TRANSACTIONS OF SECTION D. 689 



Recent Additions to the Fish Fauna of New BrunsvAcl^. 

 By Dr. Philip Con. 



9. Theories of Mimicry as illustrated by African Butterflies. 

 /)*y Edward B. Poulton, i/. J., F.li.S., Hope Professor of Zoology, Oxford. 



H. W. Bates, in his epoch-making paper ('Trans. Linn. Soc.,' vol. xxiii. 

 1862), first gave an intelligible theory of mimicry, and accounted for the superficial 

 resemblances which had been known for so long by supposing that the most 

 dominant, well-defended, and conspicuous forms in a country become the models 

 towards which natural selection leads many of the weaker hard-pressed species 

 in the same locality. The material on which Bates' theory was formed was con- 

 fined to tropical America, and his generalisation remained incomplete until it 

 could be applied to the other great tropical regions. This want, however, was 

 soon supplied by A. R. Wallace for the East ( ' Trans. Linn. Soc.,' vol. xxv. 1866), 

 and by Roland Trimen for Africa ( * Trans. Linn. Soc.,' vol. xxvi. 1870). 



In Bates' original paper a certain class of facts — frequently mentioned and 

 abundantly illustrated — cannot be explained under his theory of mimicry. This is 

 the strong resemblance which is apt to exist between the dominant forms them- 

 selves, and which is as minute and as remarkable as the resemblance of the weaker 

 for the stronger species. Bates pointed out that this was unsolved by his theory, 

 and both he and Wallace were compelled to suggest the direct action of some 

 unknown local influence as the possible cause. There the matter rested until 

 Fritz Miiller, in a paper published in ' Kosmos ' for May 1879, suggested an explana- 

 tion, viz., that the dominant forms gain an advantage by this resemblance, inasmuch 

 as it facilitates the education of their enemies by giving them fewer patterns to 

 learn. The necessary waste of life by which the education of young birds, &c., is 

 brought about is here divided between the various species of a closely convergent 

 group, instead of being contributed by each member independently. The chief 

 sub-families of butterflies which in tropical America appear to be specially dis- 

 tasteful to insect-eating animals, and which are specially mimicked by others, are 

 the Danaince, Ithomiince, JfeliconiiKP, and Acraince. Of these the second and 

 third are confined to this part of the world. The resemblances which Fritz Miiller 

 explained are those which occur very commonly between the Danain(e,Itho)mmee, 

 Heliconina, and less commonly the Acraince of any locality. In order to complete 

 this theory it was necessary to test its application in other parts of the world. 



In the East the butterflies which take the place of the four above-named sub- 

 families belong almost exclusively to the Danaintx, the Acr<ei7i(e being represented 

 by very few species. The DanairKe are, however, extremely rich in species, and 

 F. Moore first pointed out in ' Proc. Zool. Soc.,' 1883, p. 201, that there is the same 

 relationship between the species of this dominant group that obtains between those 

 of tropical America. Not only do Danainm of very difl'erent genera closely re- 

 semble each other, but there is often a strong likeness between the species belonging 

 to the two chief divisions of the sub-family — the Danaina and Euploeina. As in 

 America, these resemblances are always between the species of the same locality. 



While, however, Muller's theory received full confirmation from the facts observed 

 in India and the tropical East generally, no attempt has been made until now to 

 apply it to the African lepidopterous fauna. I have therefore examined this 

 fauna from the Miillerian standpoint, and find that in it too the same relationships 

 can be traced. 



The dominant distasteful groups of Africa are the AcreBitKB, which have their 

 metropolis here, and the Danaince. The latter are chiefly represented by the 

 species of the peculiar African genus Amauris, and by the abundant and wide- 

 spread Danais \Limnas) chrysippus. I first looked for evidence of convergence 

 between the Acrmna: and Limiias chrysippus, and soon found what appeared to 

 be evident traces of it. Such species as Planema esebria (certain forms of), Acrcra 

 petrcea (female), A. oppidia, and, above all, A. encedon {lycia) bear a considerable 



1897. Y Y 



