or mainly Oriental, and most of them are genera which range widely over the region, only two (Philentoma and Rollulus) being exclusively Malayan, and two others (Megalurus and Malacocircus) more especially Indian or continental. Five other genera, though having a wide range, are typically Palearctic, and have reached the islands through North China. They are, Monticola, Acrocephalus, Phylloscopus, Calliope, and Passer; the two first having extended their range southward into the Moluccas. The peculiarly Australian genera are only 12, the majority being characteristic Papuan and Moluccan forms; such as-Campephaga, Alcyone, Cacatua, Tanygnathus, Ptilopus, Janthanas, Phloganas, and Megapodius. One is peculiar to Celebes (Prioniturus); one to the Papuan group (Cyclopsitta); and one is chiefly Australian (Gerygone). The beautiful little parroquets forming the genus Loriculus, are characteristic of the Philippines, which possess 5 species, a larger number than occurs in any other group of islands, though they range from India to New Guinea. There remain six peculiar genera-Rhabdornis, an isolated form of creepers (Certhiidae): Gymnops, a remarkable bareheaded bird belonging to the starlings (Sturnida); Dasylophus, and Lepidogrammus, remarkable genera of cuckoos (Cuculida); Penelopides, a peculiar hornbill, and Phapitreron, a genus of pigeons. Besides these there are four other types (here classed as sub-genera, but considered to be distinct by Lord Walden) which are peculiar to the Philippines. These are Pseudoptynx, an owl of the genus Athene; Pseudolalage, a sub-genus of Lalage; Zeocephus, a subgenus of Tchitrea; and Ptilocolpa, included under Carpophaga.

When we look at the position of the Philippine group, connected by the Bashee islands with Formosa, by Palawan and the Sooloo archipelago with Borneo, and by the Tulour and other islets with the Moluccas and Celebes, we have little difficulty in accounting for the peculiarities of its bird fauna. The absence of a large number of Malayan groups would indicate that the actual connection with Borneo, which seems necessary for the introduction of the Malay types of mammalia, was not of long duration; while the large proportion of wide-spread continental genera of birds would seem to imply that greater facilities had

once existed for immigration from Southern China, perhaps by a land connection through Formosa, at which time the ancestors of the peculiar forms of deer entered the country. It may indeed be objected that our knowledge of these islands is far too imperfect to arrive at any satisfactory conclusions as to their former history; but although many more species no doubt remain to be discovered, experience shows that the broad characters of a fauna are always determined by a series of collections made by different persons, at various localities, and at different times, even when more imperfect than those of the Philippine birds really are. The isolated position, and the volcanic structure of the group, would lead us to expect them to be somewhat less productive than the Moluccas, close to the rich and varied Papuan district, or than Celebes, with its numerous indications of an extensive area and great antiquity; and taking into account the excessive poverty of its mammalian fauna, which is certain to be pretty well known, I am inclined to believe that no future discoveries will materially alter the character of Philippine ornithology, as determined from the materials already at our command.

Java. Following the same plan as we have adopted in first discussing the Philippine islands, and separating them from the body of the sub-region on account of special peculiarities, we must next take Java, as possessing marked individuality, and as being to some extent more isolated in its productions than the remaining great islands.

Java is well supplied with indigenous mammalia, possessing as nearly as can be ascertained 55 genera and 90 species. None of these genera are peculiar, and only about 5 of the species, -3 quadrumana, a deer and a wild pig. So far then there is nothing remarkable in its fauna, but on comparing it with that of the other great islands, viz., Borneo and Sumatra, and the Malay peninsula, we find an unmistakable deficiency of characteristic forms, the same in kind as that we have just commented on in the case of the Philippines, though much less in degree. First, taking genera which are found in all three of the above-named

localities and which must therefore be held to be typical Malayan groups, the following are absent from Java: Viverra, Gymnopus, Lutra, Helarctos, Tapirus, Elephas, and Gymnura; while of those known to occur in two, and which, owing to our imperfect knowledge, may very probably one day be discovered in the third, the following are equally wanting: Simia, Siamanga, Hemigalea, Paguma, Rhinosciurus, and Rhizomys. It may be said this is only negative evidence, but in the case of Java it is much more, because this island is not only the best known of any in the archipelago, but there is perhaps no portion of British India of equal extent so well known. It is one of the oldest of the Dutch possessions and the seat of their colonial government; good roads traverse it in every direction, and experienced naturalists have been resident in various parts of it for years together, and have visited every mountain and every forest, aided by bands of diligent native collectors. We should be almost as likely to find new species of mammalia in Central Europe as in Java; and therefore the absence of such animals as the Malay bear, the elephant, tapir, gymnura, and even less conspicuous forms, must be accepted as a positive fact.

