( Ixiv ) 



The ice came up from Lago each Tuesday, Tliursday, 

 and Saturday, always at the same hour, and was at once placed 

 in a suitable box. As the quantity never varied to any extent 

 and as the outer temperature kept fairly constant, the mercury 

 in the box followed a nearly uniform course, remaining at 5° C. 

 for about 18 hours after the ice had been put in and then gradu- 

 ally rising. On the following night at 7 p.m. it averaged 12° C. 

 and on the third day without a fresh supply of ice, i. e. on 

 Mondays, it averaged about 20° C. During the course of the 

 experiments the mid-day shade temperature averaged 32° C, 

 and the midnight temperature about 22° C, although occasion- 

 ally as low as 17° C. Two pupae were exposed to a moist 

 atmosphere at normal temperature." 



The results of this most interesting breeding experiment 

 are shown in tabular form below, pp. Ixv, Ixvi. 



It was clear from the above table that, as Mr. Lamborn 

 recognised, " these dionysus have not been produced by chilling 

 the pupae." Nor was any other efiect to be made out when 

 the imaginal colours and patterns were studied in relation 

 to the conditions of temperature or moisture to which their 

 pupae had been exposed. The hippocoon were very constant, 

 as also were the under surfaces of the males, where some efiect 

 of the varying conditions was perhaps to be expected. On the 

 upper surface the extremely variable inner black band of the 

 hind-wing was certainly not more variable than is usual in the 

 male offspring of the same parents. As regards variability 

 the 8 dionysus stood in startling contrast to their 9 hippocoon 

 sisters, especially in the development of the oblique black bar 

 crossing the fore-wing. Very minute traces of " tails " to the 

 hind-wing were also present in one or two dionysus but not 

 in others. It was in every way probable that the constancy 

 of the hippocoon and the variability of the dionysus were 

 related to the presence and predominance of an excellent 

 Danaine model, Amauris niavius, L., for the first form, and the 

 absence of a model for the second. This interpretation was 

 confirmed by Mr. Swynnerton's Family D (p. Ixiii), reared from 

 a female form related to leighi and without any model at 

 Chirinda. In this family, too, the hippocoon offspring were 

 constant and the Ze?^/w-like offspring highly variable. 



