132 



Analyses of Hybrids and Crosses Continued. 



After a number of these analyses of large samples had been made, 

 it was concluded to discontinue them, for the individual canes varied so 

 much that it was impossible to obtain samples which would represent 

 the plot, except in the few cases where the character of the plot was 

 uniform. In some cases the plot could be thrown out, where the aver- 

 age samples showed a very poor analysis. The work was thenceforth 

 confined to analyses of individual canes, selected With a view to per- 

 manence of type. A very large number of samples were taken in this 

 way, the seed heads removed, marked with a number corresponding to 

 the analysis and preserved. The juice was polarized, and from each plot 

 one or more samples which gave the best results, and which were to be 

 reserved for future planting, were subjected to complete analysis, so as 

 to have a complete pedigree of the cane. The following table gives 

 the results of some of the individual canes from the crosses ; only the 

 best samples in each plot are given, and these analyses are only a frac- 

 tion of the whole number made and recorded at the station : 



Analyses of Crosses. 



