194 Pigmentation Survey of School Children in Scotland 
hair colour certainly exists, although no statistics are forthcoming to prove its 
intensity. The average observer will have noticed that the offspring of parents, 
one fair and one dark, are not uniformly fair and dark, but have also on an average 
among: their number members of the brown-haired or medium class. What the 
proportions of each are, on an average, will be revealed by observation. What 
form the distribution takes does not affect the argument. Granted that pangamic 
mating (not excluding other forms) now exists for hair colour among the Scottish 
people and granted blended inheritance as probably occurring as one of the results, 
and the phenomenon of regression will appear in hair colour. The colour of future 
generations of offspring will tend to become brown-haired and in a few generations 
a brown type will be established breeding true to itself. Thus in densely populated 
areas where greater opportunities for random mating exist, a greater proportion of 
medium hair will arise, granting blending of hair colour as an appreciable factor, 
but not of course debarring exclusive and even particulate inheritance as operative 
as well. This alone, or together with the suggested greater fertility of the medium 
haired class, would explain the excess of medium hair found in densely populated 
areas particularly in and around Glasgow, an excess which is not explainable by 
the presence of non-Scottish or Scoto-Keltic elements in the population. As has 
been said before, it cannot be proved from the present data what is the cause of 
the excess, and the foregoing is only the probable explanation. The proof or 
otherwise of the validity of the theory will be forthcoming when the results of 
direct observations on parents and offspring have been made, tabulated and 
analysed. 
VII. Colour classes ivhich are associated geographically, (a) Hair classes 
luhich are associated with one another. — The theory that brown hair is really a 
blend of fair and dark is supported by the fact that throughout the country excess 
of the class is not generally associated with excess of other hair colour classes. In 
order to determine the extent of the association of excesses and otherwise of the 
various colour classes, the percentages of all the classes were compared with one 
another and the correlation coefficients determined. The following table (Table XL.) 
gives the numerical values of the correlations of each class with all the other 
classes. One must be careful as to the meaning of the result. Association of 
excesses of fair hair and blue eyes (a positive correlation) does not necessarily 
mean from this portion of the analysis that the blonde type predominates in the 
region of excess. All the analysis tells one is that regions of excess of fair hair 
are also regions of excess of blue eyes. This will be evident when one considers 
the other associations with fair hair. Examining the table it will be seen that 
regions of excess of jet black hair are also generally regions of excess fair and 
dark. This combination could not obviously occur in the same person. Regions 
of excesses of fair and dark indicate the presence of two types— a heterogeneous 
and not a homogeneous population. On the other hand, examine the column 
indicating the associations with excess of medium hair. Excess of medium hair as 
a ride is associated with excess of no other colour class. The negative correlations 
