CORRELATION OF CEPHALTC MEASUREMENTS IN 
EGYPTIAN BORN NATIVES. 
By MYER M. ORENSTEEN, F.S.S. 
{Ministrij of Public Works, Cairo, Er/i/pt.) 
1. Object of the Paper. A previous paper on the "Correlation of Anthropo- 
metrical Measurements in Cairo Born Natives*" pointed out that the coefficients 
of correlation resulting from the length and the breadth of the head vary much 
from one race to another. 
The significance of this variability is suggestive and the writer accordingly 
undertook another investigation. 
The present paper is intended to exhibit fresh coefficients of correlation for the 
cephalic measurements and is therefore a help, to the study of the variation in 
question. The analysis into the variability will be confined to a separate paper 
already in hand. 
2. The Material. Full particulars as to the nature of the raw material have 
been given in the above-mentioned paper. Suffice to repeat here that the statistics 
were drawn absolutely at random without any special selection whatever and that 
they are. fairly representative of the ordinary community which they stand for. All 
persons measured are adult males, i.e. individuals of 20 years of age and above, and 
each has been classified according to his place of birth. The measurements were 
taken during the period 1901 to 1906. 
3. The Computations. The total number of individuals involved in the com- 
putations is 9430. Considering the large amount of the data available, the 
observations have been grouped into frequencies where a standard unit of 3 milli- 
metres was adopted. 
This grouping was carried out throughout the statistics with the result that at 
least a dozen sub-groups was obtained for each character in each correlation-table. 
It will be seen that the frequency curves have high contact at both ends, 
therefore Sheppard's adjustments were applied to the moments before calculating 
the standard deviations. 
All constants have been computed twice and independently by removing the 
base-group by a unit of grouping. 
4. The Constants. Table I is self e.xplanatory. iV is the number of individuals 
entering in each classification, A and a are the mean and corresponding standard 
deviations of the group, and and their respective probable errors. 
* Biometrika, Vol. xi. Nos. 1 and 2, November 1915. 
Biornetrilsa xiii ■ 2 
