ature in a platinum dish, dissolved in water, and carefully mixed with the second portion weighed as above. The resulting mixture was acid in reaction to phenolphthalein, indicating that some of the bitartrate was possibly lost during the ignition. Several repetitions of the above indicated that there apparently was some loss on ignition and that the method indicated in the Pharmacopoeia required further study. Recourse was then had to normal sulphuric acid prepared by the usual methods, titrated by two independent workers using several indicators, with satisfactory results, the variations ranging from the lowest for methyl orange, 0.9890, to the highest for phenolphthalein, 1.0065. This solution was used as the basis for preparing decinormal sulphuric acid.

Mr. Parker's portion of the work was practically completed when the pharmacopoeial corrections appeared, and his results are therefore based upon the original methods, except that cochineal had been substituted for hematoxylin. The standard acid employed by him, however, was found to correspond closely with that employed by the other analysts.

The results obtained are given in the following table:

Comparison of methods for the assaying of aconite, belladonna, cinchona, coca leaves, and colchicum.

Comparison of methods for the assaying of aconite, belladonna, etc.-Continued.

Residue purified by re-solution in ether, treatment with water, etc., as in the assay of colchicum corm, U. S. P., not included in average, maximum, or minimum. Residue purified as above, substituting petroleum ether for ether, not included in average, maximum, or minimum.

DISCUSSION OF RESULTS.

ACONITE LEAVES.

The leaves were assayed by the same methods as the root. The variation by the United States Pharmacopœia method reached 51 per cent of the average value. The aliquot gravimetric method gave more uniform results, the variation amounting to 10 per cent; but titration in a few cases indicated that, as is frequently the case, the gravimetric results are high. The Squibb physiological test was obtained with a dilution of 1-80. It will be noted that the indications by this test and by assay do not correspond to the analogous results for aconite root.

ACONITE ROOT.

For aconite root, as for the leaves, the aliquot gravimetric results showed less variation than the United States Pharmacopoeia method, but the few titration results indicate that some of the former are considerably too high. The

Squibb physiological test gave a value of 1-500, or six times as great as the leaf, instead of two to three times, as might be inferred from the assays.

BELLADONNA LEAVES.

The results of Mr. Parker by the United States Pharmacopœia method for belladonna leaves are decidedly higher than those of the other analysts, which do not vary much among themselves. The discrepancy may indicate some deterioration in the drug in the three months intervening before the lower results were obtained. All the aliquot volumetric results, which were obtained about the same time, are fairly concordant. Experiments were made to determine whether, in the extraction of the drug by ethereal solvents, the presence of a small amount of water exercises an unfavorable influence. In the pharmacoprial method a few cubic centimeters are introduced in the form of diluted ammonia water. Mr. Parker varied the method by using alcoholic ammonia of corresponding strength (about 4 per cent), also by diluting stronger ammonia water with alcohol instead of water to the same strength. The results were decidedly lower, namely, 0.298 and 0.276 per cent.

Mr. Warren employed the total extraction method also, which is similar to the pharmacopoeial method, except in the use of alcohol for diluting the ammonia. He obtained higher gravimetric results, but titration of the residues indicated that the increase was due to impurities.

Mr. Rieger tried the total extraction method, using alcohol in one test and water in another, to dilute the ammonia water. His results are low in both cases, and no conclusions can be drawn except that the residues are impure.

BELLADONNA ROOT.

In the results for belladonna root, as in those for the leaves, the best agreement is obtained by the aliquot volumetric method, and the gravimetric results are evidently vitiated by impurities. Just as with belladonna leaves, the results of experiments on the influence of small amounts of water on the extraction of the root led to no conclusion. No positive advantage was shown to accrue from the substitution of alcohol.

CINCHONA (YELLOW AND RED).

The pharmacopoeial method for cinchona gave much less variation in the results on total alkaloid than the total extraction method, but, as was the experience last year, the results on ether-soluble alkaloid are not concordant.

The change in the pharmacopoeial method, doubling the amount of solvent used in extracting the drug, involves the shaking out of 200 cc of ether-chloro.form mixture with comparatively small volumes (15, 10, and 5 ce) of dilute acid, which is a tedious and difficult task. Evidently some nonalkaloidal matter passes into the acid solution in shaking out, and from that into the alkaloidal residue, which is discolored in all cases.

In determining ether-soluble alkaloid Mr. Warren filtered the ethereal solution through cotton in a small funnel.

