414 PHYSIOLOGICAL CHEMISTRY. 



when necessary, and add an alcoholic caustic-alkali solution. Dis- 

 solve the precipitate consisting of saccharates in a little water, pre- 

 cipitate the potash by an excess of tartaric acid, neutralize the fil- 

 trate with calcium carbonate in the cold, and filter. The filtrate 

 may be used for testing with the polariscope as well as in the fer- 

 mentation, bismuth, and phenylhydrazin tests. The presence of 

 grape-sugar may be detected by this same process in animal fluids 

 or tissues, from which the albumins have first been removed by 

 coagulation or by the addition of alcohol. 



For the physician, who naturally wants specially simple and 

 quick methods, the bismuth test must be especially recommended, 

 as this may be controlled when necessary by the fermentation or 

 phenylhydrazin test. 



Other tests for sugar, as, for example, the reaction with orthonitrophenyl- 

 propiolic acid, picric acid, diazobeuzol-sulphonic acid, are superfluous. The 

 reaction with a-naphthol, which is a reaction for carbohydrates in general, for 

 glycuronic acid and uiucin, may, because of its extreme delicacy, give rise to 

 mistakes, and is therefore not to be recommended to physicians. 



Quantitative Estimation of Sugar in the urine. The urine for 

 such an estimation must first be tested for albumin, and if any be 

 present it must be removed by coagulation and the addition of 

 acetic acid, care being taken not to increase or diminish the original 

 volume of urine. The quantity of sugar may be determined by 

 TITRATION with FEHLING'S or KNAPP'S solution, by FERMENTA- 

 TION or by POLARIZATION. 



The titration liquids not only react with sugar, but also with 

 certain other reducing substances, and on this account the titration 

 methods give rather high results. When large quantities of sugar 

 are present, as in typical diabetic urine, which generally contains a 

 lower percentage of normal reducing constituents, this is indeed 

 of little account ; but when small amounts of sugar are present in 

 an otherwise normal urine, the mistake may, on the contrary, be 

 important, as the reducing power of normal urine may correspond 

 to 4 p. m. grape-sugar (see page 374). In such cases the titration 

 method must be employed in connection with the fermentation 

 method, which will be described later. It is to be remarked that 

 in typical diabetic urines with considerable amounts of sugar the 

 titration with FEELING'S solution is just as reliable as with KNAPP'S 

 solution. When the urine, on the contrary, contains only little 

 sugar with normal amounts of physiological constituents, then the 

 titration with FEHLING'S solution is more difficult, indeed in cer- 

 tain cases almost impossible, the results being very uncertain. In 

 such cases KNAPP'S method gives good results, according to WORM 

 MULLER and his pupils. 



The TITRATION with FEHLING'S SOLUTION depends on the power 

 of sugar to reduce copper oxide in alkaline solutions. For this we 



