TABLE I.-Determination of pentosans and methyl-pentosans—Continued.

In interpreting these results, it must be borne in mind that gum tragacanth contains a very large amount of pentosans, larger, in fact, than almost any of the food or feed materials that the chemist will be called upon to examine; hence it was necessary to employ a small original weight of the gum, and this would result in small errors of weighing being very much magnified when the results were expressed on a percentage basis. Judging from Table I the method employed does not give as good results as was hoped. It will at once be seen from the work of both chemists that there is a tendency to get low results on methyl pentosans when the amount of gum tragacanth is increased. Expressing this in another way, there is a tendency for the amount of methyl pentosans to decrease as the weight of the furfural-phloroglucid methyl-furfural-phloroglucid precipitate increases. This is probably due to the fact that the alcohol extraction is not so complete when a large precipitate is being handled. The results obtained by the same analyst on a definite weight of gum tragacanth do not vary from one another, in the large majority of cases, more than one would expect. The same can not be said of the results obtained by different analysts, which vary from each other more than is desirable even in such a conventional method. However, the averages obtained by the two analysts for methyl pentosans are close enough to warrant the referee for next year in sending out samples for further comparative work.

Even as the method stands it is worthy of adoption by the association as a roughly approximate method to be used by those who desire to make as complete an analysis as possible of vegetable materials. As the analysts become more familiar with the manipulation of the method, which is somewhat difficult and tedious, it is possible that their results will become more uniform. Better results will also be obtained upon the substances usually examined by feeding stuff methods than upon gum tragacanth, in which the amount of furfuralphloroglucid is so large that the extraction of the methyl-furfural-phloroglucid

is somewhat incomplete if a moderate quantity of the original gum is used. I would therefore recommend that the method be further studied next year, especial attention being given to obtaining a large amount of comparative work. Since calculating the percentages of methyl pentosans (as rhamnosan) from the formula given by Ellett and Tollens, an article by Mayer and Tollensa has appeared in which is given the formula for calculating the methyl-pentosans to fukose and also a table for calculating methyl-furfural-phloroglucid to fukose, fukosan, rhamnose, rhamnosan, and methyl pentosan (average of fukosan and rhamnosan).

I would, therefore, further recommend that this table be used in calculating the results of analysis, the average figure for methyl pentosans being taken when it is not known whether the mother product is fukosan or rhamnosan.

In accordance with instructions from the association, a study was also made of the time required to dry the crude fat obtained in the ether extract determination on several different feeding stuffs. The dryings were made at the temperature of boiling water and the following feeding stuffs were employed: Wheat germ meal, ground corn, cotton-seed meal, and flaxseed. To be certain that all fat was extracted from the samples, the time of extraction with ether was increased from sixteen to about twenty-four hours. The following results were obtained by the two chemists engaged in the work:

TABLE II.-Determination of time required to dry the crude fat obtained in ether extracts from different feeding stuffs.

It is at once evident from these results that no definite period for drying the crude fat can be specified in the method for determining this substance in cattle food materials. The time necessary to obtain a minimum weight evidently varies with the individual performing the work, with the amount of substance taken, with the character of the fat extracted, with the amount of fat in the substance, and possibly with other factors. A second point brought out is one that would be expected, i. e., that there is a marked tendency for the fat to reach a minimum weight and then slowly to gain by oxidation of the fat. This is, of course, particularly well marked in the case of linseed, which contains an oil that has a high oxygen-absorbing power.

A third point made evident by this investigation is, that in every case at least a one-hour period of drying was necessary to reach a minimum weight, while in some cases this period was extended to two and one-half hours. In the large majority of cases a one and one-half hour period of drying the crude fat was sufficient to obtain either a minimum or results so close to the minimum that the difference was negligible. On the whole, then, while a definite time of drying the fat can not be specified, it may be of some service to the inexperienced to state that the time is usually between one and one-half hours.

I would recommend, therefore, that the last sentence under Crude Fat, or Ether Extract, Direct Method, page 39, Bulletin 107 (Methods of Analysis), which now reads, "Dry the extract at the temperature of boiling water until it ceases to lose weight," be changed to read, “Dry the extract at the temperature of boiling water for one-half hour, remove from the oven to a desiccator, let it cool, and weigh; continue this every alternate one-half hour, drying and weighing until a minimum weight of the fat is obtained. For most substances a period of one to one and one-half hours is required to obtain a minimum weight."

REPORT ON SPECIAL ANALYTICAL METHODS OF SUGAR ANALYSIS. By C. A. BROWNE, Referee.

The work of the referee on special methods of sugar analysis for the present year has covered the investigation of a number of carbohydrate products, but the present report will be limited to a discussion of only three of these: First, the analysis or assay of commercial dextrins; second, a comparison of methods for the estimation of glucose in honey, and third, a study of methods for detecting artificial invert sugar in honey.

ASSAY OF COMMERCIAL DEXTRINS.

a

During the past year the Bureau of Chemistry has had occasion to analyze a number of dextrins for the work of the Bureau of Engraving and Printing. As a basis for methods an extensive article by Friedrich Lippmann upon the "Examination and analysis of commercial dextrins' was used. An abstracted translation of this article has been prepared by the referee, and copies so far as available can be furnished to members of the association.

