Til] THE WINGS - 59 



resulting in a widening of it, but also tended to pull it obliquely 

 analwards, so that the cross-veins took on a slant. An obvious 

 danger of "buckling" was avoided by the strengthening of one of 

 the cross-veins near the middle of the cell (fig. 20 c, x). Thus the 

 discoidal cell became divided into two distinct parts, which thence 

 onward underwent different developments. By a slight bending 

 of the lower side at the junction with the thickened cross- vein, 

 and by a slight movement distad along M 4 of the upper end of 

 the same cross-vein, we arrive at what may be termed the Tetra- 

 themis-stage 1 of development (fig. 20 d; cf. fig. 130). The dis- 

 coidal cell now consists of (a) a more basal and upper portion, the 

 supra-triangle (s), and (b) a more distal and lower portion, the 

 triangle (t). Both these ar,eas are still quadrilateral in form. 

 From this stage, a further increase in the bend or break of C, 

 together with a further elongation of the separating cross-vein 

 so that its distal end reaches the upper distal angle of the discoidal 

 cell, and its direction becomes nearly longitudinal to the wing-axis 

 brings us to the Gomphine-stage (fig. 20 e, f; cf. figs. 117, 118). 

 Here the " triangle " becomes for the first time definitely triangular, 

 by the loss of the original costal side, while the supra-triangle 

 assumes the elongated and nearly triangular form seen in most 

 recent Anisoptera. 



From this point two main diverging lines of development start 

 off, one leading to the Aesckninae, the other to the Libellulidae. 

 In the former, fore and hind- wings followed closely similar lines of 

 evolution, the tendency in each case being towards an elongation of 

 the triangle in a direction parallel to the wing-axis (Plate II, fig. A}. 

 In the Libellulidae, however, opposite tendencies set in, in fore and 

 hind-wings. The base of the former became narrower, that of the 

 latter wider. The narrowing of the fore-wing tended to shut the 

 triangle up like a collapsed framework, and at the same time drove 

 it further away from the wing-base (fig. 20 </ 2 ). As the final 

 result of this process, we reach the Tramea-stage, where the triangle 

 of the fore-wing becomes excessively narrowed, elongated trans- 

 versely to the wing-axis, and far removed from the arculus. The 



1 The name does not imply an actual belief that Tetrathemis as we know it 

 stopped short at this stage. As the genus is highly specialized in many points, the 

 condition of its discoidal cell may be due to reduction from a more advanced type. 



