BY R. J. TILLYARD. 223 



to move the three most basal branches of M closer together. 

 Thus the Calopterygidce and Agrionidce have a common origin in 

 forms in which these three branches lay further apart than in 

 recent Caloptei^ygidce, but not so far apart as in recent Agrionidce. 

 The reduction of the prenodal area in the latter family is clearly 

 correlated with the extended condition of the medial branches. 

 For, obviously, three branches are not needed to support so short 

 a prenodal area. As regards the Lestidce^ I find that the most 

 distal branches of M arise by a common stem in all the last four 

 instars of Austrolestes, and in all instars of Synlestes (three) that 

 I have examined. Thus, I conclude that this is a primitive con- 

 dition, probably independently inherited, and not brought about 

 (as Dr. Kis suggested) by the movement distad of one branch so 

 as to become a "tributary" of the other. Support for this view 

 is forthcoming in the fact that the Liassic Heterophlehia shows 

 this condition fully formed, as in our recent Synlestes, The same 

 fossil supports the view that, in Lestidct, when the prenodal area 

 became reduced, the second branch of the media (counting from 

 the base) travelled basad away from the nodus, and took up a 

 position similar to that seen in Calopterygidce. If that be so, 

 this particular character was independently evolved in these two 

 families, and cannot serve to indicate a close relationship between 

 them. The totality of evidence tends to bring the Calopterygidce 

 and Agrionidce close together (the common character of the 

 absence of a long brndge is very important), and to separate the 

 Lestidce out as a very distinct group of independent origin. 



In now definitely rejecting Needham's explanation concerning 

 the shifting of Rs over to M, I must lay the greatest stress on 

 the ontogenetic evidence. If such a change as Needham pictured 

 had ever taken place, it is almost inconceivable that not the 

 slightest record of it should now be left in the ontogenetic stages of 

 the larval wing. On the other hand, if such evidence can be 

 found, the new theory must give way to Needham's original sup- 

 position, or some other theory in accord with the ontogenetic 

 evidence. 



If we now accept the Theory of the Unbranched Radius, four 

 main consequences ai'ise ; — 



