Variation and Correlation of Oats — Part I 17 



.93-1 ±.022 for 1910, 1.415 ±.030 for 1909, and 1.972±.047 for 1912. 

 The number of culms produced by a plant is very responsive to environ- 

 mental conditions. Under the best conditions here dealt with the number 

 of culms has been largely increased, and the variability has also increased 

 with the better conditions for growth. 



The means for 1910 are less than those for 1909 and 1912 in average 

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

 and average yield. When spikelet characters in 1909 and 1910 are compared, 

 it is seen that for 1910 the mean is less for average number of kernels per 

 spikelet and greater for average number of spikelets per culm than for 1909. 



In regard to the comparative variability of the different characters, 

 there is evidence in the coefficients of variability. From these it is observed 

 that average height, average weight of kernels, and average number of 

 kernels per spikelet, are the least variable characters, the coefficients of 

 these ranging from about 7 to a little more than 16 per cent. In no other 

 instance is the coefficient of variability less than 17, and from this it 

 ranges to nearly 63 per cent. 



The average weight of kernels per plant is higher for 1910 than for 

 either of the other two years under consideration, but, since the total 

 and the average number of kernels produced are less, the total and the 

 average yields are still less for 1910. The average number of spikelets 

 per culm per plant also is larger for 1910 than for 1909, but the average 

 number of kernels per spikelet is larger for the latter year. It seems that when 

 development is arrested by environmental conditions, yield is reduced by re- 

 duction in number of kernels per plant, per culm, and per spikelet, rather than 

 in average weight of kernels and in number of spikelets produced. The 

 actual kernel weights as given in the correlation tables show that larger 

 kernels were actually produced in 1910 than in the other two years. 



Altogether, then, the means indicate that the conditions for growth 

 were not so favorable for the 1910 crop as for those of 1909 and 1912, 

 and resulted in smaller plants and a reduced total yield of grain; but 

 that these smaller plants produced a larger number of spikelets (compared 

 with 1909 only) and larger kernels. 



Somewhat different conditions are presented in the year 1908. In the 

 first place, the data were taken on culms only — these being selected from 

 a pure line grown in a drill row — and were taken on these selected culms 

 without reference to the plants from which they came; while in the other 



