TESTS OF SELECTIONS OF OATS. 15 



steier; and 1 each from series 31, Burt X Red Rustproof; 62, Sixty- 

 Day; 120, Silvermine; 123, Welcome; 125, Silvermine; and 5938, 

 Sixty-Day. 



In 1911 a comparatively small number of the selections was grown, 

 but a large proportion of those which were dropped that year were 

 again included in the tests in 1912. Of the selections which were 

 grown in the five years, 1907, 1908, 1909, 1910, and 1912, 27 exceeded 

 55 bushels in average yield. Of these 27 selections, 10 were from 

 series 49, 8 from series 34, 3 from series 50, and 1 each from series 31, 

 42, 62, 120, 123, and 5938. 



COMPARISON OF SELECTIONS FROM HYBRIDS AND COMMERCIAL VARIETIES. 



The average yield of all the selections from hybrids tested in 1907 

 was 50 bushels per acre, while the yield for all selections from com- 

 mercial varieties was 49.4 bushels per acre. In 1908 the yield for all 

 selections from hybrids was 55 bushels per acre, while that for all 

 selections from commercial varieties was 49.3 bushels. In 1909 the 

 average yields were 36.5 and 38.9 bushels, respectively; in 1910 

 64.6 and 59.9 bushels; in 1911, 45.9 and 48.7 bushels; and in 1912 

 52.5 and 51.1 bushels. The average for all the selections from hybrids 

 tested for the years from 1907 to 1912, inclusive, was 51.3 bushels, 

 while that for all selections from commercial varieties was 49.8 

 bushels. When the results are taken year by year it is seen that only 

 in the years 1909 and 1911 were the yields of the selections from com- 

 mercial varieties greater than the yields of the selections from hybrids. 



The results, although indicating a higher yield for the selections 

 from hybrids, do not show enough difference to warrant recommending 

 hybridization over straight selection as a means of improvement. 

 One advantage of the selections from commercial varieties is that it 

 is possible to test a great many of them while one is purifying the 

 hybrids from one cross. The purification of hybrids must be con- 

 tinued at least to the third generation before comparative tests can 

 be made; meantime a great many selections could have been tested. 



A somewhat different set of figures is obtained from the average 

 yields of the different series as presented in Table IV. This table 

 shows the annual and average yields of each series, without regard 

 to the number of selections which were grown, for the years 1907, 

 1908, 1909, 1910, and 1912. The results for 1911 are not included, 

 because several of the series were not grown that year. The averages 

 for the selections from hybrids and from commercial varieties given 

 in the summary are obtained from the yields of the different series, 

 as shown in the table. For this reason they do not agree exactly 

 with the averages in the preceding paragraph, where the individual 

 selection, not the series, was the unit. 



