224 P if/mentation Survey of School Children in Scotlaiid 
of each group will be accurately ascertained. Again, there are the colour characters 
of groups of families as revealed by surnames to be considered. A tabulation and 
analysis of the colour characters of surname groups for eacli surname would show 
whether they were reall}' associated, like family groups, or were merely samples 
of the general population. The degrees of resemblance of brothers and sisters 
would be determined on numbers hitherto undealt with and would confirm or 
otherwise the measures found from the numerically smaller English data. Finally, 
the degrees of resemblance between the various kinds of cousins, an investigation 
suggested to the author by Professor Karl Pearson, await determination*, and 
the determination cannot be made until the almost overwhelming mass of data 
bearing on cousinships has been also tabulated. 
(/S) Comparisons. The correlation between hair and eye colour has been 
determined, the contingency method being used, for one Scottish group, namely, 
19,279 school children of the city of Aberdeen, and also for 1000 children taken at 
random from the entire pigmentation data. The following two tables give re- 
spectively (Table LXI.)the results of the observations of hair and eye combinations 
in the city of Aberdeen, and (Table LXII.) the values of the contingency coefficients. 
The author's results for other Scottish populations and those from British and 
continental returns are given alongside for the purpose of comparison. 
TABLE LXI. 
Hai7- and Eye Table. 19,279 Children in the City of Aberdeen. 
Hair. 
Fair 
Red 
Medium 
Dark 
Jet Black 
Totals 
Blue 
110.5 
131 
885 
348 
1 
2470 
Light 
2285 
405 
2434 
851 
9 
5984 
Medium 
1208 
360 
3242 
1601 
29 
6440 
Dark 
r 
366 
209 
1621 
2094 
95 
4385 
Totals 
4964 
1105 
8182 
4894 
134 
19279 
These results show, if it is a mark of racial purity of any race to have its 
individuals all of one hair colour and of one eye colour, that the Prussian school 
children are relatively more homogeneous than the Scottish school children, and 
that the latter in turn are more homogeneous than the British schoolboys 
generally, since the value of the correlation is lowest in the case of the Prussian 
children and highest in the case of the British schoolboys. It may be here noted 
that if two races, one of the blonde type and one of the brunette type, were 
present in a population in equal proportions, the degree of correlation between 
hair colour and eye colour would be equal to unity. On the other hand, (1) the 
* The author intends to hand over the classified data on cousinships to Professor Pearson as soon 
as they have been abstracted and tabulated. 
