26 DIPTERA OF NORTH AMERICA. [PART IV. 



forceps, and in many cases the antenna, has been noted down by 

 me, almost invariably from living or fresh specimens. 



At the end of the Tipulidse brevipalpi I place the two sections 

 which I consider as intermediate between them and the T. longi- 

 palpi (compare above, p. 19) : the Cylindrotomina and PiychojJ- 

 terina : — 



Sect. VII. Cylindrotomina. One submarginal cell ; first longitudinal 

 vein incurved at the tip towards the second, instead of ending in the 

 costa (exception : Phalacrocera replicata Lin., where the first vein takes 

 the usual course) ; four or five posterior cells ; a discal cell ; the auxiliary 

 vein is abruptly interrupted before the stigma, without ending either in 

 the costa, or in the first longitudinal vein. Eyes glabrous. Normal num- 

 ber of antennal joints sixteen. Tibise with spurs at the tip. Empodia 

 distinct. Structure of the forceps and the ovipositor peculiar and 

 characteristic. 



Sect. VIII. Ptychopterina. Only a single longitudinal vein posterior 

 to tlie fifth vein ; two submarginal cells. Labium largely developed ; palpi 

 long. Tibiae with spurs at the tip. 



After having given an account of the distribution into sections, 

 I have to add a few words on the genera. I am opposed to a 

 too great multiplication of the genera, and I believe that as the 

 contrast between large and small groups exists in nature, it 

 should also be brought before the eye in the classification. In 

 the genus Erioptei^a, for instance, the relationship of the groups 

 which compose it is a much more striking feature than the 

 characters which separate these groups. If we set up the groups 

 as genera, with only three or four species in each, the difference 

 between the large group, now called Erioptera, and smaller 

 groups, such for instance as the genera Gnophomyia, Trimicra, 

 and all the genera of the group Limnohina anomala, this differ- 

 ence, so strongly marked in nature, would remain unexpressed in 

 the system. Subdivisions of the larger genera should of course 

 be carefully marked, but less strongly than the intervals between 

 the small genera, and in such cases a subgeneric subdivision may 

 be useful. This is the course which I have followed. 



6. General remarks on the structure of the Tipulid^^e brevipalpi. 

 In this paragraph I do not intend to undertake a general com- 

 parative description of the external structure of the Tip. brevi- 



I 



