IN DEFENSIVE COLORATION 329 



species of insects equally distasteful to young birds, and 

 let it be supposed that the birds would destroy the same 

 number of individuals of each before they were educated 

 to avoid them. Then if these insects are thoroughly 

 mixed and become undistinguishable to the birds, a pr'o- 

 portionate advantage accrues to each over its former state 

 of existence. These proportionate advantages are in- 

 versely in the duplicate ratio of their respective original 

 numbers compounded with the ratio of the respective 

 percentages that would have survived without the 

 mimicry.' The difference between this hypothesis and 

 the Mimicry of Bates is well shown in the following 

 passage, in which it is supposed that B exists in smaller 

 numbers than A : — * It must be remembered, however, 

 that B does no harm to A by mimicking it ; on the 

 contrary, the act of mimicry is of advantage to A over 

 its former state of existence as well as to ii* ; but A being 

 the more numerous the advantage is less. Still, after the 

 assimilation neither has an advantage over the other. 

 Proportionally they suffer from the ravages of birds 

 equally ; the percentage of losses is the same ; they are 

 on equal terms. No matter how long they continue the 

 association, neither gains nor loses on the other ; though 

 through one being more numerous it loses more indi- 

 viduals, yet equally in proportion with the other. So 

 that, if one is twice as numerous as the other at the time 

 of assimilation, it must always — other conditions being 

 equal — remain twice as numerous.' ^ 



2. The Advantage Conferred during the Growth of 

 Milllerian Rese^nb lance. — The above statement is con- 

 cerned with a Milllerian Resemblance which has reached 

 the climax of evolution, when the constituent species 

 cannot be distinguished by their enemies. It is very 

 doubtful whether this climax is ever attained except when 

 affinity comes to the aid of mimicry. At the same time 

 it is useful to assume indistingulshability in a hypothetical 

 example, if by this means the advantages of resemblance 



^ The full slatement here quoted was published by the authors in 

 Nature, \o\. xxix, 1884, PP- 405~6 : a preliminary statement had been 

 published in Nature, vol. xxvii, 1883, pp. 481-2. 



