342 



in the case of the Loiigicorns, l)nt even more so in that of the 

 excessively difficnlt genns Proterhlnns. for withont such data 

 one may well despair of arriving at any definite conclusion as 

 to tlie validity of many of the species, the variation often being- 

 excessive and the distinguishing characters very slight and diffi- 

 cult to appreciate. At present I am myself left with a collec- 

 tion from all the Islands of hundreds or thousands of undeter- 

 mined or dubious, though well-mounted, specimens. If it were 

 possil>]e for me to receive other such consignments, carefully 

 collected witli data, I should hope to 1)0 able after a time to 

 revise the w^hole genus, and possibly to make some such tabula- 

 tion of the species as would facilitate their identification. As 

 might have been expected, some of the species formerly de- 

 scribed by me are now known not to l)e distinct, while others 

 once thought to l)e confined to one Island, I have subsequently 

 taken on others. Some of the earlier s])ecies contained more 

 than one form under the same name. In my collection these 

 mistakes have been mostly rectified, but I have not had the 

 oppornmity of making the same corrections in the other col- 

 lecti<»ns. In this paper, I have not dealt with such matters to 

 any considerable extent, as it has been advisable to complete 

 the descriptions of new species and return the specimens as 

 quickly as possible, because all my collections have to l)e packed 

 up for an almost immediate removal to. a new address. 



I would urge those, who are fortunate enough to ha\-e the 

 opportunity of collecting further specimens, to aim at getting 

 together a smaller collection of specimens with accurate notes 

 as to food plants and other data, rather than a much larger 

 one collected at random, though the former method occupies 

 much more time. T feel sure that it is only from careful col- 

 lecting of speciuiens that really definite results are to be 

 expected. In examples collected at random one can always 

 pick nut certain species, as obviously distinct, but a large 

 amount of the material will remain dubious or undetermined. 

 At most one can say that specimens without carofiil notes as 



