrews (_Sorex_), while the Star-nose Mole (_Condylura_) and the mole-shrews (_Urotrichus_) do not extend south of the Transition zone. Probable intruders from the south into the Boreal subregion are the pumas, or “mountain lions,” which just enter the subregion, the Canada Lynx (_Lynx rufus_) and one species of skunks (_Mephitis_), the Raccoon (_Procyon lotor_), Badger (_Taxidea americana_) and the American deer (_Odocoileus_). A large number of rodents are characteristically Boreal: marmots, or woodchucks (_Marmota_), the Sewellel (_Aplodontia rufa_), lemmings (_Myodes_), Jumping Mouse (_Zapus_), the Canada Porcupine (_Erethizon dorsatus_) and the pikas, “tailless or whistling hares” (_Ochotona_). Boreal rodents that enter the Sonoran are the chipmunks (_Tamias_), beavers (_Castor_), meadow-mice (_Microtus_), the Muskrat (_Fiber zibethicus_). The white-footed mice (_Sitomys_) and the wood-rats (_Neotoma_) are southern rodents that reach or enter the Boreal.

Between the Boreal subregion and the Sonoran region is the Transition zone, which follows all the complex windings of the boundary lines. It covers most of New England, New York, Pennsylvania and southern Ontario; passing through southern Michigan and Wisconsin, it bends northward over Minnesota and covers most of North Dakota, Manitoba and the plains of the Saskatchewan, then turns abruptly southward and includes eastern Montana and parts of South Dakota and Nebraska. Crossing Wyoming, it follows around the northern edge of the Great Basin to the plains of the Columbia. The three great mountain-systems carry the zone far to the south and arms of it extend along the Appalachians to northern Georgia, along the Rockies to New Mexico, and it follows the Sierras to southern California. “The Transition zone, as its name indicates, is a zone of overlapping Boreal and Sonoran types. Many Boreal genera and species here reach the extreme southern limits of their distribution and many Sonoran genera and species their northern limits. But a single mammalian genus (_Synaptomys_) [one of the field mice] is restricted to the Transition zone.... A number of species, however, seem to be nearly or quite confined to this zone” (Merriam).

[Illustration: FIG. 61.—Upper figure, European Bison (_Bison bonasus_). Lower figure, American Bison (_B. bison_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 62.—Wolverene (_Gulo luscus_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 63.—Wapiti or “Elk” (_Cervus canadensis_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 64.—Alaska Brown Bear (_Ursus middendorfi_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 65.—Moose (_Alce americanus_). Young male with undeveloped antlers.—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 66.—Beaver (_Castor canadensis_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 67.—Woodland Caribou (_Rangifer caribou_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 68.—Red Fox (_Vulpus fulvus_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 69.—Rocky Mountain “Goat” (_Oreamnos montanus_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 70.—Ermine (_Mustela erminea_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 71.—Timber or Grey Wolf (_Canis nubilis_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 72.—Boreal Mammals. _A._ Black-footed Ferret (_Mustela nigripes_). _B._ Otter (_Lutra canadensis_). _C._ Jumping Mouse (_Zapus hudsonius_).—_A_ and _B_ by permission of the N.Y. Zoölog. Soc. _C_, by permission of W. S. Berridge, London.]

[Illustration: FIG. 73.—Opossum (_Didelphis marsupialis_).—By permission of the N.Y. Zoölog. Soc.]

The most characteristic portion of North America, zoölogically speaking, is the Sonoran region of Dr. Merriam, the Warm Temperate of Dr. Allen. It crosses the continent from ocean to ocean, its northern boundary following for most of the way the 43d parallel of latitude, but over the Great Plains and Great Basin, on each side of the Rocky Mountains and the high plateaus, it extends to lat. 48°. On the south, it takes in the greater part of Mexico, covering all of the table-land of that country, the lowlands of which belong to the South American or Neotropical region. The Sonoran is invaded from the north by the long branches from the Boreal and Transition zones, which follow the three great mountain-systems in the manner already explained, and the Mexican plateau permits the similar invasion of Neotropical territory by the Sonoran fauna. Characteristic Sonoran genera, none of which extend into the Boreal, are the opossums (_Didelphis_), in the southern part a peccary (_Tagassu_) or “Wild Texas Pig,” representative of a family of swine quite different from the true pigs of the Old World, and an armadillo (_Tatu_). A very isolated form is the Prong-horned Antelope (_Antilocapra americana_); there are several species of the typically American deer (_Odocoileus_) which differ in important respects from those of the eastern hemisphere, and the Bison was very abundant until exterminated by Man. Bison, antelope and deer also reach or extend into the Boreal zone, but the former, or Wood Bison, is probably a different species from the plains animal.

