Variation and Correlation of Oats — Part II 



145 



Considering first the correlations between average yield of culm per 

 plant and average height of plant (Figs. 87, 72, 89, 114), it is seen that 

 the coefficients for series 1219 and series 1221 differ by only .006 ± .015, 

 which is a negligible difference. In series 1257 the coefficient is .087 ± 

 .017 greater than in series 1259. This is more than five times its prob- 

 able error, and may be of significance. 



Series 1259 



I I I I I I I I 



§ Totals 



0- 1 



1 















1 



1- 2 



1 









1 







2 



2- 3 



2 



2 



2 



1 









7 



3- 4 



1 



3 



3 



1 



1 



1 





10 



4- 5 





1 



9 



3 



1 







14 



5- 6 







8 



7 



5 







20 



6- 7 







6 



13 



3 



1 





23 



7- 8 







2 



17 



14 







33 



8- 9 







1 



12 



17 



3 





33 



9-10 









6 



17 



12 



1 



36 



10-11 









1 



15 



10 



2 



28 



11-12 











11 



18 



3 



32 



12-13 









1 



4 



18 



8 



31 



13-14 











2 



7 



o 



15 



14-15 











1 



6 



2 



9 



15-16 











1 



1 



2 



4 



16-17 













2 





2 



Totals 



1 



6 31 62 93 79 23 



300 



Fig. 114. — Correlation betiveen average yield of culm per 

 plant, in decigrams, subject; and average height of plant, 

 in centimeters, relative 



r= .788 ± .015 



For average yield of culm per plant and average number of kernels per 

 culm of plant (Figs. 91, 74, 93, 115), the coefficients are above .934 in 

 every case, and are so close together in the same varieties for the two 

 methods of sowing that no comment is necessary. 



For average yield of culm per plant and average number of spikelets 

 per culm of plant (Figs, 95, 76, 97, 116), the same relative conditions 

 exist as in the preceding case. All the coefficients are above .916, and 

 those for the same varieties are very near each other in value. 



