230 M!!. E. B. POULTON ON THE PROTECTIVE [Milf. 1, 



species known to be nauseous or dangerous. Tliere must cer- 

 tainly be a tendency towards a further general convergence, but 

 the existing condition of convergence round a few well-marked types 

 of pattern and colouring must be highly beneficial, and there was in 

 this case no initial uniformity due to close affinity, upon which to 

 base a general and uniform system. It was in fact a priori far more 

 likely that the convergence of remote species should have been round 

 a few successful types, while the prevention of divergence among 

 closely related species must ipso facto have tended to produce 

 concentration round a single type. It will be shown below that 

 Fritz Miiller's principle is probably attended by others, which also 

 assist in producing convergence, at any rate in some cases. 



Another result of the different origin of the two classes of re- 

 semblance alluded to above is that the uniform warning colours of a 

 large group of closely related species are less conspicuous, and in 

 themselves possess less of "warning" characteristics, than those of 

 the smaller groups into which the isolated nauseous species tend to 

 converge ; for the former depend largely upon some ornamental type 

 ot colour and marking, due to sexual selection, and prevalent before 

 the time when the nauseous attributes arose. Such a type has no 

 doubt been modified in the direction of greater conspicuousness on 

 the uppeisides of the wings, while bright colours have appeared on 

 their undersides, and the mode of fliglit has been changed into one 

 which gives the colours their maximum effect ; but still, in spite of 

 these changes, the whole appearance of such large groups presents us 

 with the ancestral sexually selected colours and patterns, which are 

 of great beauty, and are no doubt still of great significance as 

 secondary sexual characters. The success of such a stereotyped 

 ornamental appearance for warning purposes has depended upon the 

 modifications alluded to above, but principally upon the very fact of 

 its prevalence and uniformity. On the other hand the smaller con- 

 vergent groups of nauseous insects often present us with ideally 

 perfect types of warning patterns and colours — simple, crude, strongly 

 contrasted — everything subordinated to the paramount necessity of 

 becoming conspicuous. For the nauseous attributes arising inde- 

 pendently among the scattered species of many genera, or in all the 

 species of small genera, instead of being chiefly concentrated among 

 the members of some one or two dominant groups, it must have 

 become impossible to rely upon the slightly altered ornamental 

 appearance existing at tlie time when the attributes arose ; but it was 

 necessary to appeal strongly to the memory of enemies by the acqui- 

 sition of some special form of pattern and colour, in which everytliing 

 is subordinated to the " warning" characteristics. In the one class 

 tiie pre-existing ornamental appearance was sufficiently well known 

 to serve as a warning ; in the other class it was not sufficiently well 

 known. 



It is quite clear that the two classes of resemblance which have 

 been just considered must be carefully distinguished from true 

 mimiciy, in which the mimicking species is without any unpleasant 

 attribute, but shelters itself under the reputatit;n of the (nearly 



