BY R. J. TILLYARD. 185 



In Sipitheinis, the hindwing of the larva outgrows tlic forcwing 

 very iiiucli more than in other genera. This may be due to the 

 diverging position of the wing-cases, which do not He parallel 

 along the back of the insect, but resemble those of Cordulegaster. 

 Such a position must be regarded as a speciaHsation connected 

 with the form and habits of the larvie. In a nearly full-grown 

 nymph of *S'. macrostignia, T found the forewing to be oidi/ slightly 

 more than half a.s long as the hind wing. 



Having now completed the study of the anal loop, let us turn 

 our attention to the Gomphime, where no loop is formed, and 

 study the anal area there. The only genus available to me was 

 Austrogomphjis. There are, however, two distinct typ^s of larvae 

 included in this genus. A. ochi^aceus hJelys, and A. hete^'oclitus 

 Selys, represent these two types. As will be seen from Figs. 15- 

 16, the tracheation of the anal region shows considerable differ- 

 ences in the two types. 



In ^. ochraceus {Fig.lD), the anal trachea, at the end of Ac, 

 branches into three. The branch running basad is A^, and from 

 it the secondary anal vein A' is developed as a bridge- vein. The 

 branch descending directly from Ac is almost certainly A.,, since 

 it forms the distal side of the anal triangle in the male. The 

 branch running distad represents A^,.,, but does not reach Cu.^ at 

 any point. This branch divides into two rather low down. 

 These two branches inay be A., and Aj respectively, but their 

 determination is doubtful. 



In A. heteroclitus (Fig 16), the anal trachea, at the end of Ac, 

 branches as usual into two. 'I'he branch running basad is A3.4. 

 A;; descends from A4 somewhat basad from Ac, and forms the 

 distal side of the anal triangle of the male. This distal side is 

 strotigly bent in the imago (Fig. 18), and slants away from the 

 anal vein, instead of descending perpendicularly to it as in A. 

 ochraceus. The branch running distad from Ac is Aj.o. At first 

 it curves downwards parallel to Cu.^. It then gives off a strong 

 branch descendin^^ straight to the wing-border; tiiis is almost 

 certainly A2. From the point of origin of A.^, the more distal 

 branch Aj bends again towards (.'Uo; but, having britlged only 

 half the distance separating it from Cu2, it again turns rather 



