Thus writing of coral snakes, he says, on page 95 of _Through Southern Mexico_: “They are usually paraded as glaring instances of warning colouration, but I am not at all sure whether this is justifiable. Certainly these _Elaps_ are most conspicuous and beautiful objects. Black and carmine or coral red, in alternate rings, are the favourite pattern; sometimes with narrow golden-yellow rings between them, as if to enhance the beautiful combination. But these snakes are inclined to be nocturnal in their habits, and, except when basking, spend most of their time under rotten stumps, in mouldy ground, or in ants’ nests in search of their prey, which must be very small, to judge from the size of the mouth.” Dr Gadow goes on to show that although black and red are very strong contrasts in the day-time, the combination ceases to be effective in the dark. He suggests that red and black is a self-effacing rather than a warning pattern. He further points out that several kinds of harmless snakes have the same colouring and pattern. “There seems,” he says, “to be no reason why we should not call these cases of mimicry; and yet this is most likely a wrong interpretation, since such harmless snakes are also found in districts where the _Elaps_ does not occur, not only in Mexico, but likewise in far-distant parts of the world, where neither elapines nor any other similarly coloured poisonous snakes exist. To interpret this as an instance of ‘warning colours’ in a perfectly harmless snake, which has no chance of mimicry, amounts in such cases to nonsense, and we have to look for a different explanation upon physiological and other grounds.” It is, to say the least of it, significant that all the opposition to the theory of protective colouration comes from those who observe nature first hand, while the warmest supporters of the theory are cabinet naturalists and museum zoologists. In the case of nocturnal creatures, as Dr H. Robinson very sagely points out (_Knowledge_, January 1909), the value for protective purposes of any given colouration must depend very largely on the state of the moon. “It was,” he writes, “a common experience in the South African War that on overcast or moonless nights the nearly black army great-coat made a picquet sentry invisible at a distance of a few feet. In strong moonlight this garb could be seen at a great distance, whereas a khaki pea jacket, useless on a dark night, answered the requirements of invisibility very well.” It is thus evident that the dark colour of the buffalo and sable antelope cannot be protective on both dark and moonlight nights. The theory of protective colouration is based on the tacit assumption that beasts of prey rely on eyesight for finding their quarry. Raptorial birds certainly do use their eyes as the means of discovering their victims; but the great majority of predaceous mammals trust almost entirely to their power of smell as a means for tracking down their prey.