32 



Seed Wheat. 



exceptional result seems due beyond doubt to the small number of trials only two, both 

 cases giving a substantial excess to the lower grade. If we class the trials according to 

 the size of the larger of the grades used in the comparison, we may make the following 

 general statement : 



325 excelled in 100 per cent, of trials. 



300 78-6 



275 

 250 

 225 

 200 



75 

 100 

 none 

 53-8 



This statement, however, does not tell the whole story, for we find that, as in most 

 unequal contests, the victories of the better contestant are the more decisive. Thus 

 325 when excelling did so by 21 ! per cent, excess, and was never excelled. 

 300 when excelling did so by 12'5 per cent, excess ; when excelled it was only by 5 



per cent. 



275 when excelling did so by 14'7 ,, ,, ,, ,, only 12'3 per cent. 



250 ,, 13-8 



225 7'7 



200 ,, 14-5 ,, ,, ,, ,, 11-6 



From this 1 conclude that the yield of wheat-plants under the conditions of these parti- 

 cular experiments, as well as the power to grow to a harvestable size, is a function of the 

 size of the seed, and varies directly with the size of the seed. 



If, instead of taking the sum of the weights of the grain and straw as the criterion of 

 yield, we take the grain alone, we arrive at the following : 



The 325 grade excelled in 100 per cent, of the trials. 

 300 69 



275 

 250 

 225 



200 



72-2 



87'5 

 none 

 42-9 



As these general statements, however, fail to give a perfectly correct idea of the extent 

 of the superiority of the large seeds, we may add 



325 when excelling did so by 22 '7 per cent., and was never excelled. 



300 

 275 

 250 

 225 

 200 



13'8 ,, when excelled it was only by 3 '9 per cent. 

 11-8 ,, ,, ,, 12-5 



7-0 ' ,, ,, -5 



H-2 



8-2 , 13-0 



If, instead of taking the sum of the weights of the grain and straw as the criterion of 

 yield, we take the straw alone, we arrive at the following : 



The 325 grade excelled in 100 per cent, of the trials. 

 300 69 



275 

 250 

 225 

 200 



75 



100-0 

 none 



71-4 



As these general statements do not give a perfectly correct idea of the extent of the 

 superiority of the large seeds, we may add that 



325 when excelling did so by 20 '0 per cent. 



300 ,, ,, 14'6 ,, when excelled it was only by 10 '0 per cent. 



275 ,, 16-2 1T6 



250 ,, 20-6 



225 ,, 4-0 



200 ,, 20-2 ,, 9-9 



I have referred to the losses due to little understood and inexplicable soil conditions, 

 and have said that they were rendered nugatory by the row system. By this, I mean that 

 when a patch of such varying soil appeared, it invaded in about equal degrees the various 

 experiment rows, and thus caused about equal losses to the adjacent rows being compared. 

 In a few instances, where it was apparent that this invasion was not of an equal character, 



