THE CANADIAN ENTOMOLOGIST. 47 



ON THE CLASSIFICATION OF THE xMOSQUITOES. 



BY HARRISON G. DYAR AND FREDERICK KNAB. 



Now thai Professor Williston has cleared the ground and destroyed 

 the Theobaldian classification of Culicidas, let us try a little constructive 

 work. We regard it as essential that all the groups, both generic and 

 higher, should be based only on characters found in both sexes of the 

 adults; that these characters should be fundamental as generally recognized 

 by systematists, and that they should be supported by sound larval char- 

 acters. We have only one cause of difference with Prof Williston's 

 remarks, namely, his implied statement that the palpal characters are of 

 value in generic definition. They are not, in the case of the mosquitoes. 

 The differences consist in varying length and the number of joints. They 

 seem at first sight interesting, and we were much attracted to them on 

 beginning our generic studies. But they prove to be entirely secondary 

 sexual characters, not correspondingly represented in both sexes, and are, 

 therefore, ruled out. Moreover, the small terminal joint or joints of the 

 female palpi, on the presence or absence of which Neveu-Lemaire's 

 classification is based, is variable within the limits of a single species 

 [Culex tarsalis, Coq.), and is gradually evanescent in another series of 

 species {y^des, spp.), besides there being no modification in the male to 

 correspond with it. The long palpi of the male have been developed 

 independently in several groups (the short palpi being the generalized 

 condition), and are therefore a parallel development without fundamental 

 value. Therefore, the old classification, which Prof Williston advises his 

 readers to retain, is unsound, as it is based on these palpal characters. 

 We may remark that the same condition appears to obtain in the Tipulidje, 

 since Loew says, speaking of the division of the family on the long and 

 short palpi : "The division, indeed, is no natural one " (Dipt. No. Am., 

 lo, 1862). 



All the subfamilies of the Culicidse recognized by the Theobaldian 

 school are untenable, including the Anophelinje. We have found only 

 two subfamilies, the Culicinae and Sabethinae. W^e will not quarrel with 

 Prof. Williston over the terminology, but hasten to call them tribes. The 

 Culicini, then, have the metanotum devoid of setge; the larvae furnished 

 with a median ventral brush on the anal segment ; the Sabethini have a 

 group of setse on the metanotum, and the larvae without a ventral brush 

 on the anal segment These are primary and essential divisions, the two 

 groups showing a general dissimilarity in their appearance and habits, both 

 as adults and larvse, beside the structural points noted. 



February, 1907 



