Heierogynis 'paradoxa. 725 



please it, and it may even be variable in this point as in 

 others, but it is unquestionable that at all its habitats 

 where I met with it, each habitat was definitely and 

 (to the larva of H. imradojxC) impassably circumscribed, 

 by the limits of growth of its food-plants. 



Crossing by migration of the male probably occurs, but 

 it must be rarely. The male flies weakly, and never 

 voluntarily undertakes a migratory excursion, but such 

 must sometimes result from being caught by the wind. 



The two races Byarcnsis and Canddaripa live very 

 close together, not more than three miles as the crow flies, 

 but hcjarcnsis is some 1500 feet lower than candelaricV, 

 and on a S.E. instead of a north exposure ; the consequence 

 is that its flight is over before canddarix. comes out, so 

 that the chances of any accidental crossing are much 

 diminished if not entirely destroyed. 



Fiedrahitic is separated by a wide valley from the two 

 Bejar colonies, and curiously enough is even later in 

 appearing than canddaria}, for no reason that I could 

 observe, since it is at a level as low, possibly lower. At 

 any rate any crossing with the Bejar forms must be all but 

 impossible. 



The segregation of these three forms from each other is, 

 at any rate, if not complete, so nearly so as to make the 

 differences between them easily accounted for by variation 

 not swamped by crossing. There can be little doubt that 

 the differences represent something that is or has been 

 beneficial to each race, but here a longer study of the 

 conditions affecting each colony would be necessary, even 

 to plausible guessing. Why should Canddariie hide itself 

 so much more carefully than the others, as it does at all 

 stages ; as a result it appears to thrive at the one spot 

 where I found it, but also as evidence that it has powerful 

 enemies, it seems very restricted in habitat, whilst the 

 other two forms are co-extensive with their food-plants. 

 If there is anything in the suggestion that the larval 

 colour is protective amongst the flowers of its food-plant, 

 it is another item showing the need of protection in this 

 form, since the flowers of the food-plant are nearly 

 identical in colour in all cases, but with Canddarim the 

 flowers of G. lyiirgans are denser than in G. florida and 

 unmixed with leaves, but then per contra I only found the 

 insect where the plant did not bloom freely. 



The only other case in Lepidoptera where I have found 



