1897] REPRODUCTIVE DIVERGENCE 405 
The fresh example I propose to adduce is founded on actual data, 
namely, the measurements of human stature in the case of parents 
and their offspring, which are given by Galton in his work on 
“Natural Inheritance” (p. 208). Here the relations between 
the stature of 205 parents of each sex and of their adult 
offspring are classified. The mid-parents I split up into three 
nearly equal groups—viz., those below 67:8 inches in stature, those 
from 67°8 to 69°2 inches, and those above 69°2 inches. These 
groups were found to give rise to more or less equal numbers of 
children. The numbers of children of each stature which were 
given birth to by each of these three groups of short, medium and 
tall mid-parents were then determined, partly by plotting out the 
various numbers and calculating from the smoothed curves, and 
partly by direct enumeration of the numbers. By then taking 
means between the numbers of tall children produced by tall parents 
with those of the short children produced by short ones, and also 
between the numbers of short children produced by tall parents with 
those of tall ones produced by short parents, roughly speaking the 
following mean values were arrived at :— 
Short. Medium. Tall children. 
100 short parents give . 54 ol 15 
100 medium . : 5 en 38 ol 
100 tall _ : Sa 05: 50! 54 
100 100 100 
From these figures it is seen that the percentage numbers of children 
are given, and not the actual numbers. This plan was adopted in 
order that the numbers of children might be kept the same as that 
of the parents. We see, therefore, that if 100 short, medium and 
tall parents of each sex be taken, the numbers of short, medium and 
tall children will still remain at 100, in spite of the fact that the 
children of each group produced by the different parents vary from 
15 per cent. to 54 per cent. 
In addition to the children produced by short, medium and tall 
parents, it is for our purpose necessary to know the numbers pro- 
duced by intermarriages of short and medium, and of medium and 
tall parents. These numbers may be approximately obtained by 
taking means between the percentages for short and medium parents 
on the one hand and for medium and tall ones on the other. Thus: 
Short. Medium, Tall children. 
100 short and medium parents give 42°5 345 23 
100 medium and tall . : ae 34°5 42°5 
It is also necessary to assume, as Galton has shown his statistics 
warrant us in doing, that short and tall and also tall and short 
