[[p. 480]] The chain of islands situated to the east of Java, and ending in Timor, forms a natural subdivision of the Malayan archipelago, being distinguished by peculiarities of physical geography as well as by a characteristic fauna. These islands all contain active volcanoes, and are for the most part of volcanic origin. Timor, however, which lies somewhat obliquely to the rest, consists in a great measure of ancient sedimentary rocks, which seem to have been exposed for long periods to volcanic convulsions, since they are everywhere shattered and distorted in a remarkable manner. All of these islands have a climate which differs from that of the rest of the archipelago in being remarkably dry; and this has produced a characteristic vegetation, in which spiny and prickly shrubs abound, while the dense luxuriant forests of the regions nearer the equator are quite unknown. The forest-trees of Timor are chiefly Eucalypti and Acacias, thinly scattered over bare and gravelly slopes; while it is only in the damper and more sheltered spots that patches of bush and thickets occur.

     I visited Timor twice, and collected for nearly five months in both the eastern and western districts, obtaining 112 species of birds. Ill health and wars between the Portuguese and the natives prevented me from going far into the interior, where, especially on the south coast, I have reason to believe many additional species might have been obtained. Six other birds are said to be from Timor, mostly collected by the naturalists of the Dutch Government expeditions, making a total of 118 species from this island. My assistant, Mr. Allen, collected for nearly four months in the large island of Flores, and obtained eighty-six species of birds; and I am not aware that a [[p. 481]] single species was previously known from this locality. From the next island, Sumbawa, I have obtained no collections. A few are indicated by Bonaparte as existing in the Leyden Museum, the whole of which, with one exception only (Trichoglossus forsteni) are found in the adjacent islands. In the next island, Lombock, I collected myself for two and a half months, obtaining sixty-three species of birds. The island of Bali commences the Indian region of zoology, which we have not now to consider; but I may mention that, according to the Dutch naturalist Zollinger, its fauna is absolutely identical with that of Java. The total number of species known to inhabit this group of islands is 186. Taking into consideration the comparative sizes of the islands, the above numbers may be supposed to represent with tolerable fairness their respective bird populations, and will therefore furnish us with materials for some interesting comparisons.

     The Timorese subfauna, as we may conveniently name it, differs strikingly from the Moluccan in the absence of such genera as Lorius, Eos, Eclectus, and Lycocorax, and also of Tanysiptera, Alcyone, Criniger, and the subgenera Iotreron and Cyanotreron, all of which, though not peculiar to the Moluccas, are very characteristic of that group of islands. On the other hand, it shows a closer connexion with Australia than any other part of the archipelago, as evidenced by five genera, Sphecothera, Gerygone, Pardalotus, Glycyphila, and Amadina, which are found in this chain of islands, but do not pass into the Moluccas or Celebes. The number of species which appear to be restricted to the Timorese subfauna is eighty-one, many of which, however, are very slight modifications of Australian species. On the other hand, there is not a single genus confined to the group, or which can be said to have its metropolis in it--indicating that the fauna is strictly derivative, and of not very ancient date. In endeavouring to determine the origin of this fauna, we must eliminate those species and genera which, having a wide distribution and roaming habits, can give us little definite information. These are chiefly raptorial and aquatic birds, with a portion of the Passeres; and I find that fifty-seven species of this nature are identical with those of surrounding countries, while thirty-five more are representative species which cannot be referred to any one island rather than another for their probable origin. Deducting these ninety-two species, we have left a rather larger number of birds which we can trace directly either to Australia on the one side or to Java on the other.

     The species which appear to be confined to each island may be classed as follows--

     Lombock 4, of which 1 genus is Australian, 1 genus Indian,
     Flores 12, of which 5 genera are Australian, 2 genera Indian,
     Timor 42, of which 16 genera are Australian, 4 genera Indian,

showing that while there is a great preponderance of the forms of the Australian region in Timor, they decrease going westward, till in Lombock they are equalled by the Indian forms, The species which [[p. 482]] are identical with those of Java or Australia show the proportionate influence of the two countries in a yet more striking manner.

     JAVAN.

                                       Lombock.          Flores.            Timor.
     Species  . . . . . . . . . . . . 32 . . . . . . . . . 22 . . . . . . . . . 10
     Representatives  . . . . . . . 1 . . . . . . . . . . 3 . . . . . . . . . . 5
                                           33 . . . . . . . . . 25 . . . . . . . . . 15

     AUSTRALIAN.

