STUDIES ON OAT BREEDING. IO9 



all mutilated plants were discarded. The data were recorded on 

 5" x 8" loose leaf sheets like that shown in Fig. 17. 



Since the number of threshed plants varied in different rows 

 the average per plant of a given character was used in com- 

 paring different rows. 



From the notes and the data thus obtained a further selec- 

 tion was made and only the very best rows were selected for 

 propagation the next year. 



Work in 1912. 



Eighty-two rows from the 191 1 garden were regarded as 

 good enough to be tested further. In one case two of these 

 rows belonged to the same pure line. In another case three 

 rows were from the same line. There were then 79 pure lines 

 tested in 1912. Seeds from these selected rows were planted 

 by hand in small plots. These plots were four feet, six inches 

 wide and four feet, ten inches long, giving a total area of 21.75 

 square feet, or 1-2000 of an acre. The grain was planted in 

 rows three inches apart and one and one-half inches between 

 the plants in the row. There were in each plot 18 rows with 

 39 plants to the row. This makes 702 plants per plot or at 

 the rate of 1,404,000 plants per acre. This is about the average 

 number of grains of a medium sized oat sown on an acre when 

 seeded at the rate of two bushels. Each pure line was planted 

 in duplicate plots of this size and the average of the two plots 

 for each character was taken for comparison of the different 

 pure lines. 



The planting of such plots has been greatly facilitated by 

 the use of a planting board like that shown in figure 18. This 

 board contains rows of 3-8 inch pegs spaced at the proper dis- 

 tances. 9 



During the growing season detailed notes were taken regard- 

 ing each plot and when ripe the plots were harvested and later 

 threshed. On the basis of these data the pure lines were again 

 subjected to a severe selection. Out of the 79 pure lines tested 

 in these plots 34 were chosen for testing in larger field plots. 



The threshing data from these 34 pure lines are given in 

 Table 2. The measurements are recorded in centimeters and 



"The board shown is designed for marking only half a plot and varies 

 somewhat in shape from that described above. 



