difference of colour, in many cases so remarkable that the two sexes of the same species remained for many years under different names and were thought to be quite distinct species. We find, however, every gradation from perfect identity to complete diversity, and in some cases we are able to see a reason for this difference. Beginning with the most extraordinary cases of diversity—as in Diadema misippus, where the male is black, ornamented with a large white spot on each wing margined with rich changeable blue, while the female is orange-brown with black spots and stripes-we find the explanation in the fact that the female mimics an uneatable Danais, and thus gains protection while laying its eggs on low plants in company with that insect. In the allied species, Diadema bolina, the females are also very different from the males, but are of dusky brown tints, evidently protective and very variable, some specimens having a general resemblance to the uneatable Euplæas; so that we see here some of the earlier stages of both forms of protection. The remarkable differences in some South American Pieridæ are similarly explained. The males of Pieris pyrrha, P. lorena, and several others, are white with a few black bands and marginal spots like so many of their allies, while the females are gaily coloured with yellow and brown, and exactly resemble some species of the uneatable Heliconidæ of the same district. Similarly, in the Malay Archipelago, the female of Diadema anomala is glossy metallic blue, while the male is brown; the reason for this reversal of the usual rule being, that the female exactly mimics the brilliant colouring of the common and uneatable Euplæa midamus, and thus secures protection. In the fine Adolias dirtea, the male is black with a few specks of ochre-yellow and a broad marginal band of rich metallic greenish-blue, while the female is brownish-black entirely covered with rows of ochre-yellow spots. This latter coloration does not appear to be protective when the insect is seen in the cabinet, but it really is so. I have observed the female of this butterfly in Sumatra, where it settles on the ground in the forest, and its yellow spots so harmonise with the flickering gleams of sunlight on the dead leaves that it can only be detected with the greatest difficulty.

A hundred other cases might be quoted in which the female is either more obscurely coloured than the male, or gains protection by imitating some inedible species; and any one who has watched these female insects flying slowly along in search of the plants on which to deposit their eggs, will understand how important it must be to them not to attract the attention of insect-eating birds by too conspicuous colours. The number of birds which capture insects on the wing is much greater in tropical regions than in Europe; and this is perhaps the reason why many of our showy species are alike, or almost alike, in both sexes, while they are protectively coloured on the under side which is exposed to view when they are at rest. Such are our peacock, tortoise-shell, and red admiral butterflies; while in the tropics we more commonly find that the females are less conspicuous on the upper surface even when protectively coloured beneath.

We may here remark, that the cases already quoted prove clearly that either male or female may be modified in colour apart from the opposite sex. In Pieris pyrrha and its allies the male retains the usual type of coloration of the whole genus, while the female has acquired a distinct and peculiar style of colouring. In Adolias dirtea, on the other hand, the female appears to retain something like the primitive colour and markings of the two sexes, modified perhaps for more perfect protection; while the male has acquired more and more intense and brilliant colours, only showing his original markings by the few small yellow spots that remain near the base of the wings. In the more gaily coloured Pieridæ, of which our orange-tip butterfly may be taken as a type, we see in the female the plain ancestral colours of the group, while the male has acquired the brilliant orange tip to its wings, probably as a recognition mark.

In those species in which the under surface is protectively coloured, we often find the upper surface alike in both sexes, the tint of colour being usually more intense in the male. But in some cases this leads to the female being more conspicuous, as in some of the Lycanida, where the female is bright blue, while the male is of a much deeper blue, but obscured by dusky tints so as to have become the less brilliantly coloured.

Probable Causes of these Colours.

In the production of these varied results there have probably been several causes at work. There seems to be a constant tendency in the male of most animals—but especially of birds and insects-to develop more and more intensity of colour, often culminating in brilliant metallic blues or greens or the most splendid iridescent hues; while, at the same time, natural selection is constantly at work, preventing the female from acquiring these same tints, or modifying her colours in various directions to secure protection by assimilating her to her surroundings, or by producing mimicry of some protected form. At the same time, the need for recognition must be satisfied; and this seems to have led to diversities of colour in allied species, sometimes the female, sometimes the male undergoing the greatest change according as one or other could be modified with the greatest ease, and so as to interfere least with the welfare of the race. Hence it is that sometimes the males of allied species vary most, as in the different species of Epicalia; sometimes the females, as in the magnificent green species of Ornithoptera and the "Æneas" group of Papilio.

