Since the former strength is now commonly used in clarifying for the polariscope, its use is to be recommended on the score of convenience. Further study on the Canadian lead method is strongly recommended since this method is so simple and possesses the advantage that in adulterating with refined sugar the values fall off more rapidly than the percentage of maple sirup1.

ASH VALUES.

Van Zoeren and the writer have done considerable work on the ash values but as no results from collaborators were received our results are reserved for later publication.

OTHER METHODS.

The electrical conductivity method should be further tested before it receives official recognition. It is a very simple, rapid test. In the writer's laboratory, it has given results of a more restricted range in genuine sirups than any of the recognized methods. For collaborative purposes, it might perhaps be well to dilute the sirups to a more definite sugar content than is directed for the rapid test, but whether the increase of accuracy would compensate for the loss of time is a point that would require study.

The volumetric lead method3 has not given satisfactory results with the present samples. With sirups of high quality, there is sometimes room for difference of opinion as to the plotting of the graphs.

RECOMMENDATIONS.

It is recommended

(1) That work on the preparation of the sample be continued with a view to revision of the directions.

(2) That collaborative work be done on the Winton lead number. (3) That the Canadian lead number and conductivity value methods be further studied collaboratively with a view to their adoption.

(4) That work on the determination of moisture and ash be resumed when these topics are under study in reference to other saccharine products.

J. Ind. Eng. Chem., 1913 5: 995.

Ibid., 1916, 8: 331.

Ibid., 1916, 8: 241.

REPORT ON HONEY.

BY SIDNEY F. SHERWOOD' (Bureau of Chemistry, Washington, D. C.), Associate Referee.

The report includes a study of honeydew honeys, mixtures of honeydew and glucose, and normal honey and glucose with the object of ascertaining methods of differentiating between honeydew and normal honeys and mixtures of these products containing glucose (this problem is discussed at length by Browne2). It will be noted that honeydew exhibits a high plus polarization at 20°C. and at 87°C., both before and after inversion.

Since it was observed that normal honey usually gives a very slight precipitate with basic lead acetate, while honeydew usually gives a very copious precipitate (glucose giving a very slight or no precipitate), it was thought that a study of the application of the Winton lead number3 might prove this determination to be of value. The application of this determination to typical honeydew honeys gave values at first of 2.81 to 3.06. However, in one case, a value of 0.32 was found and, as this is lower than the values found in certain normal honeys, the investigation of the lead number was discontinued.

Since König and Karsch' called attention to the fact that after precipitating the dextrins with absolute alcohol, natural honeys exhibit levorotation while honeys containing 25 per cent or more of glucose exhibit dextrorotation, and since this method was found of confirmatory value by Browne2, it has been applied to various honeydew honeys and mixtures of honey with glucose with the following results:

* All polarization figures are stated on a basis of 20 grams of the original honey in 100 cc. of water; readings in a 200 mm. tube.

1 Present address, Bureau of Plant Industry, Washington, D. C.

2 U. S. Bur. Chem. Bull. 110: (1908).

J. Am. Chem. Soc., 1906, 28: 1204.

4 Z. anal. Chem., 1895, 34: 1.

Polarization* of honeydew honeys and mixtures of honeydew honeys and glucose after precipitation with alcohol.

* All polarization figures are stated on a basis of 26 grams of the original honey in 100 cc. of water; readings in a 200 mm. tube.

In view of the slight values for the plus polarizations of the mixtures and of the small variation between these values and the values for the honeydew honeys, it is thought that this method, in the case of the addition of small amounts of glucose, is of minimum value only.

Investigation of the precipitate thrown down by alcohol was begun but, owing to lack of time, was discontinued. It is believed that a study of this precipitate may prove of value in connection with the problem.

No report on sugar house products was made by the associate referee. No report on food preservatives was made by the referee.

REPORT ON COLORING MATTERS IN FOODS.

By W. E. MATHEWSON (Bureau of Chemistry, Washington, D. C.),

Referee.

The work was restricted to the consideration of tests for some of the more common natural coloring substances. The qualitative differentiation of the various natural coloring matters found in food products can scarcely be carried out satisfactorily until much more work has been done, but many tests are in use that have been found reliable and convenient. The Committee on Editing Methods of Analysis arranged some of the best known of these tests in the form of a table1. This table is somewhat incomplete, chiefly where tests are concerned in which no definite color change takes place.

