378 FIELD AND GARDEN PRODUCTS 



The first two years of the experiments, the Lightest seed pro- 

 duced least, the ordinary was next in yield, and the heaviest yielded 

 best. Since then, results have varied from year to year, and the aver- 

 age for the eight years does not show a marked advantage for either 

 grade of seed. 



The weight per hushel has been recorded in five different years, 

 but no average difference has been noted. After eight years' con- 

 tinuous selection by the fanning mill, it was not possible by careful 

 examination in 1907 to note a difference in either the quality or the 

 quantity of the crops produced from the light and heavy seed. 



Quite a little work has been carried on in both the United States 

 and Canada bearing on this subject. At the Ontario Agricultural 

 College, Guelph, Canada, where large seed has been compared with 

 medium or with small seed in the case of the principal cereal, and 

 root crops, the best results have uniformly been secured from the large 

 seed, and the poorest results from the small seed. Also in the case of 

 an experiment carried on for thirteen years in which the large, black 

 seeds of Joanette oats have been compared with the small, light- 

 colored seeds, the result has been very much in favor of the large, 

 black seeds. The seeds for these experiments were all selected by 

 hand, and not by fanning mill. The data do not show, however, the 

 yield of the selected large seeds in comparison with the crop not so 

 selected. In the United States a number of experiments conducted 

 only one or two years have been reported where large or heavy seed 

 was compared with small or light seed, with results slightly in favor 

 of the heavier seed. However, where continuous selection has been 

 carried on for more than two years the data do not indicate an 

 advantage in favor of heavy or large seed. It should also be noted 

 that in most of the experiments the comparison has been made of 

 very heavy or very large seed with small or shrunken seed, rather 

 than of the large seed with the original stock of seed, not so selected. 



It is easy to conceive that differences might be obtained in the 

 former case, particularly if wide extremes were selected, when no 

 apparent difference might be found in the latter. To compare the 

 selected seed with straight, unselected stock is the only practical test. 



At the Ohio Station, experiments have been carried on for eleven, 

 years by continuous selection, in which three grades of seed were 

 used, viz.: First grade, the large grains; second grade, the best 

 grains passing through the sieve in screening out the first grade, 

 third, unscreened wheat as it came fom the thrashing machine. The 

 average yields for eleven years show no difference in yield or quality 

 of grain. Three years' work with a greater difference in the 

 quality of the different grades has given similar results. In sum- 

 ming up, Professor Williams makes the following statements : 



Experiments extending over a series of thirteen years have failed 

 to show on the average any gain from the use of seed from which the 

 small and light grains have been removed by use of the fanning 

 mill, although three seasons out of the thirteen the first grade gave 

 largest yields. 



North Dakota, after four years' hand selection and comparison 



