,„, NORTH AMBRIC \N DIPTERA. 



On the othei hand the structure of the ndult insects 

 changes so much somewhere in the vicinity <>i the 6yr- 

 phidse, to say nothing of larval sti ucture m ml li ;i 1 »i t s . that 

 l believe, ii we must have suborders (which to me does 

 not seem imperatively necessary) the place to make 

 the division is somewhere between the Syrphidae and 

 Dolichopodidse. There are groups here which are oscul- 

 ant DO doubt B lot of ink has been wasted in their dis- 

 cussion, but so there arc in all classifications. The 

 soonei we learn, as learn we must, that dividing lines do 

 not occur in nature except as accidents, that evolution 

 never classified anything and never will, the sooner we 

 shall get over some of the bugbears of taxonomy. The 

 best we can do is to make such groupings as will express 

 most clearly and most conveniently relationships, and 

 especially the phylogeny oi organisms. 



As I have already said, a true classification ol the dip- 

 tera, as ol all othei organisms, is desirable, but not pos- 

 sible, toi that would mean the absolute perfection i I 

 dipterological science. An approximation to that finish- 

 ed perfection is oi course attainable, but that approxima- 

 tion must depend upon many factors which have as yel 

 scarcely engaged the attention ol students of diptera. 

 Their embryology, geological history, geographic dis- 

 tribution, ethology and comparative anatomj . are among 

 these factors. No dipterologisl has ever given serious 

 attention to the study of extinct forms, and no classifica- 

 tion oi any group oi organisms has ever been satisfactoi \ 



until such forms have been considered and senOUSl) con- 

 sidered, paleontology and embryology especially, are 

 the sine qua mm oi any entomological i lassification, and, 

 considering the relatively slight advances which have 



been made in the taxonomy of the diptera during the 



past hall centui y, now seems the opportune time lor such 



studies. What dipterologisl will undertake them ? 



