in time till their next calving of 5 months; while the cows on farm O were all very near to their next calving, and consequently more late milking, but they had not been on reclaimed land.

It ought to be mentioned that low contents of volatile acids seems to stand in no relation whatever to the commercial value of the butter.

MR. HEHNER said that the facts brought forward by Mr. Allen were, without doubt of far-reaching importance. Public analysts would have to acknowledge that exceptional samples of butter did occur which would analyse like mixtures containing about 25 per cent. of margarine. If only one such exceptional sample occurred in every hundred, as there was some evidence from the Italian experiment quoted by Dr. Faber, it would yet be henceforth impossible to give a satisfactory opinion on any sample which contained less than 25 per cent. of foreign fat. Such an acknowledgment would result in immense injury to the public and the butter trade; It therefore behoved public analysts at once to set to work to find out the cause of the extraordinary discrepancy, less in the physiological direction, or regarding the food or other circumstances influencing the cows, but in a chemical direction. The question had to be solved: In what points does a natural abnormal butter differ from one adulterated with, say 25 per cent. of foreign fat? He was convinced that there were such differences, and he drew, in this direction, attention to the following facts, deducible from Mr. Allen's analysis. On calculating from the Reichert results the percentage of volatile acids as butyric acid (although, of course, the totality of volatile acids is not obtained by distillation), there in normal butters almost reached the soluble acids estimated in other ways. Thus, in the "normal" butter of Mr. Allen, butyric acid, by Reichert, 5'14 per cent. soluble fatty acids, 5·54 per cent., or 93 per cent. of the total soluble fatty acid obtained by distillation. In sample O, which would be passed as genuine upon the Reichert results, 4.34 per cent. were obtained by distillation, against 4 66 total soluble acids, or also 93 per cent. But in sample B, which was quite abnormal, only 87 per cent. of the soluble acids were volatile.

These figures seem plainly to indicate that, in this particular abnormal sample at least, the breaking down of the albuminous matter from which butter-fat in all probability results, was not so complete as to form butyric acid, but other acids slightly higher up in the fatty series. This conclusion, he thought, was confirmed by an observation of his (Mr. Hehner's) upon the abnormal sample, a portion of which had been given him by Dr. Faber, namely, that towards the end of the distillation, notable quantities of solid fatty acids congealed in the condenser, similarly as in the case of cocoa-nut oil.

It appeared to him to be of the highest importance now to direct our attention to the intermediate fatty acids, and he had no doubt that thus this difficulty might be overcome. The German Committee would thoroughly investigate the methods at present at our command, and it was not necessary to work in this direction; he would, however, suggest that, unless Mr. Allen was prepared to take personally upon himself the immense labour of carrying out an investigation such as he (Mr. Hehner) had indicated, it would be well that the Society should appoint a Committee upon whom the labour would devolve.

Mr. ALLEN said in reply that the question was quite a different one from that submitted to the Milk Committee. In the latter case they knew what they wanted, and their object was to devise a method which would give accurate and reliable results in the hands of a number of chemists. In the present case they did not know exactly

in what direction to look for success. It was a question of studying more close y the composition of butter as compared with other fats. He had already indicated cert in directions in which hopeful work might be done, and he was thoroughly in accord with Mr. Hehner in his suggestion regarding the desirability in future of differentiating the

fatty acids more thoroughly than had been done in the past. In fact, he had suggested that direction, as a desirable one in which to work, to Mr. Faber, only a few days ago. His proposal was, that a small committee should be appointed to consider whether they could not devise some other process of examining butter for foreign fats. He did not personally know how far the results were affected by the time of distillation, and by neglect of the precaution to allow the fatty acids to become thoroughly molten before commencing to distil, Professor Wollny had recorded a number of experiments on this point, and, it being an arbitrary process, he had thought it his duty to follow the manipulation prescribed by Wollny as closely as possible, especially as the requirements were so easy to comply with. In conclusion Mr. Allen said that he was personally extremely sorry that the experiments on the Danish butter had turned out as they had done. He would have much preferred that they should have established the constancy of the composition of butter-fat rather than its variation beyond the accepted range of composition. But public analysts must not rest on their oars, but at once set to work to solve in a satisfactory manner the new problem raised by the experiments in question, and he had no doubt of their ultimate success. They must remember that a dozen year ago, before Messrs. Angell and Hehner devised their well-known process, they had no reliable method of detecting foreign fats in butter at all. He thought they should all combine for the purpose of overcoming the present difficulty, and he therefore proposed "That a Committee, consisting of the President for the time being, Dr. Vieth, Mr. Hehner, Mr. Faber, and himself, with power to add to their number, be appointed to consider and report on the possibility of improving the methods of analysing butter for the detection of adulteration." He relied on Dr. Vieth and Mr. Faber, not only for their great experience in the analysis of butter and other dairy products, but because they were, doubtless, in a position to obtain samples of which they knew the exact history, as regards the number of cows from which they were derived, the method of feeding, and so forth. The motion was seconded by Mr. Stokes, and carried after a short discussion. (Conclusion of the Society's Proceedings.)

