37-4 AUSTRALIAN OORDULIINiE, 



Rivei's not a specimen can be seen; indeed the Odonate fauna of 

 these streams, and of the beautiful Derwent River in the south, 

 is almost completely annihilated. The following analysis of the 

 contents of the stomach of a fine trout (about 3 lbs.) taken by me 

 in the Macquarie River may be of interest — a portion of a single 

 beetle, and thirty-seven undigested dragonfly heads, of which no 

 less than thirty-ttvo were recognisable as those of Procordidia jack- 

 soniensis, the other five being apparently Austroyomjihus giierini. 

 While trout-fishing, I carefully observed this species, and noticed 

 that it is very fond of skimming the surface of the water, like a 

 Lihelluline. This is a fatal habit, for I saw, several times, a 

 trout seize the insect while in rapid flight close to the water. I 

 also noticed that it often dips itself in the water, and then rises 

 and hovers for a short while. On one such occasion, a kingfisher 

 dashed down and seized the insect. 



One might venture the prediction that, in a veiy few years' 

 time, this insect will be confined to the lagoons, and that it will 

 become a dull-coloured uninteresting-looking species of more 

 robust build than at present, the supply of food on the lagoons 

 being more abundant. Compared with its congener, P. ajfinis 

 8elys, of Western Australia, which is only found on the running 

 brooks, it is already of a thick-set and more robust habit, and of 

 a much duller colour on the abdomen. 



I am of opinion that the two species, P. jackso7iiensis and P. 

 ajfinis, are truly geminate species, portions of an original single 

 stock whicli became separated by the great Desert Barrier. P. 

 ajfinis has followed the line of development of ffejuicordnlia 

 anstralice, which species it exactly resembles in flight and colour- 

 ation; while P. jacksoniensis, though still retaining the metallic 

 front, is following the line of development of //. tau. 



The three genera, Somafochlora, Procoj-dulia, and Hemicordulia, 

 are of great interest to students of zoogeographical distribution. 

 Somatochlora is evidently the most archaic, and is one of the few 

 genera of known bipolar distribution. In the subarctic regions 

 of Canada and Siberia, it has attained its greatest development. 

 But alongside that development, we find it represented in the 

 subantarctic regions of New Zealand and Chili, by a small 



