Eaymond Pearl 
33 
Finally we have all the characteis giving values above 10. Now as one passes 
from one end of the series to the other a definite biological relationship is plain. 
At the upper end of the seiies (coefficient of variation 10 and over) the organs and 
charactei'S tabulated are such as depend in a very considerable degree for their 
values as determined by measurement, on the general metabolic condition of the 
organism as a whole at the time the measurements are made or immediately 
before. This will be freely admitted, I think, for such characters as " Dermal 
sensitivity," "Keenness of sight," "Weight of body," "Swiftness of blow," "Strength 
of pull," " Breathing capacity " and the like. The same thing, though less apparent 
perhaps, is, I believe, true when the characters are the weights of viscera. Indeed 
Greenwood's own results show this to be the case, when we get such different 
results for means, variabilities and correlation accordingly as we deal with the 
" general hospital population," " healthy " organs, or the same viscera in diseased 
conditions of different characters. Wynn* has only recently shown that in rabbits 
there is a considerable degree of probability that continued administration of 
digitalis will in a short time raise the mean weight of the heart appreciably, while 
at the same time the body weight is lowered. It is to be regretted that his series 
of experiments included so (statistically) few individuals. 
Furthermore, it seems reasonable to assume that in this upper group the thing 
measured in the majority, if not in all cases, is not the thing natural selection has 
acted upon directly, allowing that it has acted at all. I think it may be fairly 
assumed that so far as natural selection has acted at all on these organs and 
characters, the selection has been in the direction of ability to function properly 
so as best to conserve the physiological economy of the organism as a whole, rather 
than in the direction of absolute size of organ or character. The physiologically 
balanced functioning with reference to the needs of the organism as a whole is the 
important thing in such organs as the spleen, pancreas, liver, etc. Absolute size 
of organ can hardly be a very close direct measure of ability to function well. So 
then, I am inclined to attribute the high variability observed in those organs and 
characters falling together in the uppermost part of the table to these two factors : 
(a) the value obtained by measurement depends to a considerable degree on the 
general metabolic condition of the individual at the time, and {b) the thing 
measured is not the thing with which natural selection, so far as it has acted at 
all, has had directly to do. 
On the other hand, if we turn to the lowest group in the above table (coefficients 
of variation under 5 — 7 for the males) we find the two factors just mentioned 
almost exactly reversed. Here the characters are either of the skeleton directly, 
or in the case of the anthropometrical data are characters which closely depend 
for their measured values on the size of portions of the skeleton. Obviously the 
general metabolic condition of the organism has, within wide limits, little effect 
on the length of the femur or the skull, for example, and if there is an effect 
it manifests itself very slowly. Further it is to be presumed that the absolute 
* Journal of the American Medical Association, Vol. xliii. pp. 164, 165, 1904. 
Biometrika iv 5 
