IN SUGAR OF DIABETIC URINE. 201 



3. About 4.0Z. of the urine were then placed in a 12 oz. 

 bottle, with a drachm or two of German yeast, and set 

 aside in a warm place to ferment, taking care to cover the 

 mouth of the bottle with a slip of glass or a loose cork. 



4. In from twelve to eighteen hours fermentation was 

 usually over, and at the end of twenty-four hours the froth 

 and scum had subsided or been dissipated sufficiently to 

 permit the density to be again taken. 



Operating in this way on a urine passed on the 215^ of 

 April, I obtained the following results : 



Sugar per loo parts by the volumetrical method ... 7 '69 



Density befoi'e fermentation at 60° or D io38"6o 



Density after fermentation at 60° or D' 1005-92 



Density lost or D-D' 32'68 



The relation, therefore, between the density lost and the 

 per centage of sugar in this instance was as 32-68 to y6g, 

 or as I to 0.235 ; so that by multiplying the density lost 

 into the co-efficient 0*235 we have for product the amount 

 of sugar per 100 parts which this urine contained. That 

 is, sugar per 100 parts or S = (D-D') x 0*235 . 



On repeating the experiment a great number of times 

 with different specimens of urine and different specimens 

 of yeast, the number 0*230 was found to be the more exact 

 co-efficient. 



The degree of exactitude with which the quantity of 

 sugar may be determined by this method is very great; 

 indeed with the precautions to be mentioned presently it 

 seems susceptible of nearly as much accuracy as the volu- 

 metrical method. 



The following table places in comparison twenty obser- 

 vations made by the two methods on various diabetic 

 urines, with densities ranging from 1031*52 to 1053*48 : 



SER. III. VOL. I. D D 



