126 MAINE AGRICULTURAL EXPERIMENT STATION. 1916. 



ing the results of the three years. The most important charac- 

 ter is of course the yield of grain. This will be considered first. 



In the years 19 14 and 191 5 we have based the discussion 

 mainly on the corrected yields. For reasons discussed above it 

 is not practicable to apply the correction method to the 1913 

 plots. In the tables below we have used the observed yield for 

 191 3 and the corrected yield for the other two years. This 

 method is perfectly justifiable and after careful consideration 

 has seemed to be the best method of dealing with these data. 

 While some of the 1913 yields are admittedly influenced by 

 heterogeneity in the soil, this does not seem to offer any sound 

 reason for introducing similar errors in the later years. 



Table 7 gives the 12 pure lines arranged in the order of their 

 average yield. The yield for each of the three years is given 

 together with the three-year average. In addition to this the 

 name of the parent variety from which the selection was made 

 is also given. 



Table 7. 

 Yield of Grain, Pure Lines Tested for Three Years. 



Line 

 number. 



Parent Variety. 



Bushels per Acre. 



1913. 1914. 1915. Mean 



Maine 



355 

 340 

 281 

 351 

 337 

 230 



Banner 



Irish Victor 



Banner 



Banner. 



Irish Victor 



Banner 



307!Banner 



336! Irish Victor 



346 Irish Victor 



247 Imported Scotch . 



286|Banner 



357 Banner 



Average . 



71.0 

 74.1 

 73.1 

 70.0 

 58.4 

 69.4 

 66.9 

 75.3 

 71.9 

 62.8 

 70.9 

 70.0 



69.5 



105.3 

 95.8 

 97.0 

 98.2 



103.9 

 93.8 

 95.8 

 89.2 

 89.5 

 92.4 

 87.1 

 83.1 



94.4 



83.6 

 83.6 



81.2; 



75.5 

 79.2 

 76.8, 

 77. ( 

 75.11 

 77.0, 

 79. l! 

 75.71 

 79.1 



78.6 



86.6 

 84.5 

 83.8 

 81.2 

 80.5 

 80.0 

 79.9 

 79.9 

 79.5 

 78.1 

 77.9 

 77.4 



80.8 



One of the most striking things about this table is the pre- 

 ponderance of certain parent varieties. These pure lines were 

 selected from the 188 lines grown in the 191 1 garden solely on 

 their merits, without thought or often even without knowledge 

 of the parent variety. Fourteen of the 188 191 1 rows were 

 from the Banner variety (cf. Table 1). Seven of these are in- 

 cluded in the above table. Likewise there were 14 Irish Victor 

 rows in 191 1 and 4 of these have been continued. Only one 



