acid decomposes it with evolution of sulphurous anhydride and iodine vapours. Nitric acid precipitates iodine; mercuric nitrate produces a yellowish-white, silver nitrate a dirty-white, and lead acetate a white precipitate. Hydrochloric acid liberates the free acid with formation of sodic chloride. A solution of mercuric chloride produces a white precipitate, which changes to pink and red, and is soluble in excess of the reagent. Baryta-water gives a white precipitate, which dissolves almost entirely in hydrochloric acid. Cupric sulphate produces, after a time, a greenish-white; auric chloride, a very brownish-red precipitate; metallic gold and silver even are acted upon by a solution of sodic iodosulphate.

The author determined the composition of the salt quantitatively by converting it into sulphate, precipitating with barium chloride, and weighing as barium sulphate. The water of crystallisation was likewise determined directly; not so, however, the iodine and soda. Neither were the two other salts, the potassic and ammonic iodosulphate, which he describes, analysed. Both salts are said to be isomorphous with the iodic salt, but to be more soluble in water. Exposed to air and light, they undergo rapid decomposition.

On the Detection of Iodine in Solutions containing Tannin. M. Tessier. (Journ. Pharm. Chim., 4th series, xiv., 46.) It is well known that iodine dissolves to a certain extent, but in variable quantities, either in the cold by trituration, or on application of heat, in liquids containing an astringent principle, and that it then becomes disguised in such a manner as to be no longer recognisable by the ordinary tests.

In such cases the iodo-tannic solution may be heated to ebullition in a flask, and nitric, containing some nitrous, acid then added drop by drop. In a few minutes the violet vapours of iodine will be perceived, and may be made to communicate to starch paper the characteristic blue colour.

Another more prompt and very reliable reagent is the perchloride of iron. If to a small quantity of the iodotannic solution a few drops of solution of perchloride of iron are added, vapours of iodine are at once disengaged, and may be known by their odour. The reaction is rendered more manifest by covering the vessel with a small funnel smeared over with starch paste. Layers of iodide of starch are formed immediately, the intensity of colour varying according to the amount of iodine vapour evolved.

By this method very small quantities of iodide may be detected in bromide of potassium. For this purpose two watch-glasses may be employed, the one containing a small quantity of a solution of

the salt mixed with a few drops of perchloride of iron, the other moistened with starch paste and inverted over it.

These experiments are very delicate and simple in operation, since for their execution they require only reagents which are to be met with in every shop.

Action of Iodine on Tannin. V. Griessmayer. (Journ. Chem. Soc.) By the addition of an excess of tannic acid to an iodine solution, the latter becomes gradually decolorised. Tannic acid also causes the decolorisation of iodised starch solution. In both cases the iodine is reduced to hydriodic acid.

The products of the action of iodine on tannic acid are ellagic, gallic, and hydriodic acid, and sugar. Doubtless the tannin is first split up into gallic acid and sugar, the gallic acid afterward yielding ellagic acid. The reaction is probably

A New Method of Obtaining Potassium. Professor A. E. Dolbear. (Amer. Chem., and Chem. News, vol. 26, No. 660, p. 33.) Perhaps it is not possible to discover a process for obtaining the metals potassium and sodium better than the one usually employed with the carbonates of the metals and charcoal, heated in iron retorts; but it is a very desirable thing to obtain them cheaply, on account of their use in the manufacture of aluminum.

Some white stick caustic potash of commerce was dissolved in water, and then treated with sulphuretted hydrogen in the way commonly described for making potassium sulphide, K, S. The solution was evaporated until it was solid when cool; then the yellowish mass was mixed with more than its bulk of iron filings and chips, and the whole put into an alembic for distillation. The heat of a furnace was applied, till the alembic was of a bright red heat; and the products of distillation were received in common coal oil. The produce was rather small, as some of the potassium vapour decomposed the heated vessel; nevertheless, the potassium showed itself when the oil was poured off; and the residuum burned upon water by its characteristic ignition and flame. The reaction is simple, and may be thus represented, K, S+ Fe - Fe S+ K2.

=

On the Examination of Bromide of Potassium. M. Falières, of Libourne. (Journ. Pharm. Chim., 4th series, xiv., 247.) The employment of bromide of potassium in affections of the nervous system has become so general, and the consumption of this salt so considerable, during the last few years, that competition has led to its extensive sophistication.

The experiments of chemists have shown that the bromide of potassium of commerce may contain various salts, such as iodide, chloride, hydrate or carbonate of potassium, sulphate of potassium, nitrate of sodium, and bromate of potassium. Iodide of potassium, in consequence of its high price, is not introduced fraudulently into bromide of potassium, but it proceeds from the bromine of commerce which often contains traces of iodine. M. Falières has observed in bromide of potassium the presence of chloride often in considerable quantities. M. Adrian also has found, from the examination of a number of samples, that this contamination may amount in some cases to as much as 30 per cent.

M. Ernest Baudrimont published in 1868 (Journ. Pharm. Chim., 4th series, vii., 411) a method for the estimation of the amount of chloride thus present, which was founded upon the fact understood by all chemists, that a given weight of chloride of potassium decomposes a much larger proportion of a titrated solution of nitrate of silver than the same weight of bromide of potassium. gram of this latter salt requires only 1427 of nitrate of silver for its complete precipitation as bromide of silver, whilst 2:279 gram of the nitrate are necessary to convert 1 gram of chloride of potassium into chloride of silver.

