ko. 



1331. 



DRAGON-FLY WING VENATION— NEEDHAAL 



731 



Bridge weakly or incompletely devel- 

 oped; oblique vein remote from the sub- 

 nodus. 



All principal veins straight or gently 

 curved and evenly forking. 



Media at the top of the arculus. 



Veins J/j-s and M^ at their de^iarture 

 from the arculus, separate and straight. 



Triangle, supertriangle, subtriangle, 

 etc., ordinary quadrangular areoles (per- 

 haps traversed by weak cross veins) . 



Cubitus somewhat symmetrically 

 forked. 



Anal angle of the wing unsupported. 



Cross venation dense, irregular, and 

 inconstant. 



Fore and hind wings alike. 



Bridge becoming strong and directly 

 attached to vein lfi+.^; oblique vein, re- 

 tracted toward the subnodus. 



Some principal veins becoming strongly 

 angulate at points of special l)racing. 



Media descending the arculus. 

 •Veins M^-^ and M^ becoming fu.^cd, or 

 strongly arched upward, or both. 



Triangle, etc., becoming strong and 

 highly differentiated inclosnres. 



Cubital fork becoming strongly uni- 

 lateral. 



Anal angle becoming supported by the 

 junction of veins C«2 and ^1, and, some- 

 times, by the development of an anal 

 loop. 



Cross veins becoming fewer, stronger, 

 more definite and regular, and the mem- 

 brane, tliinner. 



Fore and hind wings differentiating (1) 

 by following parallel paths with unequal 

 speed, and (2) by following different 

 paths. 



IV. LINES OF SPECIALIZATION. 



Hitherto we have been discussing wing characters more as individual 

 entities than as mutuallj^ dependent parts of a single organ. The 

 ilhistrations of the steps in the development of each, drawn from adult 

 wings, have been selected arbitrarily, and have not always been drawn 

 from a single line of development. The}^ have served the pvirpose of 

 illustrating in a general way the progressive modifications of each part, 

 confirming the ontogenetic record. In their application to this end 

 we have necessarily overlooked the lesser individual peculiarities of 

 each. Correlated characters varying independently preserve, some 

 here, some there, bits of the ancestral record, but with more or less of 

 individual alteration of it. It is probable that eveiy one of the char- 

 acters discussed in this paper would be found on closer study to pos- 

 sess distinctive features in each genus — earmarks of the genus. This, 

 of course, applies not to wing parts alone, but to every other part as 

 well. 



We come now to consider these same characters in their ensemble. 

 Their individual records, of cour.se, do not agree. Did they agree, we 

 should have a single lineal series, very well adapted to book making. 

 We should have a wing exhibiting the generalized characters just 

 mentioned with which to begin the series. But while it has been easy 

 to show by concurrent ontogeny and comparative morpholog}' that 



