304 
PROFESSOR KARL PEARSON AND MR. LESLIE BRAMLEY-MOORE, 
= -6337, M,; = '6525, 
(r^='1888, o-j = '1643, 
r,, = '0831. 
The coefficient of regression = '0945. 
The probable error of the correlation is ‘0193 and of the regression* •0195. Thus 
these quantities are four to five times their probable errors, and we conclude that 
fecundity is certainly inherited. 
The intensity is far below that suggested by the law of ancestral heredity, but it 
nevertheless exists. Its lowness is probably due to the fictitious character of the 
fecundity owing to the causes indicated on pp. 298-9. An attempt must now be made 
to eliminate some of the factors disguising the fecundity, but to do so is by no means 
so easy as in the case of fertility in man. 
(iii.) My first idea was that by taking a higher limit to the number of coverings a 
closer approach might be obtained to the true, fie., the inherited fecundity. 
Accordingly Table XIII. was formed for the correlation of 1000 mares and their 
dams, when the minimum number of coverings was eight. But I did not recognise 
that this would give far greater weight in the Table to the older mares, and that 
accordingly causes (cl) and (i) of p. 298 would now play a much larger part in disguising 
the true fecundity than before. There appears to be no limit to the number of 
times a famous old mare may go to the stallion when there is very small hope of any 
offspring. 
Table XIII. gives us the following results : 
M„ = -6300, = -0360, 
cr„, = '1633, (Tj = 'ISOO, 
r„,; = -0652. 
The coefficient of regression = ‘0708. 
The probable error of the correlation is ‘0212, and of the regression '0213, both 
less than a third of the observed values. We should agfain conclude from this result 
that fecundity is inherited, although it offers less strong evidence than the previous 
case. The influence of selection+ is at once apparent in the great reduction of the 
variabilities. The fact that we are throwing the determination of fecundity more 
on to the old age period of life appears from the reduced mean fecundities. I 
attribute the reduction in the fecundity-correlation to this source, i.c., the very diverse 
treatment which old mares receive at the hands of different breeders. 
(iv.) I made another attempt to remove screening causes by taking 1200 more 
* Peakson and Filon : ‘ Phil. Trans.,’ A, vol. 191, p. 214. 
t The effect of such a selection as the above in reducing correlation is dealt with in iny paper on the 
inHuence of selection on correlation. 
