124 



THE AMERICAN NATURALIST [Vol. XLVIII 



where the monthly means for May to September inclusive varied 

 from 20 per cent, to 27 per cent. The annual mean at Cairo, 

 Egypt, is 56 per cent, and at Ghardaia (Algerian Sahara) is 

 50 per cent, at 7 a.m. and 26 per cent, at 1 p.m. The humidity 

 at Buitenzorg, Java, during the height of the rainy season fluc- 

 tuates between 70 per cent, and 97 per cent, during the day. 

 Naturally, when dew is being deposited the humidity is practi- 

 cally 100 per cent. It wall be seen then that even Tower's ex- 

 treme averages (see below) are not beyond the range of 

 possibility in nature, although they are as great as it is possible 

 to use in experimental work, since at an average of 34 per cent, 

 humidity only 0.4 per cent, of the larvse reached the adult stage 

 and atmosphere can not be kept supersaturated. 



The beetles were seriated according to an arbitrary scale in 

 which "20 equals total melanism and total albinism." It is 

 difficult to suggest a better method of measuring the extent of 

 melanism than this, although we could wish for diagrams to aid 

 us in grasping just what the scale means. I have tabulated the 

 i and interpolated the normal data. 



It will be seen that mortality increases rapidly as the humidity 

 departs from normal but this can not account for the change 

 in color since the range of melanism is doubled and in three of 

 the experiments even the mode falls below the normal range. As 

 stated by the author: 



The results of experiments with deviations of humidity are almost 

 exactly the same as those which were obtained from experiments with 

 deviations of temperature. Such deviations from the^ normal either to- 



pigmentation and a consequent melanic tendency, but beyond this the 

 effect is reversed, pigmentation is retarded, and the tendency toward 

 alhinism becomes more and more pronounced as the deviation from the 

 normal hecomes greater. 



