BY K. J. TILLYARD. 755 



veins between So, R, and Rs may also be correlated with the 

 development of the vena triplica, as an additional strengthening 

 for the midrib. 



Let us now study the so-called anastomosis more fully. In 

 Text-fig. 3, A and B, I have contrasted the pupal tracheation 

 with the imaginal venation in this region. It will be seen at 

 once that the actual termination of the vena triplica is not 

 accomplished by means of an anastomosis between any of the 

 main veins, but by the interpolation of the two cross-veins, x^ and 

 .x'2, together with the bending-down of the portions of Sc and 

 Ri preceding them. We must contrast this strongly with the 

 condition to be seen in the Osmylidce, where Sc and Rj approach 

 one another distally at a very slight angle, and then fuse together, 

 without the iiitervention of an auxiliary cross-vein, for a consider- 

 able distance. There is then, in the Osmylidcn, a true anasto- 

 mosis between these two veins; whereas, in the Psychopsidce, 



l-i 



M; 



■Ml 



Text-fig. 4. 

 Types of anastomosis between IVI and Cu in Psychojjxis. a, usual type in 

 Ps. mimica Newman, forewing; h, usual type in Ps. elegans (Guerin), 

 forewing. In Ps. i)iso/e)is McLach., there is usuallj^ no anastomosis in 

 this region at all. Lettering as in Text-fig. 1. 

 there is no true anastomosis, and it would appear to be somewhat 

 misleading to term it so, as Navas has done. However, the term 

 has already been applied, in Perlaria and Trichoptera, to junc- 

 tions between main veins, brought about by means of cross- veins 

 in line with one another; so that we may allow it to remain, 

 provided that we understand that we are speaking of a junction 

 of this type, and not a true anastomosis between two or more 

 main veins. 