[Illustration: FIG. 74.—Prong-horned Antelope (_Antilocapra americana_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 75.—Kangaroo-Rat (_Dipodomys philippii_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 76.—Thirteen-lined Spermophile (_Spermophilus tredecimlineatus_).—By permission of the N.Y. Zoölog. Soc.]

The grey foxes (_Urocyon_), Coyote (_Canis latrans_), large Timber Wolf (_Canis occidentalis_), the Caxomistle (_Bassariscus_), the Coati (_Nasua_), Raccoon (_Procyon_), Badger (_Taxidea_), three genera of skunks, pumas, several species of lynx and some bears (_Ursus_) represent the Carnivora, though one species each of raccoon, skunk, badger, puma and lynx range into the Boreal. The American types of shrews (_Blarina_) and moles (_Scalops_ and _Scapanus_) are characteristic of the Sonoran, though partially shared with the Boreal. A great many peculiar rodents inhabit the Sonoran; cotton-rats (_Sigmodon_), pocket-gophers (_Geomys_, etc.), several genera of the beautiful little kangaroo-rats (_Dipodomys_, etc.); while the prairie-dogs (_Cynomys_), the white-footed mice (_Sitomys_), wood-rats (_Neotoma_) and one genus of pocket-gophers (_Thomomys_) are chiefly Sonoran, but have Boreal representatives. The flying squirrels (_Sciuropterus_), true squirrels (_Sciurus_), ground-squirrels (_Spermophilus_), rabbits (_Lepus_), wolves (_Canis_) and otters (_Lutra_) have a very wide range through both the Boreal and Sonoran, but have many more species in the latter region.

[Illustration: FIG. 77.—Grey Squirrel (_Sciurus carolinensis_).—By permission of the N.Y. Zoölog. Soc.]

The Sonoran region may be divided into the upper and lower Sonoran zones, which are demarcated by temperature and are of transcontinental extent. Each of these zones may, in turn, be subdivided into _arid_ and _humid_ provinces, but our purpose does not necessitate entering into such refinements.

[Illustration: FIG. 78.—Grey Fox (_Urocyon virginianus_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 79.—Prairie Wolf or Coyote (_Canis latrans_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 80.—Raccoon (_Procyon lotor_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 81.—Virginia Deer (_Odocoileus virginianus_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 82.—Skunk (_Mephitis mephitis_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 83.—Mule Deer (_Odocoileus hemionus_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 84.—Badger (_Taxidea americana_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 85.—Puma or Mountain Lion (_Felis concolor_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 86.—Lynx (_Lynx rufus_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 87.—Prairie-Dog (_Cynomys ludovicianus_).—By permission of the N.Y. Zoölog. Soc.]

The Neotropical, which is the only region of the Neogæic realm, comprises the West Indian islands, all of Central and South America and the lowlands of Mexico, extending a short distance into southeastern Texas. Of its four subregions, the most typical is (1) the _Brazilian_, which includes not only Brazil, but all of South America east of the Andes and as far south as Paraguay, and is a vast area of tropical forests. (2) The _Chilian_ subregion takes in the west coast, the high Andes and the southern end of the continent, south of the Brazilian subregion; it is a country chiefly of open plains and high mountains, and a few deserts, of which South America has less than any other continent, except Europe, which has none. (3) The _Central American_ subregion reaches from the Isthmus of Panama to Mexico, the lowlands of which are included and even a small portion of southeastern Texas. (4) The _West Indian_ subregion includes all the islands of that archipelago, except Trinidad, which is a fragment of the continent, detached at a comparatively recent date; the southern extremity of Florida also belongs to this subregion.

The two subregions into which continental South America is divided are not altogether satisfactory and will doubtless require change when the distribution of South American mammals has been more accurately determined.

