STUDIES ON OAT BREEDING. 9 



The labor of threshing each plant separately was found to be 

 so great when several hundred rows were grown that in some 

 cases rows from which we did not expect to breed were threshed 

 as a whole land the average yield per plant obtained as usual. In 

 no case were plants included which were in any way mutilated. 

 On this account the number of plants was sometimes less than 

 20. 



In the following paper all results are based upon the average 

 per plant of the given character for a given row. It is neces- 

 sary to deal with one character at a time and, as has been men- 

 tioned above, all of the characters show essentially the same 

 thing. We will deal first with the weight of grain p'-f; plant, 

 or the yield. This character is as satisfactory as ar.y of the 

 others, and it has the advantage that it is of practical' im- 

 portance. 



Selection for Yield oe Grain. 



Out of some 200 pure lines S|tarted in 191 1, 28 lines repre- 

 senting 13 varieties were chosen for carrying on this selection 

 work. Table i gives the line numbers by which they are 

 designated, the variety from which they were selected and the 

 average yield of grain per plant for each of the four years 

 that they have been grown. A description of the varieties 

 from which these lines came has been published in another 

 place." In general these pure lines show the characters typical 

 of the variety from which they were selected. In one or two 

 cases, as line Nos. 104 and 183, these pure Mnes are quite 

 different from the commercial varieties from which they came. 

 They are probably derived from mixtures of seed in the origi- 

 nal varieties. 



'"Surface, F. M. and Barber, C. W. Studies on Oat Breeding. I. 

 Variety Tests, 1910-1913. Ann. Rapt, Me. Agr. Exp. Sta. 1914, pp. 137- 

 192. (Bull. No. 229). 



