22 Euploeines forming Mimetic Groups 



A third species is compta Bob. (figs. 3, 4), belonging to the group 

 without any sexual stripe on the fore wing. This is one of the rarest 

 species in the genus and represented only l'l per cent, of the whole. 

 This species bears a closer resemblance to visenda than it does to 

 sacerdos, whilst in structure it is probably nearer the latter. 



A fourth species is peloro ides sp. nov. (fig. 8 S , 9 ? ), a description 

 of which is given in the appendix. It is rarer than compta and formed 

 only '8 per cent, of the whole. It belongs to the Stictoploea group and 

 bears a very close resemblance to sacerdos. 



Among the butterflies of other families found on Tenimber, there is 

 only one which can be said to enter into the Euploeine combination. 

 This is the ? of Hypolimnas alimena forbesi Butl. (fig. 7). This 

 formed 15 per cent, of the total of Euploea and forbesi. 



We have seen that on the Key and Aru Islands the pattern of this 

 Hypolimnas ? is accentuated in combination with a white-margined 

 group of Euploeas. 



Variation in the Spots of Sacerdos. 



When first examining the series of this species, it was noticed that 

 the spots on the fore wing were of obvious variability in size, and that 

 the discal spots appeared to be more variable than the subapical ones. 

 At the time the importance of such a variation was not considered 

 but for the purpose of the present investigation, the extent of this 

 variation requires to be ascertained more exactly. 



If one group of spots were less variable than the other we should 

 incline to think that the less variable were the more ancient and the 

 more important in preserving the mimetic likeness. Hence the spots 

 which vary most would be those which enter least into the mimetic 

 combination. 



It is evident that more opportunity for mimetic approach would 

 occur in the more variable spots. As the mimetic likeness became 

 stronger the variation would become less, and the more perfect the 

 likeness the less variation should be found in the spots composing 

 the whole pattern. In cases of close mimetic approximation little 

 variation is noticed in models and mimics from the same locality, but 

 the once variable mimic shows in the dyslegnic edges of its pattern 

 the existence of the varietal tendency. 



We hope to return to this aspect of variation in a later paper when 

 the spots of a large series have been measured. Then it may be shown 



