20 



AFRICAN MIMETIC BUTTERFLIES 



' If two species are concerned, of which the one is very common and the other very rare, 

 then the advantage falls almost entirely on the rarer species. If, for example, Acraea thalia 

 were a thousand times commoner than Eueides pavana, the latter would derive a million 

 times greater benefit from the resemblance of the two species, whilst for the Acraea the 

 benefit is practically nil. Thus Eueides pavana might by natural selection be converted into 

 one of the most exact mimics of Acraea thalia although it is just as distasteful as the species 

 imitated. On the other hand, if two or even several distasteful species are about equally 

 common, resemblance brings them a nearly equal advantage, and each step which the other 

 takes in this direction is preserved by natural selection — they would always meet each other 

 numerically, so that finally one would not be able to say which of them has served as the 

 model for the others. ... To this category Buna and Thyridia may belong, although the first 

 has probably made the greater step in passing from the former dissimilarity to the present 

 resemblance of the two species.' 



Such is the theory of what is now known as Miillerian mimicry, a phenomenon which will 

 be found to be of the greatest importance in its bearing on many of the complicated cases 

 of mimetic association which are now recognized. 



A careful study of Miiller's paper is the more interesting from the fact that he to some 

 extent suggests that the adaptations which take place between two species of distasteful 

 genera may be mutual ; that is to say, both species may be modified so that an ultimate 

 type of pattern may be produced which is not necessarily that of the original in either 

 species. This theory has been greatly developed by Dr. Dixey whose paper on mimetic 

 attraction (Trans. Ent. Soc, 1897) forms a most valuable contribution to the literature 

 of the subject. For the purpose of his argument the author takes certain forms of South 

 American Pieridae which have developed patterns closely mimetic of various species of Heli- 

 coninae. Starting with a Pieris of ordinary appearance it is shown that only a very small 

 modification, such as might easily arise as a variation, is sufficient to give the butterfly an 

 approximate resemblance to a certain Heliconine species. This slight variation persists as 

 a separate species of Pieris, but by further slight modifications, also persisting as species, 

 forms appear which come to resemble other species of Heliconinae. In this manner we 

 can see actually existing a series of modifications by which the original form of Pieris is 

 drawn along different lines of attraction, and so produces several mimetic species, the 

 extreme forms of which differ widely from each other, though reached by very gradual 

 modifications ; and it is further pointed out that the forms which in these cases appear 

 to act as models are not themselves isolated, but are members of large groups, all the 

 individuals of which have a similar type of coloration, and which may be considered to 

 form Mullerian associations. As is well known, distasteful butterflies are usually so coloured 

 as to appear conspicuous in their habitual surroundings, and in the case of a species having 

 the necessity of profiting by its distasteful qualities, it may either develop a conspicuous 

 pattern of its own, or adopt that of some already well-known and dominant form. The latter 

 development, as might be expected, appears to be the commoner, since it benefits all the 

 species of the dominant group, and the benefit being mutual, there are good grounds for 

 expecting the model to itself assist in the resemblance, by undergoing a certain amount of 

 modification in the direction of the mimic. To quote the author's own words, * Every 

 conspicuous and distasteful form is a centre of attraction for other forms, whether edible or 

 inedible; but in the former case (Batesian mimicry) the mimetic attraction is limited in 

 operation, and acts only in one direction, influencing nothing but the mimic ; while in the 

 latter case (Miillerian mimicry) the mimetic attraction is unlimited and mutual, acting 



