1910] HARRIS—FERTILITY IN CERCIS 119g 
ties. There seem to be at least four ways in which a relationship 
between the number of seeds developing and the size of the fruit 
might arise. 
a. Fruits with a larger number of ovules are apt to be larger, if 
for no other reason, simply because the placental space required is 
greater. In the long run fruits with larger numbers of ovules also 
have larger numbers of seeds,5 and an influence on fruit length at 
first attributed directly to the development of the seeds might be due 
indirectly to the number of ovules. 
b. The space required for the matured seeds might, through the 
purely mechanical effects of crowding, result in the fruits with greater 
numbers of seeds being larger in size. 
c. The developing seed might by means of some excreted product, 
or in some manner not yet suggested, directly induce a greater develop- 
ment of the ovary. 
d. Both the number of ovules developing into seeds and the size 
attained by the fruit may be to some measure dependent upon some 
other character; say, for example, the position of the fruit on the 
inflorescence axis, or the number of fruits developing per inflorescence. 
The correlation between them might then be due merely to their 
mutual dependence upon some other character. 
In the literature one finds only a few references to the relationship 
between the fertility of a fruit and its size. EWART (1) gives tables 
showing numbers of seeds and mean weight of fruit for three series 
of 125, 48, and 134 fruits of one variety of apples, which indicate 
that weight increases with’ number of seeds. According to EwART, 
Métrer-Tuurcav found that in apples and pears the size of the fruit 
and the number of seeds are interdependent. Mutrer-THURGAU 
cut off four of the five stigmas in the pear blossom, fertilized the re- 
maining one, and thus produced asymmetrical fruit. EWART con- 
cludes: “Es steht demnach ohne zweifel fest, dass den Kern einen 
Wachtumsreiz auf die zur Fruchtbildung bestimmten Gewebe 
ausiibt.”” 
In view of the four possibilities suggested above, I think it is 
quite evident that great caution should be used in asserting that the 
number of ovules developing into seeds has per se any influence upon 
‘ This statement is based on th Its ies of unpublished observations 
SCLICS 
- 
