Species o/Eiiocera in the British Museum, 69 



auterinse. Another feature seen in most, if not all, species of 

 Eriocei'a and Macromastix with elongate male antennae is 

 the reduction in the length of the abdomen in that sex. 



Turning to the other point on which Enderlein based his 

 generic distinctions, the number of posterior cells (presence 

 or absence of cell M{)^ here again it is doubtful if the 

 distinction has any phylogenetic value. Among those with 

 cell J/], as among those without it, there are a number of 

 species-groups which, if the genus were divided, might be 

 made into subgenera, but a study of the details of venation 

 and male hypopygium suggests that some of those without 

 cell J/j may be more nearly related to those possessing it 

 than to otliers which do not. Moreover, those possessing 

 the cell are certainly not all closely related among them- 

 selves. 



Rather than subdivide the genus into a number of natural 

 but small and poorly definable groups, I consider it will be 

 preferable to enlarge it by including the genus Penthoptera, 

 and also two species from the Seychelles which I referred in 

 1912 to Anisomera. One of these species shows a remarkable 

 variation in venation which I overlooked at the time of 

 description, and they both differ markedly from the typical 

 species of Hexatoma {Anisomera) in having a well-developed 

 ovipositor. Further, it is quite obvious that they are closely 

 related to the two species of Eriocera described from the 

 same islands. On the other hand, I consider that the two 

 species with a short fleshy ovipositor (the African E. pusilla, 

 Alex., and the N. American E. lonyicornis, Walker) would be 

 at least equally well placed in Hexatoma. 



The tendency to the development of local forms is strongly 

 marked throughout the genus, and there are very few species 

 which have a wide distribution. This may be accounted for 

 by the breeding-habits of the species, most of which probably 

 spend their early stages in the ground at the edges of ra[)id 

 streams, and probably do not migrate much from one valley 

 to another. 



In the following table of species, all those at present 

 known from the Palsearctic, Oriental, Australasian, and 

 Ethiopian regions are included, only American forms being 

 omitted. So far as possible, the diagnostic characters 

 have been arranged to give what appears to be a natural 

 arrangement of the species, but there are a considerable 

 number which I have not seen, and whose proper position is 

 therefore more or less a matter of conjecture. Nevertheless, 

 there are no fewer than sixty species in the National 

 Collection from the regions under consideration, and it is 



