56 Psyche [April 



of aldabrensis is almost quite identical with that of the type of 

 Ptilostena. He therefore places aldabrensis without hesitation in 

 Ptilostena which he thinks was correctly placed near Goyiomyia. 

 The upturned R3 is only a specific character of Pt. recurvinervis 

 Bergr. 



5. Alexander has for some time placed Empeda as a subgenus of 

 Erioptera. In this he differs widely from Schiner and Osten- 

 Sacken who never even thought of the possibility of associating 

 these two genera. I cannot help thinking that this innovation 

 has very little to recommend it. In the paper where this new 

 place was assigned to Empeda no reasons for the change were 

 given, but later, speaking of Erioptera brevior Brun. (which is an 

 Empeda), Alexander summarily says: "Empeda is merely an 

 Erioptera in which the fusion of R2+3 is a little longer than usual." 

 This fusion is, however, much longer in Empeda, so much in fact 

 that, in opposition to Erioptera,^ the radial cross-vein joins R2 

 before its branching and that the cell R2 assumes a quite different 

 shape. And what of the other differences? Sc2 is in Erioptera^ 

 removed very far backwards from the tip of Sci, whereas it is 

 placed close to its tip in Empeda; the middle legs in Erioptera are 

 strikingly shorter than the two other pairs, which is not the case 

 in Empeda; and so on. Osten-Sacken regarded Empeda as very 

 closely related to Gonomyia, and it comes in fact at least as near to 

 this genus as to Erioptera. In Erioptera he included several heter- 

 ogenous elements, and to add Empeda to them is only to increase 

 the confusion. " Cui tending to turn toward the apex of the wing " 

 is given as a character of Erioptera in Alexander's key to the genera 

 (Proc. U. S. Nat. Mus. 44, p. 492). This character fits some spe- 

 cies of Empeda, but not all. If we unite Empeda with Erioptera 

 and consider the fact that the length of Sc is rather variable in 

 Empeda, and that the presence or absence of the cross-vein r is not 

 perfectly constant as a generic character, we could quite as well 

 unite Gonomyia with Erioptera and say: "Gonomyia is merely an 

 Erioptera with the radial cross- vein wanting." The wing-vena- 

 tion of Ptilostena aldabrensis Edw. must be very similar to that of 

 Gonomyia blanda O. S. (figured by Needham), and if the structure 



' Excluding E. imbuta Wied. which forms the very distinct monotypic genus Chilotrichia 

 which comes nearer Empeda than any other Eriopterine genus and in no case can be referred to 

 Erioptera. 