The methods described in this abstract for moisture, ash, insoluble organie matter, and reducing sugars, were followed, and the difference between these and 100 per cent taken as dextrin, as advocated by Lippmann. It should be noted, however, that this difference includes other cold-water soluble organic materials besides dextrin, such as acids, proteid bodies, and various decomposition products of a carbohydrate nature formed during the process of dextrin

Zts. Spiritus-Ind., 25: 304, 307, 316, 317.

ization. These bodies should not, of course, be included with the dextrin, yet from the fact that there is no satisfactory method for the quantitative isolation of the dextrin from the accompanying impurities, the dextrin is taken as difference. For the specific purpose for which the dextrins are used at the Bureau of Engraving and Printing it was necessary to make a somewhat closer differentiation of the cold-water soluble matter (even if it were only an arbitrary one) which would throw more light upon the character of the products. Therefore the specific rotation of the product was first determined; the specific rotation equivalent to the reducing sugars present as dextrose was then subtracted from this and the remainder taken as the specific rotation due to the pure dextrin present. This remainder multiplied by 100 and divided by 186, the specific rotation of pure dextrin," will give the percentage of dextrin present. The difference between the sum of the constituents by this method of calculation and 100 is taken as the undetermined solubles. While the polariscopic determination of dextrin, by such a method as the one described, is of but little scientific value for the reasons given by Lippmann, it is believed that the results obtained by it have a decidedly practical value and give the chemist a more accurate idea of the character of his product than where dextrin is estimated simply by difference.

In the following table results obtained by the method described upon eleven samples of commercial dextrin are given. It will be noted that the results for dextrin obtained from specific rotation are more nearly in accord with the relationship indicated by viscosity than by the estimation of dextrin by difference.

The viscosity determination is of paramount value as a physical test, the results by this, corresponding most closely with the practical results to which the dextrin is put. Arranging the samples in order of viscosity, it will be seen that the value of a dextrin, as would be expected, is inversely proportional to the amount of undetermined solubles.

The following are offered to the association as provisional methods for the analysis of commercial dextrins:

METHODS OF ANALYSIS.

Physical constants.

Specific rotation.-Transfer 10 grams of the finely divided sample to a 100 cc flask, and after solution in about 50 cc of cold water add 5 ce of alumina cream and make up the contents to 100 cc, thoroughly shake, and filter. Polarize the

a L. Schulze, J. prakt. Chem., 1883, 28: 327.

filtrate in a 200 mm tube, using an S. and H. type of saccharimeter. Specific in which V=Ventzke reading.

rotation (a)D:

=

V 34.68

20

Viscosity.-Dissolve 100 grams of dextrin in 200 cc of cold water and determine the viscosity of the solution by any of the standard forms of viscosimeter. Comparative results should always be made by the same instrument and under similar conditions of temperature. Results are expressed upon the basis that water equals 100.

Chemical analysis.

Moisture.-Determine by drying from 2 to 5 grams of sample for four hours at a temperature of 105° C. Absolute constancy in weight can not be attained on account of the slow decomposition of the dextrin.

Ash.-Determine by the official method of the association, Bulletin 107, Bureau of Chemistry.

Soluble starch.-If a filtered hot-water solution of the dextrin gives a strong blue reaction with iodin, soluble starch is indicated. Weigh two lots of dextrin, 10 grams each, into a 100 cc flask, add 50 cc of cold water to each, and after all soluble matter is dissolved make up the contents of the one flask with cold water to 100 cc, shake, and filter. Evaporate 20 cc of the solution (2 grams) to dryness and dry for four hours at 105°, as under determination of moisture. Weight of residue less ash on incineration equals cold-water soluble organic matter. Heat the contents of the second flask to boiling, and then after cooling make up to 100 cc, shake, and filter. The weight of hot-water soluble organic matter in 20 cc of solution is determined as before. Hot-water soluble organic less cold-water soluble organic gives the soluble starch.

Unconverted starch.-If the residue insoluble in hot water shows under the microscope grains which are colored blue with iodin, unconverted starch is present. To determine the percentage, collect, the residue insoluble in hot water on a filter, wash until free from soluble matter, and determine the starch by the usual method.

Reducing sugars.-Determine in an aliquot of the cold-water soluble by the method of Allihn, the results being expressed as dextrose.

100

dextrin)

D

Dextrin.-Subtract the specific rotation of the dextrin due to reducing sugars 52.5X per cent reducing sugar as dextrin` from the original specific rotation of the sample. Multiply the remainder by 100 and divide by 186 ([a] for pure dextrin) to obtain the calculated percentage of dextrin in the sample. Undetermined solubles.-The per cent of cold-water soluble organic matter less calculated percentage of dextrin gives the percentage of undetermined solubles.

The other two studies on a comparison of the methods for the estimation of glucose in honey and the detection of invert sugar in honey are reported in full in Bulletin 110 of the Bureau of Chemistry, entitled, Chemical Analysis and Composition of American Honeys; they are, therefore, omitted from the Proceedings.

REPORT ON SUGAR AND MOLASSES METHODS.

By C. A. BROWNE, Referee, and J. E. HALLIGAN, Associate Referee. The samples sent out for cooperative work by the referees on sugar and molasses consisted of a sample of Louisiana low-grade sugar and one of molasses obtained from Dr. G. L. Spencer, Pijuan, Cuba, representing the final