F. KiDL) AND (\ WK8T 



'^( 



grain, the great divergence in yield is almost always observed only in 

 the first year (see Desprez(i5), Lyon (•■)!), also Soule and Vanatter(67)), 

 whereas if a hereditary factor for seed-size is concerned, it should take 

 many years to segregate it by selection since it would be in any case 

 almost completely masked by the effect of position u))f)n the seed-size. 



Waldron's (I.e.) statistical results with oats indicated that, if any 

 hereditary factor were involved in the selection of large grain, the size 

 of the individual grain might be increased, but the total yield would 

 probably be decreased. For in all populations examined, he found a 

 strong correlation between large-seededness and weak development, 

 i.e. small size of plant combined with small yield in number of grains. 

 As in nearly all experiments with cereals an increased yield has been 

 observed following the use of large grain, but in very few cases has any 

 increase in the average size of the grain been noted, the obvious pre- 

 sumption is, therefore, that increased yield is due in the main to physio- 

 logical, and not to hereditary pre-determination. Waldron concluded that 

 if plants from larger grains produce a better yield, then they must do 

 so by virtue of the increased vigour of the embryo and the increased 

 amount of food-material available. 



When the crop is a leaf crop, the correlation between large-seededness 

 and yield may in all probability be safely put down to physiological 

 pre-determination. 



The second difficulty that is encountered when considering the results 

 of many large-scale field experiments is that we cannot definitely decide 

 hoAv far they are due, not to differences in the vigour and yield of 

 individual plants, but to differences in number of plants. 



In many experiments the sowing has been carried out at the bushel- 

 acre rates and no determination has been made of the relative number 

 of grains per bushel, so that the real rate of sowing with regard to number 

 of grains per given area is not the same with the large and small seed 

 respectively, but is in favour of the latter. 



In the second place a difference in germination capacity between 

 light and heavy seed will introduce an unknown error from this point 

 of view. A. Miiller(55), who experimented with oats, wheat, and rye, 

 found, for example, that the heavier grains gave a higher percentage of 

 germination and produced larger seedlings than the lighter grains. The 

 light seeds of oats gave 68 per cent, germinations in eleven days whereas 

 the heavy seeds gave 92 per cent, germinations in the same interval 

 (see also Grenfell(2.5), BoerkerO), Shamel(65), Lehmann(47), Vandevelde 

 (72), Webber and Boykin(78), Eisenmenger(20),and Hicks and Dabney (4i)). 



