16 H. H. Love and C. E. Leighty 



.239±.006 for 1912, .241dz.006 for 1910, .2o7±.004 for 1908, and 

 .315±.007 for 1909. The least variation exists where the mean is 

 largest, but the decrease is not in proportion to the decrease in the mean. 



In number of kernels the same conditions are found as obtained in yield. 

 The means for total nmnber of kernels are, in order of value, 202.150± 

 2.657 for 1910, 217. 500 ± 4.124 for 1909, and 5^8. 650 ±7. 589 for 1912. 

 The means for average number per cuhn are, in order, 23. 580 ±.327 

 (total number per cuhn) for 1908, 56.020z±z.484 for 1910, 59.800±.559 

 for 1909, and 86.600±.514 for 1912. In variability the order differs, 

 being for total kernels 78.800±1.879 for 1910, 136.560±2. 909 for 

 1909, and 225 . 050 ±5 . 367 for 1912; and for average number 13 . 910± . 231 

 for 1908,14.340±.342 for 1910, 15.250± .364 for 1912, and 18. 508 ±.394 

 for 1909. The variability does not differ materially, but it is not greatest 

 where the average yield is greatest. 



The mean and the variabihty of average weight of kernels (milligrams 

 in each case) are least for 1912, the mean being 14.485±.036 and the 

 standard de\iation 1.056±.025. The mean is greatest for 1910, being 

 17. 100 ±.059, while the standard de\'iat:on for that year is 1.741 ±.042. 

 The means for the other two years are about the same, being slightly 

 above 16.1 in each case; the variation, however, is much greater for 1908, 

 being then 2. 626 ±.044, while for 1909 it is 1.311 ±.028. 



The mean for number of spikelets per culm per plant is 15. 320 ±.173 

 for 1908. This constant is again much smaller than the one for 1909, 

 which is 32.340± .247, while the one for 1910 is 35.400± .245, the reason 

 being undoubtedly the same as is given for the differences noted above. 

 The variability is greatest for 1909, when the standard deviation is 

 8. 171 ±.174. There is httle difference in this respect between the 

 other two years, the fig-ures being 7. 375 ±.122 for 1908 and 7. 270 ±.173 

 for 1910. 



The means for average number of kernels per spikelet were determined 

 for only two j'ears. For 1910 the mean is 1.506±.006, for 1909 it is 

 1 .803 ± .007; while the standard deviations are . 172 ± .004 and . 229 ± .005, 

 respectively. There is not much variability in this character, as is 

 indicated b}^ this constant. 



The mean number of culms produced by the plant varies from 3.568± 

 .031 for 1910 to 3.948±.043 for 1909, and 6.150±.066 for 1912, 

 when it is greatest. The variability increases as does the mean, being 



