12 Harris. A quantitative study of the Factors Lnfluencing etc. 



end of the pods. That is, seed weight increases towards 

 the Stigma. Tke change seems in some cases to be sen- 

 sibly linear. In other cases, however, the mean weight 

 increases fairly rapidly at first, then the rate of in- 

 crease falls off and finally the seeds become lighter 

 towards the distal end of the pod. 



Cold Spring Harbor, N. Y., U. S. A. 

 Oct. 23, 1912. 



Explanations of Diagrams. 



Figure 1. Regression straight lines and their eni])irical means for the 

 relationsbip between ovules per ]>od and weight of seeds, and for seeds per 

 pod and weight of seed. Relationships for ovules represented by solid dots 

 and firm lines; mean weight for seeds represented by circles and broken lines. 

 Note the small slope of the lines and the great irregularity of the means. 



Figure 2. Percentage of ovules developing into seeds in various positions 

 in the pod. 



Figure 3. Regression straight lines and observed means showing change 

 in weight from the base to the tip of the pod in the NH series. Note that 

 regression is clearly non-linear. 



Figure 4. Regression lines for the GG series. Regression is apparently 

 linear within the limits of the errors of sampling. 



