which contains phosphorus pentoxide, and finally through the carbon dioxide-absorption tube, K, containing Ascarite, backed by phosphorus pentoxide. The absorption tubes, K, are connected into the train by means of mercury-cup seals [18]. This type of connection makes possible a rapid exchange of the absorption tubes. Tube K is protected by a soda-lime tube, L, which is followed by a calibrated flowmeter, M, for estimating the rate of flow of nitrogen through the apparatus. The reaction flask is immersed in a vessel containing about 16 liters of hydrogenated cottonseed oil. A bath temperature of 130° C was found to be optimum for maintaining a steady but gentle boiling of the reaction mixture. The bath is brought to the operating temperature by means of two electric immersion heaters, one of 500 and one of 1,000 watts. When the desired temperature is reached, the 500-watt heater alone is sufficient to maintain thermal constancy within ±0.2° C. The time required to raise the temperature of the bath to 130° C is approximately 50 minutes. The flask is placed in position in the oil bath so that the oil level is 3 to 4 mm lower than the liquid level inside the flask. This precaution is taken to prevent the baking of small bits of sample which may be splashed against the sides of the flask. Nitrogen, at the rate of about 10 liters per hour, is passed through the apparatus until the Ascarite tube, K, shows no further gain in weight. This operation requires about 30 minutes, during which time the temperature of the oil bath is slowly raised to 50° C. When the apparatus is free of carbon dioxide, both heating units are turned on and the temperature is brought to 130° C. This procedure is always carefully followed in order to assure the same preliminary heating for all samples. The point of zero time is taken as the time at which the bath reaches 130° C. At that time the Ascarite tube, K, is removed for weighing and a second weighed Ascarite tube inserted in its place. At the end of 1 hour the second tube is removed for weighing and replaced by the first. This process is repeated at intervals of 1 hour for the duration of the analysis. When analyses are made of pure uronic acids or materials rich in uronic acids, a small amount of carbon dioxide is evolved by the time the temperature of the bath reaches 130° C. In these cases, this amount is measured and added to that evolved during the first hour. Since the rate of evolution of carbon dioxide is appreciably affected by variations in acid strength, it is essential that the same concentration (within 0.02 percent) of hydrochloric acid be used in all analyses. The acid should be accurately 12 percent, or 3.290 N. To determine the correction for the carbon dioxide evolved by decomposition of carbohydrates other than uronic acids, weigh the Ascarite containers hourly until the rate of increase in weight is constant. The constant rate of evolution indicates that the carbon dioxide is being derived solely from the uronic acid-free carbohydrates. Calculate from the determined increase per hour the total weight of carbon dioxide which was evolved during the period (3 to 5 hours) before the rate became constant, and deduct the computed weight from the total evolved during this period. The weight of carbon dioxide times 4.00 equals the weight of uronic acid anhydride. 323414°- 42-16 3. DETERMINATION OF TWO SUGARS IN A MIXTURE (a) TWO SUGARS BY COMBINATION OF TWO POLARIMETRIC EQUATIONS In many instances it is possible to polarize a sugar mixture under conditions sufficiently different to emphasize some striking difference in the properties of the two sugars. The variation in specific rotation of the two individuals under the varied conditions must be known in order to substitute in the two corresponding equations. If x and y are the respective percentages of each sugar in the mixture, a and a' the known specific rotations of one of the sugars under the two varied conditions, and b and b' those of the second sugar, in which [a], and [a]' are determined experimentally. The specific rotations can obviously be replaced by the saccharimetric constants. While it is possible theoretically to determine both constituents of a mixture by the procedure outlined, the method is most frequently used to determine one constituent selectively in the presence of an optically active impurity. Thus the Clerget method, which has been described in detail, is employed for the determination of sucrose in the presence of invert sugar. Theoretically, it should be possible to calculate the invert sugar also, but it is at present difficult to assign with confidence a definite value to the specific rotation of invert sugar, since our knowledge of the