VESPERTILIONIDE.-Cosmopolitan. Represented by the cosmopolite genus Vespertilio. EMBALLONURIDE.-Warmer parts of the world. Represented by the genus Taphozous. CENTETIDE.-Confined to Madagascar except one genus (Solenodon) in the West Indies. Represented in Madagascar by nearly a dozen species. Genera :—Centetes, Hemicentetes, Ericulus, Oryzorictes, Echinops.

SORICIDE.-The whole world, except South America and Australia. Represented in Madagascar by one or two species of Crocidura, a genus found in Africa, and the warmer parts of the eastern hemisphere generally.

MURIDE.-Cosmopolitan. Represented by several genera of African affinities, namely, Nesomys, Brachytarsomys, Hypogeomys.

VIII.-ANTARCTIC REALM.

The Antarctic Realm is geographically almost wholly oceanic, and its fauna hence consists almost exclusively of marine or pelagic species. It necessarily embraces not only the Antarctic Zone, but a large part of the cold south-temperate, since very few of its characteristic species. are wholly restricted to the Antarctic waters. It will hence include not only the few small groups of Antarctic Islands, but also Tierra del Fuego and the Falkland Islands, and perhaps also the extreme southern shores of South America, while some of its characteristic forms also extend to New Zealand, and even Australia and the Cape of Good Hope. The only mammals that can be considered as strictly characteristic of this region are Pinnipeds and Cetaceans, of which several genera of each are almost wholly restricted to it. A "South Frigid", "Antarctic", or "South Circumpolar" "Zone", "Region", or "Realm", has been recog nized by various writers for the marine invertebrates, and, by von Pelzeln for birds, with limitations much as here assigned. While the number of species peculiar to it is small, it is large relatively to the whole number represented, especially in the colder latitudes. There is, of course, a broad belt along its northern border of a transitional character, where Antarctic types overlap the range of groups characteristic of south-temperate latitudes.

One of the most important features of the South Circumpolar or Antarctic Realm is the resemblance of its life to the marine life of the Arctic or North Circumpolar Realm. While perhaps in no case are the species identical, the genera are frequently the same, not only among the mammalia, but among invertebrates. This is especially significant as regards the mammalia, since the terrestrial mammals of the extreme north and extreme south present no such parallelism, but the utmost divergence. Among Pinnipeds, most of the genera are peculiar to either the northern or southern waters, but in several instances the genera of the two regions are strictly representative. Thus, Otaria and Aretocephalus of the Southern Seas are represented in the Northern by Eume topias and Callorhinus. Zalophus and Macrorhinus are both Northern and Southern. Stenorhynchus, Lobodon, Leptonyx, and Ommatophoca are strictly Southern, while Phoca, Halichorus, Erignathus, Cystophora, Monachus, and one or two others, are strictly Northern, as are also the Walruses. The Mysticete, or Baleen Whales, among Cetaceans, have

a somewhat similar distribution. While a few genera are restricted respectively to the Northern and Southern waters, the larger unmber are common to both, though represented by different species in the two regions, while they are (in some cases at least) absent from the intervening tropical seas. A large proportion of the Denticete, or Toothed Whales (Dolphins, Porpoises, Rorquals, etc)., are either limited to the warmer seas or have there their chief development, quite a number of genera being peculiar to the tropics. Others, however, like Monodon, are eminently boreal, while others, like Beluga, are common to the colder waters both north and south of the tropics. In most cases, however, we know as yet too little respecting the range of the different species and genera of Cetacea to be able to make much use of them in determining questions in geographical zoölogy.

This similarity between the marine life of the Arctic and Antarctic Regions evidently indicates that the forms common to the two had a common origin, and, at some former period, a continuous, probably circumtropical, distribution, and that on the increase of temperature in the intertropical regions, through well-known geological causes, they sought the more compatible cooler waters toward the poles. The similarity of the Arctic and Antarctic marine life is also a feature that sharply differentiates the fauna of the South Circumpolar Realm from that of the South Temperate and Tropical Zones.

III.-GENERAL SUMMARY.

