CEREAL IMPROVEMENT SELECTION 183 



inside the tub. In the holes pine plugs are fitted. Rope 

 handles are attached near the top and one tub is set above 

 the other on a trestle. The upper tub is partly filled 

 with water, and the seed poured in, leaving enough space 

 for stirring. After stirring thoroughly, everything float- 

 ing should be skimmed off, the plug taken out, and the 

 water allowed to drain into the lower tub. The seed in 

 the upper tub is then emptied upon a clean floor or canvas 

 and spread out thinly to dry, after which the tubs are 

 reversed as to position, and the operation is repeated. 

 The seed is swollen as a result of this treatment, and the 

 rate of seeding should therefore be proportionally in- 

 creased. By substituting a formalin solution of 1 pound 

 of formalin to 40 gallons of water, in place of water, the 

 smut treatment may be combined with this grading pro- 

 cess to great advantage (Derr, 1910 a, pp. 3-4) (Fig. 60). 

 182. Heavy and light seed. Many experiments have 

 been conducted to determine the relative value of large 

 or heavy, and small or light kernels for seed. While 

 there has been some conflict in results, they have usually 

 favored the use of large, heavy seed. Particular atten- 

 tion has been given to the subject at the Kansas, Ohio, 

 and Ontario experiment stations. An 8-years test was 

 conducted at the Kansas station in which heavy seed 

 oats gave an average yield an acre of 30.90 bushels, com- 

 mon seed 29.89 bushels, and light seed 27.50 bushels. 

 The common seed used was the seed as it came from the 

 thrasher, the light and heavy grades being obtained by 

 the use of the fanning mill. The grades were taken 

 usually from the ordinary seed each year, but in the last 

 year of the experiment, when the greatest difference 

 was noted, the heavy and light seeds were taken from the 

 corresponding grades the previous year. At the Ohio 