partial rotatory powers of the constituents of sugar mixtures is incomplete. The principle of the method is used in the determination of mixtures of sucrose and raffinose by the Creydt raffinose formula, which would yield exact results for both constituents but for the complication that usually a third group of optically active substances, namely aminoacids, contaminates the product which is subjected to analysis. Other examples of the use of two polarimetric equations have been cited in the description of the determination of levulose and invert sugar by polarization at two temperatures. In some instances the second constituent of the mixture can be determined by calculation from the residual rotation obtained by deduction of the rotation of the determined constituent from the observed rotation. (b) TWO SUGARS BY COMBINED POLARISCOPIC AND REDUCTION EQUATIONS (1) BROWNE FORMULAS.- A thorough study of the determination of two sugars in a mixture by a combination of polariscopic and reducing-power methods has been made by Browne [2] and [3, page 475]. Reducing sugars are determined by the Allihn method and polarizations, observed in a 200-mm column, are stated in terms of Ventzke sugar degrees. Browne showed that by the Allihn method the reducing power of a sugar mixture is a strictly additive property of the constituents. The assumption is made tacitly that the polarizing power is also additive. If the reducing ratio of sugar A to dextrose is a and of sugar B is b, then in a mixture of x percent of A and y percent of B, the combined influence is ax+by=R, in which R is the percentage of total reducing sugars determined as dextrose. If the relative polarizing power of sugar A is a and that of B is 8, then, in the mixture, ar+By=P, in which P is the polarizing power of the mixture of sugars expressed in Ventzke degrees. Hence Relative polarizing power (a and 8) is defined as the ratio of specific rotation of the sugar in question to that of sucrose. The values of a and B for 20° C and a concentration of 10 percent are given in table 29. For levulose and galactose these values vary considerably with concentration and temperature and must be calculated by the formulas Levulose a=-1.3393-0.00166p+0.0085 (t-20), in which p is the percentage of the sugar in the solution polarized. TABLE 29.-Constants applicable to the Browne method of analysis of sugar mixtures The constants required for calculation are given in table 29. The list is capable of extension as the constants for other sugars are determined. The values tabulated under a and a refer to the sugar A and those under b and 8 to the sugar B. Example. A solution containing 4.52 percent of levulose and 4.84 percent of dextrose rotated -2.15° V in a 200-mm tube at 22° C and showed a reducing power equivalent to 9.06 g of dextrose. By the above formula, a=-1.3378. Then x (percentage of levulose) = 0.7939.06 (−2.15) 1 4.524. 0.915X0.793-(-1.3378) (percentage of dextrose) = 9.06-(0.915×4.524)=4.92. (2) MATHEWS FORMULA.-The most commonly occurring mixture of two sugars which can be analyzed by a combination of reducing and polarizing equations is that of dextrose and levulose. Mathews [6, p. 433] has derived a formula which permits a ready calculation of the ratio of levulose to total reducing sugar when the sample has been polarized in a saccharimeter and its reducing power determined by the Lane and Eynon method of titration. The method of calculation is valid under the assumptions that no optically active or reducing substance other than dextrose and levulose is present in the sample, and that the rotation of the mixture is the algebraic sum of the rotations of the constituents whose specific rotations are referred to the concentration of total sugar rather than to the partial concentration of each. While the method of determination strictly applies only to pure mixtures of dextrose and levulose, it may frequently be applied to crude mixtures, such as fruit juices, to yield a proximate analysis. At this Bureau the method has been applied to numerous samples of hydrolyzed juices of the jerusalem artichoke for rapid proximate analysis. The sugar mixture in such products consists of about 70 to 80 percent of levulose, about 20 to 25 percent of dextrose, and a small quantity of dextrorotary difructose anhydrides, which introduces an error of about 2 percent into the analysis. Application of an empirical correction diminished the error considerably. The procedure is simple. If the levulose content is high, prepare a sample containing 15 to 20 percent of sugars, or somewhat more if dextrose is the predominating sugar. Polarize in a 200-mm tube, preferably at 20° C. Dilute a measured aliquot to such volume that the resulting solution contains about 0.5 g of sugar per 100 ml and titrate against 25 ml of mixed standardized Soxhlet solution by the method of Lane and Eynon. If necessary, correct the burette reading to conform to an exactly standardized Soxhlet reagent. The method of calculation is greatly facilitated by use of table 95, p. 601. Example.