[Illustration: FIG. 88.—Map of the Neotropical region. (After Wallace.) Mexico inaccurate; cf. Fig. 53, p. 147.]

[Illustration: FIG. 89.—Fox-like Wolf (_Cerdocyon gracilis_).—By permission of W. S. Berridge, London.]

“Richness combined with isolation is the predominant feature of Neotropical Zoölogy, and no other region can approach it in the number of its peculiar family and generic types” (Wallace). Just as North America has received many immigrants from the Old World, so it has sent many migrants into South America, materially changing the character of the Neotropical mammalian fauna, but these intruders may be readily identified and almost seem to be out of place in their new surroundings. Not all of these northern migrants were able to maintain their footing in the southern continent and several became extinct during and at the close of the Pleistocene epoch, as was even more markedly the case with the southern forms which invaded the northern continent.

[Illustration: FIG. 90.—Spectacled Bear (_Tremarctos ornatus_).—By permission of the N.Y. Zoölog. Soc.]

There are two families of monkeys in the forested areas of South America, both very different from those of the Old World. One of these families, the marmosets (Hapalidæ), differs from all other monkeys in several particulars, most obvious of which are the long claws on the feet and the non-opposable thumb. The second family (Cebidæ) comprises forms which are superficially much more like those of the eastern hemisphere, but many of them have prehensile tails, which are used as efficient grasping organs.

Insectivora are entirely absent from the South American continent, but some shrews (_Blarina_) have entered Central America from the north and a very curious genus is represented by one species in Cuba (_Solenodon cubanus_) and another in Hayti (_S. paradoxus_). These remarkable animals are, strange to relate, most nearly allied to the tenrecs (_Centetes_) of Madagascar and by some authorities are placed in the same family.

[Illustration: FIG. 91.—_Solenodon cubanus._—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 92.—Argentine Skunk (_Conepatus gibsoni_).—By permission of W. S. Berridge, London.]

[Illustration: FIG. 93.—Little Skunk (_Spilogale putorius_).—By permission of W. S. Berridge, London.]

The Carnivora are quite numerous and varied and rather peculiar, but they all belong to northern families and are the more or less modified descendants of northern immigrants. The dogs (Canidæ) belong to genera not represented elsewhere and form a considerable assemblage of interesting types. There are no true wolves or foxes, but several species of fox-like wolves (_Cerdocyon_), with bushy tails, are common, especially in the plains regions. The Bush-Dog (_Icticyon venaticus_), a small, short-legged animal, is very peculiar. The musteline or weasel family (Mustelidæ) is rather scantily represented. There are no badgers and but few skunks (_Spilogale_ and _Conepatus_); weasels are absent, but their place is taken by the Grison (_Galera vittata_) and Tayra (_Tayra tayra_) and in the far south _Lyncodon patagonicus_. These animals are peculiar in having a lighter colouration on the back than on the belly. There are two or three species of otter (_Lutra_). The raccoons (_Procyon_) have a very wide range in South America, as in the northern continent, and the curious, long-snouted coatis (_Nasua_), which just enter the Sonoran region, are typically Neotropical. The Spectacled Bear (_Tremarctos ornatus_) is the only member of the family that occurs in South America and is confined to the highlands of Peru and Chili. The cat family is quite numerously represented; the Jaguar (_Felis onca_), which ranges from Texas to Patagonia, is a large spotted cat, rivalling the Leopard in size and ferocity; the Ocelot (_F. pardalis_, Arkansas to Paraguay) is smaller and streaked and blotched rather than spotted. The pumas differ little from those of North America, and there are many small cats, spotted, clouded and of solid colour, but no lynxes, which are essentially northern types.

[Illustration: FIG. 94.—Tayra (_Tayra tayra_).—By permission of W. S. Berridge, London.]

[Illustration: FIG. 95.—Kinkajou (_Potos caudivolvulus_), Central America.—By permission of W. S. Berridge, London.]

[Illustration: FIG. 96.—Ocelot (_Felis pardalis_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 97.—Jaguar (_Felis onca_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 98.—Collared Peccary (_Tagassu tajacu_).—By permission of the N.Y. Zoölog. Soc.]

