CHEMICAL SELECTION OF MOTHERS. 73 



could be kept at almost constant density. The Dervaux 

 method of classification was, that beet samples floating 

 in the bath of 105 density were rejected and those 

 which sank in that of 105.5 were subsequently sent to 

 the laboratory for further examination. On the other 

 hand, those which were of a density of more than 105, 

 and yet less than 105.5 were siloed and planted the fol- 

 lowing year and gave seed for the trade. The roots to 

 which preference was given weighed from 700 to 

 900 grams. 



Respecting the Dervaux method, we would say 

 that the results obtained by it are more reliable than 

 by the \'^ilmorin system, where several errors exist, the 

 most important being the effect of endosmosis of the 

 solution, and the atmospheric effect produced upon the 

 small cylinder in passing from one vase to another, 

 a series of solutions of different strength being used. 



The other errors were noticed by Champonnois, 

 and were due to a certain volume of gas contained in 

 the tissues of the piece of beet, or in the entire root. 

 The volume of these gases varies from 9 to 50 c. c* 

 per kilo of beets. Its composition is, nitrogen, 63; car- 

 bonic acid, 37. Th'e following table shows that the vol- 

 ume of these gases varies considerably. 



Vibran made a new departure, and instead of 

 taking the cylinder from the beet, he took the density 

 of the tip end as a basis of estimation. Rimpau 

 showed by a series of well-conducted experiments that 

 the method was not reliable. Several beets which were 

 examined sank in a saline solution of 104.8; their sugar 



'Dubrunfatit admits that it may reach 113 c. c. per kilo of beets. 

 Opinions very nuich differ respecting the composition of the gas; M. 

 A. Heintz declares that it consists of nitrogen, 66.8; carbonic acid, 32.8; 

 oxygen, 0.35; and the volume varies from 130tol50c. e. per kilo of beets. 



