LIV 



differences in this respect occur in Java only exceptionally ; in races or closely 

 allied species in the east of the Archipelago, in which the ocelli are invariably 

 larger, greater differences in this respect are met with, although the character 

 of these spots remains the same. The stage of development ot these ocelli, however, 

 is not identical in the several individuals, yet they are invariably in the same 

 mutual relation, n*^. 5 on the hind-wings being always the largest, n^. 9 coming 

 next, etc. They thus occur in the same relation in various stages of develop- 

 ment ; at the most advanced stage such an ocellus has the appearance in 

 miniature of the enlarged illustration given on PI. XV fig. ^06, but in less advanced 

 condition these colour rings are only partly distinguishable or not at all ; at a 

 still earlier stage they appear as yellowish dots or specks, and in proportion 

 to the lower stage of development they decrease in size ; the larger being reduced 

 to specks while the smaller disappear entirely but all in the same mutual ratio ; 

 if only one yellow speck remains visible it is found at the spot where in case 

 of complete development the largest of the ocelli — n^. 5 on the hind wings — 

 occurs. The action of a regular process of colour mutation is clearly mani- 

 fested here but there is no question of irregular changing caused by so-called 

 variability. 



In a species like Cvllo Leda L., where the degree of its colour develop- 

 ment as well as that of its ocelli can thus so easily be observed and which 

 at the same time occurs so generally in various districts and at various times, the 

 question whether or not the existence of ocelli characterizes the wet-season 

 variety, ought to be capable of being demonstrated. 



Now from the examination of a great number of specimens, available for 

 comparison, it is perfectly clear to me that such is not the case. It is true 

 the best developed ocelli occur in the wet season while ocelli met with in the 

 dry season are always less developed and mostly reduced to specks, wholly or 

 partly, or they may also be wanting in part or even entirely. But transitions 

 occur and, moreover, the rainy season in Java gradually passes into the dry 

 season while, finally, butterflies with very slightly developed ocelli are by no 

 means unfrequently met with in the middle of the wet season. It is, therefore, 

 clear that the process of colour evolution, in the way this, too, decidedly mani- 

 fests itself in the ocelli, operates in Cyllo Leda L. most strongly during the 

 wet season, but presumably because, above all, the food is more nutritious at 

 that time — since in the wet season the whole vegetation is much more robust 

 and the lar\'ae then met with have a much stronger appearance — and not 

 by reason of any direct specific meteorological influence. In the course of 

 this evolutionary process the individual susceptibility too undoubtedly plays an 



