100 Mr. G. A. K. Marshall on Diaposematism, with reference 



Further, as the numbers approach each other, any 

 variation from the larger towards the smaller tends to 

 produce equality, a condition which effectively prevents 

 the Miillerian selection from producing any mimetic 

 results. We see then that a Miillerian approach will 

 only take place in one direction, namely, from a rarer 

 species towards a more abundant one, and no species 

 can in this way approach another which has fewer 

 individuals (and therefore a higher percentage of loss) than 

 itself. 



If this conception of the conditions which make for 

 Miillerian mimicry be accepted as sound, it would appear 

 that the hypothesis of Diaposematism is placed on the 

 horns of a veritable dilemma. For this hypothesis differs 

 in no real essential from Miiller's principle, of which it 

 appears to be merely an extension ; and yet we find that 

 the very conditions which are most highly favourable for a 

 simple Miillerian approach are at the same time absolutely 

 fatal to anything in the way of reciprocal mimicry. On 

 the other hand, this latter conception only begins to 

 appear feasible at the other end of the series, namely, 

 where the numbers of the two species approximate equality. 

 But we have seen that this is a condition which renders 

 any occurrence of Miillerian mimicry in the highest degree 

 improbable ; and there can be no Diaposematism where 

 there is no Miillerian mimicry. In my opinion, the above 

 considerations render it difficult to accept the essential 

 idea of reciprocal mimicry even as a mere working 

 hypothesis. 



Now it may be urged that the statement as to the 

 practical impossibility of Miillerian mimicry when the 

 numbers of the two species are equal cannot be maintained, 

 and that its unsoundness can be demonstrated by some 

 such illustration as the following. Supposing that A and 

 B are each represented by 20,000 individuals and that 

 they lose 1000 apiece from experimental tasting when 

 their colours are different; then if they enter into a 

 mimetic association they will only lose 1000 out of 40,000 ; 

 in other words, their losses from this cause will be reduced 

 by one-half. Here is a clear and decided advantage, and 

 therefore it may be claimed that the Miillerian factor 

 must have scope to produce such a mimetic approach. 



But before hastening to accept this conclusion it is well 



