119 



In this way, I suppose, may liave arisen sucli species as JJeilepMla 

 euphorhiw^ which are very restricted in their diet; and from them, x)er- 

 haps without modification except in liabit, those which are omniv- 

 orcms. 



Take now a species of restricted liabits Avhich has arisen as above 

 described. It has had its origin through i)ossessing certain advantages, 

 but now that it is established it has to face new difficulties. The i^lant 

 may become extinct or so rare that the individuals of it are few and 

 far between, and dangers of this sort do not trouble omnivorous si)e- 

 cies. The means of spreading into new territories, which omnivorous 

 species have in so marked a degree, is naturally very restricted in the 

 case of those which can only live on one or a few x^lants. Added to 

 this, the enemies which troubled the ancestor of the new species would, 

 for their lives' sake, be obliged to turn their attention to the modified 

 and restricted forms, the old omnivorous tj^pe having become extinct. 

 Having done this, no doubt by a process of natural selection among 

 themselves, these enemies would become excessively troublesome, since 

 the supply of food would now be in each case more limited and more 

 local in distribution, and consequently more easy to exterminate. But 

 among the progeny of the new species there would be variations to- 

 wards omnivorousness, and such would survive if the adverse conditions 

 became sufficiently pressing, leading to the formation of a new omniv- 

 orous form, which would very i)ossibly differ only in its omnivorousness 

 from the type whence it was immediately derived. Such a form would, 

 for the time being, have advantages owing to the fact that enemies had 

 learned to look for the insect only on certain i^lants, and any tendency 

 to si)lit up would be checked by crossing and the advantage derived 

 from continuity, so to speak, until the conditions described at the be- 

 ginning of the cycle once more began to arise. 



I have put this very briefly, but I hope sufficiently clearly to be under- 

 stood. Hereafter I may go into further detail and give numerous in- 

 stances in illustration. If insect life really does present such cycles as 

 here outlined we can see how to account for many apparent anomalies 

 of habit and many apparently useless specific characters. 



According to the above hypothesis, it is clear that, although there 

 would be two distinct changes in the cycle, only one of them, that from 

 omnivorousness to a special diet, would be necessarily or probably ac- 

 companied by such changes as to lead to the formation of what we 

 should term a new species. This, I take it, is very important as ex- 

 plaining certain apparently anomalous facts. For example, it is not 

 easy to see how two closely allied Lepidoi)tera could have been de- 

 veloped, each inhabiting the same two species of trees, say the Oak and 

 the Elm, unless we suppose that originally one was peculiar to one and 

 the other peculiar to the other tree and that both have varied towards 

 omnivorousness. Let A, B^ 0, D stand for four different plants and 



i 



