OSBOEN^ — ZOOLOGIC AND PALiBONTOLOGIC EESEARCH 
9 
Consequently, differences which have been classed and lumped 
together in long tables of measurement hitherto as variations of pro- 
portion may be now analyzed as partly due to one or more of a great 
number of different causes, namely: 
(1) Proportions due to differences of habit and modes of locomotion. 
(2) Proportions due to differences of nutrition, kinds and habits of 
feeding. 
(3) Proportions due to normal differences of sex, male and female. 
(4) Proportions due to internal or endocrine secretions, e.g., of the 
male and female sex glands. 
(5) Proportions due to adaptive changes during age and growth 
correlated with precocity or helplessness in the young, and juvenile, 
mature, and senescent development of the sex glands. 
(6) Proportions due to the withdrawal of the internal secretions 
after the natural close of the activity of the sex glands. 
(7) Proportions due to compensatory growth. 
First : it is obvious that older systems of measurements, which lumped 
all measurements together as ‘‘variations,” irrespective of cause, 
lacked such analysis of the causes of proportion. Second: it is clear 
that many of the differences that have been treated as hereditary 
variations — as material for natural selection — are not variations at all 
in the true sense of the term, but are really adaptations to seven or 
more different sets of causes which vary with conditions of life. Third: 
we have reason to suspect that the mean fluctuating variations of size 
and proportion may be mere individual and ontogenetic phenomena, 
nonheritable, and consequently without bearing on racial evolution. 
Compensatory changes of proportion are often profoundly impor- 
tant. For example, it has been shown that a dog with the fore limb 
experimentally removed tends to develop saltatorial proportions in the 
hind limbs, byway of compensation for the loss of the fore limb, that is, 
to imitate the springing type of limb, e.g., the hare. It follows that 
ontogenetic changes in hind limb proportion may be brought about 
through defects in fore limb proportion. 
I have reached the opinion that if we could eliminate these seven or 
more causes of modification and variation, and measure a very large 
number of similar bones (the pelvis, for example) of animals of (1) the 
same habits, (2) the same food, (3) the same sex, (4) the same intensity 
of endocrine secretion, (5) the same age, (6) the same sexual stage, 
(7) of exactly the same strain or race, there would be a standard length 
and breadth. I believe that nature tends to standardize every hone 
in all pure breeds and to eliminate variations in proportions. Otherwise 
