by chemical methods. 109 



globe, Webb's golden king, and Carter's 1901 are not absolutely 

 the best. 



In the autumn 200 roots of each of these varieties were 

 picked out, attention being given to shape, size, and general ex- 

 cellence of external characteristics. This was kindly done by 

 Mr R. H. Biffen, M.A., of Emmanuel College, Botanist to the 

 Department of Agriculture. For this and much other assistance 

 the authors tender their best thanks. These picked roots were 

 carefully stored over the Christmas vacation, and were then again 

 picked over, weighed, numbered, and cored. Each core was dried 

 to constant weight, and its percentage of dry matter calculated. 

 Extraordinary variation was noticed in each case, as is shown in 

 the following diagrams. 



These diagrams, figs. 2 and 3, show several points of interest. 

 First that the percentage of dry matter in individual roots of 

 the same variety grown side by side and under conditions of as 

 great uniformity as is obtainable in ordinary farm practice, 

 varies very greatly — in each of the three cases from about 11 per 

 cent, to about 20 per cent. Secondly, that the variation is of the 

 usual type — that is to say a few individuals with very low per- 

 centages, a few with very high percentages, and a large number 

 intermediate between these, rising by short steps continuously 

 from the lower to the upper limit. Thirdly that (fig. 2) while 

 there is a distinct tendency for the percentage of dry matter to be 

 higher in small roots than in large ones, the inverse proportionality 

 between size of root and content of dry matter is by no means so 

 definite as is usually taken for granted. These diagrams furnish 

 the necessary information for the commencement of chemical 

 selection. Seed will this spring be grown from roots of each 

 variety which contain the highest percentage of dry matter, 

 precautions being taken to prevent cross-fertilization from plants 

 of inferior quality. 



This seed will not be ready to sow until the spring of 1904, 

 and its progeny will not be ready for investigation until that 

 autumn. In the meantime the work already done will be re- 

 peated, and extended. As yet the individual variation in per- 

 centage of nitrogen, sugar, and cellulose has only just been 

 commenced. 



