WOODHOUSE AND TAYLOR. 1 1 7 



large bulk samples from the plots from which these single 

 plants had been selected. 



The results obtained were as follows :— 



., Nitrogen. 



(0L0UR - Percent. 



Black ... ... ... ... ... 6 17 



Yellow ... ... ... ... ... 5-44 



Chocolate ... ... ... . . ... 5'74 



Although these plots were not grown from single plants, yet 

 they appeared to be perfectly even, and these results, in which 

 the black seed stood out ahead of the other two as regards its 

 nitrogen content, tended to corroborate our observations on the 

 small samples obtained from the single plants. Consequently 

 in the rains of 1910 we resowed not only the seeds of the singly 

 selected plants, but also seeds from the original plots from which 

 they were chosen. 



The descendants of the former were analysed plant by plant 

 and the means are shown in the tables (Tables I — IV), giving 

 their history for the past three years. From the descendants of 

 the latter, 150 plants were taken from each plot for analysis 

 (Tables V— VII). 



The results of these analyses are striking, and are of addi- 

 tional interest in that we were able this time to determine 

 the oil contents of a great many specimens. In order that 

 reference may be made to these tables, they are published at the 

 end of the letter-press. 



On collecting the results we were still further strengthened 

 in our belief that it was possible that certain differences in oil and 

 nitrogen content might be correlated with certain definite vegeta- 

 tive characters. 



Firstly, in considering the nitrogen contents of three 

 varieties, we can arrange them in the following tabular 

 statement showing the groups into which the samples of each 

 variety may be arranged according to their nitrogen content 

 (see page 118). 



