356 



Subfamily LIMNOBIIN^E. 



The principal characters of the LIMNOBILN^E are : the ending of 

 the auxiliary vein in the costa, instead of in the 1st longitudinal 

 vein ; the presence of a subcostal cross-vein between the auxiliary 

 and 1st veins, which occurs in the great majority of the species ; 

 the presence of the marginal cross-vein in a considerable number 

 of genera, a vein which is absent in TIPULINJE, though it may be 

 considered to be therein represented by what I term the costal 

 cross-vein, in spite of the slightly different position it holds. This 

 costal cross-vein never appears in the present subfamily, and Osten 

 Sacken's rhomboidal cell is therefore always absent also. 



Another great difference, which is practically constant, is the 

 ending of the 1st longitudinal vein in the costa, instead of turning 

 down into the 2nd vein as is practically always the case in the 

 TIPULINJE.* When (as is very often the case) the marginal cross-vein 

 is placed near the tip of the 1st longitudinal, the effect is frequently 

 produced of the latter turning down into the 2nd vein, with a cross- 

 vein joining it near its tip to the costa. The student must avoid 

 making this error, however apparent such a case may seem to be. 

 In a few abnormal genera (Antoclia and Toxorlrina, for instance), 

 the auxiliary and 1st longitudinal veins are united, in which case 

 the united veins are recognized as the 1st longitudinal, and this 

 itself in these genera becomes gradually merged in the costa, which 

 in consequence is a little thickened at that spot for a short distance. 

 The 2nd longitudinal vein is often forked, and one or other of its 

 branches again forked, thus making the number of marginal and 

 snbmarginal cells inconstant, and contingent on whether it is 

 respectively the upper or lower branch that is forked. 



A very constant and characteristic feature is the position at 

 which the lower branch of the 4th longitudinal vein forks, which 

 is at or beyond the distal end of the discal cell, and not at the 

 basal or proximal end of this cell as in TIPULIN^E. This feature 

 causes several other peculiarities, which are best studied in 

 comparison with their relative differences in TIPULIJT^. The 

 obliterative streak, so conspicuous in the genus Tipula itself, 

 and present to a less extent in some of the other k genera of 

 TIPITLIN^:, is invariably absent in LIMNOBIINJE. 



The antennae in LTMNOBIINJE are normally 14- or 16-jointed, but 

 exceptions are by no means rare ; such are the ANISOMEEINI, with 

 6- and 10-jointed antennae, and a few other abnormal genera with 

 13, 15, 17, and so on. The scape is usually much broader, espe- 

 cially the 2nd joint, than the basal flagellar joints. The nasus or 

 nose, so conspicuous a character in most of the TIPULIN^, is absent 

 in the present subfamily. The palpi are not so long and whip- 

 lash-like, although moderately elongate and generally incurved ; 



* Cylindrotoma forms the only Indian exception. 