The importance of the two principles-the need of protection and recognition-in modifying the comparative coloration of the sexes among butterflies, is beautifully illustrated in the case of the groups which are protected by their distastefulness, and whose females do not, therefore, need the protection afforded by sober colours.

In the great families, Heliconidæ and Acræidæ, we find that the two sexes are almost always alike; and, in the very few exceptions, that the female, though differently, is not less gaily or less conspicuously coloured. In the Danaidæ the same general rule prevails, but the cases in which the male exhibits greater intensity of colour than the female are perhaps more numerous than in the other two families. There is, however, a curious difference in this respect between the Oriental and the American groups of distasteful Papilios with warning colours, both of which are the subjects of mimicry. In the Eastern groups—of which P. hector and P. coon may be taken

as types-the two sexes are nearly alike, the male being sometimes more intensely coloured and with fewer pale markings; but in the American groups-represented by P. æneas, P. sesostris, and allies-there is a wonderful diversity, the males having a rich green or bluish patch on the fore wings, while the females have a band or spots of pure white, not always corresponding in position to the green spot of the There are, however, transitional forms, by which a complete series can be traced, from close similarity to great diversity of colouring between the sexes; and this may perhaps be only an extreme example of the intenser colour and more concentrated markings which are a very prevalent characteristic of male butterflies.

There are, in fact, many indications of a regular succession of tints in which colour development has occurred in the various groups of butterflies, from an original grayish or brownish neutral tint. Thus in the "Eneas" group of Papilios we have the patch on the upper wings yellowish in P. triopas, olivaceous in P. bolivar, bronzy-gray with a white spot in P. erlaces, more greenish and buff in P. iphidamas, gradually changing to the fine blue of P. brissonius, and the magnificent green of P. sesostris. In like manner, the intense crimson spots of the lower wings can be traced step by step from a yellow or buff tint, which is one of the most widespread colours in the whole order. The greater purity and intensity of colour seem to be usually associated with more pointed wings, indicating greater vigour and more rapid flight.

Sexual Selection as a supposed Cause of Colour Development.

Mr. Darwin, as is well known, imputed most of the brilliant colours and varied patterns of butterflies' wings to sexual selection-that is, to a constant preference, by female butterflies, for the more brilliant males; the colours thus produced being sometimes transmitted to the males alone, sometimes to both sexes. This view has always seemed to me to be unsupported by evidence, while it is also quite inadequate to account for the facts. The only direct evidence, as set forth with his usual fairness by Mr. Darwin himself, is opposed to his views. Several entomologists assured him that, in moths, the females evince not the least choice of their

partners; and Dr. Wallace of Colchester, who has largely bred the fine Bombyx cynthia, confirmed this statement. Among butterflies, several males often pursue one female, and Mr. Darwin says, that, unless the female exerts a choice the pairing must be left to chance. But, surely, it may be the most vigorous or most persevering male that is chosen, not necessarily one more brightly or differently coloured, and this will be true "natural selection." Butterflies have been noticed to prefer some coloured flowers to others; but that does not prove, or even render probable, any preference for the colour itself, but only for flowers of certain colours, on account of the more agreeable or more abundant nectar obtained from them. Dr. Schulte called Mr. Darwin's attention to the fact, that in the Diadema bolina the brilliant blue colour surrounding the white spots is only visible when we look towards the insect's head, and this is true of many of the iridescent colours of butterflies, and probably depends upon the direction of the striæ on the scales. It is suggested, however, that this display of colour will be seen by the female as the male is approaching her, and that it has been developed by sexual selection.1 But in the majority of cases the males follow the female, hovering over her in a position which would render it almost impossible for her to see the particular colours or patterns on his upper surface; to do so the female should mount higher than the male, and fly towards him-being the seeker instead of the sought, and this is quite opposed to the actual facts. I cannot, therefore, think that this suggestion adds anything whatever to the evidence for sexual selection of colour by female butterflies. This question will, however, be again touched upon after we have considered the phenomena of sexual colour among the vertebrata.

Sexual Coloration of Birds.

The general rule among vertebrates, as regards colour, is, for the two sexes to be alike. This prevails, with only a few exceptions, in fishes, reptiles, and mammalia; but in birds diversity of sexual colouring is exceedingly frequent, and is, not improbably, present in a greater or less degree in more

1 Darwin in Nature, 1880, p. 237.