Indian Region the possession of the Tragulida. Its peculiar genera consist largely of Anthropoid Apes, found elsewhere only in India, but also includes several each of Carnivores Bats, and Rodents. It is preeminently the tropical province of the African Region. While it contains a smaller number of genera than either of the others, it has relatively a much larger number restricted to it, having eighteen peculiar genera out of a total number of seventy-five, while the Eastern Province, with ninety-one genera, has only twelve that are peculiar, and the Southern seventeen out of eighty-two.

The number of genera represented in the African Region, and their range, is approximately as follows:

Indian Region.-The Indian Region may be defined, in general terms, as consisting of Intertropical Asia. It hence embraces Continental India

from the Lower Indus to the Formosa Straits, the islands of the Indian Archipelago, as well as Formosa, the Philippines, Celebes, and all of the Sunda Islands. As far as the mammalia are concerned, only two primary subdivisions, or provinces, seem to be recognizable, the one a Northern, or Continental, the other a Southern, or Insular ("Malayan"). The former, or Continental, includes nearly all of the Hindostan and IndoChinese Peninsulas, excepting the extreme southern border of the latter and Malacca. These areas belong to the Insular Province, which comprises not only Borneo, Sumatra, and Java, but all of the above-named smaller islands to the eastward, except Formosa, which pertains to the Continental Province.

The long, narrow Malaccan Peninsula is almost insular in position and character, and agrees far better, climatologically, and in its productions, with Borneo and Sumatra, than with the mainland to the northward, as does, in fact, the extreme coast border of the mainland, embracing Lower Cochin China, Cambodia, etc. The small outlying islands to the eastward have nothing in common with the Australian Realm (if we exclude the wide-ranging Chiroptera and a few marine forms, which are, of all mammals, of least importance in a zoögeographical point of view), except the single Marsupial genus Cuscus occurring in Timor and Celebes, while no placental mammals except Sus, a few Murine genera, the Dugong, and Chiroptera, reach any portion of the Australian Realm. Malacca, Borneo, and Sumatra form the central and typical portion of the Insular or Malayan Province, being, from their larger area and closer proximity to each other and to the tropical mainland, far richer in genera and species than the smaller and more remote islands to the southward and eastward. Even Java has a less varied mammalian fauna than either Borneo or Sumatra, and thus differs from them negatively rather than by the possession of peculiar types. Thence eastward, throughout the Sunda Islands, the differences are almost wholly such as result from the small size and isolated position of these insular areas, through a gradual disappearance of many types present in the larger islands. The Philippines, for simi lar reasons, lack a large proportion of the genera found in the central portion of the province, while those they do possess, with few excep tions, are such as are common to the larger areas. The few that are peculiar are Indian, rather than Australian, in their affinities.

Celebes and Timor contain one strictly Australian genus (Cuscus, represented by several species), but the few other mammals found there are either Indian or possess strictly Indian or Indo-African affinities. Hence I fail to see any good reason for assigning Celebes and all the smaller Sunda Islands to the Papuan Province, as Mr. Wallace and others have done, but abundant evidence that such is not their real affinity. Even Mr. Wallace's own tables of distribution show at a glance the wide disassociation of these islands from the Papuan fauna, and their much nearer relation to the Indian, there being but one typically Australian or Papuan form represented in any of them, while none of the placental

land mammals (excepting several subtropicopolitan genera of Bats and a few Muriform Rodents) are common to these islands and the PapuanAustralian division. The genera peculiar to the Philippines and Celebes (except Cuscus in the latter) have little if any more significance than the occurrence in Borneo and Sumatra of a few genera wholly restricted to one or the other of these last-named islands.

Ceylon and the adjoining low-coast portions of the Hindostan Peninsula are more tropical in character than the plateau region to the north. ward. While a few genera are restricted to this small area, and many more species occur here that are not found to the northward, the differentiation seems hardly great enough to warrant the separation of these areas as a region of co-ordinate rank with the "Malayan". It hence seems to me that Mr. Wallace has too emphatically recognized this comparatively unimportant difference in making it the basis of a distinct subregion (termed by him the "Ceylonese Subregion"). The only mammalian genera peculiar to this division are a genus of Lemurs (Loris), three genera (or subgenera) of Herpestina (Calictis, Taniogale, Onychogale), and a genus of Mice (Platacanthomys), each represented by a single species, and, so far as known, of limited distribution.

Continental Province.-As already intimated, the Continental Province includes nearly all of Hindostan and Indo-China, or the whole of the tropical portion of the Asiatic continent excepting Malacca and the southern portions of Tenasserim, Siam, Cambodia, and Cochin China. It also extends into Southern China somewhat beyond the tropic (probably to the divide between the Li-kiang and Yang-tse-kiang Rivers), and also to the southern slope of the Himalayas.*

The plains of the Upper Indus appear, however, to belong to the Temperate Region to the northward, as does probably most of the country northwest of Delhi. The greater part of the interior of the Hindostan Peninsula has a less tropical character and a less varied fauna than Bengal, Assam, and Burmah, situated under the same parallels. I cannot agree, however, with Messrs. Blyth, Blandford, and von Pelzeln,t

*"On the southern slope of the Himalayas there is everywhere, until it has been cleared, luxuriant forest up to at least 12,000 feet above the sea, inhabited by a fauna which extends, without any great change of generic forms, throughout the Malay Peninsula and into the hill tracts of some at least of the Malay Islands."-BLANDFORD, Proc. Zool. Soc. Lond., 1876, p. 632.

