The analytical process is conducted with the tenth-normal solution, prepared by Soldaini and described above, as follows: Place 100 cubic centimeters of the reagent in each of several porcelain dishes heat to boiling, and add little by little the sugar solution to one dish until the blue color has disappeared. Having thus determined nearly the exact quantity of sugar solution required for the copper in 100 cubic centimeters of the reagent the whole of the sugar solution is added at once, varying slightly the amounts added to each dish. The boiling is continued for fifteen minutes, and the contents of the dishes poured on filters. That filtrate which contains neither copper nor sugar represents the exact quantity of sugar solution which contained fifty milligrams of dextrose. =127. Relation of Reducing Sugar to Quantity of Copper Suboxid Obtained.=—The relation of the quantity of copper reduced to the amount of sugar oxidized by the copper carbonate solution has been determined by Ost, and the utility of the process thereby increased.[92] The solution used should have the following composition: 23.5 grams of crystallized copper sulfate, 250 grams of potassium carbonate, and 100 grams of potassium bicarbonate in one liter. Without an indicator the end reaction is distinctly marked by the passage of the blue color into a colorless solution. Ost affirms that this solution is preferable to any form of fehling liquor because it can be kept indefinitely unchanged; it attacks sucrose far less strongly, and an equal quantity of sugar precipitates nearly double the quantity of copper. The boiling requires a longer time, as a rule ten minutes, but this is a matter of no importance, when the other advantages are taken into consideration. The relations of the different sugars to the quantity of copper precipitated are given in the table in the next paragraph. =128. Factor for Different Sugars.=—For pure dextrose the relation between sugar and copper reduced has been determined by Ost, and the data are given in the table below. The data were obtained by adding to fifty cubic centimeters of the copper solution twenty-five cubic centimeters of sugar solutions of varying strength and collecting, washing, and reducing the cuprous oxid obtained in a current of hydrogen in a glass tube by the method described further on. The boiling in all cases was continued, just ten minutes, although a slight variation from the standard time did not produce so great a difference as with fehling reagent. In the case of dextrose, when fifty milligrams were used with fifty cubic centimeters of the solution, the milligrams of copper obtained after six, ten and twenty minutes’ boiling were 164.6, 165.5, and 166.9 respectively.[93] The data differ considerably from those obtained by Herzfeld, but in his experiments the boiling was continued only for five minutes, and this is not long enough to secure the proper reduction of the copper.[94]