The results obtained by this method are as follows:

Determination of known amounts of benzoic acid in ketchups by Method III.

An examination of this table shows that when there was present less than 0.15 gram of sodium benzoate no precipitate was formed, and in such cases as did give a precipitate all results were from 0.1 to 0.15 gram low. This seemed to indicate that there was something in the ketchup extract which held back the precipitation and showed conclusively that the method could not be used for such materials in combination with present extraction methods.

It appeared probable that since the completeness of precipitation varied with the alcoholic strength it might be interfered with by sugars, higher alcohols, oils from spices, etc., and the following experiments were performed. The alcoholic extract from 1 gram of spice was added to solutions containing 0.1 gram of sodium benzoate and to blanks. Also 0.1 gram of sugar, glycerin, and dextrin were added to similar solutions with the following results:

Determination of sodium benzoate in the presence of spices, sugars, etc., by Method III.

It is pos

These figures show that it is the presence of the spices which causes the failure of this method for the determination of benzoic acid in such products. sible that it may be applicable to other materials.

PRECIPITATION AS SILVER BENZOATE (METHOD IV).

This method was suggested and elaborated by Mr. W. E. Hillyer.

The sample is extracted by means of ether as directed in Bulletin 107, page 179, or by the method given by Dubois for the extraction of ketchups. The amount

a J. Amer. Chem. Soc., 1906, 28: 1616.

of substance used should be such that the portion subsequently extracted with ether will contain approximately 0.1 gram of sodium benzoate. The ether extract, after washing with water, is allowed to evaporate to dryness spontaneously, or the first portion of the ether may be distilled and recovered. After drying completely the residue is taken up with a small amount of absolute alcohol, for the purpose of separating interfering substances as far as possible, and filtered into a small beaker. The alcohol is neutralized with sodium hydroxid, evaporated to dryness, and redissolved in a few cubic centimeters of alcohol saturated with silver benzoate. The solution is filtered if not clear, washed with a few drops of aldehyde-free alcohol, saturated with silver benzoate, and treated with from 10 to 15 cc of a saturated solution of silver nitrate in aldehyde-free alcohol. The precipitate is collected in a gooch, care being taken that the asbestos filter be so constructed as to afford as rapid filtration as possible. The precipitate is then heated in a water-jacketed oven until the ether is driven off, cooled, and weighed.

Care must be taken to perform all operations as quickly as possible in order to prevent the separation of silver oxid. The aldehyde-free alcohol mentioned above is about 95 per cent by volume, and is prepared according to the directions given in Bulletin 107, page 96, with the additional precaution of distilling over soda after treatment with meta-phenylene-diamin hydrochlorid. This method involves the use of a considerable quantity of ether, which is objectionable because of its inflammability and the tendency to dissolve sodium chlorid and other interfering substances. Notwithstanding this, very satisfactory results are reported by Messrs. Hillyer and Flanders in the following table. No other results obtained by this method, using ether as solvent, were reported, though the precipitation of benzoic acid as silver acetate, using chloroform as a solvent, was included in the work of several other collaborators.

Determination of benzoic acid as silver benzoate in tomato ketchup (Hillyer and Flanders).

a Last three results by F. F. Flanders; others by W. E. Hillyer.

These results were obtained by extracting and precipitating as silver benzoate in the ether residue following Method II as given on page 70. The figures seem to be in every way comparable with those obtained by Method II (see p. 71). Extraction with ether appears to be much less satisfactory than extraction with chloroform, owing to the removal of interfering substances by the solvent. These bodies are partially removed by means of absolute alcohol, but this introduces an additional operation and the results obtained are not as satisfactory as by extracting with chloroform from a solution saturated with sodium chlorid. In the table comparing Methods I and II (p. 71), is given the percentage of sodium benzoate precipitated as silver benzoate from the residue from Method II; that is, the liquid titrated under Method II was evaporated to dryness and used as a starting point for the silver benzoate method. In the following table are given the results obtained by the examination of a number of samples of commercial ketchups using this method. In all cases the benzoic acid was extracted by chloroform from a saturated sodium chlorid solution. Here, again, it will be seen that the results obtained by weighing the residue are in all cases slightly

higher than those given by titration, whereas the amounts determined by precipitation as silver benzoate are almost identical with the amount obtained by titrating the chloroform residue. This method is evidently worthy of further study. It is much more tedious than Method II, but is of value for the purpose of checking the results obtained by that method when a further confirmation seems desirable.

