has gone off, without the tendency to bumping which causes inevitable loss.

Boiling the amyl alcohol is continued until half has evaporated. Toward the end the flame may be slightly raised and it is well to hold a thermometer in the vapor; it should read 129°-130° C., showing that the water has been driven off.

The amount of amyl alcohol remaining for the Gooch correction is judged by adding water to a flask of the same size to the same level and measuring the water.

The amyl alcohol is decanted when cold through a small filter into a weighed platinum dish; the filter is first moistened with anhydrous amyl alcohol and the residue in the flask and the filter are carefully washed with the same until free from lithia. will require 20 cc.

This

The amyl alcohol is expelled from the platinum dish in a hot air-bath, under the hood, at a temperature not higher than 125° C. At this temperature it rapidly evaporates but does not boil or spatter.

After complete dryness, o.5 cc. of concentrated sulphuric acid is added to the dish, which is warmed and rotated until all the residue has been moistened and all the chloride decomposed, and the excess of sulphuric acid is now carefully driven off by heating over a small flame or by rotating over a moderate flame. Haste in doing this will cause loss.

The lithium sulphate in the platinum dish is ignited to burn off all carbon left by the amyl alcohol, which takes but a few moments. The sulphate fuses and is kept one minute in fusion, then the dish is allowed to partially cool and covered with a watchglass or filter-paper, to return any fragments of the fused sulphate, which sometimes splits off when it contracts.

The dish is then cooled in a desiccator and weighed ; after fusing again the weight is usually found to be constant. The deduction for sodium and potassium is made as directed by Gooch.' That rubidium and caesium chlorides are virtually insoluble in amyl alcohol is shown by the following:

0.1022 gram rubidium chloride dissolved in water, with a drop of concentrated hydrochloric acid, treated as in the analysis with 15 cc. amyl alcohol and boiled down to 9 cc., left 0.0009 gram 1 Deduct 0.0011 gram for each 10 cc, amyl alcohol exclusive of washings.

residual sulphates, which showed sodium strongly, also potassium, lithium and rubidium in the spectroscope.

0.1037 gram caesium chloride in the same way, with the amyl alcohol boiled down to 7 cc., left 0.001 gram residual sulphates, showing sodium, lithium, caesium, and potassium.

If the sulphate in the dish is pure, it dissolves readily in a small amount of hot water and gives a good lithium flame. Το prove that the sulphate residue obtained in the analysis was lithium sulphate, the following parallel was made: An equivalent weight of pure lithium carbonate (99 per cent.) was weighed into a second platinum dish, transformed into sulphate, ignited, weighed and the lithia in both dishes determined as phosphate as directed in Fresenius, using exactly the same amounts of sodium phosphate, sodium hydroxide and ammonia wash-water, and collecting the precipitates but twice.

It will be seen that there are a number of precautions to be observed in order to avoid loss of lithia, either by leaving it behind, driving it off, or mechanically carrying it off. Good results will, however, be obtained, if care is taken to use a low red heat in the fusion; to thoroughly leach and triturate the fused mass; to avoid high heat in driving off the ammonium chloride; to avoid bumping when boiling the water out of the amyl alcohol; to filter the amyl alcohol cold; to evaporate the amyl alcohol in the dish at a temperature below its boiling-point; to drive off the sulphuric acid without any foaming or boiling; and to cover the dish while the lithium sulphate is cooling.

The analysis can be completed in two days.

LABORATORY OF SCHIEFFELIN & Co.

November 22, 1901.

[CONTRIBUTION FROM THE BUREAU OF CHEMISTRY, U. S. DEPARTMENT OF AGRICULTURE, No. 43. SENT BY H. W. WILEY.]

A NOTE ON THE USE OF THE BECHI OR SILVER NITRATE TEST ON OLIVE OILS.

THE

BY L. M. TOLMAN.

Received January 20, 1902.

HE author, in working with a large number of salad oils, found that nearly all of them gave a brown coloration with the Bechi test, as ordinarily applied, unless the sample was first purified as described below. The modification of the reagent used was that proposed by Pearman and Moor,' and later by Wesson.*

3

Two grams of silver nitrate are dissolved in 200 cc. of 95 per cent. alcohol, to which 40 cc. of ether and two drops of nitric acid are added. Ten cc. of oil, 10 cc. of amyl alcohol, and 5 cc. of the reagent are mixed in a test-tube. Half the mixture is poured into another test-tube and kept for comparison. The other half is heated for ten minutes in a boiling water-bath and compared with the unheated portion. The brown coloration and reduced silver show the presence of cottonseed oil. Treated in this way, some oils which were especially rancid gave a strong test with the Bechi reagent, but from the physical and chemical constants and the Halphen reaction cottonseed oil could not have been present.