In the other islands there are still vast tracts of forest in the hands of natives and utterly unexplored, and any similar absence in their case will prove little; yet on making the same comparison in the case of Borneo, the most peculiar and the least known, of the other portions of the sub-region, we find only 2 genera absent which are found in the three other divisions, and only 3 which are found in two others. A fact to be noted also is, that the only genus found in Java but not in other parts of the sub-region (Helictis) occurs again in North India; and that some Javan species, as Rhinoceros javanicus, and Lepus kurgosa occur again in the Indo-Chinese sub-region, but not in the Malayan.

Among the birds we meet with facts of a similar import; and though the absence of certain types from Java is not quite so certain as among the mammalia, this is more than balanced by the increased number of such deficiencies, so that if a few

should be proved to be erroneous, the main result will remain unaltered.

Java possesses about 270 species of land birds, of which about 40 are peculiar to it. There are, however, very few peculiar genera, Laniellus, a beautiful spotted shrike, being the most distinct, while Cochoa and Psaltria are perhaps not different from their Indian allies. The island has however a marked individuality in two ways-in the absence of characteristic Malayan types, and in the presence of a number of forms not yet found in any of the other Malay islands, but having their nearest allies in various parts of the Indo-Chinese sub-region. The following 16 genera are all found in Malacca, Sumatra, and Borneo, but are absent from Java: Setornis, Temnurus, Dendrocitta, Corydon, Calyptomena, Venilia, Reinwardtipicus, Caloramphus, Rhinortha, Nyctiornis, Cranorhinus, Psittinus, Polyplectron, Argusianus, Euplocamus, and Rollulus. The following 9 are known from two of the above localities, and will very probably be found in the third, but are absent from, and not likely to occur in, Java: Trichicos, Eupetes, Melanochlora, Chaptia, Pityriasis, Lyncornis, Carpococcyx, Poliococcyx, and Rhinoplax. We have thus 25 typically Malayan genera which are not known to occur in Java.

The following genera, on the other hand, do not occur in any of the Malayan sub-divisions except Java, and they all occur again, or under closely allied forms, in the Indo-Chinese subregion: Brachypteryx (allied species in Himalayas); Zoothera allied species in Aracan); Notodela (allied species in Pegu); Pnoëpyga (allied species in Himalayas); Allotrius (allied species in the Himalayas); Cochoa (allied species in the Himalayas); Crypsirhina (allied species in Burmah); Estrilda (allied species in India); Psaltria (allied genus-Egithaliscus-in Himalayas); Pavo muticus and Harpactes oreskios (same species in Siam and Burmah); Cecropis striolata (same species in Java and Formosa, and allied species in India).

Here we have 12 instances of very remarkable distribution, and considering that there are nearly as many birds known from Sumatra and Borneo as from Java, and considerably more from

the Malay peninsula, it is not likely that many of these well marked forms will be discovered in these countries.

There are also a considerable number of species of birds common to Malacca, Sumatra, and Borneo, but represented in Java by distinct though closely allied species. Such are,

Now if we look at our map of the region, and consider the position of Java with regard to Borneo, Sumatra, and the IndoChinese peninsula, the facts just pointed out appear most anomalous and perplexing. First, we have Java and Sumatra forming one continuous line of volcanoes, separated by a very narrow strait, and with all the appearance of having formed one continuous land; yet their productions differ considerably, and those of Sumatra show the closest resemblance to those of Borneo, an island ten times further off than Java and differing widely in the absence of volcanoes or any continuous range of lofty mountains. Then again, not only does Java differ from these two, but it agrees with a country beyond them botha country from which they seem to have a much better chance to have been supplied by immigration than Java has, and to have (almost necessarily) participated, even more largely, in the benefits of any means of transmission capable of reaching the latter island. Yet more; whatever changes have occurred to bring about the anomalous state of things that exists must have been, zoologically and geologically, recent; for the strange crossaffinities between Java and the Indo-Chinese continent (in which Sumatra and Borneo have not participated), as well as that between Malacca, Sumatra, and Borneo (in which Java has not participated) are exhibited, in many cases by community of species, in others by the presence of very closely allied forms of the same genera, of mammalia and birds. Now we know that