It was requested that the collaborators make tests with such samples. of coloring matters as were available, so that for every case a definite

1 Assoc. Official Agr. Chemists, Methods, 1916, 166.

stat rea C2

ight be given in the table, concerning the behavior of the d coloring matters with each other. Attention was also ests depending on the treatment of an acetic anhydrid soluhe coloring matter with concentrated sulphuric acid and with agents; and to the work of Palmer and Thrun1 relative to the avior of the natural coloring matters of butter and oils with ferric chlorid and other reagents.

The following comments were received from Leonard Feldstein, U. S. Food and Drug Inspection Station, Tabor Opera House Building, Denver, Colo.:

The results obtained in the cooperative work on food colors at this laboratory are as follows:

Hydrochloric acid with alkanet produces no change. Sodium hydroxid with annatto causes fading of the yellow color and the production of a light brown color.

Ferric chlorid decolorizes alkanet, leaving a turbid dark solution. It produces no change with carotin, and turns anthocyans (cherries) deep purple.

Alum solution produces a slight purple color with alkanet; turns cochineal and anthocyans (cherries) purple-red, and has no action on annatto, carotin or caramel. Uranium acetate with Brazil wood produces a deep purple-red color; no change is noticed with annatto or caramel.

Acetic anhydrid and sulphuric acid produce a green fluorescence in the red color of Brazil wood; a yellow color is produced with cochineal. No action is noticed with caramel.

It would appear that reference should be made in the tentative methods to the tests for the detection of carotin in butter since the action of ferric chlorid and carotin, whether alone or in the presence of fats and oils, appears to be different from that in aqueous solution. Water seems to retard the reducing action of carotin on ferric chlorid; at least no change is visible. When a crystal of ferric chlorid is added to dry carotin extract, the ferric chlorid is changed. If water is then added to the mixture, a wine colored solution is produced.

REPORT ON METALS IN FOODS.

BY DAVID KLEIN2 (Division of Foods and Dairies, Illinois Department of Agriculture, 1410 Kimball Building, Chicago, Ill.), Referee.

TIN.

A study was made of a volumetric method for tin, the essential features of which were suggested by W. B. D. Penniman, Baltimore, Md. Broadly outlined, the method consists of extracting the tin with hydrochloric acid, the tin is precipitated from this solution by zinc, the mixed metal residue is dissolved in hydrochloric acid in the absence of air, and this solution is titrated with standard potassium iodate. The advantages over the provisional methods are the elimination of the acid digestions

1 J. Ind. Eng. Chem., 1916, 8: 614

'Present address, The Wilson Laboratories, Chicago, Il.

and of the sulphid precipitation. Furthermore, potassium iodate solution maintains a constant strength, and is, therefore, preferable to the variable iodin solution used in the provisional methods.

The oxidation of stannous chlorid by potassium iodate is an interesting reaction, which may be assumed to occur in the following steps:

(1) 4KIO +12 SnCl + 28 HCl = 12 SnCl, +4 HI+12 H,0+4 KC

If equation (1) goes to completion before the action represented by equation (2) begins, then it will be possible to use starch as an end point indicator, for, as soon as reaction (1) is completed, further addition of potassium iodate will liberate free chlorin. This, in turn, will displace the iodin from the hydriodic acid of equation (1), according to equation (3). Continued addition of potassium iodate will convert the liberated iodin into iodin monochlorid. This action can be traced by the decolorizing of chloroform, since iodin monochlorid does not impart color to chloroform. Thus two end points may be used. However, it has been found that the chloroform end point is not always satisfactory. With certain lots of zinc, the pink color of the chloroform was not discharged, even when a large excess of potassium iodate was added. At times the coloring matter of the original food material was carried along, and imparted a decided color to the final solution. In such cases, the chloroform end point was unreliable. For these reasons, the chloroform end point was discarded.

The sensitiveness of the starch end point is dependent upon the concentration of the iodin liberated in equation (3), since an appreciable concentration of iodin is necessary to develop the color in starch1. Where a small amount of tin is being titrated, it is conceivable that the amount of hydriodic acid formed would be insufficient to liberate the minimum quantity of iodin necessary for the development of the blue color with starch. This condition should be remedied by increasing the concentration of hydriodic acid, as by the addition of potassium iodid. Then the sensitiveness will be dependent only upon the concentration of the chlorin derived from equation (2). These theoretical considerations have been fully verified experimentally. For example, no blue color is formed, when solutions containing as much as 4 mg. of tin in 125 cc. of liquid are titrated with potassium iodate (1 cc. = 0.001 gram of tin). On the other hand, when potassium iodid was added, a sharp end point was obtained with as little as 0.0001 gram of tin.

1J. Am. Chem. Soc., 1908, 30: 45.