BY ALFRED SENIER, Lecturer on Chemistry in St. John's Training College, Battersea; late Demonstrator in the Laboratories of the Pharmaceutical Society.

This paper was written several years ago, since which time no important additional experiments have been made. The author sees no immediate prospect of the work receiving at his hands the further attention which it deserves, and meanwhile the results here described are of sufficient interest to justify their publication.

*

The action of solution of ammonia upon mixtures of silver chloride and iodide was investigated by Wallace and Lamont, who proposed a method for their separation based upon the results obtained. The respective solubilities of silver chloride, bromide and iodide have been determined under various circumstances by different observers; but, so far as I have been able to ascertain, the action of solution of ammonia upon mixtures of silver chloride and bromide has not been examined. To obtain, if possible, an approximate method by which chlorides and bromides might be separated, or at all events to give precision to our knowledge of this subject, was the object in view when the experiments

here detailed were undertaken.

The solution of ammonia used throughout had a specific gravity of 959, corresponding to 10 per cent. of NH, the "liquor ammonia" in fact of the British Pharmacopoeia. The silver salts were used in the moist, freshly precipitated state, quantities of dry sodium chloride and potassium bromide being employed to give the desired weight of silver salts. The alkaline salts were dissolve either together or in some cases separately, in about 50 cubic centimetres of water, and were converted into silver salts by a slight excess of silver nitrate. The precipitate was allowed to subside and was washed with water twice by decantation. The ammonia solution was then added and the mixture thoroughly agitated and set aside to subside if anything remained undis

solved.

In the first place the solubility of moist, freshly precipitated silver chloride and silver bromide in ammonia solution by themselves was determined. 4076 gram of sodium chloride was converted into silver chloride (1 gram) and the ammonia solution added with brisk agitation until all was dissolved. Seventeen cubic centimetres of ammonia solution

Chemical Gazette, 1859, 137. THIRD SERIES, No. 680.

Action of Different Proportions of Ammonia Solution on equal Mixtures of Silver Chloride and Bromide.-An equal mixture of 1 gram each of silver chloride and bromide was prepared by dissolving equivalent quantities of sodium chloride and potassium bromide in water and converting the mixture into silver salts. These were washed and while still fresh were treated with 20 cubic centimetres of ammonia solution, or rather more than would be required to dissolve the silver chloride present if its solubility remained 1 in 17 as when treated with ammonia solution by itself. The result was an insoluble portion, which from its colour was presumedly silver bromide, and which when washed, dried and weighed in the usual manner gave 1.25 gram, or again 1.26 gram (see experiments (1) and (la) in Table I.). Evidently the solubility of the chloride, assuming for the moment the soluble portion to be chloride, was modified by the presence of bromide, otherwise there should have remained insoluble in each of the two experiments only 1 gram at the most-the amount of silver bromide present. In the next four experiments, successively, larger proportions of ammonia solution were employed. They show that in order to dissolve of bromide, 50 cubic centimetres of ammonia solution gram of chloride, when mixed with the same weight are required. In the presence of an equal weight of silver bromide the solubility of silver chloride, for the present amount, is then 1 in 50; but it must be observed that in experiment No. 1 the solubility is 1 in 27, and in No. 2 it is 1 in 33, while the last portion in experiment No. 5 dissolves 1 in about 400. Thus, while, as will be shown presently, too much importance must not be attached to these exact numbers, it is sufficiently evident that the solubility of the mixture decreases by degrees, of which the average solubility only of the silver chloride is 1 in 50. In order to determine whether the insoluble

