Variation and Correlation of Oats — Part II 139 



while the greatest relative differences are found when yield is correlated 

 with weight of straw and number of kernels and of spikelets. 



•The coefficients here reported are usually fairly close together for the 

 different varieties. No striking differences occur, such as exist between 

 the results of Waldron and those of the writer. Since, however, the variety 

 used by Waldron is not here concerned, it cannot be said that the difference 

 in his results was not due to varietal causes. Some differences occur in 

 these results which are evidently du« to varietal causes, hence we must 

 conclude that differences in correlation coefficients may result from the 

 use of different varieties. 



Comparison of hiometricdl constants determined for oat plants grown in hills 



and in drills 



Some statistical work has been done on cereals in which the plants 

 used were grown a fixed and uniform distance apart, the plants each being 

 thus allowed the same amount of space in which to grow. Other work 

 has been done with plants that were growm in rows, in which the grain 

 was drilled in the ordinary method of seeding. The work of Roberts 

 (1911) and of Myers (1911) on wheat was with plants grown in hills 

 several inches apart. The work of Noll (1911) on wheat was with plants 

 or culms that grew in drilled rows. The work of Atkinson (1912) was 

 with wheat that was seeded broadcast in plats, the same quantit}- of 

 different varieties being used. The work of Waldron (1910) on oats was 

 with plants from drilled rows. 



What differences, if any, exist in the constants determined for plants 

 groT\m in these ways? In which case is there greater varialoility, if 

 difference exists? In which is there greater correlation, if difference 

 exists? 



In order to test the effect of growing plants in separate hills and in 

 drilled rows on the biometrical results, the following plantings were 

 made : 



Sixty Day selection, pure line 62-II-6-3, and Early Champion selection, 

 pure Ime 137-6, were each sown in drills, fifteen grams being sown uni- 

 formly m eighteen-foot rows, the rows being one foot apart. Separate 

 kernels of the two Imes were also planted in hills three inches apart in 