Determination of benzoic acid in commercial samples of tomato ketchup by precipitation as silver salt from chloroform extract.

The residue

These samples were treated according to Method II, given on page 70. obtained by the evaporation of the chloroform extract was first weighed, then dissolved in about 5 cc of neutral alcohol, the solution so obtained diluted with water and titrated with saturated alkali solution. This solution when exactly neutralized is evaporated to dryness, after which the benzoic acid was determined by precipitation as silver benzoate.

THE DISTILLATION OF BENZOIC ACID FROM SULPHURIC ACID SOLUTION (METHOD V). This method was suggested and elaborated by Mr. R. M. West, a and depends on the distillation of benzoic acid with steam after the addition of sufficient concentrated sulphuric acid to insure the complete charring of vegetable tissue and prevent volatilization of coloring matter and oil. The distillation is conducted by means of a flask shown on page 21, the procedure being as follows:

About 10 grams of the sample are weighed into the inner flask of the apparatus, 1.5 to 2 grams of paraffin added, and the flask connected with the condenser. Ten cubic centimeters of strong sulphuric acid are added through a drop funnel at a rate sufficient to complete the addition at from two to three minutes, the flask is gently agitated, to mix the contents thoroughly, and allowed to stand from five to ten minutes after all apparent action of the sulphuric acid has ceased. About 150 cc of distilled water are placed in the outer flask of the apparatus and the water slowly brought to a boil and the boiling continued until 100 cc of the distillate have been collected. The stopcock in the outer flask is left open until the water has heated sufficiently to prevent the contents of the inner flask being drawn into the outer flask.

The distillate is filtered into a separatory funnel and the original receiver and filter are washed with two portions of water of about 10 cc each. The distillate is then extracted with three portions of ether of 50, 30, and 20 cc, respectively. The combined ether extracts are washed repeatedly with water until a 25 cc portion requires not more than 0.10 cc of decinormal alkali for neutralization. The ether extract is then distilled to small volume, after which it is evaporated before a blast of air, dried in a desiccator to constant weight and weighed. The residue is also dissolved in neutral alkali, using phenolphthalein as indicator.

The results obtained by titration agree closely with those obtained by weighing. Excessive foaming is likely to occur when the steam begins to pass into the inner flask. This may be caused by distilling too soon after the addition of the acid, by an insufficient amount of paraffin, or by an unusual amount of sugar in the ketchup. Care must be exercised to prevent the foam passing into the condenser.

a J. Ind. and Eng. Chem., 1909, 1: 190.

The distillation should be conducted at such a rate that 100 cc of the distillate may be obtained in from twenty-five to thirty-five minutes. Occasionally some paraffin is carried over mechanically, and this may usually be removed from the surface of the distillate by means of a wire or glass rod.

The following results were obtained on ketchups containing a known amount of sodium benzoate:

Determination of sodium benzoate in ketchup by Method V (West).

The statement has frequently been made that cinnamic acid is being used for the preservation of foods, especially in the case of tomato ketchup. The claim has often been made by those interested in the preservation of ketchup with benzoic acid that the presence of cinnamic acid could not be detected and that firms claiming to use no preservative were preserving with that substance. Two qualitative methods for the detection of cinnamic acid, differing slightly from each other, were elaborated by P. B. Dunbar. Both of these methods depend upon the well-known fact that cinnamic acid is oxidized to benzaldehyde by dilute chromic acid mixture.

Method 1.-One hundred grams of ketchup were treated with 100 cc of water and 5 cc of sulphuric acid (1 to 5) and the mixture extracted directly with three portions of chloroform, using 50, 25, and 25 cc, respectively. The chloroform extract was made alkaline with ammonia and evaporated to dryness on the water bath. The residue was dissolved in a small amount of hot water, filtered, again evaporated to dryness, and heated to boiling with 5 cc of dilute chromic acid mixture (1 part of dilute sulphuric acid saturated with potassium bichromate and 7 parts water). The odor of benzaldehyde is strongest when the mixture is cooled until the fumes of sulphuric acid are no longer apparent.

Method 2.-Two hundred grams of the ketchup are diluted to 500 cc with water, allowed to settle, and filtered. An aliquot portion of the filtrate, 250 cc or more, is acidified with 5 cc of sulphuric acid (1 to 5), extracted with chloroform, and the remainder of the operation conducted as under Method 1.

