90 LIFE-HISTORIES OF AUSTRALIAN ODONATA, iv , 



secured seven larvae, ranging from about lialf-grown to nearly 

 full-fed. Two of these were exceedingly weak and flabby, 

 and died soon afterwards. On examining them, I found the 

 whole abdomen hollow, and, in one case, the alimentary 

 canal practically destroyed : but I was unable to find the 

 parasite that caused this, if such were the case. 



After three and a half hours' work, we finished up at the 

 part of B opposite to A in the diagram, where we found 

 passages opening both on the top of B (bj, in diagram) and 

 at its side, below the water-level (bo). These were also ex- 

 plored as far as possible, and the last larva was taken about 

 a foot below the entrance of bo. 



For examination, two larvae were taken, and killed. The 

 gizzards of these were cut out, and carefully examined. 

 Firstly, as to contents, — the only definitely recognisable por 

 tion of the food, which had not been triturated, was, in each 

 case, an Agrionid labium, which I had no difficulty in recog- 

 nising as that of A rgiolestes gj-isea Selys. The larvae of this 

 little dragonfly are very common in these swamps, but not 

 easy to obtain with a dredging-net. They cling to the 

 matted outcrop of fine roots at the sides of the depressions in 

 the swamps. In the diagram, the letters Y indicate the 

 probable spots where the larva would occur. (None were 

 actually found in C, but they are small and inert, and we 

 were not looking for them; I have always taken them in 

 positions akin to Y). This seems to suggest that the Petalura 

 larvae use the channels, such as a 5 and bo, for obtaining a 

 supply of this Agrionid larva for food. However, if that be 

 the case, it further sviggests that they are nocturnal feeders ; 

 for they are so clumsy and slow-moving that they would 

 have no chance of capturing the Argiolestes larvae in the day- 

 time. The latter, though by no means active, are fairly 

 quick at dodging out of sight, behind the matted roots. It 

 seems to me, therefore, that the Petalura larvae must possess 

 a distinct advantage over them, viz., a better power of 

 nocturnal vision. That is probably the case. It also sug- 



