supposition; for, according to this theory, protective resemblance, as we now see it, has been arrived at by the gradual accumulation and improvement of colour variations which make for concealment, and the protective value of such variations must essentially be, or have been, of only a partial character. Admitting that the gambols of the Whip Butterfly (presumably of a sexual character) lead the insect into a certain amount of danger, yet, to ask us to believe that it thereby "invalidates" the protection afforded it, when at rest, by its assimilative colouring, against other enemies, and perhaps even against the "Suruqua" itself, is, as Mr. Bateson puts it, referring to a different assumption, "to ask us to abrogate reason." Further, the not unusual fact, that animals exhibiting a very high grade of resemblance are yet subject to a considerable amount of persecution, in no way invalidates, but rather strengthens, this principle; for it is evident that such a degree of resemblance can only have been developed in response to a similarly high degree of persecution, acting either now, or within recent times.
It will thus be seen that, on general considerations alone, the above objections to the principle of protective resemblance must be at once ruled out of court. It may be as well, however, to discuss the case of Cicadas in more detail. In the first place, I cannot agree with Mr. Distant that these insects are easily captured owing to their shrill cries. All high-pitched, vibrating sounds of this kind are very difficult to localise exactly, and with Cicadas I have noticed very frequently, both with myself and others, that the distance of the insect is invariably much underestimated.(n14) But even when the tree on which the Cicada is sitting has been ascertained, it must be very cautiously approached, for many species are able to detect one's presence at a distance of fifteen to twenty yards in open country, and, on so doing, they will at once cease their call; and although they will generally permit a much closer approach than this, yet it is always extremely difficult to locate the exact position of the sound on the tree. Their habits, however, vary in this respect, and among the dozen or more species which I have observed in various parts of South Africa, I have found it to be a very general rule that their wariness is inversely proportionate to their protective resemblance; those species which live on rough, knotted bark, or among dense foliage, permitting one to approach much nearer than do those that rest on bare, smooth trunks or small twigs. The above remarks apply to the calling of a single insect; but, when a number are calling together, it is still more difficult to localise any particular cry; and, indeed, I have on several occasions been driven out of a patch of machabel bush by the con-
