Calvert: Odonata of the Neotropical Region. 



117 



Postnodals, front 14 do. 12 



wing 

 Black ray on sec- 

 ond lateral thoracic 



suture very narrow narrow 



Pale anteterminal 

 yellow ring on ab- 

 dominal segments very indistinct, al- 



3-6 most absent narrow 



Black rings on (feet yellow;' in- 



femora very indistinct or terior of femora and two 



absent of tibiae absent) 



Length of abdomen 



in mm. (broken) 



Length of hind 



wing in mm. 21-24 



40-42 



41 



22 



It seems fairly certain that the pair which de Selys doubtfully re- 

 ferred to fragilis Hagen in 1886, and for which he suggested the alter- 

 native name of cornuta^ are not fragilis. It is possible that the Porto 

 Rican male above described may also be c or nut a ; it agrees better 

 therewith than with fragilis. Finally attenuata Selys may not be spe- 

 cifically distinct from the true fragilis Hagen. 



Genus Argia. 



As is to be expected in any considerable collection of Odonata from 

 tropical continental America, large numbers of specimens belonging 

 to this genus are before me, some of them representing species, which 

 seemingly were not described by de Selys. On the other hand, some 

 of the forms which he and Hagen enumerated in 1865 are not to be 

 found among the present material, so far as I can see. 



To assist in the identification of members of this difficult genus and 

 to correlate the South American species with those from Mexico and 

 Central America, a synopsis of the principal color characters and of 

 some of the structural features is herewith presented. As far as 

 possible the same expressions and the same letters for indicating the 

 divisions of this key have been employed as were used in the Biologia 

 Centrali-Americana, volume Neuroptera, pages 69-74 and 358-361. 

 One difference must be noted, however. In the l Biologia,' the black 

 humeral stripe was said to be " forked above " when " a dark mesep- 

 imeral stripe is fused with the humeral stripe proper at its lower, but 

 not its upper, end" (/. c, p. 69). It has seemed better, here and 



