l6 MAINE AGRICULTURAL EXPERIMENT STATION. I915. 



deviations from the means of the respective lines it is possible 

 to lump together all of' the pure lines regardless of varietv' or 

 absolute yield and thus- obtain sufficient data upon which to base 

 more definite conclusions. 



On account of the lack of space the detailed tables for each 

 pure line will not be given. Instead summary tables showing 

 the effect of the plus or minus selections of each year upon the 

 yield of each of the succeeding years will be given. 



It must be remembered that in dealing with these data in 

 the present way two factors are entirely neglected, (i) No 

 account is taken of differences in the size of the deviations of 

 selected plants. All plus deviations whether of 20 grams or one- 

 tenth of a gram are given equal weight and similarly for the 

 minus selections. (2) Xo account is taken of possible dif- 

 ferences in the relative variability of the different pure lines. 

 These two factors w411 be taken into account in a later section of 

 this paper. It is desirable to consider first the results obtained 

 by the method outlined in the preceding paragraphs. 



Table 3 shows the effect of the plus and minus selections 

 inade in 191 1 upon the rows grown in each of the three suc- 

 ceeding years. Thus the 1912 rows show the immediate effect 

 of the selection. The rows of 1913 and 1914 have each been 

 subjected to other plus and minus selections in the later years. 

 It is to be expected that the results, if any, of the 1911 selection 

 would be somewhat obscured in these latter years. The results 

 of successive -plus and of successive minus selections will be 

 discussed in a later paragraph. 



Table 4 shows the effect of the 191 2 plus and minus selec- 

 tions upon the rows grown in 1913 and 1914. Similarly Table 

 5 shows the effect of the 1913 selection upon the 1914 rows. 



