476 
Selective Elimination in Staphylea 
somewhat higher than that of the matured series, C, but the difference in the 
relative variability may be in part due to the fact that the mean of A is low as 
compared to that of C. 
TABLE XV. 
Series 
Compared 
Differences fou 
ov0les per locule 
Differences for Total 
Ovules per Ovary 
Positive 
Differences 
Negative 
Differences 
Positive 
Differences 
Negative 
Differences 
All Differences 
B— A 
18 
10 
16 
12 
C— B 
11 
17 
8 
20 
C— A 
7 
21 
7 
21 
Significant 
Differences 
B— A 
10 
8 
8 
7 
C— B 
6 
16 
5 
17 
C-A 
5 
14 
2 
15 
The results for variability are not so conclusive as those for type, but I think 
we may safely conclude : 
(a) That both the eliminated ovaries and the matured fruits have a lower 
absolute and relative variability than the original population of ovaries. 
(6) Thi^ is due to the ovaries with the lower number of ovules failing to 
develop, and to those with the higher numbers having the better opportunity to 
reach maturity. 
Problem 3. Is there a difference in the radial asymmetry of the eliminated 
ovaries and those which develop to maturity ? 
From the nature of the thin- walled, inflated "bladder" fruit we could not 
expect it to be other than somewhat irregular in form. The externally visible 
irregulai ities in the evaginations of the three locales are of such a nature that 
no biologist would describe the fruit as asymmetrical. Externally there is no 
conveniently measurable character from which the amount of ii'regularity might be 
measured. The only character which can be determined with certainty is the 
number of seeds developing or the number of ovules formed per locule. 
Opening the three locules of a fruit quite at random — there being no external 
character to indicate any differentiation whatever between them — we may find 
such numbers of ovules per locule as : 
11—11—11, 
10— 11—10, 
8— 9—11, 
9_ 7_ii, 
11— 7— G. 
