92 



University of California Publications in Botany [Vol. 5 



Grade Length Breadth Length-breadth index 



1. 0.85 mm. 0.60 mm. 70.58 



2. 0.75 mm. 0.58 mm. 77.33 



3. 0.66 mm. 0.43 mm. 65.15 



Measuring ten fields, indiscriminately chosen, under the low 

 power of the dissecting microscope and including in each field 

 from 90 to 115 seeds, the proportion of the various grades per 

 field was as follows: 



Grade 1 26 % 



Grade 2 46 % 



Grade 3 22 % 



Grade 4 6 % 



The separation of seed according to size was combined with the 

 separation according to weight and will be taken up in this 

 connection later on. 



2. Color 



In color there were practically^ no variations from the typical 

 very dark brown coloration of the seed coats. A few seeds at 

 times seemed to present a somewhat lighter appearance, but their 

 number was so small and the normal color so often found to be 

 present on some one of the other faces of the seed, that no 

 separation according to color was attempted. 



3. Weight 



For the separation of the various weights of seed an im- 

 provised "grader," on the general style of that described in 

 the work on wheat separation in the Kansas Agricultural Experi- 

 ment Station (Lill, 1910), was employed for the first step in 

 the process. A large centrifugal fan on a 110-volt alternating 

 current supplied a constant draft of air which was caught in 

 a large paper funnel and conveyed through a three-foot section 

 of glass tubing, one inch in diameter. The tube was tilted 

 upward from the source of the air-current at an angle of fifteen 

 degrees and the seed introduced through a tin funnel at its lower 

 end. The proper point, near the lower end of the tube, at which 



