A good example of a mimetic resemblance attained in many ways is to be seen in the large tropical American association of diverse butterflies and day-flying moths which have developed transparent areas on their wings in mimicry of dominant Ithomiine models. Transparency has been attained by the following different methods in different mimetic species or groups of species :—(1) the scales of the wing may be so reduced in size that the light passes between them; (2) remaining the same size they may become much fewer ; (3) they may stand up on edge; (4) remaining in the usual over lapping position they may become themselves transparent. Fur thermore in the models themselves, the scales of the transparent parts have become minute vestiges, reduced to a much greater extent in one of the two genera to which the species belong. It may be added that the wings of many moths which mimic bees are, on emergence from the chrysalis, covered with loosely attached scales which blow off during the first flight. Such different methods of attaining the same end are to be expected if the development has been brought about by natural selection, for it is unreasonable to believe that similar variations would appear in very different species with very different natures.
The most convincing evidence of the development of mimicry and protective resemblance by natural selection is to be found in the Homoptera which, in addition to the cicadas, aphides, etc. also include the Membracidae, a family of small insects found nearly all over the world but most commonly in tropical America. These insects have, on the body-ring behind the head, a projection which grows backwards and expands into a shield covering every part except the head, wings and legs. The form of the concealed body is much like that of the allied greenfly and other aphids of our gardens. Five tropical American membracids are represented in figs. 4-8A. Hemikyptha, like a hooked seed, seen from above in 4A, from the side in fig. 4; the thornlike Umbonia in figs. 5, 5A; the smooth-seedlike Hebetica in 6, 6A; the ant-like Heteronotus in 7, 7A; finally Oeda, resembling the orange-coloured, freely exposed cocoon of a distasteful moth, in 8, 8A. Thus in the Mem bracidae the concealing or the mimetic appearance is developed on the covering shield, and not, as in so many other insects, on the body. Mimicry is only developed where it can be seen. The value of the ant-mimicry in figs. 7, 7A has been questioned because the Membracidae, like the allied froghoppers (Cercopidae), can escape by jumping, but this is only one of the numerous examples in which different methods are adopted by a species so that if the first fails there is still a chance of success by the second. It has also been objected that ants have many enemies and are therefore dangerous models. Haase is the only naturalist who has thought it necessary to assume that the special protection of models confers absolute immunity, even believing that it is efficient against the insect enemies of insects. We know, however, that the species which are most distasteful to the higher insect-eating animals are espe cially liable to be attacked by parasitic and predaceous insects. Admitting the existence of numerous enemies, ants nevertheless possess qualities which render them the most successful insects in the world. Analogous reasoning would attempt to show that the form, colour and pattern of grass-feeding caterpillars are valueless because so many animals feed upon grass! The mimicry of ants is a vast subject which cannot be treated even superficially here. They are models not only for insects of many diverse kinds but also, in movement as well as appearance, by numerous spiders. Certain species of the carnivorous mantises
and also of long-horned grasshoppers mimic ants when small but in their later stages become flower- or leaf-like. Some idea of the advantage which may be conferred by resembling these models is suggested by a study of their guests. Since 1891 H. Donisthorpe has discovered in British ants' nests 150 species of insects, spiders and mites new to the country, including 70 new to science. Of these guests 28 species are mimics of ants and thus would be protected outside the nest or in a disturbed nest against enemies which fear the ants. In addition to these British ants are mimicked by 34 species living independently, and by 15 species possibly guests, possibly independent.
• A very different kind of mimicry is adopted by another tropical American homopterous insect, one of the large lantern flies (Ful goridae). The front part of the head of this insect, Laternaria lucifera, projects forward as a hollow mask which resembles in remarkable detail the head of an alligator as may be seen from 1 the side view represented in fig. 13. The true eye of the insect is shown behind the last of the apparent teeth, the alligator's eye and nostril are represented on humps in their characteristic position. The rows of long sharp teeth, not hidden by lips as in other ani mals, form the most terrifying feature of the alligator's head, and these are convincingly suggested by a series of brilliantly white elevations. The mask of this lantern-fly in fact reproduces the appearance of the reptile's head not only in colour but in form, as it would be reproduced by a sculptor. Alligators are excessively abundant in tropical America and a terrible menace to animals which come to the riverside to drink, among them the arboreal monkeys which are probably some of the chief enemies of the lantern-fly. It is not supposed that the insect is mistaken for the reptile—but that an appearance, even on a very small scale, asso ciated with a near escape from death, causes such a shock to an intelligent animal that the lantern-fly is given the chance of flight. This remarkable example is only one of many in which reptiles are mimicked by insects, especially snakes by caterpillars. Exper iments have proved that the deceptive appearance is effective in terrifying insectivorous monkeys, birds and lizards.
Another remarkable form of mimetic resemblance is illustrated on Plate II., fig. 3, representing the cocoon of a west African moth Deilemera, covered with the apparent cocoons of a hymenopterous parasite. Such cocoons are woven by larvae which have hatched from eggs laid in the body of a caterpillar, lived and fed without injuring any vital part until the host is full-grown, then devoured the whole of it or its chrysalis, and bored their way out. The nutriment is then divided up and enclosed in numbers of tough cases, each containing very little and very difficult to open. It is therefore an advantage for a caterpillar to mimic the appearance of such an attack, and Lamborn has shown how this is achieved. The caterpillar while spinning produces from the end of its body a number of pale yellowish frothy spheres, each encircled by a silken girdle, dragged off and fixed to the cocoon by a few strands of silk. Other examples of the kind are known and probably num bers will be found now that Lamborn's discovery has directed attention to this kind of mimicry.
It has been impossible to describe more than a very small pro portion of the recorded examples of mimicry. It is believed, how ever, that those which have been selected will serve to illustrate the chief aspects of the subject.