38 
Data for the Problem of Evolution in Man 
In other words, reproductive selection may tend to preserve the threescore years 
and ten set to the duration of man's life by the Psalmist. But the American 
statistics show that this tendency is not universal, and further, whether it exists or 
not, does not seem to depend on the relative density of population and the 
character of the environment. 
C. On the Stature of Man between the Ages of 15 and 85. 
Tables VI. and VII. give the heights of male and female criminals, with their 
ages based upon New South Wales experience. The criminals from 20 years 
upwards are for the years 1890-9, and under 20 years for the years 1895-9, each 
year's record being for distinct persons. Mr Powys writes : " I can imagine no 
solid reason why the heights of criminals should differ in any material way from 
those of the ordinary population, and the figures shown can therefore be regarded 
as typical of the latter." 
In Australia probably a majority of the population live an outdoor life with 
plenty of food and exercise, and the differentiation of the criminal from other 
classes may be small. Iti Great Britain there is undoubtedly in strength and 
stature a considerable class differentiation. The upper middle classes are dis- 
tinctly taller than the labouiing classes, and in particular than the town hand- 
workers*. The criminal here is probably of inferior physique, not because he is 
a ci'iminal, but because in the bulk he is drawn from a differentiated class-f. The 
mean stature of 1077 Cambridge undergraduates is 68"'9, and of 811 persons 
measured in Mr Galton's South Kensington Laboratory in 1884, 67""9. The mean 
stature of the New South Wales male criminals is 67"'?), taking ages 25 to 30 
for comparison. In the case of the women we have Cambridge 63""8, South 
Kensington 63""3, and New South Wales not exceeding 62""8. It seems im- 
probable that the general population of New South Wales is shorter than that 
of England; hence we shall be justified iu believing that to some extent the 
New South Wales criminals are drawn from a class or classes less fully developed, 
owing to want of nourishment or to general environment, than the popylation at 
large |. But looking at the curves for the male criminals grouped in either 5 or 
10 year periods we see that the material is extremely homogeneous. In every 
case the curve is of Type IV. §, in every case the mode is nearly identical with 
the mean, and in every case the fit is extremely good. For the males we should 
have got very good results by simply using the normal curve, and thus we have 
* See Pearson: The Chances of Death, Vol. i. pp. 310—316. 
f The school of Lombroso has jumped without a satisfactory statistical investigation at the 
existence of a definite ' criminal ' type. Further light on the matter will be found in Dr W. E. 
Macdonell's paper on Criminal Anthropometry which will be shortly published iu Biometrika. K. P. 
X For 1000 British middle-class from my own measurements (without boots) I find 68"-04 for 
males and 62"-y5 for females, both sensibly taller than the New South Wales criminals. K. P. 
§ Phil. Trans., Vol. 186 a, p. 370. 
