is made for the wide range in the invert polarization of individual honeys (-29.26 to +14.96 in the table), so that a considerable error may be introduced into the final result. Another objection, which also holds true for the method of the official chemists, is that no account is taken of the possible variation in polarization of the glucose used for adulteration. The value +175 holds only for a certain class of products; with glucose sirups of high conversion the polarization may sink below +150, and with honeys adulterated with materials of this description the methods just described would give too low results.

A third possible method of estimating the glucose in honey is based upon the lowering of the difference in invert polarization between 20° and 87. For the honeys in the table the average value for this constant was found to be 26.71. Since this difference in polarization of pure honey is due entirely to the percentage of levulose, the addition of any non-levulose-containing material whose polarizing power is not affected by change of temperature will cause a depression in the polarization difference between 20° and 87° proportional to the amount of adulterant used and irrespective of its specific rotation. The polarization difference of pure honeys, however, may be considerably influenced by variations in moisture content or through an excess of honey dextrins and other nonsugars. It is therefore better to reduce the polarization difference to the constant basis of reducing sugars after inversion before making the calculation for added glucose. The method of doing this is as follows:

The difference in invert polarization between 20° and 87° is multiplied by 77, the average percentage of invert sugar after inversion for pure honey, and this product is divided by the percentage of invert sugar after inversion found in the sample. This quotient is the corrected polarization difference which, multiplied by 100 and divided by 26.7, will give the percentage of pure honey in the sample. One hundred minus the per cent of pure honey equals the per cent of glucose.

In the following table the polarizations of five honeys of considerable range in composition, selected from the list of 100 analyses, and of mixtures of the same with 20 per cent of commercial glucose, are given, together with the percentage of glucose as calculated by the three methods described.

Number.

Kind of sample.

Polarization of honeys and glucose mixtures, with calculated percentages of glucose by different formulas.

°V. •V. -22.66 + 3.52

Ꮴ.

P. ct. °V.

P. ct. P. ct.

Ꮴ.

Alfalfa +20 per cent glucose. +19.4

+16.88 +35.82

60

Hop vine..

Hop vine +20 per cent glucose +24.9

84

Whitewood..

- 4.9

Whitewood + 20 per cent glu

cose.

Basswood+ 20 per cent glu

cose.

+3.48

White oak + 20 per cent glu- |

cose...

+43.8

+39.14+55.88 16.74 63.84 20.20 34.28 29.35 24.35

It will be seen from the results in the table that with admixtures of low purity honeys and glucose there is a considerable error in the calculation of the percentage of added adulterant. The results obtained by any method of glucose estimation have only an approximate value, and in no case ought such analytical results as those obtained for the pure basswood or white-oak honey to condemn a sample as being adulterated. In all suspicious or doubtful cases confirmatory tests such as that with iodin should be employed.

GRAPE SUGAR OR COMMERCIAL DEXTROSE.

There is no absolute method for the estimation of added grape sugar in honey, yet the addition of this may often be inferred with a fair degree of certainty. A dextrorotatory honey which gives no reaction with iodin and is, at the same time, comparatively low in dextrins and other solids not sugar, contains, in all probability, added grape sugar. The "Vereinbarungen "a set the minimum for the non-sugars of honey at 1.5 per cent; such a standard has but little value, however, since it would not cover the case of low-grade honey and dextrose mixtures. If the difference in the invert polarization of a honey between 20° and 87° falls below 20 and the percentage of reducing sugars is normal and no reaction for amylo- or erythrodextrin is obtained with iodin, then commercial dextrose has, in all probability, been added.

a Part 2, 1899, p. 121.

Considerable caution is required before drawing positive conclusions as to the addition of grape sugar to a honey, since, as has already been noted, the crystalline deposits in partially granulated honey consist entirely of dextrose, and if the liquid part has been poured off, or otherwise removed, results indicating an apparent addition of grape sugar may be obtained.

DISTINCTION BETWEEN HONEYDEW AND GLUCOSE HONEYS.

Although the dextrorotatory or honeydew honeys are excluded by the standards from the class of pure honeys, they are not adulterated, in the strict sense of the word. It is important, therefore, that the chemist be able to differentiate between honeys which are dextrorotatory as the result of purely natural conditions and honeys which are dextrorotatory as the result of adulteration.

Numerous methods have been devised for this purpose. One of the first of these proposed by Haenle is based upon the supposition that honey dextrins of all kinds are easily diffusible through parchment, while dextrins derived from starch are not. This supposition has not, however, been verified by subsequent experiments. A second method recommended by Sieben is based upon the difference in fermentation of the two kinds of dextrins, the dextrins of honey being readily fermented, while those of glucose are not. The observations of Sieben, however, have not been fully confirmed, except in the case of some special yeasts, such as Saccharomyces ellipsoideus.

a

Another process given by Beckmann' has given better satisfaction and is based upon the fact that the dextrins of starch sirup or glucose, more especially their barium derivatives, are readily precipitated by methyl alcohol, and the dextrins of floral and coniferous honey are not. The dextrins of honeydew are, however, more or less precipitated by this treatment, so that the mere formation of a precipitate is of but little value as a test with very abnormal honeys. In such cases the precipitate may be dissolved and tested with iodin for amylo- or erythro-dextrin, although, as has been previously noted, this test may fail completely with a certain class of glucose products.

