678 Mr. R. C. L. Perkins on the 



between inverted commas and dated. I have not neces- 

 sarily kept the extracts from the note-book in their original 

 order and have ventured to condense certain parts. Beyond 

 the point where extracts from the note-book cease, the 

 quotations from the correspondence are no longer placed 

 between quotation marks and are dated at the head 

 instead of at the foot as in the earlier parts of the paper. 

 In the concluding pages the passages are grouped under 

 three separate heads. The few slight additions of my 

 own are placed between the square brackets. Species 

 quoted without an author's name were described by Dr. 

 Perkins himself. — E. B. Poulton.] 



EUMENIDAE. 



[This family is considered first because of the number 

 of the Hawaiian species and the dominant position taken 

 by them in the Colour-groups of these islands.] 



The whole of the species, to the number of 102, belong 

 to the almost ubiquitous genus Odynerus, sensu latiori. 

 From this interesting complex I have split off three small 

 groups of species and considered them as distinct genera, 

 as indeed they are, although they appear to be derivations 

 of the same stock as the Hawaiian Odynerus proper. The 

 Hawaiian Eumenids are, I now think, clearly descendants 

 of two quite distinct forms of original immigrants, one 

 of which, a yellow-banded form, gave rise to the bulk 

 of the species, as well as to the endemic genera that I 

 have separated from these, while the other has pro- 

 duced but four distinct species, as at present discovered, 

 viz. 0. nigripennis, Holmgr., and its three allies. This 

 little group has now been traced to an Asiatic ancestor 

 which is, I suspect, an ancient or primitive type, showing 

 some affinity to the genus Rhynchium, in which nigripennis 

 itself was originally placed by Holmgren. 



" I have not yet identified the Oriental species (just 

 lately discovered while mounting some insects) that is 

 allied to the 0. nigripennis group. It is the closest 

 approach I know to the genus Rhynchium, but it is not 

 that genus." Nov. 13th, 1911. 



Species of Odynerus are almost ubiquitous throughout 

 the islands, though some of the densest and wettest boggy 

 forests are absolutely devoid of them. At the same time 

 a slight change in these, made by the incursion of cattle, 



