Ovules or seeds developing per pod 
122 
BOTANICAL GAZETTE 
[ AUGUST 
The regression line and the empirical mean length for the different 
numbers of ovules per pod are shown in a diagram (fig. 1). To the 
eye the agreement is very good, except at the upper extreme, where 
the observations are relatively few. We may obtain some idea 0 
55 
Mean length of pod in millimeters 
60 65 70 75 80 
1 o 
2 <i 
3 ve ee 
~ 
4 ck 
Ts 
Y\ 
5 ss 
~ 
6 \ e 
7 
Fic. 1 
seeds; solid line, theoretical length of pod 
pod for 
for seeds 
.—Diagram of slope of regression: straight lines, length on 
85 
ovules and 08 
for ovules; dotted line, theoretical length of 
seeds; solid dots, observed length for ovules; circles, observed average length 
the amount of discrepancy between the theoretical line and empirical 
means by taking the average deviation on both sides of the line 
by the equation. Considering the weight to be attached 
given 
to a deviation 
to be proportional to the number of observations on which it is based, 
