nv R. .T. TILLYARD 371 



only slightly disturbed, and returns again to its old haunt, being 

 thus easily captured. On the Horton River at Fallal, a broad 

 flowing stream, iS'. regina does not occur, but the males of //. 

 superba patrolled the banks, generally keeping just out of reach 

 of the net. Here they flew swiftly, stopping frequently at some 

 fancied spot, and remaining poised and stationary, with eyes 

 gleaming and all their brilliant colouring well displayed. The 

 only way to catch them was to note where they turned into some 

 small bay or inlet to capture gnats, and then to hide behind some 

 bush ready to strike at the insect on his next visit. Failure at 

 the first stroke sent the insect far away, probably to take up some 

 other " beat " on the river. 



These insects appear to be rather fiercely disposed towards the 

 dragonflies on the river. Generally three or four II. intermedia 

 would "patrol" past before an H. superba came in sight, and the 

 former kept closer in shore and were easier to capture. The 

 larger insect, however, would often dart in and attack the smaller, 

 driving it away just as I was about to strike. This was done 

 with such rapidity that it was almost impossible to follow what 

 took place. 



This fine insect is easily the most distinct and beautiful member 

 of the genus. It is probabi}' most closely related to H. intermedia, 

 but can be at once distinguished from that species by its much 

 greater size, more brilliant colouration, by the peculiar pattern of 

 the front (which may be considered as the intermediate stage 

 between the T-mark and the bright metallic crest), also by the 

 very long appendages of the male. The actual colour-pattern of 

 the abdomen is very similar to that of //. intermedia, but whereas 

 in the latter the join of the black and yellow on segment 7 is 

 quite straight, in //. superba it is very irregular. The very 

 distinct antehumei-al yellow bands of the thorax distinguish it at 

 once from all other members of the genus, and the beautiful 

 colour-pattern of the sides of the thorax is also unique. 



Omitting //. nov(e-hollandi(r Selys, for the present, from the 

 following table, since nothing certain can be gathered about this 

 species (unless, indeed, as seems very probable, H. continentalis 



