io6 



AFRICAN MIMETIC BUTTERFLIES 



variations and reject others were both simple in nature and regular in action, such a con- 

 summation would be difficult of attainment ; but when it is considered that the factors which 

 guide evolution are infinitely complex in their nature and varied in their incidence, it is 

 plain that the probability of such influences following identical Unes through a long course 

 of evolutionary change, and so arriving ultimately at the same result is to the last degree 

 remote. Slight resemblances may doubtless be found between butterflies which inhabit 

 widely separated regions of the globe. It would be rash to assert that in the multiplicity 

 of patterns which these insects present, no butterfly, say in South America, bears the least 

 resemblance to any Asiatic form. Such a resemblance, if it exists, is not, however, of the 

 kind of which this study treats. It is not governed by the patterns of dominant species 

 forming mimetic associations coincident in time and place. To attempt to account for the 

 known instances of mimicry by parallel variation, would be to argue that through the whole 

 course of the evolution of model and mimic, the influences which have on the one hand 

 produced profound differences of anatomical structure, have on the other hand contradicted 

 their own evident diversity by producing the same pattern, and combined this unlikely result 

 with the strange coincidence that the pattern so selected happens also to be that of a species 

 which is more or less distasteful to its enemies. In spite of the large amount of proof now 

 obtained of the comparative distastefulness of many insects, there is not yet any instance 

 recorded where the dominant form of a mimetic association has proved to be highly relished 

 by insectivorous foes. 



In the course of the discussion referred to, the danger was pointed out of mistaking for 

 mimicry resemblances which were in reality due to consanguinity. It was also urged that 

 too much stress had been laid on the supposed destruction of butterflies in the imago state, 

 and that the greatest destruction took place in the earlier stages, also that proof of the 

 inedibility of the so-called models was still required. The rarity of observed cases of the 

 attacks on butterflies by birds was insisted upon by many of the speakers, especially those 

 qualified by considerable collecting experience to express an opinion on this point, and 

 finally climatic influence was suggested as a potent cause for the development of similarity 

 of pattern. As regards proof of the inedibility of protected species, ample evidence has now 

 been accumulated, thanks to the extensive experiments of G. A. K. Marshall, Professor 

 Poulton, Professor Weismann, and others. Many of these experiments are described later. 



A very extensive criticism of the Batesian and Miillerian hypotheses appeared in 1904 

 in a paper by the late Professor Packard.^ 



This essay was criticized at some length by the present writer in a paper communicated 

 to the Entomological Society of London in 1906,^ when it was pointed out that the author's 

 arguments were to a great extent based on wrong premisses. His conception of Miiller's 

 hypothesis was incorrect owing to his having mistaken the suggestion of sexual selection 

 made by Miiller for the accepted Miillerian hypothesis, an error which was pointed out by 

 Professor Meldola in a letter to ' Nature ' in 1905. 



The view of Mr. Abbot H. Thayer, that the colours of animals are such as to cause 

 the creatures to become invisible, and which Professor Packard supports, is merely a universal 

 adaptation of the theory of cryptic coloration, and cannot be applied to the numerous 

 instances of warning colours with which we are familiar. 



* Proc. Amer. Phil. Soc, 1904: 'On the Origin of the 

 Markings of Organisms due to the Physical rather than 

 to the Biological Environment, with Criticisms of 

 the Bates-Muller Hypotheses,' A. S. Packard, Ll.D. 



2 On the late Professor Packard's paper ' On the 

 Markings of Organisms ', Eltringham, Proc. Ent. Soc, 

 1906, p. xx.xvii et seq. 



