IN THE LARV^ OF ANISOPTERID DRAGONFLIES. 183 



in each o£ which a number of complete capillary loops can be clearly seen. 

 The epithelium of the gills appears to be of a slight yellowish tinge. 



The study o£ the two pairs of lateral gills is not easy, as it is difficult 

 to persuade the larva to rest in suitable positions. However, I was able to 

 make camera-lucida drawings o£ a larva^ not only from the dorsal and ventral 

 aspects (text-figs. 17, 18), but also in profile (text-fig. 19). The actual 

 distribution of tracheae in the gill-basket was most carefully studied in the 

 case of dorsal and ventral gills, so that I will deal with these first. 



From the dorsal aspect (text-fig. 17) the dorsal gill-fold is seen to be a 

 long undulating strand of tissue lying in the middle line. From the main 

 dorsal trunk (DT) on each side, six efferent tracheae come off slantingly and 

 run towards the gill-fold. Just before reaching it, each eff'erent trachea 

 divides into several secondary branches. These branches pass into the gill 

 vertically downwards in alternate sets to right and left. Hence arises that 

 slight undulatory outline of the base-line of the gill-fold ; for each set of 

 trachese may be supposed to exert a very slight pull on the gill towards one 

 side^ so that each set enters the gill by a slight convexity on its own side. 

 It will thus be seen that the system of gills present in the larva of Anaa^ 

 soon after hatching is very different from that seen in the older larva. It 

 is, in fact, a primitive Simplex Gill-system of Undulate Type^ very similar to 

 that found in the well-grown larvse of Petalura, Corduleg aster, and Austro- 

 gomphus, but without any system of cross-folds. This fact fixes the 

 Undulate Simplex type of gill definitely as the most primitive of all the types 

 studied in this paper. 



The last of the six efiferent tracheee (pmb) on each side of the dorsal gill 

 is longer and thicker than the preceding five. Instead of breaking up into 

 a few secondary branches, it turns to run posteriad alongside the gill for 

 some considerable distance, giving off at short intervals some eight or nine 

 secondary branches. Thus these last two efferent tracheee come to supply 

 between them nearly one-third of the whole length of the gill. By referrino- 

 to text-fig. 17, it will be seen that these tracheae, which I have designated 

 the postero-median hrancJies {pmb) of the main dorsal trunk, correspond 

 exactly with the strong trunks of the same name which supply the narrow 

 posterior portion of the gill-basket in the full-grown larva of ^"Eschna 

 hrevistyla. 



Turning now to the ventral gill-fold (text-fig. 18), we notice a very similar 

 arrangement of efferent tracheae supplying the gill alternately from left to 

 right. There are, however, some important differences. Corresponding to 

 the greater length of this gill-fold, we find altogether eiglit efferent trachea 

 on each side. Of these, the first four on each side arise from the corre- 

 sponding visceral trunk (VT). Just posterior to the fourth, however, the 

 visceral trunk bends round upwards to enter the oblique branch (o) of the 

 main dorsal trunk at the point which I have called the first junction (ji). 



I'd* 



