undergone by Insects in the Pupal State. 307 



pupa by desiccation — the comparisons extending in all cases 

 over equal periods of time. They prove also, what is far more 

 important to our present inquiry, that the variation in loss is 

 sufficient in amount to lead us to expect the size, or at least 

 the specific gravity, of the imago to be sensibly influenced by 

 it. 



Trusting that other observers may be induced to take up 

 the inquiry next season (for in the multiplicity of observations 

 there is strength), I propose to point out in concluding the 

 various disturbing factors which would interfere with the ex- 

 perimental results. In any species, then, the size of an indi- 

 vidual may be influenced by : — 



1. The natural difference in size between the two sexes of 

 many species. This factor is obviously eliminated by com- 

 paring only individuals of the same sex. 



2. The force of heredity — large parents tending to give rise 

 to large individuals and vice versa. Eliminated by comparing 

 only the offspring of a single pair. 



3. The varying supply of food to different individual larvce. 

 Eliminated by supplying each larva with excess of its food- 

 plant. 



4. The different amount of substance-waste in different indi- 

 vidual larvos due to dissimilar thermal conditions. Eliminated 

 by keeping the experimental larvae at the same temperature. 

 The amount of carbonic acid excreted by different individual 

 larvse of a species exposed to the same temperature is proved 

 by Newport's experiments to vary but little. This factor 

 might, in fact, by itself be neglected altogether, but becomes 

 eliminated when eliminating the next factor. 



5. The different amount of substance-gain in different indi- 

 vidual larval^ due to individual variation in assimilative power. 

 Eliminated by periodically weighing the experimental larvae, 

 and comparing only the imagines from those which show a 

 similar proportional increase during the same periods of time. 



6. The different amount of substance-waste in different indi- 

 vidual pupo3 7 due to dissimilar thermal conditions. Eliminated 

 by keeping the experimental pupae at the same temperature. 



7. The different amount of substance-waste in different pupo& 

 (kept at the same temperature) due to individual variation in 

 respiratory activity. Newport's experiments 'and my own 

 weighings prove that this is a variable factor. Its elimination 

 can therefore only be effected by periodically weighing the ex- 

 perimental pupae, and comparing only the imagines from those 

 which in given periods of time undergo the same proportional 

 amount of loss. 



In these observations an ordinary chemical balance will be 

 found most trustworthy for making the weighings. 



