J. A. Harris 
321 
seeds developing than upon the number of ovules formed. We are dealing, how- 
ever, with a form in which the correlation between the number of ovules formed 
TABLE 11. 
Comparison of Correlation for Fruit Length and Number of Ovules 
and Fruit Length and Number of Seeds. 
Kelationship 
Correlation in 1906 
Correlation in 1907 
Length and Ovules 
Length and Seeds 
Difference 
■527 ±-015 
•576 ±-014 
•656 + -019 
•688 ±-017 
•049 ±-021 
■032 ± -026 
and the number of seeds developing is high. Under such circumstances these 
results would be expected. The problem should be worked out on some form in 
which there is only a moderate correlation between the number of ovules and 
number of seeds. 
5. The Relationship between the Number of Ovules Formed and the 
Number of Seeds Developing. 
Our knowledge of the physiology of seed production is very limited. We know 
practically nothing of why one ovule in a locule develops into a mature seed and 
another beside it fails to develop. General explanations, such as " lack of 
nutrition," " failure to be fertilized," etc., are thoroughly unsatisfactory. Experi- 
mental evidence would be the most desirable, but until it is forthcoming I think 
we shall have to depend upon the biometric analysis of large bodies of data to 
enable us to get some insight into the problems. 
TABLE 12. 
Correlation for Fertility Characters in Sanguinaria. 
Eelatiouship between Number of Ovules 
Formed and Number of Seeds 
Developing 
Correlation and Probable Error 
1906 
1907 
Difference 
For First Placenta 
For Second Placenta 
For First and Second Placenta ... 
For the Fruit 
■7901 + •0080 
•8037 ± -0076 
•7970 +'0055 
■8141 + ^0072 
•7998 + -0121 
■7905 ±-0127 
■7951 ± ^0088 
•8106f0116 
-•0019 ±-0104 
- -00351 •O 136 
The first question which suggests itself for investigation is whether an ovule 
formed in a pod producing a large number of ovules is more, or less, likely to 
