mi MilJkrian Mimicry and JXfqwsematism. 567 



Still keeping to the supposed case, and bearing in mind 

 that A and B are each of them originally suflfermg the 

 same percentage of loss, we find a certam number of A 

 varyino- in the direction of B, that is to say, exhibitmg an 

 aposeme which is sufficiently like that of B to be confused 

 with it. B and the variety of A, which wo will follow Mr. 

 Marshall in calling A', now form, so far as B's aposeme is 

 concerned, a homogeneous and mutually protective assem- 

 blaf^o But in adopting more or less of the aposeme of B, 

 A' lias not necessarily lost hold of its original aposeme, 

 and in every case where this is retained in recognisable form, 

 A' will share in the protection afiforded by both aposemes, 

 and will therefore have an advantage over both A and B, 

 which by hypothesis are not mutually protective. 



It will probably occur to any one who considers this 

 point, that there must be a strong tendency towards the 

 production and preservation of intermediate lorms, stronger 

 in the first instance than that towards the complete assi- 

 milation of one form to another. No doubt this is the case, 

 and on examining actual instances we find plenty of indi- 

 cations of the operation of this principle. I shall have 

 more to say on this head later on (see page 571), but it is 

 incumbent on me, in the first place, to show how completely 

 a recognition of the factor I am now discussing alters 

 the whole aspect of reciprocal approach. I have implied 

 already that I do not greatly favour the attempt to solve 

 problems of this kind by means of numerical calculation ; 

 but Mr. Marshall has appealed to arithmetic, and to arith- 

 metic he shall go. r ■, ,, 1 J. r 



We will suppose then, as Mr. Marshall does, two dis- 

 tasteful species, A and B, equal in numbers and distinct in 

 appearance. We will also eliminate the effect of disturbing 

 factors by supposing that the two species are equally dis- 

 tasteful, equally conspicuous and eciually given to selt- 

 advertisement. Under these conditions the aposemes ot 

 A and B respectively will be learnt by the sacrifice of an 

 equal number of A and of B ; and as A and B are equal 

 in population, this will mean that the percentage loss ot 

 each is the same. This is the state of things, reduced to 

 its simplest expression, in which Mr. Marshall thinks that 

 equilibrium will occur, and " the Miillenan principle will 

 practically cease to operate altogether " (p. 99). 



We will now express the case arithmetically. The actual 

 numbers we take are immaterial, the only essential point 



