380 STUDIES IN THE LIFE-HISTORIES OF AUSTRALIAN ODONATA, 



Note on oviposition. — The method of oviposition in the tissues 

 of reeds under water is not the only way employed by the species. 

 During a visit to Leura, on December 5th, 1908, I distinctly saw 

 a female alight on the almost vertical face of a mossy rock, over 

 which the stream was trickling, and insert her ovipositor at least 

 a dozen times into the moss-tissues. It then flew off up stream 

 and repeated the operation further on. 



In his work on the life-histories of American ZygojHera^ Prof. 

 J. B. Needham gives, for classification-purposes, the following 

 differences between CalojJterygid and Agi'ionid n3'mphs : — 



** Basal segment of antennae very large, as long as the other six together, 

 median lobe of labium ivith a very deep cleft] gills thick, the lateral ones 

 triquetral , Calopterygidcu.^^ 



"Basal segments of antenneo not longer than succeeding segments; 

 labium with a very shallow closed median cleft, or no cleft at all; gills thin, 

 lamelliform AgrionidcB.'^ 



This classification was adopted on the knowledge afforded by 

 the study of the aymphs of Calopteryx and Hetcerina only. It 

 will be seen at once that the nymph of Diphlebia differs in 

 important respects from those of these two genera. 



Firstly, as regards the antennae. In Diphlebia, it is the second 

 joint which is greatly elongated, the basal joint being short and 

 thick. The elongation of this joint, however, is not so great as 

 that of the basal joint of CalajHeryx. I am inclined to regard 

 elongated antennae in dragonfly nymphs as a primitive character. 

 The reduction in length has been carried to a great extreme in 

 the imagines, and one can see no reason at all, even in the 

 nyniphal stages, for the continuance of elongated antennae. In 

 all burrowing nymphs they have become very short, and in the 

 Gor)ip}iincB the number of joints is actually reduced to four. The 

 possession of long antennae may then merely mark out the more 

 primitive families, but it will also mark isolated groups which 

 may have been placed outside of these families for more import- 

 ant structural reasons. I shall show, in a later part of this 

 paper, that the long basal joint is possessed by at least one Agrionid 

 found commonly in parts of Australia. This suggests that it 



