F. A. Woods 
77 
The four-fold coiTelation table shows us at once, that the inheritable influence 
in the tendency to produce an excess of males must be very slight. Working 
it out carefully we find the coefficient, r, practically zero, and well within the 
probable error. 
Parental. 
.3 
Fraternities showing 
an excess of males 
Fraternities showing 
no excess of males 
Totals 
Fraternities showing an exce.ss of males 
Fraternities showing no excess of males 
291 
303 
423 
448 
714 
751 
Totals 
594 
871 
1465 
/i = -2392224, /r= -3876579, 
-0316591, A'= -3987424. 
This gives the equation : 
•006628 =r + -003787'r2-f--15698r3 + ... the root of which is ■/■= -0066 ±-0305. 
I have also selected those cases in the foregoing tables in which an excess of 
males happened on both sides of the house in the ancestral (parental) generations, 
and have sought to find if here an excess of males might not be shown among their 
children. Instead of an excess of males there were but 334 males against 351 
females born in such families. Similarly the families with an excess of females in 
both sides of the ancestry produced but 357 female children against 402 males. 
Thus we may conclude that the determination of sex, in man at least, can be 
shown to be unaffected by hereditary influence. This agrees with the statistical 
conclusion of Simon Newcomb (4) obtained by a different method. 
Nor does it seem probable that any Mendelian principles control the determin- 
ation of sex in man, for then we should expect some correlation in the distribution 
of the sexes in successive generations due to the union of dominants with each 
other, and also due to the union of recessives with each other. 
These statistical proofs which lead us to a definite conclusion of non -inheritance 
have an important bearing upon several theories regarding the determination 
of sex. If sex is largely determined by agencies acting .upon the young and 
supposedly indifferent embryo, even if these were largely external (nourishment, 
temperature, etc.), the constitutional peculiarity of the mother would have, under 
ordinary circumstances, a large share in forming these differences of environment. 
As we know that constitutional peculiarities are to a measurable degree in- 
herited and capable of giving us a correlation coefficient, and as we here find no 
such coeflGcient, we see an argument in favour of the view that sex is not deter- 
mined during gestation. 
