2 BULLETIN 85, UNITED STATES NATIONAL MUSEUM. 



less variation. It was soon found that by Loew's system very closely 

 related species were thrown into different genera and even different 

 subfamilies. Sometimes, too, specimens of the same species were 

 separated into different subfamilies. This was true especially of 

 several species of ApJudara and of related genera. 



As the work progressed and other large collections were studied, 

 many unsatisfactory features were found in this system and a great 

 many difficulties encountered. Some of these have been mentioned 

 in more or less detail in my former papers on this group. It became 

 apparent that any classification based principally on wing venational 

 characters would be artificial and unsatisfactory.. At the same time 

 other characters were noted that seemed to be more fundamental 

 and whose use in classification seemed to place together certain 

 obviously related genera. 



The form of the head, for instance, was found to be very significant 

 in diagnosis. In some genera the frons is a visible sclerite between 

 the gense, while in others it is completely covered by the gense which 

 in the latter case are nearly always produced into conical processes. 

 Genera related or separated on this character were found to be 

 distinctly related or separated in other structural characters, as well. 



The form of the vertex, and that of the pleurites of the prothorax 

 and of the pronotum, the armature of the hind tibiae, and several 

 other structural characters, have been found to correspond closely 

 in certain groups of genera. These, naturaDy, do not divide the 

 genera into sharply defined groups, but the groupings are, at least, 

 more sharply defined and less artificial than by Loew's system. 



As a result of this study, a complete rearrangement of the genera 

 of Psyllidse is herein presented. While only the American genera 

 are treated in this monograph, yet most of the genera of the world 

 have been considered in connection with this new system of classifi- 

 cation, and it appears that this system will be very satisfactory 

 for aU. 



The American genera have been arranged in six subfamilies — 

 Liviinse, PauropsyUinae, CarsidarinsB, Ceriacreminse, Triozinse, and 

 PsyUinse. Some of these are new names and those which have 

 already been used are here used with different limits, representing 

 a different group of genera. Each is described and discussed under 

 its o%vn subhead and need not be discussed here. It should be noted, 

 however, that the old subfamily Aphalarinse is not recognized. Its 

 typical genus, Aphalara, has been transferred to the group Liviinse, 

 and the other genera formerly placed in it have been referred to 

 Psyllinae or some other subfamily. 



Some confusion has arisen in the past over differences in the 

 nomenclature of the external sclerites, and these differences have 

 been due largely to our imperfect knowledge of the homologies of 



