OF THE GENUS HYPOLIMNAS. 343 



red Danais in these islands. The mimic is here much larger 

 than the mimicked. This is the only case I know of, in which 

 this species of Sypolimnas mimics a red insect and thus gains 

 itself a considerable patch of this colour. 



Next we turn to Africa, and we invariably find that both sexes 

 of what we may fairly call the African forms of Hypolimnas 

 holina mimic various species of Danaince^ the normal form 

 of the male having entirely disappeared. Hence, from the 

 systematist's point of view, the specific characters having been 

 lost in both sexes, they bear as many specific names as there are 

 local forms mimicking the accompanying species of Danais. 



In quest of these mimetic forms, I searched through Mr. 

 Crowley's magnificent collection of African butterflies at Croy- 

 don, where I found very many examples, from which I selected, 

 three. In every locality where the forms occur, the mimicry 

 seems to be remarkably perfect, but there are local peculiarities 

 in the patterns of both mimic and mimicked in many places. 

 The localities are as widely separated as Natal in the South-east, 

 and the Cameroons in the West of Africa. 



From Natal, I have obtained Hypolimnas marginalis (PL XVII. 

 fig. 3), w^hich mimics A)nauris dominicanus (fig. 4). Prom Grra- 

 hamstown, S. mina mimicking A. echeria ; from the Cameroons, 

 S. dubia (fig. 5) mimicking A. egialea (fig. 6). 



Conclusions. 



Having thus brought together all the facts I have come across 

 and those which have been previously published, it remains to 

 ascertain their bearing upon the theory of mimicry, for this 

 theory has never been subjected to the evidence derived from 

 the systematic study of a small group of wide-ranging, mimetic 

 insects, carefully traced through all the localities included in their 

 range. This has, however, been done for the Papilio merope 

 group, so admirably worked out by Eoland Trimen (Trans. Linn. 

 Soc. xxvi. p. 497, and South-African Butterflies, vol. iii. 1889, 

 pp. 243-55), but the total range of these butterflies is far more 

 limited and the number of diff"erent forms much smaller than is 

 the case with the Hypolimnas group. 



bearing upon general Theory of Mimicry. 



In the first place, we find the strongest support to the general 

 theory of mimicry as originally suggested by H. W. Bates. The 



