512 MR. II. W. BATES ON THE LEPIDOPTETA 



on the individuals, because, in limited districts where these conditions are the same, the 

 most widely contrasted varieties are found existing together, and it is inexplicable how 

 they could have produced the nice adaptations which these diverse varieties exhibit. All 

 the varieties figured on PL LV. figs. 2, 7, 9, and on PL LVI. figs. 1, 2, 3, 4, 6, are 

 found at St. Paulo, within a mile of each other, in the same humid forest. Neither can 

 these adapted races, as before remarked, have originated in one generation by sports or a 

 single act of variation in each case. It is clear, therefore, that some other active prin- 

 ciple must be here at work to draw out, as it were, steadily in certain directions the 

 suitable variations which arise, generation after generation, until forms have resulted 

 which, like our races of Leptalls Theonoe, are considerably diff'erent from their parent 

 as well as their sister forms. This principle can be no other than natural selection, the 

 selecting agents being insectivorous animals, which gradually destroy those sports or 

 varietias that are not sufficiently like Ithom'ue to deceive them. It would seem as 

 though our Leptalls naturally produced simple varieties of a nature to resemble 

 Ithoin'uB ; it is not always so, as is proved by many of them figured in the places above 

 quoted. There is some general resemblance, it is true ; and this is not purely accidental ; 

 for it is quite natural that the parent Leptalls should produce offspring varying in the 

 direction of Ithoiiihv, being itself similar to an Ithomla, and having inherited the pro- 

 perty of varying in this manner through a long Hue of ancestors. We cannot ascertain, 

 in this case, whether changed physical conditions have had any effect, quantitative or 

 qualitative, on the variability of the species after migrating to a new district. At any 

 rate, the existing varieties of our Leptalls show that the variations of Leptalls and 

 Ithomla are not quite coincident, and that the agency of natural selection is required to 

 l>ring the slowly forming race of one to resemldc the other. I do not forget that at each 

 step of selection the forms of Leptalls must have had sufficient resemblance to an Itho- 

 III I <i to lead to their preservation, or, at least, to prevent theii' complete extinction : as, 

 however, the two analogues so much resemble each other at the commencement of the 

 process, these steps would not be numerous. In many cases of mimetic resemblance, 

 the mimicry is not so exact as in the Leptalkles. This would show either that the imi- 

 tator has only inherited its form from remote ancestors who were actively persecuted, 

 the persecution having ceased during the career of its immediate ancestors ; or it would 

 show that the persecutor is not keen or rigid in its selection ; a moderate degree of 

 resemblance sufilces to deceive it, and therefore the process halts at that point. I leave 

 out of consideration all resemblances which can only be accidental, or which are resem- 

 l)lanccs of affinity. 



[fa mimetic species varies, some of its varieties must be more and some less faithful 

 imitations of tlie o1)jeet mimicked. According, therefore, to the closeness of its ])ersecu- 

 tioii l)y enemies, who s(H!k the imitator, but avoid the imitated, will be its tendency to 

 Ix'conie an exact counterfeit, — the less perfect degrees of resemblance being, generation 

 after gem-ration, eliminated, and only the otliers left to propagate their kind. Tlie actual 

 slate of Leptalls Tlicoiioe is not the same in all of its tlu'ce districts. A few varieties, or 

 sjiorts, are seen at Ega (();")' W. long.) and St. Paulo (09' W. long.), namely, those figured 

 I'l. LV. figs. 1-, '), 7, S, and 9, which have an indeterminate resemblance. On the Cupari 



