2 8o MIMICRY AND NATURAL SELECTION 



perhaps the most powerful argument of all in favour of 

 an interpretation based on the theory of Natural Selec- 

 tion. If these resemblances are attained by selection 

 because they are advantageous in the struggle for life, we 

 should expect to find that they are produced in a great 

 variety of ways ; for one species would reach the bene- 

 ficial end by one path pointed out to it by the structure 

 it possessed at the beginning and by the trend of its 

 variation, while another species with a very different initial 

 structure would reach the same end by a widely different 

 path. Thus many Diptera, for example species of Ceria, 

 gain a superficial resemblance to wasps by a narrowing 

 in the anterior abdominal region which suggests the 

 characteristic peduncle of a Hymenopterous insect. On 

 the other hand, Longicorn beetles of the genus Oberea 

 gain the same effect by a patch of white which obliterates 

 the anterior abdominal region with the exception of 

 a small linear remnant representing the peduncle. 1 In 

 brilliant illumination the white marking is not seen as 

 part of the insect. The resemblance of the Locustid 

 Myrmecophana fallax to an ant is produced in the same 

 manner. 2 The Homopterous family Membracidae are 

 characterized by an enormous growth of the dorsal region 

 of the pro-thorax, which spreads backwards and in many 

 species covers the insect like a shield. In the American 

 species which mimic ants, this shield, and not the insect 

 beneath it, becomes ant-like. Some of the larval Mem- 

 bracidae are laterally compressed, becoming in the dorsal 

 region as thin as a leaf, and the body is green like a leaf, 

 while the head and legs are brown. The whole appear- 

 ance is singularly like that of the tropical American ant 

 Atta (Oecodoma) cephalotcs, carrying its leaf vertically in 

 its mandibles, and thrown over its back, so that the brown 

 head, legs, and part of the body are seen beneath the 

 green burden. 3 It is manifestly absurd to attempt to 



1 See R. Shelford in Proc. Zool. Soc. Lond., 1902, pp. 238-40, plate 

 xix, figs. 13, 14, 15. 



2 See pp. 256-7 of the present volume, together with Fig. 5, p. 258. 



3 See description and figure of a specimen found by Mr. W. L. Sclater 

 in British Guiana. Poulton, in Proc. Zool. Soc. Lond., 1891, p. 462, 

 pi. xxxvi. See also pp. 258-60 of the present volume and Fig. 7, p. 259. 



