94 Mr. G. A. K. Marshall on Dicqwscmatism, vnth reference 



although M tiller evidently conceived, and briefly men- 

 tioned, the possibility of a mutual approach between two 

 such species (Proc. Ent. Soc. 1879, p. xxviii), yet his 

 theory was generally interpreted as involving only a one- 

 sided approach from one species to another. On the 

 other hand the hypothesis of Diaposematism specially 

 emphasises the probability of a mutual simultaneous 

 approach. To use the author's own words : " There seems 

 to be no reason why, especially if there is no conspicuous 

 inequality of numbers, there should not be a kind of 

 ' give and take ' arrangement between mimicker and 

 mimicked, the latter advancing some way to meet the 

 former for their mutual benefit. In other words, when 

 two species, A and B, form an association of this kind, it 

 need not be supposed that the form of A remains fixed, 

 while B assimilates itself to it, or vice versa ; but the 

 association may really be formed by both A and B con- 

 verging to a point between them, or, in short, mimicking 

 each other. The acceleration of the process, which in 

 many cases would result, must of itself be an advantage." 

 (Trans. Ent. Soc. 1894, p. 297.) And again : " In the 

 latter case (Miillerian Mimicry) the mimetic attraction is 

 unlimited and mutual, acting reciprocally in both direc- 

 tions, and influencing each member of the group." (Trans. 

 Ent. Soc. 1897, p. 325.) This line of argument was also 

 advocated in the Transactions for 1896 (p. 72), and since 

 that time various specific cases have been published by 

 Dr. Dixey, Professor Poulton and Mr. S. A. Neave, which 

 are considered to furnish convincing proof of the validity 

 of the hypothesis of reciprocal mimicry. Indeed, it has 

 been held that these proofs are so cogent that Diapose- 

 matism should no longer be regarded as a mere hypothesis, 

 but that it is now entitled to rank as an established law 

 (Proc. Ent. Soc. 1906, p. lxxi). 



But before looking into the details of these cases it may 

 be well to examine the whole conception from a more 

 general standpoint. It has already been remarked that 

 reciprocal mimicry is merely a complication of Miiller's 

 theory, and it is therefore important to have a very clear 

 idea of the essential nature of that theory, of the factors 

 which make for a Miillerian Association, and of the 

 conditions which are most favourable to its development. 

 The real starting-point for the Miillerian hypothesis lies 

 in the proposition (sufficiently well established by now) 