1

portions in these experiments contained really all the silver bromide, except, of course, in those cases in which the insoluble matter was less than 1 gram, it was not thought sufficiently promising to undertake any of the ordinary methods of quantitative separation, and, of course, in experiments Nos. 1-3, one would be certain to find some chloride qualitatively. I, therefore, tested the insoluble residues from experiments No. 4, No. 4a and No. 5, qualitatively by fusion with sodium carbonate and distillation with potassium dichromate and sulphuric acid. In the case of No. 4 and No. 4a much bromide with a trace only of chloride was found, while in the residue from experiment No. 5 no chloride could be detected. It is, therefore, at least approximately true that the chloride dissolves first, then the bromide. The next two experiments (6 and 6a) were conducted with silver salts precipitated separately and afterwards mixed together in order to prevent any possible production of double salts when, during the precipitation with silver nitrate, as would happen towards the middle of the operation, the two salts were being precipitated together. They gave insoluble portions weighing 98 and 100 gram, thus showing, when compared with No. 3, that no appreciable difference occurs whether the silver salts are precipitated together or separately. Thus it is gratuitous to suppose either that a double salt or other compound is formed during the precipitation, or that the ammonia is prevented from acting upon the chloride by a physical coating of its particles by the less soluble bromide. All the experiments, so far, point conclusively to the fact that the solubility of silver chloride in ammonia solution is modified by the presence of silver bromide, so that when the latter is present in equal amount the chloride dissolves 1 in 50, and if sufficient ammonia is present the bromide dissolves 1 in 250, its solubility when separate. The next two experiments (7 and 7a) seem decidedly to confirm this view. A solution of chloride in ammonia 1 in 50, which is capable of dissolving nearly 2 grams more chloride, dissolves no bromide whatever; and again, a saturated solution of bromide in ammonia dissolves chloride 1 in 50, afterwards slowly precipitating a substance which

Insoluble. Gram.

1.25

1.26

1.11

1.01

.86

⚫87

⚫82

•98

1.00

*98

8

10

11

12
13

14

15

•12

Number Silver of experi- chloride ment. Gram.

Soluble (difference). Grain.

75

.74

⚫89

⚫99

1.14

1.13

1.18

1.02

1.00

1.02

in the experiment proved to be practically all bromide, and, moreover, a great part of the bromide previously in solution. These experiments show also that if, as is quite possible, new compounds are formed in these reactions, they are not produced by precipitation of the silver salts, but within narrow limits the ammonia solution may itself be the medium of such combination.

Action of Solution of Ammonia upon Unequal Mixtures of Silver Chloride and Bromide.-The following table exhibits the results of a series of experiments, first, silver chloride, and, secondly, when silver bromide is in excess:

TABLE II.

Silver bromide. Gram.

75

The hypothetical numbers given in the table were calculated from formulæ depending upon the following hypothesis, a development of that already stated. When the proportion of silver bromide present is equal or in excess of the chloride, and when at least sufficient ammonia solution is present to dissolve the chloride 1 in 50, the chloride dissolves 1 in 50, any excess of ammonia solution dissolving the bromide also, 1 in 250, till saturated. Again, when silver chloride is in excess, and the quantity of ammonia solution is sufficient, the excess of chloride dissolves 1 in 17, the remainder 1 in 50, and the bromide 1 in 250, to the extent of the ammonia solution employed. It is clear from the table that the latter part of the hypothesis is inadequate, and an attempt was made to correct it with the view to

a quantitative analytical method for separation of chlorides and bromides. But after a long series of experiments it was found to be practically useless. For example, the solubility of the silver salts in ammonia solution is altered by varying the degree of dilution-the solutions from the above experiments were rendered turbid by the addition of water-and under the circumstances of the case the proportion of water could not be kept uniform. Again, doubtless variable physical states of the preeipitate which could not be avoided contributed to the results which were afterwards obtained, and which led to the abandonment of the hoped for quantitative analytical method. The exact limits of the solubility of mixtures of silver chloride and bromide in their bearing on qualitative testing I have investigated and incorporated in a separate paper.

To sum up, these experiments, taken as a whole, show:-that the solubility of silver chloride is not the same when mixed with silver bromide; that the solubility of moist freshly precipitated silver chloride in ammonia solution (10 per cent. NH3) is 1 gram in 17 c.c., and of silver bromide is 1 gram in about 250 c.c.; that the solubility of the chloride in presence of bromide is much less, so that when the proportion of bromide is one half or more, it is, on the whole, 1 in 50; that silver bromide is insoluble in a solution of silver chloride in ammonia, 1 in 50; that silver chloride displaces silver bromide from its solution in ammonia, but that the unavoidable errors of experiment preclude the use of these facts in quantitative analytical separation of the two acid radicles. Finally, while the probability of the formation of double salts or other compounds as an explanation of these phenomena has not been disproved, still the prospect of getting at their composition, if such bodies exist, of making sure that they are not products of the processes employed, seems too remote and uncertain to encourage further investigation at present.

