332 STEPHEN G. RICH 



postdorsal trachea. The three primary branches divide at once, 

 as in the form referred to. The last branch is unusually large; 

 the three branches are closer together than in any other member 

 of the family and are separated far from the postdorsal trachea; 

 in this the condition approaches that found in the Libellulidae. 

 Figure 46 indicates this. The branching on the ventral side of 

 the rectum is more highly specialized than in any other form 

 in either family. The postventral trachea, the ten branches, and 

 the trachea passing to the Malpighian tubules at the cephalic 

 end of the small intestine, are all fused together for a short 

 distance from the point where they leave the ventral trunk. 

 They then separate suddenly; there is no dichotomy, but this 

 big trachea suddenly breaks up into a fan of radiating branches, 

 the most caudal of which is the postdorsal and the most cephalic 

 of which passes to the Malpighian tubules. Even the ventral 

 tracheation of Pantala flavescens, the extreme form of the Libel- 

 lulidae, does not equal this. Figure 47 indicates the structures 

 described. 



The gill bases are in the usual six double rows, three on the 

 ventral and three on the dorsal side. The gill bases are oval, 

 set at an angle of only 20° from the transverse position. There 

 are fifteen of the bases in each single row. 



The interior of the rectum of Lanthus represents the greatest 

 perfection of the longitudinal fold as a respiratory mechanism. 

 The six longitudinal folds, each in the middle of a double row 

 and homologous with those of the Aeschninae, are exceedingly 

 high in proportion to the size of the rectum. They are smooth 

 on the edge; then- tracheation is as in the aeschnid fonns that 

 I have described. The buttress folds are exceedingly small and 

 have only a slight connection at their proximal corner with the 

 longitudinal fold. Their shape — elongate and pointed — is not 

 unlike that of some libellulid gills. I was not able to detect 

 the fat in their bases save from the outside of the rectum; and 

 hence I cannot figure its extent. There appear to be the usual 

 cushions on each side of the buttress fold. There are fifteen of 

 these folds in each row, and the adjacent rows alternate in the 

 usual manner. The longitudinal folds are smoothly rounded at 

 each end of the rectum (fig. 48). 



