582 



PROCEEDINGS OF THE NATIONAL MUSEUM. 



VOL. 45. 



Triangles with 8 or 10 cells; Mj strongly arched above fork of Rs; branches of Cu 

 below the triangle in hind wing separated for a considerable distance by three 

 rows of cells; branches of media leaving arculus distinctly above middle. 



Neurseschna costalis [Williamson photo.]. 



Triangles with not more than 7 cells; Mg more gently and regularly arched above 

 fork of Rs; branches of Cu below the triangle in hind wings separated by one or 

 (Amphiaeschna) two rows of cells 2 



1. Stigma very short, bounding about 2 cells below; triangle of three cells, the basal 



one not divided; branches of media leaving arculus well below its middle. 



Calixschna microstigma [Williamson photo.]. 



Stigma longer, bounding at least three cells below; triangle of at least five cells, 



two basal, at least in hind wing 2 



2. Branches of media leaving arculus below the middle; only one row of cells between 



the Rs and supplementary vein below it; arculus close to base of triangle; 

 stigma more or less swollen, bounding 5-5J cells below; a dark patch at nodus. 



Telephlebia godefroyi [Williamson photo.]. 



Branches of media leaving arculus above the middle; many rows of cells (about 



six at widest part) between Rs and supplementary vein below it; arculus not 



so close to base of triangle ; stigma bounding 3-3 J cells below 3 



3. Branches of Cu below triangle in hind wing separated by only one row of cells. 



Amphiaeschna from Siam [Williamson photo.]. 



Branches of Cu below triangle in hind wing separated by two rows of cells for a 



considerable distance Amphiseschna ampla [Williamson photo.]. 



The Jiirassic Cymatophlehia and Morbseschna (see Bull. Amer. Mus. Nat. Hist., 

 vol. 23, pp. 133-134, 141, 142) are certainly to be excluded from the Aeshninse. 



On other grounds, E. M. Walker divides the AeshniuiE into five groups. One of 

 these corresponds with the Anacini above. The others may represent a number of 

 tribes as follows: 



This agrees well enough with Walker's phylogenetic tree, except that Neurseschna 

 can not apparently be derived from Gynacantha, nor Amphiseschna from Aeshna. 



