6 
Harris, A quantitative study of the factors infhiencing etc. 
Apparently, therefore, the pods, in which the nnmher of 
seeds is above the average, produce lighter seeds than those of 
average (or lower) fertility. But the difference is not large. The 
influence of seed number is shown by the second term of the re- 
gression equations, which are: 
of Plants 
Regression Straight 
Line Equation 
Rate of Cha: 
in Grams 
L 
W = 17.4473 — .5241 s 
.0131 
LL 
W = 14.8753 — .2508 s 
.0063 
GG 
W = 18.4166 — .2356 s 
.0059 
NH 
■ W = 9.8102 — .0719 s 
.0018 
ND 
W = 7.5299 — .0097 s 
.0002 
These straight line equations are also represented in Diagram 1. 
The absolute change in seed weight as one passes through a ränge 
of one seed per pod is clearly very slight, amounting in the ex¬ 
treme case to only 13/1000th of a gram, and averaging only 
55/10,000th gram. 
These results are in agreement with those already annonnced 
for Siapliylea and Cladastris A) 
If the correlations between number of seeds developing per 
pod and seed weight be negative, then one might expect those 
for number of ovules failing to develop per pod and seed weight 
to be positive. Bat since both the number of ovules and the 
number of seeds are concerned in determining the number of ovules 
which fail to develop, one might expect these results to be more 
irregulär than those for number of seeds. This is, as a matter of 
fact, the case. Three of the constants are positive and two are 
negative. The two negative constants taken alone are the least 
probably significant of the flve. The mean value for the five 
series is -f- .065. 
A priori, one might suppose that the closest of all the inter- 
dependences would be that between the relative number of seeds 
which develop per pod and weight. The actual correlations between 
the ratio — e( ? S ^ G1 are shown in the final column of the table. 
ovules per pod 
Three are positive, two negative. The two positive constants 
cannot be considered significant with regard to their probable errors. 
The largest is —.199. The average is —.073. Thus, the. corre¬ 
lations for relative number of seeds per pod substantiate those for 
actual number of seeds per pod, although they are more irregulär 
and on the average lower. 
The thing which impresses one about these constants is this: 
there is but a very slight correlation between the fertility characters 
of the pod and the weight of the seed which it produces. These 
results are in direct contradiction to the kind of Statements which 
are sometimes made by biologists. They show what caution must 
9 Harris, J. Arthur, Seed Weight in Staphylea and Cladastris. (Torreya-, 
Vol. XI. 1911. p. 165—169.) 
