Raymond Pearl 
21 
material so as to best get at this. I finally decided after a good deal of con- 
sideration to adopt provisionally Marchand's view stated above, viz., that the 
weight of the brain reaches a maximum between the ages 15 and 20, remains 
constant till about age 50, then declines through the old age period. Acting on 
this plan I separated the material in each case into what I have called a " young " 
and the "total" group. The "young" series included the individuals falling 
between the ages 20 to 50 (in the Hessian material, 15 — 50, and in the Bohemian 
20 — 59 were the limits, in the former case from choice, in the latter from 
necessity). The " total " series included all the individuals between ages 20 and 
80 (in the Hessian sei-ies 15 was the lower limit). The results show that on the 
whole this method of handling the material was adequate, considering the ends to 
be gained and the amount of material available. 
3. Fundamental Constants. 
As a preliminary to the discussion of the correlation of brain-weight with other 
characters it is necessary to exhibit and discuss the constants, measuring type and 
variability of the characters considered. Certain problems which are of consider- 
able interest on their own account present themselves here. Of first importance 
perhaps is the question as to the amount of differentiation in respect of brain- 
weight which exists between different sub-races of men, both in type and variability. 
Further, the rather extensive material worked over makes it possible to settle the 
problem of the relative variability of the sexes with respect to the weight of the 
brain. On this point Pearson {loc. cit.) has made a brief communication based on 
an analysis of four short series of English data, Bischoff's Bavarian series, and a 
French series. His general conclusion is that in respect of brain-weight the sexes 
are sensibly equally variable. Another interesting problem which merits discussion 
here is as to how the relative variability in brain-weight compares with the vaJ'ia- 
bility of other organs and characters of the body. 
In Table I are exhibited the values of the Means, Standard Deviations and 
Coefficients of Variation together with their probable errors for each of the 
characters studied in this work. While, of course, the primary object of study is 
the weight of the brain, yet it is necessary for the correlation determinations that 
we have the fundamental constants for the other chai'acters, age, stature, etc. To 
effect economy of space the constants for these other characters have been included 
in this first table. The frequency distributions from which the constants have 
been deduced will be found in the fundamental tables, numbered in Arabic 
numerals from 1 to 44, collected at the end of this paper. Regarding the calcula- 
tion of the constants the following explanations should be made. In preparing the 
correlation tables from the raw statistical material the following values for the 
units of grouping were chosen, and used uniformly throughout the work, except 
in certain cases where it was impossible to keep to them. 
