IMPROVED STRAINS OF ROOTS 41 



seed, are counted, the roots are then lifted and counted, the 

 number of forked roots, etc., noted, and all the roots in each row 

 are then weighed together. A character is given to each sample, 

 indicating how far the roots approached to the shape typical for 

 the variety. Finally an average sample is taken from each of 

 the 20 samples, and the percentages of sugar and of total 

 solids are estimated by chemical analysis. 



To take an average sample of a quantity of roots for chemical 

 analysis is by no means an easy matter, as the individual roots 

 vary considerably in composition. To get a true average 

 sample, is, of course, a necessary condition for obtaining a 

 reliable analysis, and as more than 10,000 estimations of total 

 solids in roots are made every year in Denmark, the matter 

 has been closely looked into, and a method has been evolved, 

 which has since been adopted in Sweden, Canada (McDonald 

 College), Holland, and Germany. 1 If a result correct to the 

 first decimal of the percentage is wanted, it has been found 

 necessary to work on at least four samples of 50 roots each. 

 The roots to be analysed should have the same average weight 

 as the roots in the total stock. 500 roots are taken at random 

 and weighed ; 50 roots are taken from these 500, avoiding 

 extreme sizes and shapes. The 50 roots are weighed, and by 

 exchanging roots the weight of the 50 roots is regulated so as 

 to be one tenth of the weight of the 500. The 50 roots are then 

 carefully washed and left to dry. 



When a sample of 50 roots has been picked out, the next 

 thing is to obtain a true average sample of these 50 roots for 

 a chemical analysis. Several methods have been proposed 

 and used, but most of them fail to give a true average. In the 

 eighties it was usual in Germany, by means of a revolving tooth- 

 edged scraper, to scrape a wedge out of the root lengthwise 

 reaching into the centre axis, the scraper removing the material 

 in the form of a pulp, from which samples were taken for 

 chemical analysis. But R. K. Kristensen has proved 2 that 



1 R. K. Kristensen," T0rstofbestemmelse i Roer," Tidsskrift for Planteavl, 

 23 vol., K0benhavn, 1916, p. 155. (Report written to the International 

 Agricultural Institute in Rome.) 



8 R. K. Kristensen, " T0rstofbestemmelse i Roer," Tidsskrift for Planteavl, 

 18 vol., K0benhavn, 1911, p. 96 (being the 55th Report of the State Committee 

 on Plant Culture). 