Hoofed animals are not numerously represented in South America. The only existing Perissodactyla of the western hemisphere are the tapirs (_Tapirus_) of Central and tropical South America, a very remarkable contrast to the ancient faunas, especially of the northern continent, as will be shown in the sequel. The Artiodactyla are more varied, though very scanty in comparison with those of the Old World; even North America, which has but a poor representation of these animals, is much richer than the southern continent, where, indeed, all the hoofed animals are the descendants of comparatively recent immigrants from the north and none are truly autochthonous. Members of three different artiodactyl suborders occur in the Neotropical region; the peccaries (_Tagassu_) extend through Central and South America to Paraguay, though also entering the Sonoran region in Texas. Most interesting are the members of the camel family, which are very distinct from the true Camel of Asia. Tierra del Fuego and the Patagonian plains support great herds of the Guanaco (_Lama huanacus_), which extends along the Andes to Ecuador and Peru, where it is associated with the Vicuña (_L. vicunia_), a smaller and more slenderly built species. The Vicuña does not range south of Bolivia. Just as the mountain systems of North America carry the Boreal and Transition faunas through nearly the whole breadth of the Sonoran region, so the high Andes afford a pathway by which the mammals of the south temperate zone extend their range to the equator.

[Illustration: FIG. 99.—Vicuña (_Lama vicunia_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 100.—Florida Deer (_Odocoileus virginianus osceola_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 101.—Marsh Deer (_Blastoceros paludosus_), female.—By permission of the N.Y. Zoölog. Soc.]

The suborder Pecora of the Artiodactyla is represented in the Neotropical region only by the deer family (_Cervidæ_), of which there are several genera (or subgenera), all of them North American as distinguished from the Old World type, but some are so peculiar that they must have had a relatively long South American ancestry. The Virginia Deer (_Odocoileus virginianus_) of the northern United States is a comparatively large animal, becoming much smaller in Florida and the Southwest. The type extends through Mexico and Central America to Guiana and Peru, the Neotropical forms being so small and having such weak antlers that they are referred to separate species. Another type is the Marsh Deer (_Blastoceros paludosus)_ of eastern South America, which has short, stout antlers, each beam with two double bifurcations; there are other species of the same genus, such as the Pampas Deer of Argentina (_B. bezoarticus_). In the Andes of Peru and Chili and the forests of western Patagonia are two species of a genus which bears the preposterous name of _Hippocamelus_ and in which the antlers are simply forked. The vernacular name of these animals is “Huemul.” Peculiarly Neotropical are the little brockets, which hardly exceed a height of two feet at the shoulder, with simple spike-like antlers not more than three inches long; the genus, _Mazama_, has several species, one of which occurs as far north as the state of Puebla in Mexico. “The smallest of all deer is the Chilian pudu (_Pudua pudu_), a creature not much larger than a hare, with almost rudimentary antlers” (Lydekker). Old World types of deer, such as the Wapiti, Moose and Caribou, of the Boreal and Transition zones of North America, are entirely absent from the Neotropical region.

[Illustration: FIG. 102.—Wood Brocket (_Mazama nemorivagus_).—By permission of W. S. Berridge, London.]

[Illustration: FIG. 103.—Brazilian Tree Porcupine (_Coendou prehensilis_).—By permission of the N.Y. Zoölog. Soc.]

South America has an astonishingly rich and varied assemblage of rodents, both indigenous and immigrant, but the former are much the more important, varied and abundant. Of the four divisions of the order, all of which are represented, three are immigrants from the north and the fourth is autochthonous, but this far outnumbers the other three combined. The hares and rabbits have but very few species, one of which occurs in Brazil and is separated by a very wide interval from the one in Costa Rica, while the pikas are absent. Of the squirrel division, only the true squirrels are found, and of these there are many species, the ground-squirrels, marmots, prairie-dogs and beavers all being lacking. In the same way the rat and mouse division is represented by a single family. The vesper or white-footed mice (_Sitomys_) have invaded the southern continent and a number of peculiar genera have arisen there, but all of northern ancestry, such as the groove-toothed mice (_Rheithrodon_) and the fish-eating rats (_Ichthyomys_). The voles, or meadow-mice, the muskrats, jumping mice, kangaroo-rats and pocket-gophers of the northern continent are all absent. While the immigrant suborders have thus but one family each in South America, the case is very different with the fourth or porcupine group, of which that continent is to-day, as it has been for ages past, the headquarters. No less than six families and twenty-nine genera are known, all of the genera and four of the families being restricted to the Neotropical region. Contrast this assemblage with the extreme scantiness of this group in North America, where but a single genus, the Short-tailed or Canada Porcupine (_Erethizon_) represents it, and that is a late immigrant from the south.

