222 



The directions of instrument makers as to the use of bichromate and the strength of 

 solution to be employed are not at all explicit. One maker directs merely that the 

 bichromate cell must always be filled with bichromate solution, which, however, can 

 be more or less concentrated, according to the character of the liquid under examina- 

 tion. 



The purpose of the bichromate solution insaccharimetric work is, of course, to correct 

 the difference in rotation dispersion between cane sugar and quartz. The rays of 

 light in the bljue and violet which cause the greatest Amount of rotation dispersion are 

 absorbed by the bichromate. To gain more exact information as to the effect of elim- 

 inating the bichromate solution in the polarization of raw sugars I have recently com- 

 pared the absorption spectra of bichromate solution with those of different clarified 

 sugar solutions. Molasses sugars when clarified give a brownish yellow liquid, which 

 absorbs practically all of the light in the blue and violet part of the spectrum. Solu- 

 tions of such sugars act themselves as light filters and absorb the rays producing the 

 greatest dispersion disturbances. They show upon polarization but little difference 

 between filtered and unfiltered light. Clarified solutions of several low-grade beet 

 sugars were found to absorb all of the violet but only a part of the blue. Slight rotation 

 dispersion was obtained without the bichromate cell. Ninety-six degree centrifugal 

 sugars give usually straw-colored solutions, which absorb most of the violet, but prac- 

 tically nothing of the blue. Rotation dispersion with these sugars is usually well 

 marked without bichromate. Java and other high-grade sugars give upon clarification 

 nearly colorless solutions which show very pronounced rotation dispersion without 

 the bichromate. With such sugars the difference in reading with and without bichro- 

 mate was in some cases nearly 0.2 per cent for the same observer. 



The error due to rotation dispersion was found by Schonrock to be variable with dif- 

 ferent observers, a circumstance due perhaps to some physiological difference in the 

 pigment of the eye. Comparisons which I have made on five sugars polarizing over 

 96, using no bichromate and 1 and 3 per cent solutions of bichromate in a 3 cm cell, 

 showed that the discrepancies in the readings of the same solution between four 

 observers were augmented six and one-half times, when no bichromate was used, as 

 compared with the 3 per cent bichromate, and two and one-half times when 1 per 

 cent bichromate was used, as compared with the 3 per cent. Using a 3 per cent solu- 

 tion of bichromate in a 3 cm cell the average difference between the readings of the 

 lowest of the four observers and the other three was only 0.03 V., using a 1 per cent 

 solution the average difference was 0.08 V., and using no bichromate 0.22 V. The 

 3 per cent bichromate in a 3 cm cell gives the same effect as the 6 per cent bichro- 

 mate in a 1.5 cm cell advocated by Schonrock. The use of bichromate of the above 

 concentrations according to the length of cell should therefore be prescribed and 

 rigidly adhered to in the polarization of sugars. 



These concentrations apply, however, only to cane sugar. With substances of 

 greater rotation dispersion such as commercial glucose, dextrin, malt products, etc., 

 it will be found necessary to increase the strength of the bichromate considerably, as 

 may be seen from the following: 



Polarizations of starch conversion products with and without bichromate (V.). 



