ki33i. DRAGON-FLY WING VENATION— NEEDHAM. 737 



blaced as to cause vein 3£^ to appear forked (see ^. calif arnica, Plate 

 XL, fig-. 1), when they appear as the upper branch of that fork; finally 

 ihey appear as the true continuation of the vein. As the anal vein is 

 fwitched upon the second cubito-anal cross vein in the Libelulidae, 

 so here vein J/^ is switched into an entirely new position upon two 

 i^ross veins. 



It is worthy of passing remark that this shifting is an accompani- 

 •nent of differentiation between longitudinal veins in very highl}^ 

 ,;pecialized ^Eschnina?. This will be much more clearly seen in actual 

 wings than in the best of figures. A glance at such wings as those of 

 Anax jn?iius or ^Eschna californ'ira will discover that Ra is strong, 

 M^ is weak; M^ is strong, J/, is weak; Cu^ is strong, Cu^ is weak. 

 The strong vein bounds posteriorly the space in which the weaker one 

 ies. The cross veins just discussed, which join vein J/j so solidly to 

 rein J/,, together with several lines of cross veins descending from the 

 j>itrongest part of the radial sector, complete the boundary of the space 

 n which the weakest part of vein J/j is included. 



; But to return to our theme, we have seen that the ])ehavior of radial 

 pector and supplement, of trigonal supplement, and the distal ends of 

 reins J/, and J/^ furnish cunmlative testimony to two divergent lines 

 :)f development, which, starting with forms a little less ^lilschna-like 

 than Boyeria, have e^'olved the groups of Brachytron and ^Enchna. 

 And we have shown that, with respect to venation at least, the groups 

 of Neurx^clina^ A/ia,/', and OontphcWc/uiaavG successively more remote. 



GompMnse. — In this group we come upon another type of vEschnid 

 venation, a type which is chiefly characterized by the similar shifting 

 of the anal vein in both wings, but only to such extent that that vein 

 tvppears more or less symmetrically forked. There are few venational 

 •specialties. Neither the shifting of veins for advantage of position 

 jtior the reduction of cross veins have proceeded very far. That vena- 

 tion is at a standstill is indicated by the unusual constancy of mechan- 

 ically unimportant cross veins, such, for instance, as the one traversing 

 ithe supertriangle. Variation from the type is slight, considering the 

 large luimber of genera in the group; and such as there is, it does not 

 lend itself readily to . serial arrangement. The very moderately 

 widened anal angle of the hind wing is supported mainly by straight 

 and parallel anal and cul)ital Ijranches, as in the more generalized 

 ^Eschninfe. An anal loop is developed in many genera. ])ut has so 

 different aspect in several of them it seems quite ])ossil)le it may have 

 ibeen developed several times independentlv. (See Lanthiis^ Plate 

 XXXV, fig. 3; CyclophyUa, Plate XXXlV,'fig. Land Hagenhix, tig. 

 2.3.) A strong trigonal supplement is developed in Jlageinux. The 

 declined portion of the cubitus (inner side of the triangle) becomes 

 elongated in the fore wing more than in the hind in CyvlojthyUa and 

 ApJiylla (Plate XXXIII, fig. 3). Accessory sectors of unusual strength 



