centration of the sample is negligible. The objecton to its use is the disagreeable odor and corrosive effect of free bromine.

In order to make use of the Stanek-Babinsky principle, Schlemmer [49] studied the clarification with Aktivin or Chloramin-T, which reacts with water as follows:

In combination with sodium bromide, the oxygen set free releases two atoms of bromine from the sodium bromide. The released bromine clarifies the solution which, after filtration, remains clear for about 1⁄2 hour. The precipitate from beet molasses weighs about 0.15 g. No odor of bromine is appreciable.

The following solutions are required:

(a) Hydrochloric acid, 18.38 percent; d=1.092.

(b) Sodium acetate, 400 g; and potassium bromide, 50 g in 1 liter. (c) A solution containing about 15 percent of Chloramin-T.

(d) A mixture of solutions (a) and (b) in the ratio 20 ml of (b) to 10 ml. of (a).

Procedure for beet molasses.-Transfer 52 g of molasses to a 200-ml flask and fill to the mark at 20°. Mix thoroughly, and pipette two 50-ml portions to 100-ml flasks. To one add 30 ml of solution (d). Add 10 ml of solution (c). Adjust to 20°, mix, filter, and polarize at 20°. To the other solution add 10 ml of solution (a), invert according to Schrefeld's method (page 129), cool, and add 20 ml of solution (b). Add in 3 portions 10 ml of solution (c), make to volume at 20° C, filter, and polarize at 20°.

Schlemmer determined the value of the divisor for one-half-, onefourth-, and one-eighth-normal solutions of sucrose and, surprisingly enough, found no variation with concentration. He reported the values 131.98 for pure sucrose and 131.75 at 20° for final beet molasses. Apparently no measures are taken in these methods to evaluate the raffinose content of beet products.

Steuerwald [50] devised a method, extensively used in the Dutch East Indies, in which the inversion is carried out at room temperature by hydrochloric acid of such high concentration that the reaction is completed without the attention of the analyst within 2 or 3 hours. The direct polarization is observed in the usual manner.

For the invert polarization, measure 50 ml of the clarified filtrate with a 100-ml flask, and add 30 ml of hydrochloric acid of 1.1 sp gr (acid of 1.188 sp gr diluted with an equal volume of water). Set aside for 3 hours if the temperature is between 20° and 25° C or for 2 hours if above 25°. Dilute the solution to 100 ml and polarize at a carefully observed temperature. Calculate both polarizations in terms of the normal weight of the sample in 100 ml.

Calculate the percentage of sucrose by the formula

Jackson and Gillis [3, p. 168] showed that the high basic value of the Steuerwald divisor was consistent with their own and other values of the divisor if the effect of the acid is considered.

If the sample contains a considerable quantity of invert sugar it would seem probable that the Steuerwald method would yield high results for sucrose, since the effect of the acid is to increase greatly the negative rotation of original invert sugar in the invert polarization. This effect is uncompensated.

5. REFERENCES

[1] M. T. Clerget, Ann. chim. phys. [3] 26, 175 (1849).

[2] C. A. Browne, J. Assn. Official Agr. Chem. 2, 134 (1916).

[3] R. F. Jackson and C. L. Gillis, Sci. Pap. BS 16, 141 (1920) S375.

[4] A. Herzfeld, Z. Ver. deut. Zucker-Ind. 38, 699 (1888).

[5] J. Dammüller, Z. Ver. deut. Zucker-Ind. 38, 746 (1888).

[6] H. S. Walker, Sugar 17, 47 (1915).

[7] L. M. Tolman, Bul. Bur. Chem. 73, 73 (1903).

[8] L. G. L. Steuerwald, Int. Sugar J. 16, 82 (1914).

[9] R. F. Jackson and C. L. Gillis, Z. Ver. deut. Zucker-Ind. 70, 521 (1920).

[10] O. Schrefeld, Z. Ver. deut. Zucker-Ind. 70, 402 (1920).

[11] O. Spengler, K. Zablinsky, and A. Wolf, Z. Wirtschaftsgruppe Zuckerind. 86,

670 (1936).

