NATHAN BANKS. 23 



The Hemerobiidse were formerly united to the Chrysopidse in one 

 family. The restricted family has commonly been divided into 

 three groups. One of these is based on the peculiar genus Dilar. 

 Another (Osmylinje) has often been separated by having but one 

 radial sector, while the true Hemerobiinse have two or more sectors. 

 This, I consider, a poor character, since the number of radial sec- 

 tors varies in the species and in the opposite wing of the same speci- 

 men. Therefore I have separated the two groups on the point of 

 termination of the subcosta ; whether it ends in the margin of the 

 wing (Hemerobiinse) or whether it runs into the radius (Sisyrinse). 

 This makes a slightly different alignment, and I think more natural. 

 We do not have the genus Osmxjlus in this country, and I expect 

 that it, together with various allied exotic genera, will form another 

 subfamily. Our eleven genera are based almost wholly on vena- 

 tional characters; while specific characters are also found in the 

 venation, the shape of wings, the male genitalia, and to a less ex- 

 tent in coloration. The pattern of coloration is usually constant for 

 each species, but the extent of its development varies greatly ; too 

 much stress has been laid upon the markings by many authors. 

 Striking differences in markings, unaccompanied by structural dif- 

 ferencesj should be looked upon with much suspicion. 



This paper is based principally on my own collection, but I have 

 had free use of the material in the National collection, and, through 

 the kindness of Mr. Henshaw, have examined the material in the 

 Museum of Comparative Zoology at Cambridge containing nearly 

 all of the Fitch types. 



The three subfamilies known to occur in our country may be 

 tabulated as below : 



SYNOPSIS OF SUBFAMILIES. 



1. Male with pectinate antenna- ; female with an exserted ovipositor. ..Dilarin^, 

 Antennse not pectinate in either sex ; female without exserted ovipositor. . .2, 



2. Subcosta runs into radius near pterostigma; usually but one radial sector, at 



least none arising beyond basal third of wing Sisyrin.s:. 



Subcosta runs into margin of wing; normally at least two radial sectors, the 

 last beyond basal third of wing Hemeeobiin.^, 



DlLARIN^. 



Contains but the one genus — Dllar, 



TRANS. AM. ENT. SOC. XXXII. DECEMBER, 1905. 



