BY R. J. TILLYARD. 



173 



forms of ana] loop. Without attempting to homologise the cor- 

 responding parts of the jEschnine and Italian loops, it was a 

 protest against the assumption of a double origin for two essen- 

 tially similar formations; and, in particular, it was a special 

 protest against the assumption that the narrow-winged forms, 

 such as Tetratheinis, Cordulephya^ Afjrior/omph^is, lay anyvvliere 

 close alongside the main line of descent of the Anisojdera. In 

 brief, I considered all such nari'ow-winged forms to be highly 

 specialised asthenogenetic offshoots from the main stem, while the 

 main line of advance had consisted of bi'oad-wim/ed forms from 

 the very beginning; and hence, there could be only one origin 

 for all kinds of anal loop. The notation that agrees with this 

 theory is given in Fig. 4b. 



Fh 



Fig.4. — Scapanea frontalis Buin\.,S. Diagram to show tlie notation of 

 the anal loop. A, as given by Ris; B, as worked out in this paper. 

 Compare Figs. 11-12. 



These two theories are, in this paper, put to the crucial test by 

 a careful study of the tracheation of the developing anal loop. 

 Putting aside the Gomjyhince, i)i most of which no anal loop is 

 formed, the twelve other genera studied by me can be arranged 

 in four distinct groups or stages, which show a distinct pliylo- 

 genetic connection as regards their anal loops, and may be 

 arranged in ascending order of development, as follows : — 