THE VALUE OF SILVER NITRATE AND AMMONIA AS A TEST FOR BROMIDES IN PRESENCE OF CHLORIDES.

BY ALFRED SENIER,

Lecturer on Chemistry in St. John's Training College, Battersea late Demonstrator in the Laboratories of the Pharmaceutical Society.

It is well known that in cases where chlorides and bromides occur together the solvent action of ammomia upon their silver salts is of little value when applied in the usual manner of qualitative testing. The fact, however, that their silver salts differ so much in solubility in solution of ammonia, as seen in the preceding paper, induced me to find out by experiment exactly what value the method had when applied to definite proportions.

The results in the following table indicate that when mixtures of silver chloride and silver bromide are treated with definite proportions of ammonia solution, the presence of the bromide may be detected when its proportion is not less than about two per cent. of the silver salts or two and a half per cent. of the potassium salts. In these experiments enough sodium chloride and potassium bromide to give the required amount of silver salts was dissolved in a few cubic centimetres of water, the measured volume of ammonia solution (10 per cent.

0

⚫025

⚫02 015

⚫01

Ammonia solution (10 p. c. NH3) c.c.

10

10

10

10

10

*005❘ 10

Result.

No precipitate.

| Abundant precipi

tate.

These experiments led to the following method of testing. Weigh enough of the salt under examination to give approximately half a gram of silver salt. If it is a potassium salt, weigh 0:25 gram, or 0-2 gram if a sodium salt. Dissolve in about ten cubic centimetres of water and mix with ten cubic centimetres of ammonia solution (10 per cent. NH3). To the mixture add a few drops of solution of silver nitrate, and agitate. A permanent precipitate indicates presence of bromides equal to at least two per cent. of the silver salts. Working in this way upon mixtures of known proportions I have obtained perfectly reliable results.

Precipitate.
Precipitate.
Slight precipitate.
No precipitate.

ABRUS PRECATORIUS AND ITS THERAPEUTIC USE IN OPHTHALMIC DISEASES.*

Abrust precatorius, Willd, or Indian Liquorice (Jamaica Wild Licorice; Liane de Réglisse, etc.) is a small woody twiner, with a long, woody, tortuous, branched root,

inch or more in diameter. It has slender, branched stems and a brown bark. Its leaves are alternate, shortly stalked, spreading, 2 to inches long, abruptly pinnate, leaflets in about eight to fifteen closely-placed pairs, to

inch long, oblong, very blunt at both ends. Flowers are pale rose-tinted, small, in small clusters arranged on large tuberosities along the outer side of a stiff, curved rachis. The fruit is a pod, about 1 inch long, broadly oblong, shortly beaked, somewhat compressed, two-valved, with imperfect septa between the seeds. Seeds four to six, globular ovoid, about 4 inch long (of the size of a small pea); testa hard, bright and shining, brilliant scarlet, with a black patch at one end round the hilum ; cotyledons plane-convex; no endosperm.

The plant is very common in all parts of India, where

*From New Remedies, June, 1883. A portion of the description is after Bentley and Trimen's Medicinal Plants,' No. 77, where there is a coloured figure of the plant.

For further information on Abrus, the reader is referred to Pharmacographia' (2nd ed.), p. 188; Bengal Dispensatory,' p. 297; Drury, Useful Plants of India,' p. 3; Waring in Madras Quart. Med. Jour., 1860, p. 61; Moodeen Sheriff, Supplement to the Pharm. of India,' Madras, 1869, p. 17; Pharmacop. of India,' p. 74, 446; Dalzell and Gibson, Bombay Flora,' p. 76; Bolton, Medicinal Plants of Mauritius,' p. 43; Roxburgh, Flora

Ind.,' iii., 257.

Abrus is given by Prosper Alpinus (1592) as the name of the plant in Egypt, where the seeds were used for necklaces. Theis derives it from aßpos, delicate.-Bentley and Trimen.

Jamaica Wild Licorice; Liane de Réglisse (Fr.); Liane à Reglisse; Faginolo Corallino, or Semi di Corallo (Ital.); Bejuco peronilla, B. peonilla (Peurto Rico); Orozuz abro de cuentas de rosario, or Abro de cuentas (Spain).