( 10:; •) 



and there is, indeed, no donlit that croesxs, !i/(l'iua, and priamua have developed 

 from one ancestral form, which was similar either to the orange croesus and lydim, 

 or to the green pn'amus. But the orange and the green Troides seem to me to be 

 already po fiir separated in their characters that they cannot be united under one 

 specific name. My reasons for treating croestis as a distinct species are as follows : — 

 (1) The caterpillar of crocsus has on each side two wliite oblique streaks, while 

 in the races of pridinus there is only one or no stripe. Ribbe, Iris 111. p. 42 (1890), 

 says, however, of T. prlataus poseidon (Doubl.) that the caterpillars " mostly " have 

 only one stripe, from which I must conclude that they have sometimes two. If Kibbe's 

 observations are correct, there is no great difference between the caterpillars of croesus 

 and priarims. Considering, however, that the caterpillars of T. victoriae ((Jray) and 

 pjaradiseus (Standing.) do also not differ from those of priamus except in having no 

 white streak at all (in which character they agree with the Australian races of p-iamii^), 

 and that the chrysalides of urvillianus and victoriae, which I ha\e before me (collected 

 by (.'aptain \Velister), do not exhibit any difference worthy of note, we come to tlie 

 conclusion that we cannot derive any essential specific characters from the early 

 stages of these Troides : indeed, the similarity in larvae and pupae proves that all 

 these species are close relatives; that the characters distinguishing the imagines 

 have been most probably acquired in comparatively recent ages ; and that perhaps 

 there exist species, still unknown to us, which stand intermediate iu nem-ation between 

 victoriae and paradiseiis on one side, and priainus, croesus, and lydius on the other. 



(2) There are no intermediate specimens between the males of croesus and 

 priamus. Sometimes tlie hiudwings are partly green above, but such specimens 

 have been reared and were killed too soon after the emergence from the ympa. There 

 exist certainly no individuals in collections which cannot be recognised at once as 

 belonging to ci'oesus. 



(3) The /^mc(/f is constantly different from that of priamus ; truly intergraduate 

 specimens are again unknown. 



We do not know the insects from the island of Olii, which lies between the 

 Southern and Northern Moluccas this large island may be inhabited by a race of 

 Troides connecting croesus with pmamiis. 



i. Above, orange instead of green. 



There is usually no band along the hinder margin of the forewings above ; Imt 

 some individuals have a feeble orange patch near the hind angle, and in others 

 there are orange scales scattered all along the inner margin of the wing. The 

 number of the black spots on the hiudwings above varies from to J. 



5 . The wedge-shaped whitish markings of the hiudwings above reach rather 

 close to the cell and the outer margin ; they are separated from one anotlier by the 

 nervules, which are rather broadly brownish black ; the black spots within the wedge- 

 shajied markings are so increased in size that at least the two anterior of the wedge- 

 shaped marks are widely separated into a discal and a submarginal j)ortion. 



The white spots of the forewings are small ; the patch within the cell is often 

 reduced [as in T. p'''iainus euphorion ((iray)] to two small spots, or is even absent. 

 The nundjer of the spots is very variable ; sometimes there are two complete rows of 

 markings; the discal row is much more liable to obliteration than the submai-ginal 

 one. The cell of the hiudwings has occasionally a white spot at the apex. The sub- 

 marginal markings of the hiudwings below are often of a rather pure yellow colour. 



Hah. Hatjan (VV. Doherly, March 1892) (1:5 cf, 9 ? ). 



