MIMETIC ASSOCIATIONS 



63 



always be towards a reduction in the number of warning patterns, and this is what we 

 do actually find in unquestionably Miillerian associations. We are forced, therefore, to 

 the conclusion that species which in the same locality produce polymorphic mimetic forms, 

 are Batesian mimics, and that the value of the multiple varieties lies in the distribution of 

 the mimetic forms amongst a corresponding number of models, thus avoiding the risk of 

 the edible mimic becoming more numerous than its distasteful model, a state of things 

 which would lead to a disastrous increase in the amount of experimental tasting by 

 insectivorous animals. Indeed, we may suppose that this process has actually taken place, 

 and that under its influence the same species has developed new mimetic forms, thus causing 

 such phenomena of polymorphism as we now observe in the case of H. duhius. The obscure 

 and comphcated processes by which a female butterfly can produce ova which give rise 

 to several forms belong to a separate study. That they do so is an established fact, of which 

 the only feature affecting the present consideration is that when these differing forms are 

 produced they are not promiscuously coloured, but in nearly every case resemble common 

 distasteful species of other genera. 



Professor Aurivillius thus describes the larva and pupa of duhius from specimens obtained 

 in the Cameroon (Ent. Tidskr., p. 282, 1894) : — 



' The larva is black, and has at each of the segments i-ii a narrow white transverse 

 band, from which arise the dorsal and lateral spines. The head is pale and bears two long 

 spines. The first segment has two lateral but no dorsal spines ; and in the remaining 

 segments the spines are arranged as in the larva of Salamis anacardii. The spines are pale 

 at the base and darker towards the end. Legs light coloured. 



' The pupa is brown speckled with black, and has a short black cremaster. Dorsal 

 spines short.' 



As the arrangement of the spines in the larva is compared to that in S. anacardii, I 

 append the description of these also : — 



' Second and third segment on each side with two spines placed close together, the 

 fourth with two spines placed one above the other, and the fifth with three sublateral spines. 

 There are also two sublateral spines at the tenth segment, the lower of which, however, is 

 very small ; the remaining segments have, on the contrary, no sublateral spines. The 

 twelfth segment appears to have only two spines, in that the dorsal spine is so far removed 

 forward that it appears to arise from the eleventh.' 



I have examined a long series of about seventy examples of H. duhius. Of these, 

 twenty-three were typical wahlhergi from Natal, Zululand, and British East Africa. They 

 varied little beyond the fact that the Natal specimens generally, though not always, had 

 a narrower sub-apical bar on the fore-wing. Nine were of the mima form from Natal and 

 Zululand, varying only in the extent of the yellow patch in the hind-wing and the size of the 

 white spots on the primaries. Fourteen anthedon from Sierra Leone to Lake Tanganyika 

 showed some variation in the width of the black hind-wing border, always, however, much 

 broader than in wahlhergi. The remaining examples varied from typical duhius, through 

 damoclina to cerherus, the last having the hind-wings entirely black. In this form it is 

 very remarkable that whilst the white of the hind-wings has disappeared, the cell spot 

 in the fore-wing remains large, larger than in many examples even of duhius, thus increasing 

 the resemblance to the model A. tartarea. 



There is one other African Hypolimnas allied to H. duhius which is a member of the 

 echeria association, viz. H. dinar cha. Hew., which occurs from Sierra Leone to British East 

 Africa. It is nearly as large as H. usamhara, but resembles white-spotted echeria in colour 



