732 PROCEEDINGS OF THE NATIONAL MUSEUM. vol. xxvi. 



thTcharacters there presented are relatively primitive for the jrroup, 

 and while ouch of these characters is exhil)itod in some of the more 

 o-eneralized forms, it is quite impossible to liiul a single wing embody- 

 hi"- them all. Let us therefore now carefully compare the several 

 records and discover, if we may, the history of the principal Odonate 

 wing tvpes. 



■ THE FIRST DICHOTOMY. 



The first dichotomy has been already indicated, or, at least, implied. 

 Aside from the crossing of the radial sector, respecting which all 

 dragon-fly wings are in essential agreement, the cpiadrangle has cer- 

 tainly played the most important role in the evolution of these wings. 

 Upon the quadrangle two distinct tendencies have operated, producing . 

 the subt)rders Zygoptera and Anisoptera. In the former the tendency ^ 

 has been to preserve the quadrangle as a unit of wing structure, to 

 continue vein Ct/^ outward from it in direct line, and to set off Cu^ , 

 squarely upon the distal end of vein .l^. In the latter the tendency . 

 has been to divide the quadrangle into triangle and supertriangle, and 

 to differentiate between fore and hind wing. Supplements are devel- 

 oped only in the latter, while the tendency toward the matching of 

 cross veins and the reduction of the l)ase of the wing are characteristic 

 only of the former. 



SUBORDER ANISOPTER.\. 



Again, in the two great families of this sul)order two dominant tend- 

 encies are almost equally plain. In the ^Eschnidtv these are toward 

 (1) the similar elongation of the triangle in both wings; (2) the develop- 

 ment of strong supplements; (3) the hypertrophy of two antenodal 

 cross veins; (1) the development of a brace to the stigma, and (.5) the 

 angulation of the hind angle of the hind wing in the males. In the i 

 Libellurulie there is the single, more definite, and more exclusive tend- 

 ency, already discussed, toward the differentiation between fore and 

 hind wings at the triangle. 



a: ^Fsc/i liida^. -^The family ^Eschniche is a bundle of remnants. In 

 it are found no less than eight fairly distinct types of venation. The 

 types represented by IleteropJdeUa, StenojihlSia, and ^Eschnidium'' 

 are extinct. That of Petalura was much more abundant in past times 

 than now. Chlorogomphm and Cordidegaster are represented by but 

 a few isolated species. Only the groups of Gomplim and ^Eschna are- 

 dominant at the present time. In some cases, therefore, there is nc 



« Attention should be called here to the peculiar and altogether isolated type of vena- 

 tion found in .■Esrhildiiun. In winjjs, otherwise like those of the vEschnidte, the 

 shifting of the anal vein up toward the inner angle of the triangle is completely carried 

 out in both fore and hind wings. In this, and also in the secondary developments for 

 supporting the expanding anal angle, .Kfchiidlum far outran all living forms. In the 

 matter of differentiation between veins and membrane, however, and the reduction 

 of cross veins, it riMuaint-d verv generalized indeed. 



