to limitations of the Milllerian Hypothesvi of Mimicnj. 101 



to point out that unless due care be exercised in the use 

 of this kind of argument from advantage, it is liable to 

 lead to erroneous conceptions. In order to show this 

 more clearly let us apply the same line of argument to a 

 slightly different case. If we suppose that there are 

 20,000 examples of A and only 10,000 of B, each losing 

 1000 ; then when A stands alone it will lose 5 per cent, of 

 its numbers. But if A develops a mimetic tendency in 

 the direction of B, and finally becomes a mimic of that 

 species, then its loss from experiments will be reduced to 

 only 3i per cent. Here again there is a clear advantage 

 as compared with its previous condition, therefore it will 

 be claimed that the Miillerian factor must be capable of 

 converting A into a mimic of B. 



Now we have already seen that where the numbers of 

 A are considerably in excess of those of B, as in this case, 

 any initial variation (in the Darwinian sense) from A 

 towards B will be going from a lower percentage of loss in 

 the direction of a higher percentage, and that therefore 

 that variation will be at a disadvantage as compared with 

 its own type form, in relation to the factor which is 

 causing the loss. In such circumstances therefore the 

 Miillerian factor cannot convert A into a mimic of B. 



Here then these two lines of argument, based on the 

 same data, have led to diametrically opposite results. The 

 reason for this divergence is not far to seek. It will be 

 observed that the contention in favour of a Miillerian 

 approach from A to B is based entirely on a consideration 

 of the advantage which would accrue ichen the mimicry 

 hud hecome an accomplished fact, while tbe intermediate 

 stages are in no way taken into account. But the whole 

 Darwinian idea of the evolution of such a case of mimicry 

 involves the assumption that it has been built up by a 

 gradual process of selection from comparatively small 

 individual variations. Therefore the essential point to be 

 considered must be the question of relative advantage as 

 between the initial variation and its typical form ; and 

 the assumed ultimate advantasre has no real significance 

 unless it can be shown that the initial variation has some 

 definite advantage over the type in relation to the 

 selective factor which is considered to cause the mimetic 

 approach. "Where this point is overlooked, an argument 

 based merely on a supposed eventual gain may, in some 

 cases, be entirely misleading. 



