34 
Variation and Correlation in Br aiyi-W eight 
size is of greater direct selective value in the case of parts of the skeleton than in 
the case of organs like the liver, etc. In other words, size is mnch more directly 
related to proper functioning in the former than in the latter case. The skull 
may at first sight appear to form an exception here. It must be kept in mind, 
however, that the skull serves two very important — perhaps equally important — 
functions; viz., (1) the enclosing and protecting of the brain, and (2) the serving 
as a basis for the attachment of the complexes of muscles which actuate the 
masticatory apparatus and support and move the head. With reference to each of 
these factors natural selection may act on the skull. So far as the second factor at 
least is concerned, size of skull (in a broad sense) will be of selective value. So 
that here again, since size and function are interrelated, we should expect to 
find the results of selective action reflected in the size measurements of characters. 
In general, I think it can safely be maintained that the low variability shown in 
the group of characters under discussion is what might be expected to result from 
the operation of the factors just mentioned. 
In the middle group (coefficients of variation ranging from 7 to 10 in the male) 
including brain-weight and skull capacity we seem to have also an intermediate 
condition, with respect to the two factors which have been mentioned as among 
the causes which contribute to the observed variability. In the first place, there 
can be no doubt in face of the evidence both from the bio metrical and anatomical 
standpoints, that the cranial capacity is quite highly correlated with volume of the 
brain*. Admitting this, the discussion may be restricted to the brain. Now it has 
been known from the time of the earliest collection of brain-weight statistics that 
the weight of the brain is influenced to a certain extent by the general metabolic 
condition of the individual preceding death. I am inclined to think however that 
the extent of this influence has been over-estimated. My primary reason for this 
view comes from the analysis of the statistics themselves. If the metabolic 
condition of the individual preceding death influenced brain-weight to a marked 
degree, one would hai'dly expect to find the agreement shown in the constants 
of variation and correlation tabulated in this paper, when different groups of 
individuals with different distribution of ante-mortem affections were compared. 
We might expect by chance to get two groups in agreement, but on this 
supposition the odds would be great against getting four groups to agree so closely 
as they actually do. Furthermore, there is a certain amount of evidence from 
other sources indicating that the influence of ante-mortem general metabolic 
conditions on brain-weight, while always present as one factor, does not produce 
so marked an effect on the weight of the brain as on the size measurements of 
some of the organs in our most variable group. For example, Miihlmann-f studied 
microscopically the brains of 24 individuals of different ages from birth to 90 years 
to determine the amount of pigment degeneration in the ganglion cells at the 
different ages. In the adults the causes of death included such divergent con- 
* On the anatomical side, cf. Symington, NaUire, Vol. 68, pp. 539 — 544. 
+ Verhandl. d. deutsch. Pathol. Gesellsch. Bd. iii. pp. 148—157. 
