to limitations of the Mullerian Hypothesis of Mimicry. 99 



the type form. In these particular circumstances it is clear 

 that the simultaneous and mutual mimicry, postulated by 

 the hypothesis of Diaposematism, is entirely out of the 

 question. The Mullerian factor is capable of converting B 

 into a mimic of A, but it cannot cause A to mimic B. 



This simple example illustrates several interesting 

 points. In the first place, it shows that the whole idea of 

 Mullerian mimicry practically resolves itself into a numer- 

 ical computation of the relative percentages of loss. The 

 only kind of variation which can have any significance 

 will be one that shall be trending from a form having a 

 higher percentage of loss towards one having a lower 

 percentage ; so that a difference in these percentages is 

 an essential condition for the production of Mullerian 

 mimicry. But Midler's hypothesis postulates that the 

 absolute destruction is practically constant for each group 

 of different colours, and therefore the necessary conditions 

 can arise only where there is a noticeable difference in the 

 initial numbers of the two species involved. Further, the 

 greater the discrepany between these two numbers, the 

 greater will be the advantage derived by any variation 

 from the rarer species in the direction of the more 

 numerous, and consequently the more powerful and rapid 

 will be the operation of the Mullerian factor in producing 

 a mimetic approach. This therefore constitutes the most 

 highly favourable condition for the evolution of Mullerian 

 mimicry. Conversely, as the numbers approach one 

 another, so will the value and importance of the initial 

 variation from the numerically smaller species towards 

 the larger become less and less ; so too will steadily 

 diminish the effectiveness of the Mullerian factor as a 

 producer of mimicry. Finally, when we arrive at a case 

 where the two species are equal in numbers, there will 

 then be a condition of equilibrium, and the Mullerian 

 principle will practically cease to operate altogether. 



So far as concerns the variations which may arise from 

 the numerically larger species towards the smaller, these 

 will be the more ra}3idly exterminated, the greater the 

 discrepancy in the numbers of the two species. As this 

 difference diminishes, the disadvantage of such a variety, 

 as compared with its typical form, will also diminish ; but 

 in no case can its development be fostered and increased 

 through the medium of the Mullerian factor, because it 

 still remains a variation in a disadvantageous direction. 