[Illustration: FIG. 104.—Neotropical rodents. _A._ Vizcacha (_Viscaccia_). _B._ Paca (_Agouti paca_). _C._ Rock Cavy (_Cavia rupestris_). _D._ Water-Hog, or Carpincho (_Hydrochærus_). _D_, by permission of the N.Y. Zoölog. Soc. _A_, _B_, _C_, by permission of W. S. Berridge, London.]

[Illustration: FIG. 105.—Chinchilla (_Chinchilla laniger_).—By permission of W. S. Berridge, London.]

It would lead us too far to attempt a description of this horde of curious and interesting rodents, so only a few of the more striking and characteristic forms can be mentioned. There are two genera of porcupines (_Coendou_ and _Chætomys_), both arboreal, which belong in the same family as the North American _Erethizon_, but are distinguished by their long, prehensile tails, which they use, as monkeys and opossums do, for grasping and climbing. The very large family of the Octodontidæ has 17 Neotropical genera and four others are found in Africa. The Degu (_Octodon_) of Chili, Bolivia and Peru has the appearance of a large rat with tufted tail; the tuco-tucos (_Ctenomys_) are extremely abundant burrowers in Patagonia, where they honeycomb the ground over wide areas. The spiny rats (_Echimys_ and _Loncheres_) are so called from their appearance, not because they are related to the true rats; they have numerous horny spikes through the fur of the back. The Coypu (_Myocastor_) is a large, aquatic animal, remotely like the northern Muskrat, and the Hutias (_Capromys_ and _Plagiodontia_) are arboreal and found only in Cuba, Hayti and Jamaica. The chinchillas (_Chinchilla_ and _Lagidium_) of the Andes and the Vizcacha (_Viscaccia_) of the Argentine plains have somewhat the appearance of hares, but with long and bushy tails. The cavies, to which the familiar, misnamed Guinea-Pig (_Cavia porcellus_) belongs, are a very characteristic family; besides the true cavies, it includes the Patagonian Mara (_Dolichotis_), a large, long-legged, long-eared, short-tailed creature, and the Water-Hog, or Carpincho (_Hydrochærus_), an aquatic animal, as its name implies, and much the largest of existing rodents; it occurs in the warmer regions, south to Argentina. The heavy Paca (_Agouti_) and the slender-limbed Agouti (_Dasyprocta_) make up another family. Altogether, this assemblage of the porcupine-like suborder of rodents is a very remarkable one and in no other region of the earth is anything like it to be found.

[Illustration: FIG. 106.—Hairy-rumped Agouti (_Dasyprocta prymnolopha_).—By permission of W. S. Berridge, London.]

[Illustration: FIG. 107.—Three-toed Sloth (_Bradypus tridactylus_).—By permission of the N.Y. Zoölog. Soc.]

With the exception of one genus of armadillos, which has invaded Texas, the entire order of the Edentata is at present confined to the Neotropical region, the so-called edentates of the Old World now being removed to other orders. The Edentata, which were once far more varied and abundant than they now are, comprise three groups of animals so bizarre and strange that they seem more like fabulous creatures than actual, living mammals. One group, or suborder, is that of the sloths (Tardigrada), arboreal, shaggy animals, with short, almost monkey-like head and no tail; their very long legs and hook-like feet make them nearly helpless on the ground, but are very useful for hanging from the branches of the trees, in which the creatures live. Indeed, the sloths are the only mammals which habitually hang in a suspended position. Two genera of sloths inhabit the tropical forests, between which the most obvious difference is that in one (_Bradypus_) the forefoot has three toes, and in the other (_Cholœpus_) two.

[Illustration: FIG. 108.—Two-toed Sloth (_Cholœpus didactylus_).—By permission of W. S. Berridge, London.]