[12] H. S. Walker, J. Ind. Eng. Chem. 9, 490 (1917).

[13] S. Arrhenius, Z. physik. Chem. 4, 230 (1889).

[14] O. Gubbe, Ber. deut. chem. Ges. 18, 2207 (1885).

[15] R. F. Jackson and E. J. McDonald, J. Assn. Official Agr. Chem. 22, 580 (1939).

[16] C. Tuchschmidt, Z. Ver. deut. Zucker-Ind. 20,, 649 (1870).

[17] R. Gillet, Z Ver. deut. Zucker-Ind. 64, 271 (1914).

[18] Official and Tentative Methods of Analysis of the Association Official Agricultural Chemists, 3d ed. (1935).

[19] W. C. Vosburgh, J. Am. Chem. Soc. 43, 219 (1921).

[20] F. W. Zerban, J. Assn. Official Agr. Chem. 8, 384 (1925); 11, 167 (1928). [21] E. von Lippmann, Die Chemie der Zuckerarten, 11, 1188, (Vieweg u. Sohn, Braunschweig, 1904).

[22] R. F. Jackson and C. L. Gillis, Louisiana Planter 66, 380 (1921); Facts About Sugar 13, 10 (1921); Int. Sugar J. 23, 445 (1921).

[23] C. A. Browne, Louisiana Planter 66, 171 (1921); Facts About Sugar 12, 230 (1921).

[24] R. J. Brown, Ind. Eng. Chem. 17, 39 (1925).

[25] C. A. Browne, J. Assn. Official Agr. Chem. 2, 138 (1916).

[26] E. Saillard, Eighth Int. Cong. Applied Chem. Communication 25, 541 (1912). [27] E. Saillard, J. fab. sucre (May 22, 1912 and July 1, 1914); Z. Ver. deut. Zucker-Ind. 64, 841 (1914).

[28] F. W. Zerban and C. A. Gamble, Ind. Eng. Chem., Anal. Ed. 5, 34 (1933). [29] R. Creydt, Z. Ver. deut. Zucker-Ind. 37, 153 (1887).

[30] C. A. Browne and C. A. Gamble, J. Ind. Eng. Chem. 13, 793 (1921). [31] S. J. Osborn and J. H. Zisch, Ind. Eng. Chem., Anal. Ed. 6, 193 (1934). [32] F. W. Zerban and C. A. Gamble, Ind. Eng. Chem., Anal. Ed. 5, 34 (1933). [33] F. W. Zerban, J. Assn. Official Agr. Chem. 8, 384 (1925); 9, 166 (1926); 10, 183 (1927); 11, 167 (1928); 12, 158 (1929); 13, 188 (1930); 14, 172 (1931). [34] F. W. Zerban, Orig. Com. Eighth Int. Cong. Applied Chem. 8, 103 (1912). [35] J. A. Ambler, Int. Sugar J. 29, 439 (1927).

[36] C. A. Browne and R. E. Blouin, Louisiana Sugar Expt. Sta. Bul. 91, 93 (1907). [37] C. S. Hudson and T. S. Harding, J. Ind. Eng. Chem. 7, 2193 (1915).

[38] F. W. Reynolds, Ind. Eng. Chem. 16, 169 (1924).

[39] L. Michaelis and H. Davidsohn, Biochem. Z. 35, 386 (1911).

[40] R. Willstätter and R. Kuhn, Ber. deut. chem. Ges. 56, 509 (1923).

[41] M. Adams and C. S. Hudson, J. Am. Chem. Soc. 60, 982 (1938).

[42] N. K. Richtmyer and C. S. Hudson, J. Am. Chem. Soc. 60, 983 (1938).

[43] H. S. Paine and R. T. Balch, J. Am. Chem. Soc. 49, 1019 (1927).

[44] F. W. Zerban, J. Am. Chem. Soc. 47, 1104 (1925).

[45] H. S. Paine and R. T. Balch, J. Ind. Eng. Chem. 17, 240 (1925).

