612- PEOr. E. B. POULTON ON MIMETIC RESEMBLANCE. 



These and the species of Anthomyza belong to the Pericopidm 

 (Hypsidce). The row of moths on the right hand are widely separated 

 from the others, being the species of CastniidcB which belong to the 

 group. The uppermost moth is Castnia linus (Cr.), the next C. dodona 

 (Druce), the next C. heliconioides (Herr.-Sch.), and the lowest C. micha, 

 (Druce). 



Fig. 4, X about 15. The method by which transparency has been gained in 

 the genus Anthomyza. Although A. BucHeyi has been drawn for this 

 figure, the method is identical in all the species. The scales lie flat oa 

 the wing with an entirely normal ai'rangement and overlap. Their 

 size and shape are but little altered, but the scales themselves become 

 quite transparent (compare the border scales seen in the lower part 

 of the circle). There is generally a little yellowish or faintly greenish- 

 yellow pigment retained, giving a tint which closely resembles that of 

 Methona and other members of the group, none of which are entirely 

 colourless. This method contrasts in a most interesting manner with 

 all the others made use of in attaining transparency as an element in 

 superficial resemblance. 



Fig. 5. XaboutlS. The scales on the transparent part of the wing of Hyelosia 

 tiresia, part of the border being shown on the lower part of the 

 figure. The scales are greatly reduced in number, they become trans- 

 parent and are set in a far more upright position on the wing-metn- 

 brane. In these ways they interfere but little with the passage of light. 

 The wide contrast with Anthomyzo; belonging to the same family, is 

 of high interest. 



Fig. 6. X about 15. The method adopted by the species of Castnia has much 

 resemblance to that of Hyelosia, although the forms are so distantly 

 related. All the species examined have the same method, the 

 scales of C. linus being actually drawn for the figure. The 

 scales are not reduced in size and but little, if at all, in number, 

 The shape is somewhat simplified, and the pigment is lost. At the 

 same time the scales are set much more steeply on the vring-membrane, 

 and the light freely passes between their rows. This is seen to be the 

 case in fig. 6, but further from the border of the transparent patch 

 they become practically upright. Even where the slope becomes less 

 the light still passes through (although less perfectly), because of the 

 transparency of the scales. 



Hence in this large group of superficially similar South- American 

 Lepidoptera figured and studied in some of its minute details in Plates 

 42, 43, and 44, the transparency which is essential to mutual resem- 

 blance has been produced by widely difierent methods:— the same 

 end, viz., the production of a close superficial likeness, has been 

 reached from very divergent paths. Such a result, while harmonizing 

 with an explanation based on the principle of selection, and especially 

 on the theory of natural selection, is quite inexplicable on any other 

 hypothesis which has as yet been suggested. 



