Studies on Tipulidae II. 165 
In preparing the present paper I have largely used the materials 
I found in the collections of the Berlin Museum. To the authorities 
of that Museum I express herewith my sincere gratitude for their 
liberality and courtesy. 
Introduction. 
The characters upon which, in 1859, was founded my primary 
subdivision of the Tipulidae brevipalpi in sections (empodia, 
spurs on the tibiae, number of submarginal cells, nnmber of joints 
of the antennae and structure of the male forceps) have, upon the 
whole, proved remarkably trustworthy. The system based upon them 
‚has absorbed without difficulty all the additions since made to it. 
Exceptions are rare. Abnormal forms, like Chionea, found their 
right place when tried by the test of these characters; forms very 
dissimilar to the eye like Libnotes and Peripheroptera, owing to the 
same test, proved to be close relatives. — In three instances only 
has the number of submarginal cells failed to determine the natural 
position of a genus, and even these exceptions are only apparent, and 
easily explained. Goniomyia manca O. S. (N. America), has only 
one submarginal cell, instead of two; but this is owing to the obl- 
teration of the anterior branch of the second vein, which is usually 
very short in that genus. A precisely similar case is that of Oloni- 
ophora Loew, an Anisomera with a single submarginal cell. In 
Paratropesa, on the contrary, there are two submarginal cells, al- 
though this genus being closely related to Teucholabis should have 
only one; I explain this anomaly in assuming that the vein which 
encloses this pretended second submarginal cell is not a branch of 
the second vein, but only a supernumerary crossvein, inserted at its 
place in order to strengthen the interval between the second vein and 
the costa, which is unusually broad in this genus. In the genus 
Oerozodia the somewhat aberrant course of the auxiliary vein, which 
ends in the first, is due to the weakness or total obliteration of the 
crossvein, which ought to connect it with the costa, and to represent 
its real end. 
The empodia are distinet in Dieranomyia morio, while they 
are wanting in the other species of the same section; on the con- 
trary they are wanting in some Goniomyyiae. while they are present 
in others; they are wanting in Lipsothri®, where they should have 
been present, if it be true that this genus belongs to the Zriople- 
rina. Similar anomalies in the presence or absence of pulvilli occur 
in other families (Asilidae, Bombylidae). 
12* 