The total extraction method fails to exhaust the marc even after obtaining 400 to 500 ce of percolate, though weak runnings give no test for alkaloids. On testing the mare, as directed last year, by boiling with normal hydrochloric acid and shaking out the acid extract, the amounts of impure alkaloid indicated in the column under "marc" were obtained.

COCA LEAVES.

The results by all methods for coca leaves are so variable that none can be said to be encouraging. The results of the aliquot gravimetric method agree best among themselves, but the titration equivalents show that they must be regarded as too high.

COLCHICUM CORM.

The results for colchicum corm by the pharmacopoeial method are materially higher than by Method II, but also vary relatively more among themselves.

COLCHICUM SEEDS.

Results for colchicum seeds, by both methods, exhibit great variations, and experiments on purifying the residues by a process similar to the latter part of the pharmacopoeial method for colchicum corm showed that the impurities present would render the results entirely unreliable unless they were removed.

In the absence of the referee on tannin, his report was presented by Mr. Tolman, as follows:

REPORT ON TANNIN.

By F. P. VEITCH, Referee.

The referee was unable to obtain the samples desired until very late, and for this reason very few of the twelve or fifteen analysts, who had promised to collaborate, were able to do so and the work was confined to a comprehensive study of the filtration of soluble solids through folded S and S 590 and S and S 588 paper, the first a high-grade paper and the latter a cheaper paper of ordinary grade. For a number of years the No. 590 paper has been unsatisfactory for use in soluble-solids determinations, as it is not uniform and is too weak to stand the weight of the solution after the paper is folded.

In a study which the referee had made, using a large number of different extracts, it was found that the No. 588 paper gave results agreeing closely with those with the No. 590 paper, except in the case of quebracho. Reed had previously obtained satisfactory results with No. 588, but only on hemlock and lentiscus. The referee's favorable results led him to submit the paper to collaborative tests, as both time and expense would be saved if it could be used.. The following circular letter was sent out:

DIRECTIONS FOR TANNIN WORK, 1907.

DEAR SIR: I am forwarding you to-day three samples of extracts in which the following determinations are to be made in triplicate: (1) Total solids.

(2) Soluble solids, using 15 cm 590 S and S folded filters. (3) Soluble solids, using 15 cm 588 S and S folded filters.

Make all dryings on the bottom shelf of a boiling-water oven, drying for 12 hours.

Dissolve the following quantities of extracts in 2,700 cc of water at 80° C: Sample No. 1, 18 grams; Sample No. 2, 42 grams; Sample No. 3, 15 grams.

Allow to cool overnight to 20° and dilute to 3,000 cc. In these solutions determine total and soluble solids as follows:

Total solids: Thoroughly mix the solution, immediately pipette 100 ce into a tared dish, evaporate, and dry as directed.

a J. Soc. Chem. Ind., 1902, 21: 691.

Soluble solids: Add 75 cc of solution (kept at from 20° to 25° C during filtration) to 2 grams of kaolin (free from soluble salts); stir, let stand 15 minutes, decant, and discard as much as possible of the supernatant liquid. Again add 75 cc of the tannin solution to the kaolin, stir, and pour immediately on a folded filter. Keep the filter full and the funnel and receiving vessel covered. Reject the first 150 cc of filtrate, evaporate and dry the next 100 cc (which must be as clear as practicable) as directed.

If you have not the kind and size of filter paper specified please communicate with the referee before proceeding further.

The results reported are given in the following table:

TABLE I.-Determinations of total and soluble solids in three extracts, comparing S and S 588 and 590 filter papers.

a UnderTime" the first figures show the time required for the first 150 cc; the second figures for the next 100 cc.

No final conclusion can be drawn from these few results, but it is quite evident that the S and S 588 paper gives slightly lower results than the 590 paper under the conditions of filtration, which, it must be remembered, were devised for use with the 590 paper. The results agree in general with those that have been previously obtained, and it seems probable that with some modification in the method of filtration, such as the filtration of about 200 cc before collecting the portion for analysis, the results would agree with those obtained with the 590 paper. Such a procedure would, of course, lengthen the time of filtration, and this would not be desirable.

The time of filtration with the 588 paper was, as a rule, slightly shorter than with the 590 paper. Mr. Smoot observed that the temperature at which filtration was conducted affected the speed of filtration as well as the percentage of soluble solids; the cooler the solution the more slowly it filtered. Mr. Smoot also made some determinations with an asbestos-kaolin mat on a Buckner fun