an insectivorous bird; but it is a remarkable fact that most of the genera of Fly-catchers of Borneo and Java on the one side (Muscipeta, Philentoma,) and of the Moluccas on the other (Monarcha, Rhipidura), are almost entirely absent from Celebes. Their place seems to be supplied by the Caterpillar-catchers (Graucalus, Campephaga, &c.), of which six or seven species are known from Celebes and are very numerous in individuals. We have no positive evidence that these birds pursue butterflies on the wing, but it is highly probable that they do so when other food is scarce. Mr. Bates has suggested to me that the larger Dragonflies (Eshna, &c.) prey upon butterflies; but I did not notice that they were more abundant in Celebes than elsewhere. However this may be, the fauna of Celebes is undoubtedly highly peculiar in every department of which we have any accurate knowledge; and though we may not be able satisfactorily to trace how it has been effected, there can, I think, be little doubt that the singular modification in the wings of so many of the butterflies of that island is an effect of that complicated action and reaction of all living things upon each other in the struggle for existence, which continually tends to readjust disturbed relations, and to bring every species into harmony with the varying conditions of the surrounding universe.

But even the conjectural explanation now given fails us in the other cases of local modification. Why the species of the Western islands should be smaller than those further east,-why those of Amboyna should

N

exceed in size those of Gilolo and New Guinea-why the tailed species of India should begin to lose that appendage in the islands, and retain no trace of it on the borders of the Pacific,-and why, in three separate cases, the females of Amboyna species should be less gaily attired than the corresponding females of the surrounding islands,-are questions which we cannot at present attempt to answer. That they depend, however, on some general principle is certain, because analogous facts have been observed in other parts of the world. Mr. Bates informs me that, in three distinct. groups, Papilios which on the Upper Amazon and in most other parts of South America have spotless upper wings obtain pale or white spots at Pará and on the Lower Amazon; and also that the Eneas-group of Papilios never have tails in the equatorial regions and the Amazons valley, but gradually acquire tails in many cases as they range towards the northern or southern tropic. Even in Europe we have somewhat similar facts; for the species and varieties of butterflies peculiar to the island of Sardinia are generally smaller and more deeply coloured than those of the mainland, and the same has recently been shown to be the case with the common tortoiseshell butterfly in the Isle of Man ; while Papilio Hospiton, peculiar to the former island, has lost the tail, which is a prominent feature of the closely allied P. Machaon.

Facts of a similar nature to those now brought forward would no doubt be found to occur in other groups of insects, were local faunas carefully studied in

relation to those of the surrounding countries; and they seem to indicate that climate and other physical causes have, in some cases, a very powerful effect in modifying specific form and colour, and thus directly aid in producing the endless variety of nature.

Mimicry.

Having fully discussed this subject in the preceding essay, I have only to adduce such illustrations of it, as are furnished by the Eastern Papilionidæ, and to show their bearing upon the phenomena of variation already mentioned. As in America, so in the Old World, species of Danaidæ are the objects which the other families most often imitate. But besides these, some genera of Morphide and one section of the genus Papilio are also less frequently copied. Many species of Papilio mimic other species of these three groups so closely that they are undistinguishable when on the wing; and in every case the pairs which resemble each other inhabit the same locality.

The following list exhibits the most important and best marked cases of mimicry which occur among the Papilionidæ of the Malayan region and India :

We have, therefore, fourteen species or marked varieties of Papilio, which so closely resemble species of other groups in their respective localities, that it is not possible to impute the resemblance to accident. The first two in the list (Papilio paradoxa and P. Caunus) are so exactly like Euploa Midamus and E. Rhadamanthus on the wing, that although they fly very slowly, I was quite unable to distinguish them. The first is a very interesting case, because the male and female differ considerably, and each mimics the corresponding sex of the Euploa. A new species of Papilio which I discovered in New Guinea resembles Danais sobrina,

from the same country, just as Papilio Marcareus resembles Danais Aglaia in Malacca, and (according to Dr. Horsfield's figure) still more closely in Java. The Indian Papilio Agestor closely imitates Danais Tytia, which has quite a different style of colouring from the preceding; and the extraordinary Papilio Idæoides from the Philippine Islands, must, when on the wing, perfectly resemble the Hestia Leuconoë of the same region, as also does the Papilio Delessertii imitate the Ideopsis daos from Penang. Now in every one of these cases the Papilios are very scarce, while the Danaida which they resemble are exceedingly abundant-most of them swarming so as to be a positive nuisance to the collecting entomologist by continually hovering before him when he is in search of newer and more varied captures. Every garden, every roadside, the suburbs of every village are full of them, indicating very clearly that their life is an easy one, and that they are free from persecution by the foes which keep down the population of less favoured races. This superabundant population has been shown by Mr. Bates to be a general characteristic of all American groups and species which are objects of mimicry; and it is interesting to find his observations confirmed by examples on the other side of the globe.

The remarkable genus Drusilla, a group of palecoloured butterflies, more or less adorned with ocellate spots, is also the object of mimicry by three distinct genera (Melanitis, Hyantis, and Papilio). These insects, like the Danaidæ, are abundant in individuals,