+ Mr. Blyth makes "Hindostân proper, or the plains of Upper India east and south of the North West desert; Dukhun, or tableland of the Peninsula of India, and the intervening territory, inclusive of the Vindhaian ghâts; Coromandel Coast and low northern half of Ceylon" a subregion of his "Ethiopian Region" (Nature, vol. iii, p. 428). Mr. Blandford holds that the "hills of Southern India with the Malabar Coast and Southern Ceylon form a province of the Malay region, whilst the greater portion of the Indian peninsula is African in its affinities" (Proc. Zoöl. Soc. Lond., 1876, p. 632). Von Pelzeln considers India proper, from the Lower Brahmaputra River westward, a distinct primary region, which he calls the "hindostanische Region". His "malayische Region" hence consists of Warin-temperate and Tropical Asia, minus the Hindostan Peninsula, to which he adds the Philippines, Borneo, Bali, Java, and Sumatra. It includes China as far as the Yang-tse-kiang River, and the Himalayan plateau from

that the larger part of Hindostan should be joined to the African Region rather than the Indian, since only a very few African genera occur here that do not also range far to the eastward, or almost throughout the Indian Region. According to von Pelzeln,* about one-third of the genera of the "hindostanischen Fauna" are peculiar to it, while it shares almost another third with Indo-China. The remaining third (fourteen genera) are common to the African Region, but all except four of them occur also more or less generally over the Indian Region. Of these, two (Hyæna and "Ratelus" Mellivora) scarcely reach the limits of the Indian Region as here defined. Among the genera given by him as peculiar are, however, several that range beyond the Indian Peninsula.

=

There is more reason for Mr. Wallace's separation of the Hindostan Peninsula from the Indo-Chinese portion of the Indian Region, and its subdivision into two "subregions"-a northern "Hindostan Subregion" and a southern "Ceylonese Subregion". As already shown, the latter has a number of peculiar forms, while three or four genera are also peculiar to the Hindostan Peninsula at large. But the scale of division that would make the Hindostan Peninsula separable into two subregions would also require a somewhat similar subdivision of Indo-China, making four divisions of what I here term the Continental Province. While these divisions would have some natural basis, they are too detailed to come into the category of divisions for which I adopt the term "prov ince".

Continental Province.-The Continental Province, with the limitations here assumed, is nearly equivalent to Mr. Wallace's three "subregions", termed respectively "Hindostan", "Ceylonese", and "Indo-Chinese". Of about ninety-four genera represented in it, about two-thirds have a pretty general range throughout the province, while only about oneeighth are limited to the Hindostanese portion, including those already named as almost peculiar to Ceylon and the low coast region east of the Eastern Ghâts. Excluding about a dozen that range over at least half the surface of the globe, ope-third of the remainder (more than onefourth of the whole) are common to the African Region; more than onehalf (almost one-half of the whole) are restricted to the Indian Region and a little more than one-fifth (about one-eighth of all) are peculiar to the province. This shows, as already noted in discussing the fauna Burmah, Assam, and Bengal to the Kuenluen Mountains, thus embracing Nepal, Butan, and Thibet. It is divided into five subregions, the two northernmost of which belong mainly to the North Temperate Realm. (Festschrift z. Feier des fünfundzwanzigjäbrigen Bestehens d. K.-K. Zool.-Bot. Gesells. in Wien, 1876, pp. 53-74 u. Karte.) The fauna of the Thibetan plateau, as claimed by Mr. Blandford, being boreal and alpine, and having almost nothing in common with the tropical region to the southward, the artificial character of von Pelzeln's "subregions" is shown by his assuming the Yangtse-kiang River to be a natural boundary between two primary regions, and his sepa ration of Malacca from Sumatra and Borneo to form a part of his "hinter-indische Unterabtheilung", which thus consists of the whole of the Indo-Chinese Peninsula down to the very southern extremity of Malacca!

* Verhandl. d. K.-K. Zool.-Bot. Gesells. in Wien, xxv. Bd., p. 57, 1875.

of the Indo-African Realm, how strong an affinity exists between the African and Indian Regions, two-fifths of all the genera of the Indian Region which have an extralimital range occurring also in the African Region. The close affinity of the two provinces of the Indian Region is shown by the fact that two-thirds of the peculiar Indian genera found in the Northern or Continental division range also into the Southern or Insular. As will be shown later, the Insular Province is the more highly specialized of the two divisions.

Restricted (almost wholly) to the province

Other genera ranging over most of the Indian Region and restricted to it .......

Common to the African Region

Common to portions of the Europæo-Asiatic Region

Ranging over most of the northern hemisphere

Nearly cosmopolite ...

94

43

16

27

28

34

17

10

5