ON SOURCES OF ERROR IN DETERMINATION OF NITROGEN BY SODA-
LIME, AND MEANS FOR AVOIDING THEM.
By W. O. ATWATER.

(Continued from page 237.)

Loss by Dissociation and Oxidation of Ammonia.

It is frequently urged that in the determination of nitrogen with soda-lime, ammonia may be dissociated at the temperature at which the combustions are conducted, and that in aspirating air through the tube at the end of the operation some of the ammonia may be burned. The following experiments were made in the hope of getting light upon the conditions under which such loss may thus occur in determinations as ordinarily made.

For this purpose means were taken to test the effects of (1) unusually long exposure of the ammonia to heat, (2) very high heat, and (3) exclusion of air. Ammonia was furnished by ammonium sulphate in which the ammonia had been determined by distillation by the method described by Gooch, two determinations giving respectively 21-11 and 21.14, and averaging 21.13 per cent. of nitrogen. About 0.3 gramme mixed with from 0.5 to 10 gramme of oxalic acid previously proven to be free from nitrogen were employed in each of the determinations detailed beyond. The soda-lime, made from 1 part of caustic soda and 2 of lime, and the method of charging the tube, were as previously described. The method, which is that usually followed in this laboratory, involves filling the tube quite full of soda-lime, so as to avoid large open spaces which

become filled with gases and allow long sojourn of ammonia in the tube. Where the long tubes were used, the length of the anterior layer of coarse fragments of soda-lime was increased with the length of the tube, the length of mixture of soda-lime and nitrogenous substance and the charging otherwise remaining the same as in the tubes of ordinary length. The time of combustion, about thirty to thirty-five minutes, was naturally a little less than would usually be the case with animal tissues. The flames were turned off at the close of the combustion before the aspiration with air to wash out the ammonia. In No. 10 (see Table II.) the air was expelled before the combustion, and the ammonia washed out at the end by carbonic acid. In Nos. 11 to 19 the expulsion of air and cleaning out of ammonia were effected by a current of hydrogen, which was delivered by a generator so arranged as to permit easy regulation of the flow. The hydrogen was passed over heated copper foil, to remove the oxygen, before entering the combustion tube.

TABLE II.

Determinations of Nitrogen in Ammonium Sulphate under Different Conditions.

MONTHLY RECORD OF GENERAL RESEARCHES INTO ANALYTICAL

CHEMISTRY.

PURIFICATION OF SULPHURIC ACID FOR KJELDAHL'S PROCESS. G. LUNGE. Zeitschr. f. angew Chemie, No. 23.—Meldola and Moritz recommend to heat 10 c.c. of the acid for about two hours with 05 grm. of potassic nitrite. They take it, therefore, for granted that the nitrogen contained in commercial acid is always in the form of ammonia. The usual impurity is, however, nitrosylsulphuric acid, which the manufacturer tries to destroy by addition of ammonium sulphate. The decomposition is, however, never complete, so purified acid generally contains both nitrosylsulphuric acid and ammonia. The author criticises their method, which may give good results if the amount of ammonia is known and the theoretical amount of potassic nitrite is added, but which fails if any notable excess of nitrite is used. Then it is a great error to suppose that nitrous acid is completely expelled on boiling. Meldola and Moritz have overlooked the fact that the very stable nitrosylsulphuric acid is formed, which is but sparingly volatile. This fact is so well known that further experiments seemed almost superfluous, but to contradict the work of such an analyst as Meldola the author had several experiments carried out. A preliminary trial showed that 2255 grms. of sodium nitrite dissolved in 100 c.c. pure sulphuric acid gave a very strong nitric reaction after the acid had been boiling for two hours. A quantitative experiment was then conducted; 392 grms. of sodium nitrite was dissolved in 250 c.c. of pure sulphuric acid of 1-84 sp. gr., and 50 c.c. of this mixture diluted and titrated with permanganate, which showed a total of 1764 nitrous acid. The remaining 200 c.c. were now boiled for four hours, which caused 100 c.c. to distil over. The residue was now titrated, and showed a total of 175 grms. nitrous acid. The greater portion of nitrous acid was therefore still present. A second experiment gave before boiling 1938, after boiling 1842 grms. of nitrous acid. These experiments conclusively show that the purification process proposed by Meldola and Moritz is absolutely useless, and only then of avail when the quantity of ammonia is accurately known, so as to regulate the quantity of the nitrite. But in this case the simplest way is really to deduct this quantity (if not too large) from any quantity found during an analysis.

