BY B. J. T1LLYARD. 



695 



or two branches; and li> retains the terminal dichotomic forkings 

 of R 2 an(1 ^3, though generally (but not always) the original 

 dichotomy of Ilt + 5 ls suppressed. 



We must now contrast this with the somewhat different type 

 shown in Protohermes (Text-fig. 106), claimed by van der 

 Weele to be the most archaic type within the Order. 



^Tc^trTM^O 1 ' * 4t5 



Text-Fig. 105. 

 Wings of Archichauliodes gut t if cms (Walk.), (fam. Corydalidae) , 

 (x 2-7). Lettering as ou p. 535. 



The first thing to be noted in Protohermes is the clear evi- 

 dence of addition of extra branches to the main veins, on the 

 Plannipennian plan, from tl>e tip inwards. This is best seen 

 in the case of Cu ls which lias become transformed from its 

 original dichotomic condition into a pectinate vein, on the lines 

 indicated in Text-fig 1 . 40. The same thing has happened with 

 M, although it seems also quite clear that this vein possessed at 

 least four branches when the additions began, seeing that both 

 Mi 1 2 and M3 + 4 fork dichotomically at about the same level as 

 the original forking of Cu 1( Some of the branches of Rs have 

 evidently been added from the tip backwards also; but thei'e is 

 certainly no evidence to restrict the original number of branches 

 of this vein to four. 



