J. A. Harris 
317 
object to it as a working hypothesis, and I dare say that most would take it for 
granted. Now it is important to know whether the vigor of the plant shows 
itself more in the number of seeds developing — a character which is not definitely 
and unchangeably fixed until after a considerable period of physiological activity — 
than in the number of ovules formed — a character definitely and unchangeably 
established relatively early in the development of the organ we are studying. 
If vigor finds fuller expression in the number of seeds developing than in the 
number of ovules formed we should expect to find a higher coefficient of correlation 
between the length of stalk and the number of seeds than between the length of 
stalk and the number of ovules. The comparison is made in Table 6. 
TABLE 6. 
Comparison of Relationship between Stalk-Length and Number of 
Ovules and Stalk-Length and Number of Seeds. 
llelation between Stalk-Length and 
Correlation and 
Probable Error 
Ovules on First aud on Second Placenta . . . 
Seeds on First and on Second Placenta ... 
Difference ... 
Total Ovules per Fruit 
Total Seeds per Fruit 
Difference 
•315 + -019 
■,348 ±-019 
+ -033 ± -027 
•323 + -01 9 
•.363f019 
+ ^040f027 
It appears that while actually the correlation between length of stalk and 
number of seeds is slightly higher than that between length of stalk and number 
of ovules, the difference is so small as to be insignificant in comparison with its 
probable error. 
4. The Relationship between the I^ength of the Fruit and its Fertility. 
That there should be a fairly close interdependence between the size of a fruit 
like that of Sanguinaria and the number of ovules formed and the number of 
seeds developing is to be expected. The placental space in a fruit of this species 
is pretty well filled up with ovules, and if there are many ovules the placentae 
must be longer than if there are only a few. The demonstration by means of 
a " formidable mathematical apparatus " of this rather obvious fact is not in itself 
a matter of much weight, but it is of importance to measure the degree of 
resemblance instead of merely asserting that a relationship exists, and to express 
the result of the measurement in such terms that it may easily and directly be 
compared with like measures in other species. 
The constants for both lots of data are compared in Table 7. 
Biometrika vii 41 
