PRODUCTION OF SUGAR IN THE LIVEK. 461 



In ordinary examinations, Trommer's test is sufficiently delicate ; but it is not so sen- 

 sitive or so convenient as some of the standard test-solutions. We have been in the 

 habit of using, for the determination of sugar in the urine, a modification of Fehling's 

 test, which is also very convenient for examinations of the blood and liver-extract. This 

 may be used as well for quantitative examinations ; but, like all of the standard solutions, it 

 presents the inconvenience of undergoing alteration by keeping, so that it is desirable to 

 use it freshly made for each series of examinations. We have succeeded in obviating 

 this difficulty, however, by the following modification in its preparation ; and, made in 

 this way, it is probably the most convenient test that can be used in the examination of 

 any of the animal fluids for sugar. 



Fehling's Test for Sugar. The modification in this test consists simply in preparing 

 three separate solutions, which are to be mixed just before using, as follows : 



Solution of crystallized sulphate of copper, 94'73 grains in an ounce of distilled water. 



Solution of neutral tartrate of potash, 378*91 grains in an ounce of distilled water. 



Solution of caustic soda, specific gravity 1-12, or about 16|, Baume"'s hydrometer. 



These solutions are to be kept in separate bottles and used as follows : 



Take half of a fluidrachm of the copper-solution, add half a fluidrachm of the tar- 

 trate of potash, and add the caustic soda, to make three fluidrachms. It is important to 

 measure the copper-solution with accuracy, in quantitative analyses, as the quantity of 

 copper decomposed indicates the amount of sugar. 



To apply Fehling's test in ordinary qualitative analyses, heat a small portion of the 

 test-liquid to the boiling-point in a test-tube, and add the suspected fiuid, drop by drop. 

 If sugar be present in even a moderate quantity, a dense, yellowish precipitate of the sub- 

 oxide of copper will be produced after adding a few drops ; and, if the liquid be added to 

 about the same volume as the test, and the mixture be again raised to the boiling-point 

 without producing any deposit, it is certain that no sugar is present. The estimation of 

 the quantity of sugar in any liquid depends upon the fact that two hundred grains of the 

 test-liquid is decolorized by exactly one grain of glucose. To apply this test, measure oft', 

 in a glass specially graduated for the purpose, two hundred grains of the solution ; put 

 this into a flask, with about twice its volume of distilled water, and boil ; when boiling, 

 add the suspected solution, little by little, from a burette graduated in grains (raising the 

 mixture to the boiling-point each time and afterward allowing the precipitate to subside), 

 until the blue color is completely discharged ; by then reading off the number of grains 

 of the saccharine solution that has been added, the proportion of sugar may be readily 

 calculated. If the solution be suspected to contain a considerable quantity of sugar, the 

 estimate may be more accurately made by diluting it to a known degree, say with nine 

 parts of water, and adding this diluted mixture to the test-liquid. 



Examination of the Blood of the Portal System for Sugar. If starch or sugar be 

 taken into the alimentary canal, it is well known that sugar is always to be found, during 

 absorption, in the blood of the portal system ; but, in carnivorous animals, that have been 

 fed entirely upon meat, no sugar can be discovered in the portal blood. The statements of 

 Bernard are very definite upon this point, and he indicates a liability to error when the 

 operation of tying the portal vein has not been skilfully performed, and when blood, con- 

 taining sugar, is allowed to regurgitate from the substance of the liver. In taking the 

 blood just before it enters the liver, it is necessary to apply a ligature to the vessels as 

 they penetrate at the transverse fissure. This should be done quickly, and the opening 

 into the abdominal cavity should be small. Otherwise, as the vessels have no valves, we 

 are liable to have reflux of blood from the liver. We have frequently performed the 

 experiment, after the method described by Bernard, making a small opening in the linea 

 alba a little below the ensiform cartilage, just large enough to admit the forefinger of the 

 left hand ; introducing the finger, and feeling along the concave surface of the liver until 

 we are able to seize the vessels ; then passing in an aneurism-needle, and constricting 

 the vessels before the abdomen is widely opened, when a firm ligature is applied. When 