The suborder of the anteaters (Vermilingua) is more varied, and is the only one of the order to which the term “edentate” applies strictly, for they alone in the order are altogether toothless. The great Ant-Bear (_Myrmecophaga jubata_), which may reach a total length of seven feet, has an extravagantly long, slender and nearly cylindrical head, long, shaggy, black and white hair and an immense, bushy tail; the forefeet are armed with huge, sharp-pointed claws, which are used for tearing open ant-hills, and when occasion arises, as formidable weapons of defence, for the Ant-Bear can successfully repulse even the Jaguar. In walking, the claws are curved inward and the preposterous beast rests his weight upon the outside edges of the forefeet, while the hind feet apply the sole to the ground, as does a bear or raccoon. The Collared Anteater (_Tamandua_) is much smaller and mainly arboreal in habits. It has a short-haired, black body, with a white stripe down the back, white neck and limbs, a colour-pattern which gives to the animal the appearance of wearing a close-fitting black jacket; the long tail, which has some cross bars, is short-haired, very different from the extremely bushy tail of the Ant-Bear. The little Two-toed Anteater (_Cyclopes didactylus_), hardly larger than a rat, is exclusively arboreal and has a prehensile tail, like so many other South American mammals. Sloths and anteaters are forest animals and are not found west of the Andes or south of Paraguay.

[Illustration: FIG. 109.—Ant-Bear (_Myrmecophaga jubata_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 110.—Collared Anteater (_Tamandua tetradactyla_).—By permission of the N.Y. Zoölog. Soc.]

[Illustration: FIG. 111.—Six-banded Armadillo (_Dasypus sexcinctus_).—By permission of the N.Y. Zoölog. Soc.]

The third existing suborder of edentates is that of the armadillos (Dasypoda), which have a very complete armour of bony scutes, ossifications in the skin, covered with scales of horn. They are all more or less burrowers in habit and omnivorous in diet, eating roots, insects, worms, etc.; the extraordinary rapidity with which they burrow into the ground is almost their only way of escape from pursuit, but in one genus, _Tolypeutes_, the animal can roll itself into a ball, completely protected by mail all around. The armadillos are much more varied than the anteaters or sloths and have a wider geographical range, extending from Texas to Patagonia. The head, which is long-snouted, is protected by a shield made up of numerous horn-covered plates of bone, and the tail is encased in a tubular sheath of more or less regular rings, each ring of bony plates and horny scales. The body-shield, or carapace, which covers the back and sides, consists of an anterior and posterior buckler, in which the plates are immovably attached to one another by their edges, and between the two is a series of movable, overlapping bands, the number of which varies in the different genera. In the little Pichiciago (_Chlamydophorus truncatus_) the head and back are covered with four-sided plates of horn, the bony scutes being small and thin and much reduced. The carapace has no bucklers, but about 20 transverse rows of plates, and is attached along only the middle line of the back and beneath it the body is covered with silky, white fur; the rump is covered with a solid shield of bone, placed nearly vertically and covered with thin scales, and is notched below for the tail, altogether a most exceptional arrangement. Seven or more distinct genera of armadillos are found in the Neotropical region and they display a great range in size; the Giant Armadillo of Brazil (_Priodontes_) is a yard or more in length, while the little _Zaëdyus_ of Patagonia is smaller than a rabbit and, least of all, the Pichiciago is but five inches long.

[Illustration: FIG. 112.—Nine-banded Armadillo (_Tatu novemcinctus_).—By permission of the N.Y. Zoölog. Soc.]

Two families of marsupials occur in South America. The opossums are much more numerous and varied than in North America; three genera and a large number of species, some not larger than mice, range through the forested parts of the continent. Of particular interest is the little _Cænolestes_, which has two species, with two enlarged lower front teeth, the sole survivors of a group which is abundantly represented in the Tertiary deposits of Patagonia.

The fauna of the Central American subregion is less rich and characteristic than that of the Brazilian and is, to a certain extent, transitional to that of the Sonoran region of North America, several genera proper to the latter region extending into it, which are not known to pass the Isthmus of Panama, such as shrews, a fox and one of the pocket-mice. The West Indian islands are exceedingly poor in mammals, a great contrast to the East Indian, or Malay, Archipelago; only a few rodents, insectivores and bats occur in them.