J. A, Harris , 
479 
not done this is that I have used the coefficient of asymmetry in a large series of 
investigations on the relationship between asymmetry and fecundity in this and 
other species*. In them I could not well use the relative coefficient because of 
the danger of introducing spurious correlation in determining the interdependence 
of asymmetry and fecundity. It seems desirable, for reasons which will be 
apparent when my data for fecundity are ready to publish, to use the same method 
of describing the irregularity of the ovaries in both studies. 
The chief objection to the use of the coefficient of asymmetry will be removed 
if we show that the magnitude is not greatly dependent upon the number of 
ovules per locule. This must be done for every species investigated. For the 
present series I have discussed this point in Section VI (see page 493) where it 
has been shown that the correlation between the coefficient of asymmetry and the 
number of ovules per locule is very slight. 
Table XVI gives the interlocular difference for all the coefficients of asymmetry 
found in the present study. Anyone who desires may calculate all the constants 
dealt with in terms of interlocular differences by the use of this table and the 
tables of data. 
TABLE XVI. 
Interlocular Differences for Coefficients 
of Asymmetry. 
Coefficient 
of 
Asymmetry 
Interlocular 
Difference 
■0000 
0 
•4714 
2 
•8165 
4 
•9428 
4 
r2472 
6 
r4142 
6 
1-6330 
8 
1-6997 
8 
1 -8856 
10 
2-0548 
10 
2-1602 
10 
To determine whether there is a selective elimination depending upon the 
asymmetry of the fruit, the asymmetry of each individual fruit is determined, and 
the distribution of the asymmetries of the eliminated and uneliminated ovaries 
compared. The determination of the asymmetries of individual fruits, a rather 
laborious process at best, is facilitated by tabling the coefficients for all the 
different ovule formulae at the outset. 
A standard deviation calculated on only three " observations " — the three 
locules of an ovary — is of little value as a means of predicting the standard 
* I hope that some portions of these investigations on fecunJity will be ready for publication in 
a few months. 
61-2 