It has been advised to heat the oil for one hour at 100°, but this, while it reduced the reaction of the reagent with other oils, also weakened the reaction with cottonseed oil and was not at all satisfactory.

Wesson advises treatment with 2 per cent. nitric acid, but while this purified the oil in some cases it could not be depended on always, since quite a number of the oils treated in this way gave strong reactions although they contained no cottonseed oil.

The use of a dilute alkali works very well but is inconvenient, for it emulsifies and separates very slowly. A method which is much easier, more rapid, and one which gives very satisfactory results is as follows:

To about 25 cc. of the oil add 25 cc. of 95 per cent. alcohol, Allen's "Commercial Organic Analysis," Vol. II, Part I, second edition, p. 143.

2 This Journal, 17, 724.

3 Amyl alcohol is used to dissolve the fats or oils and thus secure a much better mixture with the reagent.

heat gently, and shake vigorously; allow to stand until the liquids separate, decant as much of the alcohol solution as possible, and then wash the residue with 2 per cent. nitric acid and finally with water. Cottonseed oils treated in this manner reacted with undiminished strength, while the olive oils which before treatment gave deep brown colorations, showed after treatment no coloration or reduction of the silver solution at all. The free fatty acids and other products of rancidity, which are evidently the cause of the brown coloration with the Bechi reagent which these oils give, are dissolved by the alcohol and removed while the reducing principle of the cottonseed oil is not affected. This method can be readily applied to lards or other fats which it is desirable to test by the Bechi reagent. The writer was able, by this method, to get very satisfactory and reliable results.

A STUDY OF THE BECHI TEST FOR COTTONSEED OIL.

W

BY AUGUSTUS H. GILL AND CHARLES H. DENNISON.
Received February 3, 1902.

WHILE in the case of sesame oil the substances to which the Baudouin test is due may be said to be fairly well known,' such is not the case with the test in question. Some❜ consider that it is due to a sulphur compound in the oil and others to an aldehydic body. While the evidence given by this paper is not conclusive, it was thought sufficiently interesting to publish.

The tests made by these various experimenters were repeated and confirmed; it was found that an oil heated to 260° gave neither the Bechi-Milliau test nor the Halphen test. This might indicate that the active principle of both tests was the same but for the fact that Raikow,' by oxidizing the oil with potassium permanganate and sulphuric acid, obtained an oil which gave the Bechi test but not the Halphen.

TESTS FOR ALDEHYDIC BODIES.

400 cc. of oil were shaken with an equal quantity of strong sodium bisulphite solution for twelve hours; the oil upon separation gave the Bechi test although a longer time was required for

Villavecchia and Fabris: Abstr. J. Soc. Chem. Ind., 16, 1045.

2 Dupont: Bull. Soc. Chim., 13, 696; Charabot and March: Ibid., 21, 552.

3 Holde: J. Soc. Chem. Ind., 11, 637; Wilson: Chem News, 59, 99.

A Chem. Ztg., 24, 562, 583.

its development and it appeared in less intensity. This was repeated with similar results; in this latter case crystals separated, which on being distilled with sodium carbonate gave no aldehyde reaction with the fuchsine-aldehyde reagent.

A portion of the oil was treated with dry ammonia gas for seven hours; there was no indication of any crystalline compound of aldehyde ammonia having been formed, and the oil gave the BechiMilliau test with even greater intensity.

TEST FOR SULPHUR COMPOUNDS.

In two cases only the black compound formed by the action of silver nitrate upon the fatty acid gave off hydrogen sulphide; this was recognized by its odor and blackening lead acetate paper. If, however, sulphur compounds were present in the fatty acid sufficient to blacken silver nitrate, it would seem as if other metallic salts should show characteristic colors. Salts of cadmium, lead, copper, and mercury were tried with negative results except that with the mercury salts reduction to metallic mercury took place.

In the repetition of Dupont's work both the aqueous distillate from the unheated oil and the oil itself gave the Bechi test. This was not the case with the oil which had been heated to 260°. In conclusion, in view of the fact that bodies of an aldehydic character could not be detected in the oil by either the fuchsine aldehyde reagent, by sodium bisulphite, or by ammonia, it would seem that it was not likely that they were present. This would seem to be confirmed by Raikow's oxidation with permanganate and sulphuric acid. The evidence would seem to favor the presence of sulphur compounds.

NEW BOOKS.

STUDIES FROM THE CHEMICAL LABORATORY Of Sheffield SCIENTIFIC SCHOOL. EDITED by Horace L. WELLS. Vol. I, xi + 444 pp. Vol. II, ix + 379 PP. New York: Charles Scribner's Sons: 1901. Price, $7.50 per set.

In consequence of the recent bicentennial celebration of Yale University, there has been presented to the world "a series of volumes * * * prepared ** * and issued * * * as a partial indication of the character of the studies in which the University teachers are engaged." Two volumes of this series,