The results of the experiments made with these colors proved that the temperature at which the decolorized solution began to show a return of the color depended on the amount of sodium bisulphite reagent used. It was also found that, in many cases at least, if this reagent was added in considerable excess of the amount required to bleach the solution, no color would return even on heating to the boiling point.

The following description of the results of some of the experiments made on two of the colors will illustrate in a general way the behavior of the abovenamed coal-tar colors when treated with this reagent:

Acid magenta (A).—A solution of this color was made of such a depth that printed letters one-eighth of an inch long could just be distinguished in strong daylight through a layer of 5 inches.

Experiment I: To 5 cc of this solution was added 1 cc of sodium bisulphite reagent. The solution was decolorized within a few seconds. On heating to the boiling point only a faint color returned, which quickly disappeared on cooling.

Experiment II: To 5 cc of solution 0.5 cc of the reagent was added. The color disappeared in two or three minutes. On heating to the boiling point a deep color returned, which disappeared at 50° C.

Acridin red (Grübler).—The solution used was made of the same depth as in the case of acid magenta.

Experiment I: To 5 cc of solution 1 cc of reagent was added. The color disappeared at once. Only a faint color returned on heating to the boiling point. Experiment II: To 10 cc of the solution 0.1 cc reagent was added. The color disappeared in one minute, but returned in its original intensity on heating to the boiling point and did not entirely disappear until the solution was cooled to 27° C.

Ethyl green (A).—This color is very easily bleached, and if a great excess of the reagent is used no color returns even on heating to the boiling point and continuing the heat for some time. When only a slight excess of the reagent is present a strong green color returns on heating to 60° C., which gradually disappears as the solution cools.

The temperature at which the color finally disappears from heated solutions of any of the colors given in the list depends on the rate of cooling. If cooled slowly the color will disappear at a much higher temperature than if cooled rapidly.

With the majority of the colors given in this list, a solution of sulphur dioxid in water will produce results similar to those obtained with the sodium bisulphite reagent. However, in working with a large number of coal-tar colors the latter reagent was found to give distinctive reactions in a larger number of cases than the sulphur dioxid solution. Preference was therefore given to the bisulphite solution as a reagent for general use.

THE DETERMINATION OF MOISTURE IN SIRUPS AND MOLASSES.

By W. D. HORNE.

The correct determination of water in sirups and molasses is very important both from the technical and the legal point of view-first, because this alone enables one to determine the true purity of the solution, and, secondly, because a sirup is defined by law as a solution containing not more than 25 per cent of water.

Various methods have been proposed for determining the water, but difficulties have been met with in all of them. Undiluted sirups dry slowly and frequently incompletely, necessitating prolonged heating or high temperatures, both of which tend to decompose organic matter. To dilute and mix with sand,

pumice, etc., solves this difficulty only partly. Vacuum drying is not always convenient. It necessitates elaborate apparatus and is often rather slow. The importance of finding a simple, quick, and accurate method for making this determination is apparent, and a very satisfactory procedure, gradually evolved by the writer in the course of many years of experimenting, is offered for the consideration of the association. The method reads as follows:

One gram of molasses is weighed into a flat-bottomed dish 3 inches wide and about 0.5 inch deep and containing a glass rod. About 0.8 cc of water is well mixed with the molasses and then about 30 grams of dry quartz sand, previously extracted with hydrochloric acid, is exactly weighed and added to the diluted molasses. The dish is placed on an open boiling water bath and stirred carefully and frequently during half an hour. This causes the free evaporation of about nine-tenths of the water at a moderate temperature. The dish is next placed in a water-jacketed air bath, where it is heated at the temperature of boiling water for two hours. After cooling it is weighed and reheated for one-hour intervals until the weight is constant.

This very easy method, requiring only the simplest devices and manipulations, gives, in from two to four hours, constant results that agree closely with determinations effected in partial vacuo during longer heating periods and under the influence of a fine stream of air dried by sulphuric acid. The method has been employed on refinery molasses for several years with great satisfaction.

In this work it is found that the degree Brix of the undiluted molasses plus the moisture determined as above amounts to 103.3 with great regularity, despite rather wide variations in the percentage of ash in the sirups.

Mr. Tolman submitted by title a report entitled "A study of the changes taking place in whisky stored in wood," by C. A. Crampton and himself. This report covers an investigation of eight years' duration conducted in the laboratory of the Bureau of Internal Revenue and is published in full in the Journal of the American Chemical Society for January, 1908, page 98. The conclusions reached are as follows:

(1) There are important relationships among the acids, esters, color, and solids in a properly aged whisky, which will differentiate it from artificial mixtures and from young spirit.

(2) All the constituents are undergoing changes as the aging process proceeds, and it is evident that the matured whisky is the result of these combined changes.

(3) The amount of higher alcohols increases in the matured whisky only in proportion to the concentration.

(4) Acids and esters reach an equilibrum, which is maintained after about three or four years.

(5) The characteristic aroma of American whisky is derived almost entirely from the charred package in which it is aged.