-Assume that a solution of levulose and dextrose polarized - 43.8°S at 20° C, and that 5 ml of this solution diluted to 100 ml gave a Lane and Eynon titration (25 ml of Soxhlet solution) of 26.18. Then D=100/5=20 and PT - 43.8X26.18 - 57.3. By table 95, p. 601, the approximate ratio is 89.8 percent, and the correction factor, f, is -0.80. The correction is and the true ratio is 89.8-0.6-89.2. The concentration of total sugar is calculated in the usual way from the titer (c) TWO SUGARS BY COMBINATION OF TWO REDUCTION EQUATIONS (1) GENERAL.-For analysis of two sugars in a mixture, advantage is frequently taken of differences in reducing action which the individual sugars show under different conditions of analysis. In many instances the difference in behavior between the two sugars is so marked that one sugar can be determined selectively. In most cases the accompanying sugar produces minor effects, and corrections are required for accurate analysis. Thus Jackson and Mathews in their modification of the Nyns method found that 12.4 mg of dextrose reduced as much copper as 1 mg of levulose, but that this constant. correction could be applied with certainty. The variety of combinations by which this analysis can be conducted is considerable, but quite invariably one process is the de termination of total reducing sugar. The remaining methods of analysis can be chosen from the group of selective methods, but should take advantage of some property which the accompanying sugar lacks. (2) SUCROSE AND LACTOSE IN DAIRY PRODUCTS BY TWO REDUCTION PROCESSES.-An interesting method for the simultaneous determination of sucrose and lactose in sweetened condensed milk and ice cream has been described by White [19]. In outline, the clarified solution is subjected to the Munson and Walker method of lactose analysis and the copper referred to the appropriate column of lactosesucrose mixtures. The filtrate from the cuprous oxide, which is then free from lactose, is collected quantitatively, acidified, and heated to invert the sucrose, which is then determined in the form of invert sugar by a second reducing-sugar analysis. Inasmuch as a portion of the sucrose is destroyed during the lactose analysis, an empirical correction is applied to the cuprous oxide precipitated by invert sugar. The method is given in the following brief example: Weigh 10 g of condensed milk (20 g of ice cream) into a 250-ml volumetric flask and dissolve in 125 ml of boiling water. Mix for 3 minutes, cool to 20° C and add gradually 10 ml of Soxhlet coppersulfate solution and 6 ml of 0.5 N sodium hydroxide. Make to 1.5 ml over the mark (3.2 ml for ice cream) and filter. Determine lactose in 50 ml of the filtrate by the Munson and Walker method, using the "1 lactose-4 sucrose" column. Collect the filtrate from the cuprous oxide precipitate in a 250-ml flask and wash with 80 ml of hot water. Add 34 ml of 1+1 hydrochloric acid and invert in a boiling-water bath for 5 minutes. Cool and neutralize with 50-percent sodium hydroxide. Determine invert sugar by Munson and Walker method. Add 1.6 mg to the weight of cuprous oxide (1.0 mg for ice cream). Refer both weights of copper to the Munson and Walker table 78, p. 564. 4. DETERMINATION OF THREE SUGARS IN A MIXTURE (a) GENERAL The analysis of mixtures containing three sugars requires the application of analytical processes which yield three equations. Special methods in great variety have been brought into use for the analysis of these complex products. The combinations of methods which have proved most successful are those which include at least one process which is selective for one of the constituent sugars. The number and nature of the possible combinations of methods is large, but for the present purpose it will suffice to illustrate the principles by a few examples given in detail. In a very few instances one equation can be evaluated for total sugar in a mixture by using a physical method. Such a mixture can consist solely of pure sugars, but it is of such infrequent occurrence that the methods of analysis will not be described here. They can be found in Browne's Handbook of Sugar Analysis [3]. (b) THREE SUGARS BY COMBINATION OF POLARIMETRIC AND REDUCTION METHODS AND ONE SELECTIVE METHOD Wherever this combination can be applied, it is the simplest method of analysis of a complex mixture, involving as it does but three stand |