F. KiDD AND C. West 



•Jl) 



and dark colour on the one hand, and for light weight and light colour 

 on the other. Equal numbers of seeds of the heavy and light samples 

 were used for the yield tests. We may quote his final results obtained 

 in the thirteenth year. "'The large plump seed produced 65-5 bushels, 

 the light seed 44-7 bushels per acre." Tn this year the average size of 

 the seed from the crop produced by the large plump seed was greater 

 than that from the crop produced by the light seed. In the first case 

 it required 1149 seed-grains to weigh 1 oz. and in the second case 2066. 

 In this experiment it will be seen that Zavitz, in contrast to most 

 other workers, records a marked increase in average size of grains as well 

 as an increase in total yield. If plants from large grains are more vigorous 

 and produce a larger yield as a result of physiological pre-determination 

 an increase in the average size of the grains yielded would also probably 

 occur^, but in these results of Zavitz's it seems more probable that 

 hereditary factors are concerned. We have to note the large number of 

 years over which the experiment was continued and which would allow 

 time for the slow separation of a possible hereditary factor for large-seeded- 

 ness. As has been pointed out, selection for seed-size would have little 

 effect in one generation owing to the fact that seed-size as an hereditary 

 character would be almost entirely masked by the influence upon size 

 of the position of the seed upon the parent plant, so that in a first separa- 

 tion of large and of small seeds the sample of large seed would probably 

 only contain relatively very few more seeds large by heredity than 

 the sample of small seed. In the second place his selection was not 



1 Desprez (15) with wheat found that the use of large seed gave a croii with grains 

 larger than those grown from small seed. In Lyon's (51) work this was also noted. 

 Am. Biol. V 9 



