16 MOSQUITOES OF NORTH AMERICA 



Theobald's extensive monograph was based essentially upon the work of his 

 predecessors. He brought forward the excellent character of the presence or 

 absence of setae upon the postnotum, but did not rank it sufficiently high. He 

 studied carefully the shape and distribution of the scales, and came to depend 

 more and more largely upon them as the principal character in the separation 

 of genera. Wiedemann, Loew, and Williston rightly discarded Desvoidy's 

 genera Psorophora and Sahcthes as being founded upon superficial characters of 

 vestiture. The same objection, only with increased force, applies to all the 

 genera founded on scale-characters. We now recognize Psorophora and Sahethes 

 on other characters, but this does not invalidate the principle laid down above 

 that superficial adaptational characters should not be made the basis of generic 

 definition. They are of specific value only. It is true that these characters may 

 in many cases show true affinities, but discrepancies would not become apparent 

 to a student working with only one set of characters. Nevertheless discrepancies 

 are likely to occur at any time whenever the character in question becomes 

 affected by adaptation, so that a natural system will not result. 



Theobald's comprehensive work stimulated many students to attack the sub- 

 ject, but for the most part they revolved in the same orbit, and the weeding out 

 of the unreliable characters proceeded very slowly until Dyar and Knab ap- 

 proached the subject from a new standpoint, that of the larvae. 



The subject of the characters shown by the larvae was independently taken up 

 by Christophers, who discussed the larvae of India. The species there seem to 

 exhibit the same problems as the North American ones, all the principal types 

 being represented, both vegetable-feeding and predaceous, except the sabethine 

 forms, which seem not to have come under his observation. Christophers duly 

 comments on the numerous unlike types which have been heretofore included 

 under the name Culex, but does not separate them out under definite names. 

 His new genus, founded for Culex concolor and G. tigripes, seems undoubtedly 

 valid, for the larva are of a most peculiar type, unlike anything familiar to us, 

 and the adults should be examined for generic characters. 



At about the same time Felt, and also Dyar, used the modifications of the male 

 genitalia as the basis of generic separation and definition. It is obvious that 

 genera founded upon larvae or male genitalia violate none of the principles laid 

 down above, and the only objection to be urged against them is one of con- 

 venience. Not all the specimens received for determination are bred, nor are 

 both sexes always represented. Moreover, the proper association of male and 

 female is not always certain, while the proper examination of the structure of the 

 male genitalia requires the practical destruction of the specimen and consid- 

 erable time. It seems therefore advisable to restrict the characters used in 

 generic definition to those found in both sexes of the adult. 



Stimulated by this work, Coquillett produced a classification in which the 

 objectionable characters were almost entirely eliminated as far as the sub- 

 families were concerned. They would have been completely so except for the 

 attempt to recognize the subfamilies proposed by Mitchell. In the genera he 

 failed to introduce new and valid characters, for the reason that it was im- 



