gio THE AMERICAN NATURALIST. [Vou XXXII. 
instance, at present in Cordulegaster). In the fore wing the 
anal vein has come to connect with the antero-internal angle 
of the triangle through the deflection of the second cubito-anal 
cross-vein, and the triangle has proceeded farther from the arcu- 
lus. Successive steps are shown in the accompanying diagram 
(Fig. 66). In the hind wing the triangle has receded to the 
level of the arculus, or even a little farther, by the easy stages 
shown in the accompanying diagram (Fig. 67), and the second 
cubito-anal cross-vein has atrophied. 
The Anal Loop. — There is also in the Anisoptera a strong 
tendency toward the development in the hind wing of a broadly 
expanded anal area — an aéroplane. + This region remains still 
Fic. 68. — Forms of the anal loop in the Anisoptera: 1, anal loop of Cyclophylla diphylla ; 
2, of Gomphoides stigmatus; 3, of Gompheschna furcillata; 4, of Gomph mia 
paradoxa ; 5, of Syncordulia gracilis ; 6, of Agrionoptera insignis ; 7, of () Nannophya 
maculosa ; 8, of Ephidatia longipes ; 9, of Hydrobasil traneus. 
unexplored territory. It will furnish, however, at least one 
character of much systematic importance. This is a space 
included between the first and second principal branches of 
the anal vein, which we designate as the anal loop. Its develop- 
opment is shown in Fig. 61. When developed in the Æschni- 
dæ as a distinct enclosure, it is always compact in form, but in 
the more specialized of the Libellulidæ it becomes elongate, 
then gland-shaped, and then foot-shaped. Fig. 68 shows in 
more characteristic forms, and gives an idea of its variability 
within the group. i 
We have now indicated the homologies of the principal 
veins: we have briefly discussed the development of a few of 
the distinctive venational characters of this interesting group ; 
