IN THE LARV^ OF ANISOPTERID DRAGONFLIES. 187 



(projecting into the rectal cavity in a slanting direction, somewhat posteriad) 

 a small but very distinct oval leaflet filled vidth tiny tracheal loops. These 

 leaflets are unmistakably similar both in appearance and position to the 

 lamellce o£ the Libellulidse. They also correspond fairly closely to the cross- 

 folds in the Simplex St/stem of gill-folds. 



With these leaflets in front of us, we stand at the parting of the ways- 

 between the Undulate Simplex and the Lamellate Duplex Tjpes. If the 

 leaflets, in their further development, joined up to the main fold as a series 

 of supports, we should obtain the Undulate Simplex type of gill-fold seen in 

 Petalura, Cordidegaster, and Austrogomphus. If, on the other hand, the 

 main folds became aborted, and the leaflets developed into lamellae of con- 

 siderable size, we should then have the Lamellate Duplex type seen in the 

 Libellulidse. 



In order to study the formation of the gill-basket at the fourth instar 

 more fully, several larvse of y^sclina and Anax were taken immediately after 

 ecdysis and prepared for sectioning. Text-fig. 1 shows a cross-section 

 through a larva of JEsclina hrevistyla at about the middle of the gill-basket. 

 Owing to the semi-collapsed state of the main folds, especially on the 

 ventral side, the undulations are flattened down, and several are cut through 

 in the same section. This accounts for the zigzag course of the sections 

 across these main folds. 



Between each pair of main folds two small triangular prominences (c/) 

 are to be seen, appearing as slight evaginations of the rectal epithelium. 

 Their size depends, of course, on whether the particular section studied cuts 

 the leaflets through their middle or near one of their edges. The largest 

 cross-section of a leaflet seen in this series of sections was somewhat larger 

 than the two marked c/in the figure. 



A careful examination of sections of a larva in the third instar shows that 

 these leaflets really begin to develop at that stage as tiny outpushings of the 

 rectal epithelium, between, and fairly close up to, the main folds. They are 

 not, however, large enough to be noticed in an examination of the living 

 larva. Sections of larvse in the second instar, on the other hand, show no 

 sign of these outpushings. The only other noticeable differences between 

 sections taken from larvse of the second, third, and fourth instars are the 

 gradual increase in definiteness and depth of the undulations in the main 

 folds and the rapid increase in size in the larger tracheae. 



From the fourth instar onwards the small leaflets rapidly increase in size, 

 while numerous tracheoles are seen to ramify in them. On the other hand, 

 the trachese entering the main folds become inconspicuous and are soon lost 

 to view, in the live larva, in the mass of yellowish-brown epithelium which 

 seems to be raised up along the course of the original main folds during the 

 rapid growth of the leaflets. At the sixth instar these leaflets are very 



