18 H. H. Love and C. E. Leighty 



years data are given for the entire plants and the culm is dealt with only 

 in averages for the plants. But, by comparing the data for the culms 

 of 1908 with the averages of the other years, it is seen that in 1908 the 

 culms were not so tall as, and the number of spikelets, number of kernels, 

 and yield were all less than, in the other years, but that the kernels 

 were practically as large as, or larger than, in any other year except 1910. 

 This condition offers evidence corroborative of that given above in regard 

 to the relation of environment to variation. 



Coupled with the less plant development and yield of 1908 and 1910, 

 there is less variability, as indicated by the standard deviation, in number 

 of culms, total and average number of kernels, total yield, average number 

 of spikelets, and average number of kernels per spikelet. In average 

 weight of kernels, however, there is greater variability for 1908 and 1910. 



COREELATION 



The coefficients expressing the correlation between the different char- 

 acters under observation are given in Table 4. These coefficients have 

 been determined independently for the four years 1908, 1909, 1910, and 

 1912. The coefficients expressing the correlation between the same 

 characters are directly comparable for the different years, except in so 

 far as the results may be affected by (1) the use of culms as the unit of 

 measurement in 1908 and of plants as the units in other years, and 

 (2) the methods of growing the plants. The effect of the former has 

 been shown to be small by a study of oats, by Dr. C. E. Leighty, soon 

 to be published as Part II of this series. With reference to the methods 

 of growing the plants, however, it is shown in the study just cited 

 that rather large differences occur in the same variety between the co- 

 efficient of correlation determined for plants grown in hills and in drills. 



By reference to Table 4 it is seen that the correlation between average 

 height of plant and average weight of kernels per plant varies from 

 .555± .016 for 1908 to — .023± .034 for 1910 — a considerable variation, 

 which indicates that the correlation between these two characters is not 

 stable nor constant, but fluctuates from year to year. 



In contrast to this variation is the correlation between the total yield 

 of plant and the total number of kernels per plant. In this case the 

 coefficient for the different years varies from .980 ±.001 for 1912 to 

 .918±.005 for 1910, a difference which is comparatively small. 