The second method appears to be slightly more delicate than the first, although with either it was possible to detect cinnamic acid in tomato ketchup when present to an extent of 25 mg per kilogram.

This reaction is also given by cinnamic aldehyde. The method, therefore, does not distinguish of itself between cinnamic aldehyde, resulting from the use of cinnamon as a flavor and cinnamic acid used as a preservative, except that the amount of cinnamic aldehyde present in the commercial ketchups examined was not sufficient to give a reaction. If cinnamic acid were present in the ketchup, it would be detected by the methods used for the detection of benzoic acid. Cinnamic aldehyde, on the other hand, would not be detected by the methods suggested for benzoic acid. The benzoic-acid residue obtained by the evaporation of the chloroform extract may be examined by the cinnamic-acid methods described.

The germicidal and antiseptic properties of cinnamic acid were investigated by G. W. Stiles, who found them to be very much lower than those of benzoic acid. The preservation of a food, therefore, would require a much larger percentage of cinnamic

acid than benzoic acid. In fact, the antiseptic properties of a saturated solution of cinnamic acid are so slight that this substance would probably not serve as a preservative for foods.

A method for the separation of benzoic acid and cinnamic acid by precipitation of the latter with manganous salts a was tried unsuccessfully by Mr. Dunbar, who was unable to secure a precipitation of either benzoate or cinnamate of manganese in dilute solution. As is to be expected, Mohler's and Peter's reaction also give the same end reaction in the presence of cinnamic acid.

REPORT ON TEA, COFFEE, AND COCOA.

By A. G. WOODMAN, Associate Referee.

The work of the referee for the past year has been limited to a study of methods for the determination of caffein and caffetannic acid in coffee, extract in tea, crude fiber and starch in chocolate, and sugars in milk chocolate. Twenty-two samples were prepared and sent out to those who had expressed a willingness to collaborate, ten on tea and coffee and twelve on cocoa products. These were accompanied by the following directions and a letter of transmittal:

CAFFETANNIC ACID.

(a) Krug's method.-Proceed as directed in Bul. 107, p. 155. (Note that the formula for lead caffetannate should be Pb(CHO), as in Bul. 107, Rev.) Save the filtrate for the determination of caffein. After weighing the lead caffetannate determine its lead content as follows: Digest with aqua regia, add sulphuric acid, heat to fumes, cool, dilute, add alcohol, settle, filter, ignite, and weigh as lead sulphate. Calculate as per cent of lead.

(b) Method of Trillich and Göckel.b-Boil 3 grams of coffee one-half hour with water, filter, and repeat this treatment on the residue three times. The united filtrates are made up to 1,000 cc. To 400 cc add 1 ce of basic lead acetate solution and allow to stand overnight. Filter, wash, decompose the precipitate with sulphuretted hydrogen, filter from lead sulphid, evaporate to dryness, and weigh.

CAFFEIN.

In the filtrate from the lead caffetannate precipitate the lead with hydrogen sulphid, filter, and remove the excess of hydrogen sulphid by boiling, concentrating the solution, if necessary, to about 100 to 150 cc. Add tenth-normal potassium iodid solution of iodin in excess, filter through a little glass wool and determine the excess of iodin with tenth-normal sodium thiosulphate.

1 cc tenth-normal iodin equals 0.00485 gram caffein, c

EXTRACT IN TEA.

(a) Follow the provisional method as described in Bul. 107, p. 149.

(b) Follow the method proposed by Doolittle and Woodruff (Bul. 105, p. 48).

CRUDE FIBER (SAMPLE A).

Proceed as directed in Bul. 107 under "VI. General Methods," 11, page 56, except that the fiber is filtered and weighed on a paper. The sample should be pulverized by grinding with ether as described in the succeeding paragraph.

CRUDE STARCH (COPPER-REDUCING MATTERS BY DIRECT ACID HYDROLYSIS), SAMPLE A.

Weigh 4 grams of the material if unsweetened, or 10 grams if sweetened, into a small wedgewood mortar, add 25 cc of ether and grind with a pestle. After the coarser material has settled decant off the ether together with the fine suspended matter on

a Scoville, Amer. J. Pharm., 1907, 79 [12]: 549–551.

bZts. Nahr. Genussm., 1898, 101.

c Gomberg: J. Amer. Chem. Soc., 1896, 18: 331.