(6) The rye whiskies show a higher content of solids, acids, esters, etc., than do the Bourbon whiskies, but this is explained by the fact that heated warehouses are almost universally used for the maturing of rye whiskies, and unheated warehouses for the maturing of Bourbon whiskies.

(7) The improvement in flavor of whiskies stored in charred packages after the fourth year is due largely to concentration.

(8) The oily appearance of a matured whisky is due to material extracted from the charred package, as this appearance is almost lacking in whiskies aged in uncharred wood.

(9) The "body" of a whisky, so called, is due largely to the solids extracted from the wood.

DETECTION OF THICKENERS IN ICE CREAM.

By G. E. PATRICK.

The thickeners commonly used for ice cream to-day are gelatin, certain vegetable gums or jellies, and various forms of starch. Of the true vegetable substances gum tragacanth is most used, and it is believed that at least one other substance of this class is employed, but it was impossible to tell whether it was agar or some less common member of the group.

A test for the detection of such thickeners has been formulated which it is believed will prove useful if the conditions are closely observed, the main features of the method offering little that is new from a chemical point of view. Picric acid is used for precipitating gelatin, and alcohol for the gums. Preparatory to the test the liquid is clarified by coagulating the proteids with acid and heat and filtering. The details of the procedure, elaborated with the help of H. S. Bailey and B. McClelland, of the Dairy Laboratory, are as follows:

To 50 cc of ice cream add 25 cc of water, boil for half a minute to dissolve any thickener that may be present, add 2 cc of a 10 per cent solution of acetic acid, heat again just to boiling, add two or three heaping teaspoonfuls of kieselguhr, shake well, and filter immediately through a plaited filter. To 3 cc of the clear filtrate add 12 cc of 95 per cent alcohol and mix; this precipitates the milk proteids not removed in the clarification, together with the gums, and some of the gelatin, if much be present; add 3 cc of acidified alcohol prepared by mixing 95 cc of 95 per cent alcohol and 5 cc of concentrated hydrochloric acid; this dissolves the milk proteids completely. If the liquid is clear after this treatment, no gums or vegetable jellies are present. But, on the other hand, turbidity or a precipitate does not necessarily indicate the presence of a thickener; for this may be caused by the large amount of gelatin present or by eggs (more than three or four per gallon), or because the ice cream has become sour. Fortunately the precipitate due to gelatin or to the substance derived from eggs can be readily dissolved by a moderate dilution of the alcohol with water, for instance, by 3 cc of water in the test just described. This amount of water does not sensibly dissolve any precipitate due to vegetable gums or jellies, but entirely dissolves that due to gelatin and eggs. If gum tragacanth is present, the precipitate will be cohesive and stringy, or will become so upon shaking; while if the undissolved matter is due to other vegetable thickeners (possibly agar), it is finely flocculent and devoid of cohesive property.

If the ice cream is sour, there is sometimes a precipitate of another character, which must be derived chiefly, if not entirely, from the cane sugar present, since it appears very faintly, if at all, in tests upon unsweetened milks that have been allowed to sour. But this substance does not always develop in sour ice creams or in sweetened milks that have become sour; its appearance seems to depend upon some special property or condition of the milk, probably upon the presence of certain kinds of bacteria. Its nature is now being investigated, and Mr. C. A. Browne suggests that it may prove to be dextran. Whatever it may be, the precipitate which it yields with alcohol is not sensibly dissolved either by the acidified alcohol or by the water added to dissolve the gelatin and egg substance; therefore it must remain mixed with the vegetable gums and jellies. It does not resemble gum tragacanth, as it is not stringy, but does closely resemble the other vegetable thickener. It is therefore a serious obstacle in the test. Fortunately, however, its formation appears to be prevented by formaldehyde (experiments on this point are still in progress), and it is believed that if fresh ice cream is treated with a liberal dose of formalin there will be no trouble from this annoying substance even if the test is carried out several weeks later.

The test for the detection of gelatin was made on a small portion of diluted cream after boiling but before acidifying to precipitate proteids. The provi

sional method of the association, originally published by Stokes (Analyst, 1837, p. 320), was used, which consists in clarifying the milk or cream with dilute mercuric nitrate solution and precipitating the gelatin in the filtrate with picric acid. It is in general a satisfactory test. But the interesting fact was observed by Mr. Bailey that in testing ice cream which had been sour for a week or more, or very sour milk, whether previously sugared or not, picric acid produces a yellow precipitate easily mistaken for that produced with gelatin. No way of distinguishing between the two has been found, but, as would be expected, formaldehyde prevents the formation of this "pseudo gelatin" and therefore samples thoroughly preserved with formalin present no difficulty from this source. Starch, often used in a mixture with gum tragacanth, and sometimes alone, was detected in the usual way with iodin, using a portion of the boiled diluted sample.

REPORT ON THE DETERMINATION OF WATER IN FOODS.

By F. C. WEBER, Associate Referee.

