Raymond Pearl 
71 
each a constant regardless of the absohite size of the individual dimensions 
themselves. In other words, it is contended that these ratios are not sensibly 
Fig. 6. 
correlated with the absolute size of the system. From this assumed independence 
Driesch deduces rather far-reaching generalizations, as the quotations show. 
But, as has been brought out above, when the matter is subjected to quan- 
titative test it is found that, in the case of two protozoan forms at least, there 
AB 
is a sensible and definite correlation between such a ratio (Fig- 7) and the 
B 
Fig. 7. 
AB 
absolute size of the system. Now clearly the ratio ^jjj is an index of the pro- 
portionality of the two chief dimensions of the body, or, in a word, of the shape 
of the body. It seems to me that the facts given demonstrate that in Paramecium 
and Ghilomonas size and form of body are correlated, and hence, in so far, 
experience does not agree with Driesch's generalization. It is probable that the 
same thing will be found to be generally true. It has been demonstrated for the 
principal indices of the human skull by Miss Fawcett (1902) and Macdonell (1904). 
Unpublished material on other and widely different organisms gives the same 
result. If it holds generally that the proportionality of the parts and the absolute 
size of a differentiated system are sensibly correlated, it seems to me that the 
analysis on which Driesch's first " proof" of the " Autonomic der Lebensvorgange " 
is based will have to be considerably modified. 
Simimary. 
A comparative study of variation and correlation in the flagellate Infusorian 
Ghilomonas paramecium when living on the one hand under the optimum 
environmental conditions, and on the other hand under extremely unfavourable 
conditions, has led to the following results. 
1. The individuals in the unfavourable environment are markedly smaller 
than those in an optimum environment. 
2. The individuals under the two sets of conditions are significantly different 
in shape, those living under poor conditions being relatively narrower. 
3. There is no marked difference in variability or correlation between the two 
groups, though there is a slight preponderance for both variability and correlation 
in the group living in the unfavourable environment. 
