738 



PROCEEDINGS OF THE NATIONAL MUSEUM. 



VOL. XXVI 



are developed upon the posterior side of veins Rs and M^ in Gomjyhid'h 

 (Plato XXXIV, tjo-. 3). Possibly these features indicate the tips o 

 numerous short developmental twigs. The brief records of the sovem 

 parts arc certainly contradictory; and the parts themselves, which on 

 mig-ht use as bases for the first divisions of the g-roup, seem so ncarl 

 of equivalent importance that one may hardly choose l)etween thei 

 with assurance. Probably the records of other organs will, bo mo' 

 easy of interpretation. 



The South American genus Agrlogomphm (tig. 27) exhibits vena 

 tional characters that are quite as generalized as are to be found in th 

 tn-oup. Tritmgle and su])triangle are still four-sided, the cubital an 

 anal veins before them })eing but moderately augulated, the stigma i 

 weakly braced, there is no anal loop and there are no supplements. 



Fig. 27. — Wings of Agrmgomphus sp'i 



We ma}^ remark, in passing, that the Gomphina?, having attained 

 fair measure of advantageous vein arrangement, and a fair (or, in tt 

 strongest species, even a superior) degree of vein diti'erentiation, see) 

 to have turned to a specialization of another sort. The imagoes are o 

 the wing less constantly than other Odonata and fly for a shorter perio( 

 During this time their chief concern is with reproduction, hence ^^ 

 find the external genitalia highly specialized. The shortening of tl 

 period of imaginal life is correlated, also, with extreme specializatio 

 of the nymph. In fact, the struggle for existence has })een mainl 

 transferred to nymphal life.'* 



" I liave discussed this niatttcr more at length in my paper, PreUininary studies 

 North American Gomphina^ Can. Ent., XXIX, 164-168, 1897. 