As stated at the beginning of the present paper, one of the chief topics here proposed for discussion was the influences and laws which govern the distribution of life,-whether it is or is not co-ordinated with climatic zones, and governed in a large degree by climatic conditions, and especially by temperature. In fact, so generally is temperature recognized by the leading writers on the distribution of marine life that it seems superfluous to reiterate or emphasize this principle. That the zones of life should be perhaps a little less obvious over the land-areas,—in consequence of the diversity of contour resulting from differences of elevation, and the interruptions and exceptional conditions due to mountain chains and high plateaus,-than over the oceanic expanses, is naturally to be expected. That there is, however, a similar correspondence between climatic belts and the zones of life seems to me abundantly evident. As has been already shown, the broader or primary zones are, first, an Arctic or North Circumpolar Zone, embracing the arctic, subarctic, and colder temperate latitudes of the northern hemisphere, throughout the whole of which area there is a marked homogeneity of mammalian life, as well as of animal and vegetable life in general; secondly, that below this there is a broad belt of life, which, in its general facies, is distinctive of the temperate and warm-temperate latitudes, and that these two zones of life are far more closely related inter se than with the life of the intertropical regions, with which regions they may be collectively contrasted, and together receive the appropriate name of "Arctogœa”;

thirdly, it has been shown, so far as the northern hemisphere is concerned, that the life of the tropical and temperate regions of the same continent is more widely different than is the life of corresponding por tions of the temperate and colder parts of the (so-called) Old World and the New; fourthly, that the life of Tropical America has very little in common with that of the tropical portions of Asia and Africa; fifthly, that the life of the South Temperate Zone presents a facies distinct from that of the tropics, and bas still less in common with that of the North Temperate Zone; sixthly, that Australasia is so highly differentiated as to form a distinct primary region, having little in common with other lands, even with those of contiguous regions, or those having a similar geographical position; seventhly, that Madagascar and its contiguous islands, while to some extent African in affinity, form also a highly specialized region; lastly, that the antarctic and cold south-temperate oceanic regions are recognizable as a primary region, characterized by a peculiar general facies of life that more strongly recalls that of the corresponding portions of the northern hemisphere than of any other portion of the earth. It has been further shown that the Australian Realm is divisible into temperate and tropical portions, and also that the land surface is separable into zones of even still narrower limits, corresponding in a general way with those recognized by Dana for marine life.

The almost total absence of identical genera, or even of families, excepting such as are essentially cosmopolitan, in the American and Old World tropics, as well as the distinctness of the Lemurian Realm, and the almost total isolation of the Australian Realm, evidently require for their explanation other causes than merely the existing climates. The geological history of these land-areas and their faunæ must be of course considered in order to understand their present relationships. As the northern hemisphere at present most clearly shows, nearly continuous land surface and similarity of climatic conditions implies identity of fauna, while isolation, especially when joined with diverse climatic conditions, implies diversity of life, and a differentiation propor tionate to the degree of isolation, and the length of time such isolation has existed; in other words, that the present want of affinity between the life of the Lemurian and Australian Realms and that of the rest of the world is due rather to their long geographical isolation than to present climatic conditions, and that we here find, for reasons perhaps not wholly apparent, the remnants of a somewhat primitive or early fauna that was formerly shared more largely by other areas than at present, that these regions became isolated before the development of many of the higher and now prevalent types of the larger and more diversified land areas, and that here differentiation has proceeded less rapidly and along fewer and narrower lines than elsewhere; further more, that the present highly diversified fauna of the chief tropical areas, in comparison with the fauna of the north-circumpolar lands, is due in part to the southward migration, near the close of the Tertiary

period, of forms adapted to a high temperature, and in part to the high rate of differentiation favored by tropical conditions of climate. Hence, given: 1. Arctic and cold-temperate conditions of climate, and we have a fauna only slightly or moderately diversified; 2. A moderate increase of temperature, giving warm-temperate conditions of climate, and we have the addition of many new types of life; 3. A high increase of temperature, giving tropical conditions of climate, and we have a rapid multiplication of new forms and a maximum of differentiation. Again, given: 1. A long-continued continuity of land surface, and we have an essential identity of fauna; 2. A divergence and partial isolation of land-areas, and we find a moderate but decided differentiation of faunæ; 3. A total isolation of land-areas, and we have a thorough and radical differentiation of faunæ, proportioned to the length of time the isolation has continued. Hence, the present diversity of life is correlated with two fundamental conditions: 1. Continuity or isolation, past as well as present, of land surface; and, 2. Climatic conditions, as determined mainly by temperature.*

