Glucose and Mannitol hy B. coli communis. 101 



the addition of sodium hydroxide and iodine, and it is possible that the 

 substance causing this reaction may also have reduced mercuric chloride. It 

 was observed that the precipitate of mercurous chloride obtained in this 

 experiment in the estimation of formic acid was blackened. The figures for 

 this fourth analysis are derived by the consecutive subtractions from the 

 previous analyses, and it would therefore not be safe to insist too much upon 

 small differences here, especially as the total amount of sugar fermented 

 during this last interval is small. 



It is against the principles indicated in the commencement of this com- 

 munication to prolong the period of a fermentation beyond the time during 

 which the main primary changes occur, and in the discussion of these results 

 therefore conclusions will alone be considered in as far as such can be 

 supported by the evidence of the first 48 hours. Even this period is far too 

 long and it will be the object of future researches to reduce it. 



The balance sheet of this experiment is made out in the following Table. 



Table II. — Balance Sheet for the Experiment on the Phases of the 

 Decomposition of Grlucose by B. coli comraunis. 



Time from the 

 heginrdng. 



Carbohydrate decompoeed. 



Products recoTered. 



Glucose. 



Non-reducing 

 carbohydrate. 



Of degradation. 



Of synthesis. 



hours. 

 12 

 24 

 48 

 72 



10 -78 

 10 -54 

 6-70 

 3 -41 



4-98 

 0-67 



11 -30 

 4-86 

 9-32 

 4-08 



7-40 





31 -43 



5 '65 



29 -56 



7-40 





37 -08 



36 



•96 



For the sake of easier comparison the results are grouped together in 

 Table III, the products being represented as percentage of the total weight 

 of products of degradation in each period. The balance sheet shows that 

 there is no total loss during these four parts of the experiment, but, owing 

 to the difficulty of being certain as to the amount of glucose as determined 

 by reduction of copper solution, it may be that the proportion between 

 reducing and non-reducing carbohydrate during the intermediate periods 

 is not of the same order of accuracy as the determination of the ultimate 

 degradation products. Since, however, the non-reducing substance produced 