L. DE K.

DETERMINATION OF NITRATES IN WATER. S. C. HOOKER. Berichte, Dec. 10, 1888.The author makes use of the property of the solution of carbazol in conc. H,SO, of turning deep green on the addition of small quantities of oxidising agents. The reaction is delicate enough to detect nitric acid in water containing one 2-millionth part of this substance. The following is the method used: A measured quantity of the water, 2 ccm. or less, is mixed with 4 ccm. of conc. H2SO,, and after cooling, a small quantity of carbazol in conc. H2SO, added. The colouration is compared with that obtained with varying quantities of KNO, solution treated in exactly the same manner. detailed publication is promised.

A more

A. L. G.

MONTHLY RECORD OF ANALYTICAL RESEARCHES INTO FOOD. RIGHT-HANDED POLARISING HONEY. VON LIPPMAN. Zeitschr. f. angew Chemie, No. 22.-The author confirms the idea of Benseman (see ANALYST, 1888), and finds that when in the neighbourhood of sugar-refineries bees swarm there in their thousands, and regale themselves on the sugar. Their honey is then very clear and thin, but almost without the valued aroma. Four samples of such honey, analysed by the author, contained respectively 4.88, 3.92, 16-38, and 9.93 per cent. of saccharose. Samples showing such percentages cannot, therefore, be legally called adulterated. L. DE K.

INFLUENCE OF DIFFERENT CATTLE FOODS ON THE COMPOSITION OF BUTTER. A. MAYER. Landw Vers, 35.-The author has made a series of experiments to ascertain the composition of butter, and has arrived at the following conclusions:1. There is a close relation between the specific gravity of butter-fat and the percentage of volatile acids. An increase in the one gives an increase in the other. 2. The melting point of the fat is not much influenced by the tributyrin, as it chiefly depends on the amount of triolein. 3. The amount of volatile acids in the butter of the milk of any particular cow is not so constant as was formerly believed, but is influenced by the quality of the food the cow has been fed on. 4. The amount of volatile acids is gradually increasing during the period of lactation. 5. When the food greatly consists of mangelwurzel the butter shows the highest percentage of volatile acid. Grass and clover will cause a larger amount than ensilage. 6. Hay and ensilage will cause the butter to have a high melting point, whilst green fodder will cause the lowest one. Mangelwurzel gives an intermediate melting point. 7. The solidifying point rises or falls with the melting point. 8. Cattle grazing in the fields will give the best milk and the richest butter. The author finds the quantity of deci-normal soda wanted for the distillate of 5 grms. of butter to vary from 20 to 30 c.c. L. DE K.

THE OCCURRENCE OF BORIC ACID IN WINE. GEORGE BAUMERT. Ber., Dec. 10th, 1888., p. 3290.-The author's attention was drawn to this subject by his finding boric acid in every sample of a number of Californian wines analysed by him three years ago.* Recently, Professor Rising, in San Francisco, stated that boric acid is a characteristic constituent of all Californian wines.† Soltsient found that boric acid is a very frequent constituent of wines, and also occurs in cultivated and wild vines (Ampelopsis quinquefolia) in Saxony, whilst Rippers not only found it in 1,000 samples of German and other wines, but also in various parts of different kinds of vines. In one case he separated it as fluoboride of potassium from 21. of a Riesling wine. The author has, for three years, tested every German, French, and Spanish wine which came into his hands, and found boric acid in them all; and also in must (fresh juice) from the Freyburg and Naumburg districts. He also systematically examined the growing vines in Thuringia in Saxony, and invariably found boric acid in the leaves, tendrils, wood, grapes, and grape-stems, examining 0.2 to 04 grms. of the ash in each case. The same results were obtained on examining a number of leaves and woods from the Freyburg district. A. L. G.

* Nobbe's Landwirthschaftliche Versuchsstationen, 33, 39-88.

↑ Report of the Sixth Annual State Viticultural Convention, San Francisco, 1888.

Pharm. Zeitung, 33, No. 40, p. 312; ibid., No. 90.

§ Weinban u. Weinhandl, No. 36, 1888, and ANALYST, Dec., 1888.