tate and potassium oxalate is generally unnecessary except for very low-grade raw sugars.

The sucrose content may be found from the invert-sugar content before and after hydrolysis, which is carried out as described below.

Determination of reducing sugars in refined or raw sugars containing less than about 0.4 percent of reducing sugars.-Twenty-five grams of the sample is dissolved with a quantity of neutralized standard invertsugar solution containing 0.1 g of invert sugar, and the whole is made up to 100 ml. A convenient procedure is to wash 25 g of the sample into a 100-ml flask containing the requisite quantity of neutralized invert-sugar solution. After solution and completion to volume, the solution is titrated against 10 ml of Fehling solution. The percentage of invert sugar in the sample is found by deducting 0.1 from the percentage of invert sugar in the solution and multiplying the remainder by 4, or it may be found directly from table 87, page 592.

Determination of invert sugar in molasses.-If the sucrose in the sample is to be determined polarimetrically, it is convenient to weigh 65 g of the sample and dilute to 250 ml, using 50 or 100 ml of this solution for determination of reducing sugars, as described below.

Cane molasses.-A solution containing 13 g of the sample is treated with 25 ml of a 10-percent solution of normal lead acetate, made up to 250 ml, shaken and filtered. Fifty milliliters of the filtrate is treated with 5 ml of a 10-percent solution of potassium oxalate, a little alumina cream is added, and the solution is made up to 250 ml, shaken, and filtered. The filtrate, which is free from lead and calcium salts, represents a 1.04-percent solution of the sample and is titrated against 10 ml of Fehling solution. The percentage of invert sugar in the sample is found from table 85 by interpolation between the values given in the columns headed "no sucrose" and "1 g of sucrose per 100 ml.'

NOTE. For purposes of interpolation, a 1.04-percent solution of cane molasses may be taken as containing 0.3 percent of sucrose.

Beet molasses. One hundred milliliters of a solution of the sample (26 g in 100 ml) is treated with 30 ml of a 10-percent solution of normal lead acetate, shaken, and filtered. The filtrate, representing a solution of the sample of 20 g in 100 ml, is treated with enough solid potassium oxalate (e. g., about 2 g) to remove lead and calcium and filtered. The filtrate is titrated against 10 ml of Fehling solution, the percentage of invert sugar in the sample being found from table 85, column headed "10 g of sucrose per 100 ml." If the sample contains more than 1 to 1.5 percent of invert sugar, the deleaded and decalcified solution should be further diluted with distilled water before titration.

NOTE. A 20-g sample of beet molasses may be taken as containing 10 g of

sucrose.

Determination of sucrose in cane or beet molasses.-Fifty milliliters of the clarified, deleaded, and decalcified 1.04-percent solution of cane molasses obtained as above is treated with 15 ml of 1.0 N hydrochloric acid, diluted to about 150 ml, heated to boiling, and kept boiling for 2 minutes. The solution is then cooled, neutralized with sodium hydroxide, and made up to 200 ml. The solution, which now represents a 0.26-percent solution of the sample, may be titrated against 10 ml of Fehling solution, the invert-sugar content being found from table 85,

column headed "no sucrose." The sucrose content of the sample is calculated from the difference between the invert-sugar content before and after hydrolysis. Sucrose in beet molasses may be determined by a similar method.

Determination of reducing sugars in starch, dextrin, and glucose.— For most purposes it is sufficient to assume that the reducing sugars in starch dextrin and starch sugar consist of dextrose. This is determined by titrating a solution of the sample of suitable concentration against 10 or 25 ml of Fehling solution and finding the dextrose content of the solution from table 85 or 86.

If it is desired to determine dextrose and maltose in starch products, the procedure devised by Morris [37] may be used.

Determination of lactose in sweetened condensed milk.-In the analysis of sweetened condensed milk, the determination of lactose by Fehling solution is somewhat affected by the sucrose present. The effect of the sucrose is to reduce the volume of lactose solution required, and it may be allowed for by adding the requisite correction to the burette reading. Tables 85 and 86 give the values for lactose itself, and table 88 shows the burette corrections, in milliliters, for sucrose-lactose ratios of 3:1 and 6:1. At any given part of table 88 the correction is practically proportional to the sucrose-lactose ratio, and the proportionality holds up to a ratio of about 10:1.

(b) SCALES METHOD

(1) GENERAL.-Scales [38] devised an iodometric semimicro method of analysis which requires the use of samples of less than 20 mg of monosaccharides or less than 30 mg of diaccharides. The entire analysis is conducted in a single reaction vessel. Scales confined his experimental analyses to dextrose and concluded that each milliliter of iodine represented a constant weight of sugar irrespective of the concentration. He prescribed a 3-minute period of boiling over a free flame or an electric hot plate.

Isbell, Pigman, and Frush [39] employed a modification of the method for determining the reducing powers of a wide variety of sugars. They found that greater precision of analysis was obtained if the period of boiling was increased to 6 minutes. Shaffer and Somogyi (see p. 169) showed that relatively large differences in reaction rates occurred between certain sugars. Hence in order that a uniform method of analysis might be adopted, the time of reaction must be adapted to the most slowly reacting sugar. If the rapidly reacting sugars, dextrose and levulose, are analyzed, it is suitable to shorten the period of boiling to the 3 minutes prescribed by Scales.

