424 
Miscellanea 
If x-2, x 1 and x {) denote the positions of the means of the (A A), (Act) and (act) rows we find 
_ m + 1 2 
?n + 3 1 u m + 3 
The above value of r will be reasoriably accurate only as long as these two expressions do 
not greatly differ, i.e. in the neighbourhood of m = \. Actual values of the expressions are : 
m 0 
■5 
1 
2 
10 
00 
x-2 — X\ "33 
•43 
•50 
•60 
•85 
1-0 
•57 
■50 
•40 
•15 
0 
Thus, except when in is very close to unity, the expressions differ considerably, and the above 
formula for r cannot be taken as giving a true value of the correlation for such a table. If we 
approximate by taking the mean of the values of x 2 — X\ and x x — x 0 weighted according to the 
frequency on which each is based, we reach results which in the present case are not far different 
from the values given by Dr Brownlee's formula. In another case, however, which we have 
worked out the difference is greater. If we do not weight, the difference may be very great. 
We have by no means exhausted the possible criticisms of Dr Brownlee's work. Sufficient 
has been done, however, to show that his results must be interpreted with very great caution. 
Finally, in all cases of the determination of correlation, we would emphasize the necessity 
for a clear understanding of the nature of the distributions dealt with. 
III. A Mendelian View of racial Heredity. 
By H. J. LASKI, New College, Oxford. 
The application of Mendelian principles to the problems of inter-racial heredity has not thus 
far met with any very conspicuous success. The work of Davenport on the mulatto is quite 
inconclusive in character. The incomplete pedigrees of Mudge merely defeat the end which 
doubtless they were intended to serve. It is not, indeed, too much to say that the endeavour to 
make man a complex of sharply-defined unit characters has failed, and failed completely. Even 
the researches of Hurst, which were received with an almost lyrical enthusiasm by the adherents 
of Mendelism, are not above suspicion; for post-mortem microscopic examination has revealed 
the presence of anterior pigment in eyes or parts of eyes which, after most careful examination, 
were judged free from it in life. Undeterred, however, by previous failures Dr Redclifi'e Salaman * 
has endeavoured to prove the applicability of Mendelian theory to the problems of Jewish 
heredity. His paper is so characteristic an example of much recent Mendelian work that it is 
not entirely purposeless to examine it in some detail. 
Its conclusions may, at the outset, be briefly summarised. Dr Redcliffe Salaman investigated 
the transmission of a Jewish facial characteristic which, somewhat mysteriously, he calls 
Jewishness. His method was to collect cases of intermarriage between Jew and Gentile, and 
thus to discover whether the resulting offspring were Jewish, non-Jewish, or intermediate in 
appearance. He was assisted throughout by Jewish observers who, he states, seemed desirous 
whenever possible of claiming the child as Jewish +. This is, perhaps, a little curious in the 
light of our knowledge of the Jewish aversion for intermarriage and their known desire to 
dissociate its consequences from anything Jewish. It is remarkable that the bias of the 
observers influenced their judgment so little if the personal equation is so high as Dr Salamau 
* Journal of Genetics, \Yo\. i. pp. 273—290. 
f Loc. cit. p. 280. 
