on Millkrian Mimicry and Diaposematism. 571 



by a gradual process of selection from comparatively small 

 individual variations" (p, 101). The intermediate form 

 Math its double aposeme at once brings the Miillerian 

 factor to bear throughout the whole assemblage which it 

 unites. It is not always realised how easily the gap 

 between A and B may be bridged over. Take the case of 

 Ficris demojyJiile (both sexes) and Aoria agna,a. Pierineand 

 an Ithomiine from Brazil. All that is necessary is to insert 

 the yellow form of P. demophile $ between the normal 

 types of the two species, and the chain is complete from 

 end to end ; moreover, it becomes linked on to the great 

 assemblage of which A. agna is a representative. I do 

 not assert that this particular association is Miillerian ; I 

 only adduce it to show how easily a Miillerian couple 

 might be established. It illustrates some other points as 

 well, viz. (1) that the rise of a mimetic variation does 

 not necessarily involve the extinction of the parent form 

 (this survives in typical P. demophile), and (2) that dis- 

 tastefiil intermediates may be conceived of as mutually 

 protecting and being protected by the distasteful forms 

 not on one side only, but on each side of them. This last 

 point is of course only another way of putting what I 

 have already spoken of as the function of the double 

 aposeme. It is obviously of the first importance for the 

 right understanding of Miillerian mimicry. 



It may be said, in reference to the foregoing, that I 

 have taken a case where mimic and model are already 

 somewhat alike. Let me therefore now show how the 

 two hypothetical species A and B may be brought into 

 association with one another, though originally very dis- 

 similar in aspect. If we were to confine ourselves to 

 theory, I admit that the process might be somewhat 

 difficult of conception. But when we turn to the actual 

 facts of such a case, we see how the passage may be helped 

 along by the existence of other species, each of which is 

 capable of forming a collateral association with the transi- 

 tional forms in turn. Thus, a very slight modification of 

 the yellow form of P. demophile ^ gives us another Pierine 

 form, that of P. viardi $, which aligns itself with Heliconius 

 charithonia ; while a short step onwards from P. viardi 

 brings us to the' form now called P. tithoreides % the mimetic 

 relation of which with Tithorea pavonii, H. atthis, and the 

 peruviana form of H. charithonia will only be questioned 

 by those who do not accept the doctrine of mimicry at all. 



