326 BIONOMICS, ETC. [CH. 



the case of the little Hemiphlebia mirabilis. This metallic green 

 Dragonfly is almost invisible on green reed-stems, except for its 

 long ribbon-like white appendages (fig. 14 L). These are displayed 

 as a sign to the female, by raising the abdomen and bending it 

 slightly sideways, while walking up the reed-stem. The female 

 replies by moving the whitened end of her abdomen from side 

 to side in a peculiar manner. Finally, the little creatures fly 

 out from the reeds, and engage in a miniature "pas de deux" 

 before pairing [173]. 



Comparative Abundance of the Sexes. 



The causes of the capture of more males than females by most 

 collectors have been already indicated (p. 321). It is quite 

 possible to reverse the proportion by choosing the right locality. 

 Thus, an afternoon's collecting along a deep railway-cutting in 

 the tropics (including the mouth of a tunnel, in which I secured 

 several females of Gynacantha) yielded me over sixty females of 

 various species, but only two or three males. An even more 

 extraordinary result than this was obtained by working three or 

 four small streams in dense palm jungle in Queensland. Argiolestes 

 amabilis was to be seen on every rock. In a week I had taken 195 

 females, but only a single male ! Though I sought for the males 

 high and low, they were not to be found. Probably they lived in 

 the sunlight, on the tops of the giant trees, more than a hundred 

 feet above my head. 



As the result of breeding out many species during the last 

 ten years, I find that two conclusions may safely be drawn. 



(i) The number of males and females is approximately equal. 

 My records shew about 5 per cent, excess of females. The most 

 vigorous larvae were nearly always females. 



(ii) The females emerge, on the average, a few days before 

 the males. They should therefore be sought for early in the 

 season, before pairing has begun. 



Habits of the Larvae. 



The habits of the larvae are, if anything, more varied than 

 those of the imagines. A general resume of this question has 

 already been given, as far as the correlation between habitat and 



