STUDIES ON OAT BREEDING. 



31 



Indices of Successive Selections in the Same Direction. 



As a matter of completeness it will be well to examine the 

 indices for those rows which have been subjected to several 

 successive selections in the same or different directions. 



Table 10 shows the mean indices for the four classes ot 

 selection possible in the 1913 rows. Table 11 shows the mean 

 indices for the possible kinds of selections to which the 1914 

 rows have been subjected. 



Table 10. 



Mean Indices^ of Selection for Each of the Four Kinds of 



Selection for the 19 13 Rows. 



Character of 



Number 

 of rows. 



Mean 



Index. 



Mean Index Omitting 

 Line No. 262. 



the selection. 



1911 

 selection. 



1912 

 selection. 



Number 

 of rows. 



1911 



selection. 



+ + 



+— 

 — + 



67 

 65 

 59 

 67 



.0.7350 

 0.8097 

 1.9525 

 0.9698 



0.9252 

 0.8615 

 1.0716 

 0.9766 



65 



64 

 57 

 66 



1 . 1044 

 0.8241 

 1.1481 

 0.9698 



Table ii. 



Mean Indices of Selection for Each of the Possible Kinds of 

 Selection for the 1Q14 Rozvs. 



Character op Selection. 



Number 

 of rows. 



Mean Index. 



1911 

 selection. 



1912 

 selection. 



1913 

 selection. 



+ + +■ 

 + +— . 

 — + +. 

 +— +. 

 +— . 

 — +— , 

 +. 



40 

 26 

 16 

 6 

 12 

 10 

 30 

 38 



0.9734 

 1 .0833 

 1 . ] 394 

 . 5460 

 1 .0912 

 1.2460 

 0.7536 

 0.7335 



1.0453 

 1.0128 

 1 .0320 

 1 .0731 

 1.0574 

 0.7777 

 1 .0576 

 0.9726 



1.0821 

 0.9546 

 1.0051 

 0.9776 

 1 . 3085 

 1.1770 

 0.9797 

 0.9063 



From Table 10 it is seen that the apparent marked effect of 

 the two plus selections in 191 1 entirely disappears, if we omit 

 Line No. 262 in which there is one selection only very slightly 

 above the mean. In a similar imanner the very large mean 

 index for the -\- selection in 191 1 is reduced to a figure com- 

 parable with other indices. However, with two minus selec- 



