24 MAINE AGRICULTURAL EXPERIMENT STATION. ipiS- 



determine whether these four points deviate farther from the 

 theoretical values than could be expected. By this test we 

 find that 



P = .6561 



Or, in other words, if there is no effect of selection we should 

 expect to get a fit as bad or worse than that observed in 65 out 

 of every 100 trials. This is a reasonably good fit and we may 

 conclude that these deviations do not behave differently than 

 might be expected if there was no effect of the selection. There 

 is thus no positive evidence that the selection has produced any 

 effect even in the first year after the selection. 



SUMMARY. 



Summarizing the results of this section it may be said thai 

 from the data so far examined there is some evidence of an effect 

 of a selection upon the rows grown the next year. In no single 

 case is the eft"ect great enough to be significant in itself. How- 

 ever, the fact that there is a reasonably large difference in the 

 same direction after each of the three selections indicates that 

 there may be some immediate effect of the selection. 



It is quite possible that the apparent eff'ect of selection upon 

 the rows grown the following year is to be explained as physio- 

 logical rather than genetic. It is possible that the larger and 

 consequently more vigorous plants produce grain which has 

 mo.re nourishment or w^hich for some other reason gives the new 

 plant a better start. This brings us to an old queston which it is 

 extremely difficult to answer. It will be considered in a later 

 paragraph. 



The Effect of Successive Selections in the Same; 

 Direction. 



If the slight eff'ect of selection upon the following generation, 

 as noted above, is due to changes in the germinal substance, then 

 a much more marked effect ought to be produced by two or 

 three successive selections in the same direction. A portion of 

 the 19 1 3 rows have been subjected to two such selections and 

 some of the 19 14 rows to three such selections. The results are 

 shown in table 7. 