Thus one

M. Falières has adopted a process based upon the same principle. Starting with the fact that one gram of bromide of potassium requires 1427 of nitrate of silver, and one gram of chloride of potassium requires 2:279, for complete precipitation, it is obvious. that with a mixture of 9 of bromide and 1 of chloride the amount of nitrate of silver required would be

Consequently if the bromide contained 10 per cent. of chloride, it would be necessary after having treated the mixture with 1427 of nitrate of silver to add further 0852 of this reagent. If then 852 of nitrate of silver is dissolved in 100 cubic centimetres of distilled water, and if after having treated one gram of the bromide by 1·427 of nitrate of silver it is found necessary in order to complete precipitation, to add 5, 10, 20, 30 c. c. of the titrated solution, the bromide analysed contains 5, 10, 20, 30 per cent. of chloride of potassium.

Before proceeding with this operation, it is essential to ascertain by means of appropriate tests that the bromide under examination contains no salts, such as iodide, carbonate, sulphate, nitrate, which might interfere with the accuracy of the process.

M. Falières recommends the following mode of proceeding: One

gram of bromide of potassium is dissolved in 30 or 40 c.c. of water in a stoppered bottle, and the solution of nitrate of silver containing 1.427 gram of that salt is then added. When the precipitate has subsided, the "homometric" solution is added, drop by drop, from a burette. If the bromide is pure, the addition of a drop of this solution causes no turbidity. If, on the contrary, a precipitate is formed, the volume of the test solution employed indicates the quantity of chloride of potassium.

The latter part of M. Falières' paper is devoted to the preparation of pure bromide of potassium.

If the salt contain iodide nothing is necessary but to add an excess of bromine water, and evaporate the solution to dryness.

When the bromine to be employed contains chlorine, it may be treated with a small quantity of solution of bromide of potassium, from which the chlorine displaces the bromine, and gives rise to chloride, which is carried off in the supernatant solution.

Pure caustic potash being somewhat expensive, M. Falières proposes to employ bicarbonate for the preparation of the bromide of potassium.

For this purpose 100 grams of bicarbonate of potassium are dissolved in 500 grams of distilled water; 80 grams of pure bromine are then added, and when the evolution of carbonic acid has ceased, the solution is poured into an ammoniacal solution composed of 90 parts of water and 30 parts of ammonia (875). The whole is then evaporated to dryness, and the residue gently heated until white fumes of carbonate of ammonia cease to be evolved, and finally the salt is fused so as to convert the bromate into bromide. The product is then redissolved in distilled water, some bromide water added, and the solution is evaporated and crystallised. In this process the bromide of ammonium reacts upon the unchanged carbonate of potassium and converts it into carbonate of ammonia, which volatilises. The bromide thus obtained appears to be entirely free from carbonate.

New Process for the Detection of Bromide in Iodide of Potassium. E. Van Melcke beke. (Journ. Pharm. d'Anvers, xxviii., 49, 1872.) This process is based upon the property which a liquid saturated with one salt possesses of dissolving another salt, provided no chemical reaction resulting in the formation of a precipitate comes between them.

The process is conducted as follows: A saturated solution of bromide of potassium is first prepared by dissolving a quantity of that salt in warm distilled water, allowing that solution to cool and

the excess to crystallise out. 10 c. c. of this solution contained in a test-tube are mixed with 10 drops of distilled water, and about one gram of the iodide to be tested, coarsely powdered, is then projected in small quantities at a time into the liquid. The liquid is shaken after each addition in order to assist the process of solution. If the iodide is free from bromide it disappears almost immediately, whilst if contaminated by that salt the latter remains undissolved.

It is necessary to agitate the liquid sufficiently after each addition, and not to add too large a quantity of the iodide at once, or a part of the bromide might be displaced locally and the result rendered doubtful. By previously adding 10 drops of water to every 100 c.c. of the solution of the bromide, the risk of this accident is done away with.

Detection and Estimation of Bromide in Iodide of Potassium M. Lepage. (Journ. Pharm. Chem., 4th series, xv., 103.) In 1844, M. Personne published in the Journal de Pharmacie a process for the determination of the presence of bromide of potassium in the iodide. At that time the intermixture of the two salts could be attributed only to imperfections in the process of manufacture, since the price of the bromide was higher than that of the iodide. At the present day the contrary is the case, the cost of the iodide having been for some time treble that of the bromide. Under these circumstances it becomes a matter of importance to be able to detect the latter salt.

The process recommended by the author is based upon the property possessed by mercuric chloride of precipitating the iodine from a mixture of iodide and bromide of potassium, to the exclusion of the bromine, the mercuric bromide being soluble. Having ascertained, by the ordinary method, that the iodide to be tested is free from chloride, carbonate, and iodate, one gram of it is taken, and dis solved in 30 grams of water.

In the next place one gram of bichloride of mercury in 20 cubic centimetres of water is prepared, and this solution is then run from a burette into that of the iodide of potassium until it ceases to cause a precipitate. It is advisable to exceed slightly the quantity of bichloride necessary for the precipitation of the whole of the iodide. If the iodide is pure, 16 cubic centimetres at least of the solution will be required; if not, a volume greater than four cubic centimetres remains, being the greater the larger the proportion of bromide present, and from this the proportion of this latter may be calculated.

In order to prove the presence of bromine in the liquid from