IN THE LARV^ OF ANISOPTKRID DRAGONFLIES. 179 



Odonate larvse. In fact, I have already seen it in several genera o£ Zygoptera 

 as well as in ^schna. Now we have already shown that the main gill-folds 

 of the Simplex System are homologous with the " rectal glands.'"' Their 

 hypobranchial tissue, therefore, corresponds in position with the supporting 

 tissue of the " glands." 



The differences between the hypobranchial tissue and the supporting tissue 

 of the glands are obvious enough. The former stains with eosin, the latter 

 does not. After treatment with soap-alcohol, the former is unaffected, the 

 latter has its contents extracted and shows up as a reticulated network of 

 more or less empty spaces, much resembling the " skeleton " of the fat-body 

 obtained by the same reagent. Thus, while the supporting tissue is closely 

 analogous to true adipose tissue, and probably contains considerable stores of 

 fat, the hypobranchial tissue is of a very different nature, and the contents 

 of its lobules are impervious to the action of soap-alcohol. This suggests 

 that the nature of the tissue has become chemically altered by the assump- 

 tion of some new function connected with respiration, possibly the extraction 

 of nitrogen from the gas passing along the tracliese. That question, how- 

 ever, cannot be decided here, but must be left for future physiological 

 investigations. 



I have figured the hypobranchial tissue (liy) in the gills of Dendroa^sclma 

 (text-fig. 16), Austrogomphus (Plate 22. fig. 25), ^Esdina (Plate 22. fig. 2Q), 

 and Cordulephya (Plate 22. fig. 27). In Uendroceschna it is, like the basal 

 pad, very narrow in cross-section. In Austrogomphus it usually fills uj) the 

 hypobranchial space fairly completely ; while in ^dEscJma, where its mass is 

 greater and more irregular, it often leaves a considerable space both above 

 and below. In the Libellulidse it may appear as a single row of lobules 

 (possibly separate cells), or may be more irregular, with two or more layers 

 of lobules in some cross-sections. 



Ontogeny of the Gills in the Larv^ of ^schna and Anax. 

 (Text-figs. 17-20.) 



The development of the gills in the rectum of the growing larvse of 

 yEsclma and Anax is a process of very great interest, and supplies the 

 necessary evidence for linking up the structure of the Simplex and Duplex 

 Systems. 



A number of eggs of Anax papuensis were obtained by me in February 

 last. Those that hatched out were taken on to the different larval instars as 

 far as the fifth, two or three being killed immediately after ecdysis for the 

 purposes of sectioning. Unfortunately the supply was not large enough to 

 carry the investigation beyond the fifth instar. But the deficiency was here 

 supplied by the discovery of a number of larvse of ^:EscJma hrevistyla in very 

 early stages, easily seen to correspond exactly with the fourth and fifth 



