1913.] PUBLIC DOCUMENT — No. 31. 155 



plant from the fleshj roots. The pLants were selected by the 

 size of their tops, and the effort was made to get average plants 

 for each type of fertilization. The plants were carefnlly exca- 

 vated, and as nearly as possible all the fleshy roots were collected. 

 In some instances growth had been so extensive that adjacent 

 roots were much interlaced, and it was impossible to secure one 

 plant without destroj-ing all others around it. Therefore, our 

 largest, heaviest roots M'ere more or less incomplete. 



On the day following the collection of the specimens the roots 

 were shipped from Concord to the experiment station, where 

 they were placed in a cool cellar. The process of washing free 

 from soil, weighing when free from surface moisture, drying 

 and grinding, was very laborious and required several weeks, 

 owing to the weight and bulk of the material. There is always 

 danger of respiratory changes during the period of preparation, 

 between collecting and drying ; but it is believed that they were 

 reduced to a minimum by keeping the roots in a cool place at 

 about 10° C. Washing was performed within a few days after 

 digging, and the weights of the fresh roots were obtained before 

 there could have been any respiratory losses. 



The material was prepared for drying by passing the crown 

 and roots through a hand-lever fodder-cutter by which they were 

 cut into pieces about 2 centimeters in length. Drying was done 

 in a large oven at a temperature between 50° and G0° C, and 

 the material was dried until sufficiently brittle to be easily 

 ground. The weight of the specimen was taken at this stage, 

 and then the entire lot was coarsely ground in a drug-mill, after 

 which it was sampled by quartering and the sample reduced to 

 a powder which would pass though a 1 millimeter mesh sieve. 

 Moisture determinations were then made in the powdered sam- 

 ples, from which were calculated the weights of absolutely dry 

 matter contained in the specimen roots. 



Three different series of samples have been studied during 

 this investigation. 



The first series consisted of 44 roots, or 4 specimens from 

 each of 11 plots, used for testing different quantities and dif- 

 ferent seasons of application of nitrate of soda. They were 

 collected in I^ovember, 1908, two years after the field had 

 been set with plants, by Mr. E. F. Gaskill of the agricultural 



