124 



BREEDING CROP PLANTS 



correlation coefficient. The coefficients for the years 1914, 

 1915, 1916, and 1917 were 0.143 O.C38, 0.088 0.028, 0.445 

 0.020, and 0.478 0.024, respectively. In this investigation 

 each plant was given the same space for individual development. 

 The results show that under these conditions relatively large 

 amounts of stored plant food in the germinating seed may or 

 may not give the resultant plants an advantage, depending on 

 environmental influences other than the amount of endosperm. 



Several investigators have attacked this problem from a prac- 

 tical viewpoint. Seeds were separated into light, medium, and 

 heavy by means of a fanning-mill. The productivity of the 

 plants coming from the various classes of seed was compared 

 under field conditions. Some investigators procured a slightly 

 greater yield from plants produced by heavy seed than from those 

 coming from light seed. Others obtained no such difference. 

 Plants from medium or ungraded seed in almost all cases proved 

 as productive as those from heavy seed. The work carried on 

 at the Ohio Station may be taken as a typical example of, these 

 investigations. 



Table XXIV presents the average results (Williams and Welton, 

 1911) of an experiment with weight of seed wheat over a period of 

 seven years. The grades are first, second, and third, represent- 

 ing heavy, medium, and light seed, respectively. Two methods 

 of seeding were practiced, namely, a uniform rate by weight and 

 a varied seeding to obtain approximately an equal number of 

 plants on equal areas. 



TABLE XXIV. THE RELATION OF WEIGHT OF GRAIN TO YIELD IN WHEAT 

 Seven-year Average Results 



In the case of oats (Williams and Welton, 1913) a greater 

 difference was obtained between light and heavy seed, but the un- 

 screened seed yielded only a little less than the large seed. Table 

 XXV presents the average data of a four-year period. 



