( 44 ) 



all on the work and conclusions of such a master of the 

 Hawaiian fauna as the author had proved himself to be. 

 Nevertheless he ventured to make a few remarks bearing 

 upon the origin and present distribution of the Colour-groups 

 in the islands and on one or two other points. Dr. Perkins 

 had brought forward strong evidence for the conclusion that 

 the first immigrant Odynerus was an ordinary-looking yellow- 

 banded species — viz. one that had previously been an insigni- 

 ficant member of one of the largest and most widely dis- 

 tributed of the Aculeate combinations, containing many of 

 the most formidable and dominant species, and bearing pro- 

 bably the simplest and most effective of warning patterns. 

 The immigrant ancestor had behind it endless generations in 

 the course of which its pattern had been rendered stable by 

 selection ceaselessly exercised on some unknown continental 

 area. Thus it was possible to understand the remarkable fact 

 that so much of the original pattern should have survived or 

 should still be revealed by reversion, at the close of a period 

 long enough to have produced all the Eumenid Structure- 

 groups in the islands except that associated with the later 



[lxi 

 Immigrant 0. nigripennis. Prolonged isolation, in the 

 Hawaiian islands, from all the other dominant bearers 

 "i tlie yellow-banded pattern also belped us to understand 

 the ultimate loss of the original pattern in so many of the 

 Bpecies. 



The mention of this great dominant Aculeate pattern made 

 it appropriate to refer at this point to a question raised b) 

 Dr. Perkins in his paper — " Why Bhould Colour-groups be 

 formed at all I Why is not the fact that an insect belongs 

 to the Aculeates sufficient warning by itself?" It might be 

 replied that the Aculeates themselves are probably avoided 

 for different reasons and in different degrees, and that, for 

 securing the advantages of Miillerian association, colour 

 and pattern are probably the most easily recognised and 

 remembered of all the characters that can be seen at a little 

 distance when an insect i- at rest. There was furthermore 

 much, but not nearly enough, experimental evidence that 

 insect-eating animals were greatly impressed by the patterns 



