NATURAL SELECTION 261 



selection of warning patterns has brought about colour con- 

 vergence. In another scries of examples it is very difficult 

 to see how selection could have led to the observed effects. 

 In the majority of cases of mimetic resemblance, however, it is 

 impossible at present to estimate to what extent, if at all, selection 

 has been active. We are left, in fact, in a state of suspended 

 judgment : it is probable that selection has played some part 

 in the evolution of mimetic convergence, but it is usually 

 impossible to say how large a part in any particular case. 



(5) If two species of a genus enter into two different 

 mimetic associations, then the colour differences between them 

 will be adaptive in so far as the mimicry is due to Natural 

 Selection. Similarly, in a polymorphic mimetic species, the 

 differences between the various forms may be adaptive, and 

 if these differences are analogous to those observed between 

 species in other cases, then we can obtain some evidence as to 

 the extent to which specific characters are adaptive. To 

 assess what proportion of the differences observed is actually 

 due to adaptive change is very difficult and usually impossible. 

 We shall first have to consider the evolution of warning colours 

 amongst models. We are on very uncertain ground in trying 

 to decide which patterns are most conspicuous and therefore 

 most efficient in warning enemies (especially birds) against 

 making attacks. It is scarcely possible (except in the broadest 

 way) to arrange insects in a scale of distastefulness to see if 

 this corresponds in any way with the apparent scale in con- 

 spicuousness. An even greater difficulty is the lack of adequate 

 systematic knowledge. A few genera, such as Acraea and 

 Heliconius, have received thorough monographic treatment 

 (Eltringham, 1912, 1916), but even here the species are so 

 variable, have been so little reared, and many are still so 

 imperfectly known that it is still often impossible to come to 

 any very definite conclusions as to the limits of species. Further, 

 it has been a common systematic procedure to unite under 

 one species all forms connected by more or less clear inter- 

 mediate colour-forms : yet genetical experiments show that 

 a more or less continuous range of phenotypic variation may 

 be the expression of distinct genotypic composition, and the 

 occurrence of apparent intermediates is not necessarily signi- 

 ficant unless the connecting forms are ranged along a definite 

 geographical gradient. 



