^16 ON THE GENUS CORDULEPHYA, 



to three or four cells in C. montana shews us still a ^vell- 

 defined anal loop, especially in the female ; from which we 

 may infer tliat the original larger loop was also fairly well defined, 

 as in other Eucordulina. Another interesting point is the 

 reduction of the niembranule. In C. montana 9 it is still 

 present in both wings ; in C. montana d* it is absent from the 

 forewing, and present, though much narrowed, in the hind- 

 wing. In ('. i)y(jnimi, both sexes, it is completely eliminated, 

 being replaced in the hindwing by a small tuft of hairs 

 (Plate xii., figs. 3-6). Finally, it should be noted that some 

 of the parts vmder discussion are still variable, a long series of 

 G. 'pyanKfa shewing considerable differences in the position 

 of the "break" of the costal side of the triangles. A few 

 specimens have normal hindwing-triangles, and a nearly 

 normal one in the forewing ; but by far the greater number 

 incline to the opposite extreme. 



D. Comparison of the two species apart from venation. — 

 Two points here are worthy of note. Firstly, C. montana 

 has the more corduJiform abdomen of the two. This shews 

 that the species, before reduction, may have had a typically 

 Eucordulian abdomen. I do not think the original abdomen 

 was broadly corduliform, but a very good idea of its probable 

 appearance may be gathered by comparing the abdomens of 

 the Synthemina; the form in Synthemis corresponding to the 

 original form for Cordulephya, that of Metathernis being 

 similar to that of ('. montana, and that of Choristhemis 

 similar to that of ('. jnjynuea. Secondly, in the females of 

 the two species, the vulvar lamina is more reduced in C . pyg- 

 mcra, leaving the underside of segment 9 widely open, so 

 that larger egg-masses can be more easily exuded. Bearing 

 in mind the extreme rapidity and timidity of the movements 

 of the female during oviposition, the advantage of this to 

 C pyymaa is evident. 



E. The comparative sizes of the larvce. — Though the imagi- 

 nes of C. pyynum and C. montana do not differ very greatly 



