REPORT OF THE STATE ENTOMOLOGIST I907 3O7 



to hold this as an adequate explanation of the origin of these 

 structures, when one considers the very tortuous courses taken 

 by these fili in the male Asphondylia for example, or the bizarre 

 form presented in Winnertzia. Furthermore, if these structures 

 were originally the thickened margins of lamellae, we would 

 expect traces of a membrane in some of the lower forms such, 

 for example, as the females of Dasyneura, Rhabdophaga and 

 Lasioptera. There appears to be no trace of any such remnant, 

 and we are therefore led to believe that these organs may be 

 hypodermal structures which, through a process of development, 

 have migrated from the interior of the antennal segment, be- 

 coming external, and thus greatly increased their value as 

 auditory organs. An alternative explanation is that the cir- 

 cumfili are simply specially modified setae which have become, 

 in some unknown manner, most intimately connected one with 

 the other. 



STUDIES IN CECIDOMYIIDAE II 

 The following matter relating to the Cecidomyiidae represents 

 only a small portion of our work upon this group. It is pub- 

 lished as a preliminary contribution to the classification of our 

 American forms. The tables for the separation of the major groups, 

 genera and species have been sorely needed in the past, and will 

 undoubtedly have a most important influence on all subsequent work 

 in this extremely interesting and very important group. 



PRELIMINARY KEYS FOR THE RECOGNITION OF CER- 

 TAIN CECIDOMYIIDAE 



The following diagnoses and keys should prove of consider- 

 able service in identifying many of the American species be- 

 longing to this exceedingly interesting group. They are to be 

 regarded as tentative, since it has been impossible for us to 

 go over the entire material in a thorough manner, and further 

 study may lead to considerable modification in our views. 



Keys to subfamilies and tribes 



a Metatarsus longer than the following segment; five tarsal segments and 

 at least four long veins 



h Fourth long vein present Subfamily LESTREMIINAE 



c Fourth vein forked Tribe LESTREMIINARIAE 



re Fourth vein simple Tribe CAMPYLOMV ZARIAE 



aa Metatarsus usually longer than the follovi/ing segment; at least three 



long veins Subfamily HETEROPEZINAE 



aaa Metatarsus always shorter than the following segment ; wings with three 

 or four long veins Subfamily CECIDOMYIINAE 



