0.1331. DRAGON-FLY WING VENATION— NEEDHAM. 735 



Doubtless, the most generalized group within the Aeschnine series 

 |S composed of Ci/matophleh!a (fossil), and Goinj)hvi<chna (Plate 

 j?CXXVII, tig-. 1), which have the triangle little elongated, all .supple- 

 inents but little developed, the anal loop small and compact, with no 

 uppleniental loop, and the anal angle of the hind wing, therefore, not 

 ^reatl}^ dilated, but which always have the stigma braced against vein 

 t^T/j opposite a forward bond in vein M.^. These all exhibit affinities 

 vith the preceding and the next following groups. Even the upward 

 )end of vein J/j at the stigma is hardly more marked in Gomjjhseschna 

 ;-han in Pkyllopetalia (lig. 26). 



Passing by these two genera, we next come upon a number of 

 )thers which have a more pronounced /E.schnine aspect, as Boyeria 

 Plate XXXVII, fig. 3), Basiseschna (Plate XXXVII, fig. 2), and 

 HoploniVKchna (Plate XXXVIII, fig. 1). These have vein J/, more 

 strongly bent toward the stigma, the triangles longer, radial and 

 'nedian supplements well developed, and trigonal supplement and sup- 

 plementary anal loop developing. In these genera the radial sector is 

 lot forked. 



In uEscJina and its nearest allies (Plate XL) we see the further 

 [progress of these tendencies. There is this added feature — the radial 

 ?ector has become forked. It will i)e observed that the anterior 

 branch of this fork is separated from vein M^ by a single row of cells, 

 ^md that in the same place in Basia^schna (Plate XXXVIl, fig. 2) there 

 IS a line of cross veins tending to straighten out. The anterior liranch 

 bf the fork is developed out of this line of cross veins. In the Aus- 

 tralian ^EseJina hrevistyJa (not figured herein) all stages of its com- 

 pleteness and incompleteness may be found in a series of specimens. 

 In the nj^mphal wing of Anaxjunhis^ a strong trachea is seen to pre- 

 cede only the posterior branch of the fork. Thus the anterior branch 

 of the fork is clearly a relativel}^ recent acquisition of such genera as 

 ^c/uid ,( J yn(fca ti f ha and Atia^i\ the dominant genera of the subfamih^ 



In tracing this relatively simple developmental line that leads to the 

 highest specialization of this kind to bo found, we have purposely passed 

 by several lateral offshoots. The one of these which seems nearest 

 dlied to yEschmi is represented by Neura^sclnKt and Staurophleb'ia 

 (Plate XXXIX, fig. 2), which, having retained the tip of the subcosta 

 which extended beyond the nodus and many cross veins in important 

 spaces, are not to be derived directly from forms in which those have 

 been lost; lacking a brace to the stigma, can not have sprung directly 

 from such genera as those of the Boyeria group in which this useful 

 >tructui-e is well developed, l)ut may well have sprung from the com- 

 mon ancestor of Boyeria and ^EscJma. 



Furthermore, Anax (Plate XL, fig. 3), while very highly speciali/ed, 

 has retained the most primitive position of the media at the arcuhis to 

 1)6 found in the group, and has the hind angle of the liind wing rounded, 