In accordance with these principles, which rest on incontrovertible facts of distribution, it follows that the nearly united lands of the North present a continuous, almost homogeneous, arctopolitan fauna; that farther southward, in the warmer temperate latitudes, we begin to find a marked differentiation on the two continents; that this differentiation is still further developed in the tropical continuations of these same land-areas, till an almost total want of resemblance is reached, except that there is what may be termed, in contrast with the more northern regions, a "tropical facies" common to the two. The small amount of land surface belonging to these primary land regions south of the tropics have no more in common (a few marine species excepted) than have these two tropical areas, but it is hardly possible for them to have much less. The Antarctic (mainly oceanic) region has a fauna strongly recalling the marine fauna of the Arctic, but has no resemblance to that of the intervening area.

The northern circumpolar lands may be looked upon as the base or centre from which have spread all the more recently developed forms of mammalian life, as it is still the bond that unites the whole. Of the few cosmopolitan types that in a manner bind together and connect the whole mammalian fauna of the globe (the Lemurian and Australian Realms in part excepted), nearly all have either their true home or be long to groups that are mainly developed in the northern lands. A few

In illustration of the above, it may be added that the circumpolar lands north of the mean annual of 36° F., or, in general terms, north of the fiftieth parallel, with approximately an area of about 12,500,000 square miles, have representatives of about fifty-four genera of mammals; Tropical America, with an approximate area of about 5,000,000 square miles, has about ninety genera; the Indo-African Realm, with an approximate area of about 15,000,000 square miles, has about two hundred and fifty genera. Hence the tropical lands are four to five times richer in genera, in proportion to area, than those of the Cold-temperate and Arctic regions.

have been pressed a little to the southward by the extreme rigor of an Arctic climate, but are still characteristic elements of all boreal faunas. The very few truly tropicopolitan mammalia are either Chiroptera, or marine, or at least aquatic, and have thus exceptional means of dis persal.

The primary regions and their subdivisions, recognized in the preced ing pages, are enumerated in the subjoined schedule.

1.-Primary divisions, or "Realms".

1. An ARCTIC, or NORTH CIRCUMPOLAR.

II. A NORTH TEMPERATE, divided into two regions and eight prov inces.

III. An AMERICAN TROPICAL, with three regions. (Provinces not characterized.)

IV. An INDO-AFRICAN, with two regions and five provinces.

V. A SOUTH AMERICAN TEMPERATE, with two provinces. VI. An AUSTRALIAN, with three regions and two provinces. VII. A LEMURIAN.

VIII. An ANTARCTIC or SOUTH CIRCUMPOLAR.

2.-Secondary divisions, or "Regions".

II. North Temperate Realm: 1, American; 2, Europæo-Asiatic. III. American Tropical Realm: 1, Antillean; 2, Central American; 3, Brazilian.

IV. Indo-African Realm: 1, African; 2, Indian.

VI. Australian Realm: 1, Australian (Australia, Tasmania, and New Guinea); 2, Polynesian; 3, New Zealand.

3.-Divisions of third rank, or "Provinces".

II, 1. American Region: a, Boreal*; b, Eastern; c, Middle; d, Western. II, 2. Europæo Asiatic Region: a, European; b, Siberian; c, Mediter ranean; d, Manchurian.

IV, 1. African Region: a, Eastern; b, Western; c, Southern.

IV, 2. Indian Region: a, Continental; b, Insular.

V.

South American Temperate Realm: a, Andean; b, Pampean. VI, 1. Australian Region: a, Australian; b, Papuan.

*A "Boreal" province has not been distinctly recognized in the preceding pages as a division belonging to the same category as the other so-called or commonly recognized provinces, and is not at all recognized in the table of distribution given at p. 339. It is nearly equivalent to what is there implied by "Cold Temperate". I hope soon to be able, in a paper to be devoted especially to a consideration of the geographical distribution of North American mammals, to define and characterize it more definitely.