Hetcrogynis imradoxa. 721 



brooms also growing close by, nor on any other plant, 

 differing therein from pcncUa, which prefers common 

 English broom, has no objection to almost any leguminous 

 plant, and is even almost omnivorous. 



The form candelarix appearing a full fortnight later 

 than hcjarensis, was a trifle smaller than that variety, and 

 had not so great a range in size, varying from 22 — 26 m.m. 

 in expanse, and averaging 24 m.m. The colour was 

 distinctly lighter, a smoky-brown rather than a translucent 

 black. Their habit of resting so as to be unseen, con- 

 trasted strongly with that of hcjarensis and of piedraliitss ; 

 the food-plant is the same as that of picdrahitm, and 

 though it does not afford such spray-like branches to rest 

 on as the G. Jlorida does, 2ncdraliitse always rested in the 

 same manner as hejarensis. It is to be noticed that 

 candelarise thus hides itself as an imago, as well as a 

 pupa ; and so far as ray observation went, as a larva also, 

 as I found very few of these, but this might be due to my 

 being too late for them. As compared with j^^^^drahiLv on 

 the same food-plant, it is to be observed that candclariiv, 

 as an imago to be stirred up out of its food-plant, but not 

 to be seen at rest, was, on a certain small area of a few 

 hundred acres, very much more abundant than either of 

 the other varieties in a similar space, but outside this 

 small area it was not seen at all, although its food-plant 

 extended a thousand or two feet higher up in the greatest 

 profusion, as well as over other ground at the same level; 

 whilst picdraliitx occurred wherever the food-plant 

 grew, over a large area extending both vertically and 

 horizontally. 



One is inclined to suggest that the bright lemon-yellow 

 of the larva of van candelarLv- is correlated with the 

 denser yellow masses of the flower of its food-plant, and 

 that we saw nothing of it higher up where the broom was 

 in flower, because it was so well hidden. This may very 

 well be so, and it may have little bearing on the point 

 that pt'^^drahitm on the same food-plant was not at all 

 so hidden in any stage, nor was the larva modified in 

 colouring like that of canddariiB. 



The larva is quite yellow, the black-green markings 

 being much reduced, only about a dozen larva3 were taken, 

 but were all constant in this coloration. The $ moths 

 bred from them agreed with them exactly in coloration 

 as is the case in all specimens of the genus. This pale 



