474 Mr. G. A. K. Marshall on 



very inuch smaller than those bred from larvai which fed 

 on the more luxuriant food-plant in Mombasa. Examining 

 the two series, we are driven to the conclusion that the 

 Machakos larvae were partially starved, probably by feeding 

 on parched food-plant. Interesting and important inferences 

 may be drawn from the comparison. 



In the first place the specimens, so far from supporting 

 the conclusion often arrived at from incomplete and, as I 

 think, ill-regulated experiments, that males are produced 

 by starvation, actually show a larger number of females in 

 a smaller total of specimens than the set from Mombasa, 

 viz. 6 out of 13, as against 5 out of 15. Even if the 

 females had been very scarce at Machakos, nothing would 

 have been proved in the direction of the determination of 

 the sex of the individual by diet, for starvation pushed to 

 the extreme of preventing the completion of development 

 of many individuals is certain to kill off the heavier sex 

 far more freely than the lighter. The results, however, 

 show no abnormal excess of males, and in every way 

 support a prediction firmly founded on the anatomical fact 

 that the essential organs of sex, the testis and ovary, are 

 already present, rudimentary, but perfectly distinct, in the 

 larval stage. 



A comparison of the two series furthermore indicates 

 very strongly that the various forms of the species are not 

 in any way due to environmental causes, but are inherent 

 and hereditary. It is believed that Idugii is due to 

 drought, but there is a larger proportion of this form in 

 the series bred in the moist heat of the coast than in 

 that reared at high and dry Machakos. The great differ- 

 ence in conditions which is manifest in the different 

 average size of the two series was powerless to effect any 

 change in the inherent hereditary tendency of the indi- 

 vidual to become either Idn.gii or its modificatioxi dorvpims, 

 the type-form or its modification cdcijjpoides. 



This comparison of forms from adjacent localities under 

 different climatic conditions leads to an inference which is 

 precisely the same as that drawn from the comparison of 

 forms from different localities under the same climatic con- 

 ditions. The tropical forests of West Africa and the 

 Malayan Islands are very similar as regards climate : in 

 the first chriidppus occurs as the white-hind-winged 

 alci2}2ncs, in the second as the type-form, a peculiar dark 

 form inhabiting Java {L. hataviana). It is not necessary 



