REPORT OF THE CHEMIST, 67 



nate, as ^Till appear from the following, by omittiug the second and third 

 members of the series. Thus : 



( 2 Ce Hi, Oe ) _ ( 5 Cu2 O ) _ ( 10 Fe S O, ) _ ( K. Mu, Os \ 



\ oGO parts ) " \ 715 parts i ~ \. 1,520 parts ) ~ ( 31G.2 parts ) 



• 



In other words, 316.2 parts by weight of x)otassium permanganate are 

 equivalent to 3G0 parts of glucose, or one part of permanganate corres- 

 ])onds to 1.13S5 parts of glucose. If, then, the amount of permanganate 

 decolorized be multiplied by 1.1385 it vv ill correctly represent the amount 

 of glucose present. So much for the theoretical explanation. In prac- 

 tice it is found that each chemist must determine for himself his titra- 

 tion error by estimations made upon sugar of known purity. 



This individual error is due to the difficulty in determining the exact 

 end reaction ; experience has shown, in the course of this work, that 

 the point where the color of the permanganate barely appears in the 

 rapidly agitated hquid is nearly identical with the true end reaction. 

 Some operators carry the titration a little further until a faint rose tint 

 is permanent for about two seconds. Each man who has done this work 

 has carefully determined his titration error, and all figures submitted 

 have been corrected therefor. The iron solution works best if very 

 strongly acidulated with sulphuric acid. The most convenient strength 

 for the permanganate solution is 4.392 grams to the liter, equal to .005 

 grams glucose for each cubic centimeter. 



In the earlier i)art of these determinations it was not considered 

 necessary to thoroughly wash the precipitated suboxide of copper before 

 dissolving it in the ferric sulphate soliition. Carefully performed ex- 

 periments, however, showed that washing was best, and that the results 

 obtained on unwashed suboxide would equal those on the washed if 

 multiplied by .9076 for glucose and by .9'138 for sucrose. 



As the results of much careful work, it appears that if the suboxide 

 be well washed, and if each operator determines his titration error, the 

 determination of glucose by this method is very accurate. 



The amount of glucose found was divided by the weight of 8 c. c. of 

 the juice aualyzedfor percentage of glucose. The sucrose was found 

 by subtracting from the total glucose after inversion the amount origi- 

 nally present in 4 c. c. of the juice, and multiplying the remaining glu- 

 cose by .95. The percentage was then calculated in the usual way. 



SUGAR BEETS. 



vSeven samples of beet-roots in their natural condition and one sample 

 of sliced and dried root have been received and tested. Table No, I 

 gives the names and addresses of senders and dates of analyses. No 

 information has been received concerning the kind of seed sown, except 

 from tlic sender of samples 1, 2, and 3, and none in any case as to fer- 

 tilization, cultivation, or crops obtained. Table ¥o. II gives the results 

 of analytical tests, aud shows that only two roots, samples 1 and 3, come 

 u]) to the requirement for a good sugar-beet, these being the only ones 

 which contain not less than 80 per cent, of cane sugar in the soluble 

 matter of the juice. The juice of sample 2 contains 0.8 per cent, more 

 of cane sugar than sample 3, but also contains 1.9 per cent, more of mat- 

 ter not sugar, tluis bringing the proportion of cane sugar in the soluble 

 matter down to 73.3 per cent., and causing this sample to rank tixe poor- 

 est but one in the list in this respect. 