Isbell, Pigman, and Frush were unable to verify Scales' statement that each milliliter of iodine represented a constant weight of glucose, irrespective of the concentration of sugar, but found the function

f=1.0511+0.0021C+0.000086C2,

in which ƒ is the factor for glucose and Cis the titer in milliliters of 0.04 N iodine. The factors for glucose are shown in tables 22 and 23. The various sugars under strictly comparable conditions reduce different quantities of copper. It may be observed by comparing the reducing values given in table 24 that epimeric sugars give approxi

903232 O-50-14

TABLE 22.-Factors for calculating glucose by the Scales method

mately like reducing values-that is, the configuration of carbon 2 does not materially influence the reducing power. The similarity of epimeric sugars probably arises from the rapidity with which they revert to the common enolic form. Since the groups about carbons 3, 4, and 5 are held more firmly, sugars which differ in the configuration of these carbons are not interconvertible in the alkaline solution and exhibit characteristic properties. Sugars in which the hydroxyl on carbon 3 is trans to the hydroxyls on carbons 4 and 5 give the highest reducing powers, while sugars which have cis hydroxyls on carbons 3 and 4 give lower reducing values. The differences in the values for lactose, maltose, cellobiose, and turanose show that the configuration of the nonreducing component in the disaccharide molecule influences the reducing power and that the reducing power of the disaccharide may be higher or lower than that of the reducing sugar component. If the glycosidic union is on carbon 3, as in turanose, the molecular reducing power is less than that of the corresponding monosaccharide. If the glycosidic union is on carbon 4, the molecular reducing power is about 1.4 that of the corresponding monosaccharide; and if the glycosidic union is on carbon 6, the molecuTABLE 23.-Factors for various sugars for use with the modified Scales method 1

! The factors given in this table are the averages of the values obtained in measurements conducted in this laboratory over a period of 8 years. Each value represents not less than 3 determinations. The individual determinations by the same operator did not vary over 1 percent, but the results of different operators varied by as much as 2 percent.

TABLE 24.-Relative molecular reducing power1 (modified Scales method)

1 Ratio of the reducing power of the sugar to the reducing power of glucose.

lar reducing power is about 1.2 that of the corresponding monosaccharide. It may be observed from the results given in table 25 that the molecular reducing powers of the pentoses are slightly lower and of the heptoses slightly higher than the molecular reducing powers of the corresponding hexoses.

TABLE 25.-Molecular reducing power for configurationally related substances

The configurations of carbons 2, 3, 4, and 5 are indicated by plus and minus signs according to whether the OH lies to the right or left when the formula is written in the conventional manner. For example, the configuration of d-glucose is indicated by +++. In the case of the pentoses, the sugar has been classified with the configurationally related hexose. Thus d-xylose, d-lyxose, l-arabinose, and l-ribose are placed with groups having the (+) configuration for carbon 5.

(2) METHOD FOR MAKING SUGAR DETERMINATIONS.

Reagents. (1) Dissolve 16 g of copper sulfate (CuSO4.5H2O) in 125 to 150 ml of water. Dissolve sodium citrate, 150 g of 2NазC ̧H5O7. 11H2O or 124 g of Na,C,H,O7.2H2O; 130 g of anhydrous sodium carbonate; and 10 g of sodium bicarbonate in about 650 ml of water, while warming them slightly to accelerate solution. Cool and combine the two solutions while stirring, make up to 1 liter, and filter. (2) 0.04 N solution of iodine containing 4 percent of potassium iodide.

(3) 0.04 N sodium thiosulfate solution containing 0.1 g of sodium carbonate.

(4) Hydrochloric acid solution containing 60 ml of concentrated HCl per liter.

(5) Acetic acid solution containing 24 ml of glacial acetic acid per liter.

Procedure.-To 10 ml of a solution containing 10 to 20 mg of a monosaccharide (or about 30 mg of a disaccharide) and contained in a 300 ml Erlenmeyer flask, add from a fast-draining pipette 20 ml of the copper reagent. Stopper the flask with a two-hole rubber stopper and place over an electric heater or gas flame so regulated that the solution comes to boil in 4 minutes. Allow the solution to boil for 6 minutes and then cool it in an ice-water bath for 4 minute while keeping the solution in gentle circular motion. Remove the flask from the ice-water bath, draw up 25 ml of 0.04 N iodine solution into a pipette, and while holding the pipette, pour into the flask from a graduate 100 ml of the acetic acid solution, mix gently, and add the iodine solution from the pipette. Then pour 25 ml of the hydrochloric acid solution down the walls of the flask and into the solution. Mix with a gentle circular motion and titrate the excess iodine with 0.04 N sodium thiosulfate, using starch as the indicator. Subtract the back titration from the iodine originally added and multiply this value by the sugar factor to give the milligrams of sugar in the sample. For best results each worker should determine his own factors by applying the method to known quantities of sugars. If a large number of determinations is to be made, charts in which the factors are plotted against the titrations may be constructed so that the factor for any titration may be obtained readily.

(c) SCHOORL METHOD FOR INVERT SUGAR IN CANE MOLASSES [40]

Solutions. Soxhlet solution, p. 170. Deleading solution. Dissolve 7 g of Na2HPO4.12H2O and 3 g of K,C2O4.H2O and make to 100 ml. Procedure.-Dissolve 6 g of molasses in water in a 250-ml volumetric flask and defecate with 15 ml of a 10-percent neutral lead acetate solution. Make to volume and filter. Transfer 50 ml of the filtrate to a 100-ml flask and add 5 ml of deleading solution. Make to volume and filter. Pipette 50 ml of the filtrate containing 0.6 g of molasses into a 300-ml Erlenmeyer flask and add accurately 50 ml of Soxhlet reagent. Add one or two fragments of washed and ignited pumice, and place the flask on a wire gauze resting on an asbestos card with a central hole 6.5 cm in diameter. Heat to boiling in 4 minutes and continue the boiling for exactly 2 minutes. Cool rapidly without agitation and add 25 ml of KI solution (20 g in 100 ml) and 35 ml of H2SO, (1 volume of concentrated acid to 5 volumes of water). Titrate