                                       Lombock.         Flores.           Timor.
     Species . . . . . . . . . . . . . 4 . . . . . . . . . . 5 . . . . . . . . . . 10
     Representatives   . . . . . . 3 . . . . . . . . . . 9 . . . . . . . . . . 26
                                            7 . . . . . . . . . 14 . . . . . . . . . . 36

     This table shows how two streams of immigration have entered the islands, the one from Java diminishing in intensity as it flowed on to Timor, the other from Australia diminishing in about the same degree towards Lombock. The total number of species which have entered on the two sides seems nearly equal, with only a slight preponderance in favour of Australia; but there is this remarkable difference, that whereas the great majority of the species derived from Java are identical with the present inhabitants of that island, those derived from Australia are for the most part representative species, less than half of them being identical with birds still living in that country. We shall see this perhaps more clearly by treating the islands as a whole, and dividing that portion of their birds which have exclusive relations to Java or Australia in a similar manner; thus the Timorese avifauna contains

     Javan species . . . . . . . . .  35
     Javan representatives . . .  11
                                             46

     Australian species  . . . . . . . . . . 13
     Australian representatives . . . . . 35
                                                      48

showing that, though the total number of species derived from the two districts is nearly the same, the identical species and representatives are divided in exactly reverse proportions. This fact is one of the most important that we can draw from a consideration of these lists of species, since it gives us a clue to the manner in which this little group of islands was first stocked with animal life, and, in connexion with geological considerations, will enable us to form a general idea of their early history.

     Change of species is a slow process. On that we are all agreed, though we may differ about how it has taken place. The fact that the Australian species in these islands have mostly changed, while the Javan species have almost all remained unchanged, would therefore indicate that the district was first peopled from Australia. But, for this to have been the case, the physical conditions must have been very different from what they are now. Nearly 300 miles of open sea now separates Australia from Timor, which island is [[p. 483]] connected with Java by a chain of broken land divided by straits which are nowhere more than about twenty miles wide. Evidently there are now great facilities for the natural productions of Java to spread over and occupy the whole of these islands, while those of Australia would find very great difficulty in getting across. To account for the present state of things, we should naturally suppose that Australia was once much more closely connected with Timor than it is at present; and that this was the case is rendered highly probable by the fact of a submarine bank extending along all the north and west coast of Australia, and at one place approaching within twenty miles of the coast of Timor. This indicates a recent subsidence of North Australia, which probably once extended as far as the edge of this bank. I do not think Timor was ever absolutely connected with Australia, because the representation of the forms of that country is not sufficiently perfect. There are no Kangaroos in Timor, nor indeed any Marsupials whatever, except a Cuscus, which is a Moluccan and not an Australian genus. Many highly characteristic genera of birds are also absent which we should certainly expect to find had the countries ever been connected, such as Calyptorhynchus, Malurus, Cracticus, Anthochæra, Poëphila, Falcunculus, Colluricincla, &c. Nor do any of the characteristic Australian groups of insects occur in Timor. Everything indicates therefore that a strait of the sea has always separated it from Australia--a supposition which is confirmed by the deep gulf that still runs between its rocky southern coast and the edge of the before-mentioned submarine bank.

     But at the time when this narrowing of the sea took place in one direction, there must have been a greater separation at the other end of the chain, or we should find more equality in the numbers of identical and representative species derived from each extremity. It is true that the widening of the strait at the Australian end by subsidence would, by putting a stop to immigration and intercrossing of individuals from the mother country, have allowed the full action of the causes which have led to the modification of the species; while the continued stream of immigrants from Java would by continual intercrossing check such modification. This view will not, however, explain all the facts; for the character of the fauna of the Timorese group is indicated as well by the forms which are absent from it as by those which it contains, and is by this kind of evidence shown to be much more Australian than Indian. No less than twenty-nine genera, all more or less abundant in Java, and most of which range over a wide area, are quite absent; while of the equally diffused Australian genera only about fourteen are wanting. This would clearly indicate that there has been till recently a wide separation from Java; and the fact that the islands of Baly and Lombock are small and are almost wholly volcanic, and contain a smaller number of modified forms than the other islands, would point them out as of comparatively recent origin. Here probably existed a wide arm of the sea at the time when Timor was in the closest proximity to Australia; and as the subterranean fires were slowly piling up the now fertile islands of Bali and Lombock, the northern shores of [[p. 484]] Australia would be sinking beneath the ocean. Some such changes as these will enable us to understand how it happens that, though the birds of these islands are on the whole almost as much Indian as Australian, yet the apparently endemic species have such a preponderating Australian character, and why such a very large number of characteristic Indian forms, which are common in Java and are known in most instances to extend into Bali, have yet never transmitted a single representative to the islands further east.

     The following is a list of all the birds known to inhabit this group, with their distribution in the several islands. Those marked with an * are not known from any other localities. The descriptions of twenty-eight new species are afterwards given . . .