J. A. Harris 469 
Designating the parental generation seeds by x and the offspring seeds by y, I find 
from Table XV, 
% ( x ')= 15416, 2 (</') = 1^097, 
S (a?'«) = 145400, 2 (v/' 2 ) - 12.3273, 
S OY) = 130794, 
whence, with Sheppard's correction, 
r = (130794/1668 -9-242206 x 8-45143!>)/r29l727 x 1/547424 = -1519. 
From Table XIV I find for the ascendant-descendant correlation 
8(pq) = 00750, 
8 [qX (.//)] = 564066, 8 [pi (y')] = 514990, 
8[qS («' 2 )] - 5351080, <S [p% ( v/' 2 )] = 4518074, 
S[Z(a>') X (y')] =4798171. 
These give 
r = (4798171/60756 - 9 284120 x S470304)/l-34041 2 x 1-559539 = -1334. 
But obtaining the summations about 0 for the avuncular relationship by simple 
subtraction I get for the population N 
# [q(p- l)] = S (pq) - S(q) = 00756 - 1608 = 59088, 
and for physical constants, for " aunts " 
x = (564066 - 15416)/59088 = 9-285303, 
tr x = V(53510S0 - 145400)/59088 - (9-285303)- - "083333 = 1-341743, 
and for " nieces " 
y = (514990 - 14097)/59088 = 8-477068, 
cr y = V(4518074 - 123273)/59088 - (8477068) 2 - -083333 = 1-559872, 
and for the relationship between them 
r = [(4798171 - 130794)/59088 - ^/]/a x a y = -1329. 
a child's weight at birth. Second, all the seeds produced by the same plant are under closely similar 
environmental conditions. Thus while they are (so far as the embryo to which weight is chiefly due) 
sister seeds, they are far more nearly alike than would he expected on the basis of heredity alone, for all 
are borne in and nourished by the maternal tissues. 
Under these circumstances one may consider either (a) the correlation between the individual seed 
planted and the individual seed produced, or {b) the correlation between the average character of all (or 
of a representative sample) of the seeds of a parental and filial individual to be the true correlation. 
If the latter view is taken one considers the characteristics of the seeds to be determined exclusively 
by the plant on which they are borne (the parent plant). 
If the former alternative be adopted one neglects for the moment the important environmental factor 
introduced by the fact that the whole array of parental seeds are subjected to the limited environment 
of one plant while the offspring array (produced by any single plant) is subjected to the environment of 
another plant. 
It has seemed to me safest to follow strictly the biological homologies in terminology. 
Biometrika ix 60 
