THE CAUSE OP MIMETIC REaEMBLlNCE. 605 



observer, and, so far as I am aware, the inquiry has never been 

 approached in so thorough a manner. Even if collectors would 

 pay attention to the worst specimens instead o£ the best, some 

 evidence of the nature and amount of attack would be forth- 

 coming. During the visit of the British Association to Canada 

 last year (1897) I made a point of captiiring butterflies which 

 had evidently been pecked by birds. In this way, although I 

 did not witness a single attack, I obtained the proof that butter- 

 flies are not nearly so immune as has been asserted. On 

 another occasion I hope to deal with this evidence in detail. 

 Similar observations have been made by Fritz Miiller. 



The review of the whole subject during the past 36 years 

 increases our confidence in the theories of Bates and Fritz 

 Miiller, while it disposers of all alternative hypotheses. Even more 

 than this, — it will, I believe, be claimed by all who take a broad 

 view over the whole field of evidence, that the explanation of 

 these deeply interesting facts, which form so fascinating and 

 important a department of natural history in the tropics, is one 

 of the most notable triumphs ever won by the great theory of 

 natural selection. 



EXPLANATION OF THE PLATES. 



Plate 40. 

 Cei'tain aspects of Mimetic Eesemblance in Lepicloptera. 



Fig. 1. Natural size. This figure represents the appearance of the young larva 

 of Stauropus fagi in the first stage, upon a twig of birch. The drawing 

 was made from living specimens in the Hope Department at Oxford 

 by P. J. Bayzand. The ant-like appeai-ance is produced by the 

 immensely elongated second and third pairs of thoraciclegs, which are 

 kept in a state of active movement upon the slightest disturbance. 

 The caudal shield with its two processes resembles the head and 

 antennffi of an ant. This figure should be studied in connexion with 

 figures 1 to 7 in the text (Section 11), in which the resemblance to 

 ants is shown to be brouglit about in many very difierent ways, thus 

 supporting an interpretation based on the theory of natural selection. 



Fig. 2. Natural size. A group of the young larvjE of Eoidromis versicolor upon 

 a twig of birch. Drawn from life in the Hope Department by P. J. 

 Bayzand. The larvae are represented in the attitude they assume 

 when disturbed, the anterior part of the body being thrown back and 

 the orange-coloured thoracic legs exposed to view. In this position 

 they strongly suggest the appearance of a group of young larvje of 



LIN>'. JOURN. — ZOOLOGY, TOL. XXVI. 43 



