BY R. J. TILLYARD. l99 



tion can be detected between Rs and R, but Rs in all these cases 

 appears as a branch of M. It can generally be detected by its 

 peculiar method of parting from M, which is at a considerably 

 greater angle than that made by a true branch of M, such as 

 Mg (Plate xii., fig.l). This distinct curve in Rs, on leaving M, 

 is most probably due to the original manner in which it cut 

 across and under Mj and M.^, before it became permanently 

 hitched on to Mi.o. As Needham has justly observed, there can 

 be no difficulty in understanding this cutting-off of Rs from R, 

 and its subsequent permanent attachment to M. For with the 

 setting-in of the asthenogenetic process, and the consecjuent nar. 

 rowing of the wing-rudiment in width, any difference of level 

 which originally existed between M and R (and the fact that Rs 

 still passes undey^ Mj.^ in Anisoptera shows that there was once a 

 difference in the level) must inevitably become lessened. Thus 

 the trachea Rs must gradually become pressed near its base by 

 the stronger overlying M1.2, and if it could not effect a union with 

 the latter by the abortion of its original base, it would inevitably 

 perish for lack of oxygen. 



Let us now examine the condition seen in the Protoneurini. 

 The arrangement of tracheae in the region below the nodus seems 

 to be the same in Isosticta (Plate xii., figs. 11-12) as it is in Cali- 

 agrion. Two tracheae branch off from Mj.o close together. One 

 would naturally suppose that the more proximal of these is M3, 

 and the more distal is Rs, as labelled in the plate (with a query). 

 But if we turn to the closely allied genus Neosticta (Plate xii., 

 figs. 9-10), we find Rs actually descending from R below the 

 nodus, and crossing both M1.2 and Mo. Having done so, it runs 

 along only for a very short distance, and then ends up; so that 

 the rest of the imaginal vein Rs is not found about a trachea at 

 all. 



It seems very probable that we have, in Neosticta, the condition 

 described above, viz., that Rs has failed to coiniect basally on to 

 M, and is in process of perishing for lack of oxygen, owing to 

 pressure from the overlying branches of M. If this is really so, 

 then we might well consider that Isosticta shows the next step 

 in advance, viz., that Rs has attained a basal fusion with M, 



