674 ON SOME EXPERIMENTS WITH DRAGONFLY LARViE, 



applied, which amounted to a severe drought of ten weeks' 

 length, without even the help of a single fall of dew through 

 the whole period, is, I think, sufficient to prove that the memV)ers 

 of this genus can easily survive the driest conditions they are 

 ever likely to meet with in their natural habitat. This is of 

 great interest, since I have noticed that most dragonfly larvje, 

 such as nearly all the ^schnidoi and Agrionidce, with not a few 

 of the LibeUulidcn, perish as soon as the pond or river containing 

 them dries up. It may go far to explain the peculiar success of 

 the Sy^Uhemis-type in Australia, and the manner in which tliey 

 have survived and successfully held their own against more 

 highly evolved types in all parts of the continent. As regards 

 those few Libellulidce which appear to have established them- 

 selves permanently in the desert belt of Central Australia (and 

 whose larv8e therefore we might expect to possess a drought- 

 resisting power superior to that of Synthemis), we are not yet in 

 a position to know whether it is really their larvae which with- 

 stand drought, or whether the imagines themselves may not 

 scatter far and wide over the heated sandy plains, and so always 

 keep the race alive over a large area, by ovipositing in whatever 

 waterholes may happen to contain water at any given period. 

 As far as I know, only Orthetrum caledonicum and Diplacodes 

 heematodes occur commonly over the whole of Central Australia; 

 and these two dragonflies are noted above all others in Australia 

 as being sand-lovers. Even in well watered regions they seek 

 out the dry patches of sand and settle on them in hundreds. 

 Also, whereas the seasonal range of the imago in Synthemis is 

 comparatively short (from the end of November to the end of 

 February), that of the above-mentioned species is long. They 

 occur all the year round in tropical parts, and in New South 

 Wales they may be taken from early in September right up to 

 the beginning of June. 



The experiment also shows that the larvse of Synthemis, when 

 conditions are unfavourable to development, can and do remain 

 alive for more than one year. As »S'. eustalacta is only on the 

 wing from November to February, and most of the ovijiositing is 