The referee regrets that a full and definite report on this subject can not be given this year owing to lack of cooperation and of sufficient time to devote to the work. Circular letters were sent to 15 chemists asking for their collaboration; only 5 replied, stating that they could not take up the work this year. The results obtained this year showed that drying in a vacuum of 0 to 5 mm of mercury over sulphuric acid compared favorably with drying in an oven with partial vacuum at 100° C. The results, however, in the vacuum oven were obtained in ten hours, while nine days were required to get the same results in the vacuum desiccator. The addition of phosphoric anhydrid as an auxiliary drying agent in the vacuum-desiccator method did not appear to shorten the time of drying. From the somewhat limited amount of work which has been done so far, the referee does not feel justified in condemning the method, though it is doubtful whether it could, even when modified, come into general use. The importance of further study on this subject is recognized, and it is therefore respectfully recommended that the work be continued next year.

THE CARBON DIOXID VALUE OF PURE COMPRESSED YEAST AND COMPRESSED YEAST AND STARCH COMPOUNDS.

By T. J. BRYAN.

The examination of numerous samples of compressed yeast in this laboratory showed that the majority of them contained added starch, usually either potato or corn starch.

The question arose, What is the value of the pure compressed yeast to the consumer as compared to the article mixed with starch? Some preliminary experiments were made, using samples of pure compressed yeast and yeast containing corn starch, such as were found on the market. In these preliminary tests the different yeasts were allowed to act upon a 10 per cent sugar solution and the volume of gas generated was measured. The results showed that the pure yeast had a greater carbon dioxid value than yeast mixed with starch. Owing to the size of the apparatus necessary for collecting and measuring the volume of the gas produced, it was thought better, in subsequent experiments, to determine the amount of carbon dioxid gas produced by the loss in weight. Furthermore, as the different yeasts tested were made by different manufacturers from different cultures, it was decided that the results secured were of little value in deciding the effect of the presence of starch in compressed yeast, and that samples of yeast from the same culture must be

secured, a portion of which should be kept in the pure state, and other portions mixed with potato and corn starch.

Seventy pounds of this pure yeast were mixed with 10 pounds of corn starch and 16 pounds of water (it was found necessary to add the water to make the mixing uniform and in order to produce a cake). To another portion of 70 pounds of the yeast were added 10 pounds of potato starch and 19 pounds of water. To the remaining portion of pure yeast no starch was added in the mixing. These three samples were then pressed separately and cut into 1-pound cakes which were each dipped in water before being wrapped, as is the custom of the manufacturer. These three samples were then allowed to act upon 100 cc of a 10 per cent sugar solution to which 75 cc of water had been added in an Erlenmeyer flask of about 190 cc capacity. This flask was provided with a calcium chlorid tube adjusted by means of a cork in order to prevent loss of weight by evaporation. Two grams of yeast were used in each test and duplicates were run. The flasks were put in a box, the opening of which was covered with a towel and placed against the chimney where a temperature varying between 20 and 30° C. was maintained. During the last three days the box was replaced by a section of a bookcase and the temperature maintained ran a little higher than on the preceding days, varying between 25° and 30° C. Owing to the varying temperatures on different days only the results for the same periods on the same day are comparable. The flasks were weighed when they had been filled with the reagents and once every hour after that for ten hours, and again at the end of each twenty-four hours. Acknowledgments are due to Messrs. Nehls and Gardner and Mrs. M. B. Shulda for assistance in making the tests.

Table I shows the average loss in weight for the periods specified. Table II shows the average per cent of loss in weight (or the percentage of the weight of carbon dioxid gas produced for the specified periods), considering that the average weight of carbon dioxid produced by the pure yeast samples for each period is 100 per cent. It will be noted that after two hours the average loss in weight of the flasks containing pure yeast is, with a single exception, greater at every weighing.

The average percentages of the yield of carbon dioxid gas for twenty-four hours and for six hours, as given in Table II, show conclusively that the effect of starch upon yeast is to reduce the carbon dioxid value and that this percentage reduction is greater than the percentage of starch used in preparing the samples. It has been claimed by some manufacturers that it is necessary to use starch in compressed yeast in order to preserve the same. The data in Table II show that on the fourteenth day the value of the pure yeast is greater than on the first day as compared with both of the starch yeast mixtures. The differences, however, are not sufficient, in the writer's opinion to justify the statement that the yeasts containing starch had deteriorated, though the data point in that direction. As fourteen days is a longer time than compressed yeast is kept before being put on the market and used the contention that starch is necessary to preserve yeast is seen to be absolutely false. Nineteen days after the yeast was prepared all of the samples were perfectly sweet.

It was thought desirable to test the action of these different yeasts in the making of bread. The bread was made with 650 grams of flour, 500 grams of water, and 10 grams of yeast or yeast starch mixture. On the second day pure yeast yielded bread having a volume of 2,225 cc; corn-starch yeast mixture, bread 2,100 cc in volume. On the fifth day pure yeast yielded bread of a volume of 2,000 cc; corn-starch yeast yielded bread 1,860 cc in volume. On the twelfth day pure yeast yielded bread 2,150 ce; corn-starch yeast mixture, bread 2,015 cc. On the thirteenth day pure yeast yielded bread 3,000 ce in volume and corn-starch yeast mixture, bread of 2,575 cc. On the fourteenth day the