NECESSITY FOR FIELD PLOT TESTS. 





in yielding power, tliiis illustrating the fact that all but the few in 

 which the new quality persists must be discarded. In trying to raise 

 the yield above the average we must eliminate not only the poor but 

 the average blood, retaining only the blood of the few which have 

 the greatest al)ility to produce progeny with exceptionally large yield. 



But the important result is that some of these new stocks of seeds 

 did give large average plant yields, showing their promise of large 

 yields when grown under field conditions. 



Even when those are chosen which yield well in centgeners there is 

 yet need of further elimination by testing them for a number of years 

 in the uniform'test field plots. In figure 11 are shown the field yields 

 of the three Fife, three Blue Stem, and the four hybrid stocks which 

 show the highest centgener yields in figure 10. Even here some of 

 those which yielded well in the nursery centgeners yielded less when 

 grown crowded together in the grain field than the varieties used as 



"bo. 



Fig. 11.— Yields, in bushels, of ten new strains of wheat developed in experiments, which have 

 been illustrated in figs. 2-l(i. The curved line intersects ten vertical lines, representing the 

 yields in averat?ed field trials of the nursery stocks, which show the highest centgener aver- 

 ages in fig. 10. The three solid vertical lines represent .3 Fife stocks, the three broken hues 

 represent :> Blue Stem stocks, and the four lines marked with a- represent 4 hybrid ( Blue Stem- 

 Fife ) stocks. The line representing the best hybrid strain rises almost to the horizontal line, 

 representing a yield of 27 bushels per acre, while several strains fall considerably below the 

 yields of the parent varieties. 



foundation stocks as shown by their dropping below the standard 

 line. The records of wheat breeding in the Minnesota experiment 

 station show that the yields of new strains and varieties do not neces- 

 sarily correspond with the yields of their respective mother plants. 



In a general way two facts are illustrated here: (1) That we can 

 improve wlieat by selecting the best from our standard wheats; (2) 

 that still HKjre can be accomplished if we create new qualities by 

 hybridizing and then seek from among very many those few plants 

 whieii will best perpetuate tlie desired <iuality. 



The real value of variation lies in the ability of the plant to produce 

 plants which individually and in the aggregate yield more and better 

 grain than the average of the same variety. Tlie yield of the mother 

 plant is a very uncertain indication of its use for the mother of a new 

 strain, as is shown in figures 9, 10, and 11, just as Messenger's record 

 as a trotting horse is no index to his great value as the progenitor of 

 the American breed of trottiug horses. This fact is shown with force 

 by the history of the progi'ii\ of the Jersey bull, Stoke Pogis. He 



