Temperature Experiments on two Tropical Butterflies. 453 



marked on the hindwing. From the island of Formosa 

 there are two males like the above, and the tornus beneath 

 is red. Whether this is the usual form of the insect in 

 that island I am unable to say. In constructing the 

 ancestral type, we should probably be on safe ground by 

 assuming that it had more white and a certain amount of 

 red on the forewing, a lunulated band coiupletely round 

 the outer margin of the hindwing, and more red on the 

 underwing. Anything beyond this is conjectural. 



Referring for a moment to the females, it will be seen 

 that in the forced specimens there are in the blue costal 

 margin of the cell, two spots, sometimes red, sometimes 

 white, in precisely the same position as the two white 

 spots in the cell of the males. These are absent, or 

 nearly so, in normal specimens, and we may conclude that, 

 at one time in its history, the female had more white on 

 the forewing than it has at present. This would rather 

 incline ns to the view that diocippus is the earlier form, 

 but, as in chrysippiis, we are confronted with the difficulty 

 that shock throws back the insect to the earlier stage, in 

 which case, judging by these intermediates, inaria is the 

 more ancient type, and we must account for these addi- 

 tional spots by the not improbable conjecture that the 

 evolution of such a variable butterfly has not been uniform. 



The study of the closely allied species Hypolimnas holina 

 may help us in our determination of this question. It is 

 difficult in a few words to give a brief, and at the same 

 time lucid, account of the innumerable variations of this 

 protean butterfly. The male, throughout its immense 

 range, is very fairly constant, being very similar in general 

 appearance to that of H. misippus. In Fiji the spots are 

 very small, and a very deep blue. The females in their 

 western area do not vary greatly, being generally plain 

 brown and slightly blue on the costa, with a variable 

 number of marginal yellow spots. In Formosa the colour 

 is also plain brown, sometimes tinted with blue, and with 

 a white band as in misippus. Further east, in the 

 Loochoo Islands, the brown is replaced with glistening 

 blue. But it is in Australia and the Fiji Islands that the 

 butterfly reaches its maximum development both of size 

 and variability. In the Godman-Salvin Collection, now 

 in South Kensington, there is a series of some two dozen 

 females, taken at Suva, Fiji, on the same day and on the 

 same flower bed. All are different, and vary from plain 