A fourth process used by König and Karsch depends upon the fact that, after precipitating the dextrins by absolute alcohol, the resulting solution after inversion in case of natural honeys will show a levorotation; with honeys containing 25 per cent or more of glucose,

a For a description of this and other methods for distinguishing between abnormal and adulterated honeys the reader is referred to König's Untersuchung landwirtschaftlich und gewerblich wichtiger Stoffe, 3d. ed., p. 593. bZts. anal. Chem., 1896, 35: 263; Zts. Nahr. Genussm., 1901, 4: 1065. cZts. anal. Chem., 1895, 34: 1.

a dextrorotation. This method has been employed considerably in the sugar laboratory of the Bureau of Chemistry and has been found to be of value as a confirmatory test. The method as described by König is as follows:

Forty grams of honey are dissolved in a cylinder with water to 40 cc. Of the homogeneous solution 20 cc are transferred to a 250 cc flask and filled up to the mark with absolute alcohol with slow addition and constant shaking, and then allowed to stand two or three days, with occasional agitation. At the end of this time all the dextrin has settled out. After shaking the solution is filtered and 100 cc of the filtrate evaporated until free from alcohol. The liquid residue, with the addition of a little subacetate of lead and sodium sulphate, is made up to 20 cc with water and the filtered solution polarized. Dextrorotatory natural honeys show by this method a levorotation; honeys adulterated with dextrose or glucose to the extent of 25 per cent or more, a dextrorotation. In case the honey contains a large amount of sucrose the solution should be inverted with hydrochloric acid before polarizing.

The points upon which most reliance has been placed in this laboratory in the differentiation of abnormal and glucose honeys are, first, the difference in invert polarization between 20° and 87° (corrected to 77 per cent of invert sugar); second, the reaction of the honey and of its precipitated dextrin toward iodin (Beckmann's test); third, the polarization of the inverted solution after precipitation of the dextrin with absolute alcohol (method of König and Karsch). In addition to the above, the quantity and character of the ash, acidity, and the microscopic examination of pollen, starch, etc., are all of value in determining the origin of suspicious samples.

INVERT SUGAR.

The adulteration of honey with invert sugar is much more difficult to detect than the addition of glucose or starch sirup. Invert sugar, i. e., a mixture of the sugars levulose and dextrose in equal amounts, constitutes nearly three-quarters of honey itself, so that the detection of added invert sugar becomes a possibility only in so far as the nonsugars of the honey are modified, either in quantity or through the addition of extraneous substances produced in the manufacture of the invert sugar used for adulterating.

Before discussing methods for the detection of invert sugar in honeys a short description should be given of the methods used for the preparation of invert sugar sirup.

PREPARATION OF INVERT SUGAR SIRUP.

The methods generally employed for the preparation of invert sugar sirup consist in heating to gentle boiling concentrated canesugar solution with a very small amount of mineral or organic acid (hydrochloric, tartaric, or formic). The inversion is completed very

rapidly at the high temperature, and the residual acidity is so small (no greater than in pure honey) that no further treatment of the sirup is required.

By the method of Herzfeld," which is used quite extensively in Germany for preparing artificial honey, 1 kilogram of refined sugar is heated with 300 cc of water and 1.1 gram of tartaric acid to boiling and maintained at this temperature until the solution acquires a golden yellow color (one-half to three-quarters of an hour). This preparation is said to improve greatly the appearance and flavor of low-grade honeys. An invert sirup prepared by Herzfeld's process had a rich golden color and a mild, pleasant flavor. On standing for a few weeks it began to granulate, the dextrose soon crystallizing to a solid mass. The composition of this product was as follows:

This analysis, except for the deficiency in ash, agrees very closely with that recorded for pure honeys. It conforms in every respect with the limits set by the various standards. The practical impossibility of estimating the amount of such a product in honey mixtures from the polarization or chemical analysis is at once apparent. As compared with pure honey the difference in invert polarization between 20° and 87° is rather low, owing to a slight destruction of the levulose at the high temperature of inversion. In fact, this low difference is characteristic of all invert sugar sirups prepared at a high temperature.

Invert sugar may also be prepared by inverting sucrose with larger quantities of acid at a lower temperature. With such products, however, the excessive acid has to be removed and this increases considerably the cost of manufacture.

a Zts. Ver. d. Zucker-Ind., 31: 1988.