[46] F. W. Zerban, J. Assn. Official Agr. Chem. 12, 158 (1929).

[47] V. Stanek, Z. Zuckerind. Böhmen 38, 429 (1914); Int. Sugar J. 16, 387 (1914).

[48] J. Babinski and W. Ablomowicz, Gez. Ceur. 1914-15, T44, S10, 147.

[49] J. Schlemmer, Z. Zuckerind. čechoslovak. Rep. 53, 13 (1928).

[50] L. Steuerwald, Arch. Suikerind. 21, 831 (1913). Int. Sugar J. 15, 489 (1913).

IX. CHEMICAL METHODS FOR THE DETERMINATION OF REDUCING SUGARS

1. THEORETICAL AND GENERAL

(a) INTRODUCTION

The history of the growth of reducing-sugar analysis begins in 1815, when Vogel showed that the reddish precipitate produced by boiling copper acetate with honey was not metallic copper, as had previously been supposed, but was cuprous oxide. From this small beginning the development was slow, with the major steps in progress decades apart. In 1841 Trommer found that, by making the copper solution alkaline, not only was a differentiation of sugars made possible, but the sensitivity was increased. In 1838 a French society offered a prize for a successful method of quantitative estimation of sugar, and an award of a portion of the prize was made in 1844 to Barreswil, who adapted Trommer's qualitative method to a quantitative method of analysis. He also showed that cane sugar could be determined by observing its reducing power before and after inversion.

In 1849 H. Fehling [1] worked out with great care the details of the method, giving some account of the stoichiometrical equivalents. Fehling believed that one molecule of glucose reduced five equivalents of copper, not recognizing that the reaction is quantitative only within narrow limits of concentration and time of reaction. Fehling's method proved satisfactory in respect to sensitivity and reproducibility of analysis, but the copper solution was unstable.

Soxhlet [2] effected still further improvements, utilizing the same reagents in the same proportions as Fehling, but preserving the copper solution and the alkaline tartrate solution in separate containers until required for analysis. This solution and method have been utilized up to the present day.

(b) REDUCING SUGARS IN ALKALINE SOLUTION

When glucose, levulose, or mannose is subjected to the action of dilute alkali in aqueous solution, the three sugars undergo a mutual conversion into each other until an equilibrium is established. The composition of this mixture is the same regardless of the sugar taken as the starting material. These relations were shown in a very striking manner by Lobry de Bruyn and van Ekenstein [3], who applied the same conditions to other sugars and found equilibria between galactose, talose, and tagatose, and in many other systems. This reaction is a perfectly general one and is of practical value for the conversion of readily available sugars into new sugars or into sugars of less common occurrence.

The mechanism of the reaction has been studied extensively. Nef [4] showed that a hexose in alkaline solution was converted into a 1,2-enediol (I). He accounted for the final products by assuming that, to this enediol, water can be added in three different ways: The hydroxyl may attach to the terminal carbon atom, yielding the aldehyde group; the hydrogen attaching to the second carbon atom

The reactions occurring in alkaline solution, and briefly indicated in this and succeeding paragraphs, are too involved to describe in detail. The purpose here is to show the nature of the reactions rather than their accurate course.

by rupture of either one of the bonds, yielding either of two aldoses; or the hydroxyl may attach to the second carbon atom and yield, with elimination of water, the ketose.

Wolfrom and Lewis [5], working with tetramethylglucopyranose, observed that true equilibrium was apparently established between glucose and mannose but that no ketose appeared to be formed. On this basis they considered the reaction to consist of a simple enolization and regeneration of the carbonyl group. Thus

On more prolonged action of the alkali, the enediols may descend to the 2,3 position (II), or to the 3,4 position (III). The latter yields a 3-ketose, known as glutose (IV), which, since the hydroxyl on the second carbon may previously have undergone transformation, occurs in two forms, alpha and beta glutose. As a result of overliming or of high local concentrations of lime in the processes of cane-sugar manufacture, glutose is frequently found in cane molasses to the extent of several percent. It has about one-half the reducing power of dextrose and is unfermentable by yeast.

If the alkalinity of the solution is greatly increased, still more deepseated changes occur in the reducing-sugar molecule, and saccharinic acids and their lactones are formed, of which there are 24 isomers theoretically possible.

In the presence of oxidizing agents, such as cupric salts in alkaline solution, the enediols are strongly reducing and can take up oxygen at the expense of copper, thereby reducing it to cuprous oxide. According to Nef's theory [4], a momentary dissociation of the molecule occurs at the position of the double bond and each fragment takes up oxygen to yield the corresponding hydroxy acid. Since the enediol may, at the time the molecule is ruptured, be either at the 1,2; 2,3; or 3,4 position, and since the hydroxyls may have altered their relative position, the number of different acids produced is large. Mannose, glucose, and levulose all give the same oxidation products in the presence of sodium and cupric hydroxides, namely carbon dioxide, formic, glycolic, d,l-glycerinic, l-threonic, and d-erythronic, d-mannonic, d-gluconic, a-hydroxy-methyl-d-arabonic, and the pentonic acids.

(c) COPPER-TARTRATE COMPLEXES

If a solution of copper sulfate is added to a chemically equivalent solution of sodium tartrate, cupric tartrate is precipitated and may be

isolated. If to this cupric tartrate one equivalent of sodium hydroxide is added, the insoluble cupric tartrate dissolves to form a deep-blue solution that is neutral to litmus, indicating that the whole cupric tartrate residue has behaved as an anion to neutralize the alkali. This is further shown by the migration of copper to the anode upon electrolysis. Cupric salts behave in a similar manner with many other substances, such as citrates, oxalates, salicylates, carbonates, glycerol, and cane sugar.

Fortunately, this property of forming soluble complexes is confined to the cupric salts, for, when under the influence of reducing reagents the copper is reduced to the cuprous state, the latter, being unable to form such complexes, precipitates in the alkaline solution in the form of cuprous oxide.

For the preparation of reagents suitable for sugar analysis, the alkali must be in excess of one equivalent of alkali to one of cupric tartrate, for the enolization occurs only in alkaline solutions.

(d) CLASSIFICATION OF REDUCING SUGARS

The sugar group may be classified with respect to reducing power into three classes: Nonreducing sugars, such as sucrose and raffinose; monosaccharides, such as glucose, levulose, and xylose; reducing disaccharides, such as lactose and maltose. The nonreducing sugars lack the free aldehyde or lactonyl structure which is characteristic of reducing sugars. The reducing disaccharides possess the reducing group on only one of the hexose residues. During the reduction it is this residue which is mainly subject to oxidation; consequently the disaccharides have but little more than half the reducing power possessed by the monosaccharides.

(e) MODERN INVESTIGATIONS

Subsequent to the work of Fehling and of Soxhlet, which laid the foundation for accurate analytical processes, variations of procedure and of composition of the alkaline copper solutions were proposed in great number in an effort to effect still further improvements. Many of these modifications served a useful purpose in their day, but almost all have been displaced in modern times by the unified methods. The early tendency was to devise a particular method for each sugar under examination. This necessitated the use of different reagents and different procedures and, when the sample contained a mixture of sugars, rendered an interpolation of copper equivalents impossible. These difficulties led to the establishment of unified methods of procedure, whereby the same reagents and procedure were used, regardless of the nature of the sugar. Empirical copper equivalents were then determined for the sugars of common occurrence. This unification of methods has caused most of the older methods to become obsolete. However, the latter are of historical interest and, in some instances, of intrinsic value, and they are frequently useful to the specialized worker.

Quisumbing and Thomas [6] have investigated the reduction reaction in detail, studying in particular the results caused by varying the

A list of the many alkaline copper solutions which were devised during this early period is given in Wiley's Agricultural Analysis, 1st ed. vol. 3, p. 183. Some of the more valuable early methods are described in Browne's Handbook of Sugar Analysis, 1st ed., p. 388 (1912).