are contained in cerite, in addition to those of the cerium-group, though of course not in large quantities, and that oxide of didymium if prepared in the usual way as the least basic of earths present in cerite, is not a homogeneous body, but a mixture of earths.

The investigation of the cerite earths, which were left untouched in the present memoir, is in progress, and its results will form the object of another communication.

XXXIV. Some Compounds of Antimony and Bismuth containing two Halogens.

By R. W. ATKINSON, B.Sc. (Lond.), F.I.C.

AMONGST So-called "double-salts," one of the best characterised is that produced from antimonious chloride and potassic chloride. Discovered by Jacquelaine about 40 years since, it has been followed by other salts built up in a similar manner, though beyond a more or less detailed description of their physical properties no attempt seems to have been made to subject them to such an examination as it is now universally the custom to submit organic compounds to for the purpose of ascertaining their constitution. Such an examination ought to be synthetic as well as analytic, but evidently little information could be gained by combining two such salts as potassic chloride and antimonious chloride, for the reason that the chlorine in the former salt could not be distinguished from that in the latter after combination had taken place; in order, therefore, to be able to distinguish the halogen in combination with the potassium, from that combined with the antimony, I used chlorine in the one case, and bromine in the other. By this means, on the theory of a molecular combination between the respective salts, two distinct compounds ought to be produced, viz., (1) SbCl,.3KBr, and (2) SbBr3.3KCl. On mixing together antimonious chloride and potassic bromide in the proportions in No. (1), adding just sufficient water to ensure solution, and slowly evaporating over oil of vitriol, yellow rhombic pyramids were obtained, possessing the composition SbCl,Br,K, + 1Aq.

Similarly, by mixing together antimonious bromide and potassic chloride in the proportions required for (2), under like conditions, yellow pyramidal crystals were obtained, which, from the crystallographic examination kindly undertaken by Mr. R. H. Solly, of the Mineral Museum, Cambridge, are identical with the former. The

formula found also was identical, viz., SbCl,Br.K, +14Aq, from which we may conclude that the same body is produced in both cases. Their chemical behaviour also is the same. Over calcic chloride they effloresce at ordinary temperatures, and finally lose the whole of their water of crystallisation; at a slightly elevated temperature they lose their water very readily. In ordinary air they gradually deliquesce to a clear yellow liquid. The crystals also dissolve in a small amount of water, giving a solution when saturated, having a specific gravity of 1.9, and containing 120.5 grams in 100 c.c. A larger quantity of waterdecomposes the salt, forming a mixture of SbOCI and SbOBr, and on the addition of a still larger amount of water a mixture of Sb,O¿Cl2 and Sb.,Br2 is formed. In the former case the mixture consists of the compounds in equal molecular proportions, but in the latter the proportions are not quite equal, more chlorine being removed than bromine.

Strong hydric chloride dissolves the crystals at first, and almost immediately deposits a crop of crystals of potassic chloride, the solution doubtless containing the hydrogen salt, SbCl, Br,H3, which, however, I have not succeeded in isolating.

The action of heat upon the crystals is of considerable interest. The dried salt, which is of a pale lemon-yellow colour, on heating to 100° C., becomes darker in colour, but on increasing the heat to between 200-300°, it loses its yellow colour, and leaves a white mass, whilst antimonious fumes escape. The loss in weight is a little more than 50 per cent., and the residue contains chlorine and bromine in equal atomic proportions, so that the antimony in being volatilised has carried away with it one-half of the chlorine and one-half of the bromine. Taken in connection with the amount of water in the crystals, we must give them at least the formula Sb2K,Cl¿Br ̧ + 3Aq.

Nicklès (Ann. Pharm., 40, 191) has described a chlorobromide of bismuth and ammonium formed by adding bromine to powdered bismuth, suspended in a solution of ammonic chloride. He attributed to it the formula Bi(Cl, Br,)N2H. + 5H2O, which, however, does not agree so well with his analytical data as a formula similar to the one above given, viz., BiCl,Br ̧(NH)3 + H2O. I have not prepared this salt, but I have little doubt but that it corresponds exactly with the antimonio-potassic chloro-bromide described above. I hope to be able to compare them on a future occasion.

The following are the analytical numbers obtained, but on account of the impossibility of freeing the crystals completely from the motherliquid, they are not so closely concordant as might be desired. The numbers for bromine and chlorine refer to the dried salt.

From the facts above communicated we must conclude, I think, that in the molecule (SbCl, Br,K3)2, there is no evidence of any grouping into two parts, as the theory of molecular combination supposes. If it were otherwise we should expect to find that on heating (no fusion taking place) the residue would consist wholly of potassic chloride or potassic bromide, as the case might be, not of a mixture of the two in molecular proportions. This remark also answers sufficiently the objection that by the act of solution a redistribution of elements takes place, and that therefore we must obtain the same compound from whichever pairs we start. Granting that it were so, the crystals resulting would have one or other of the two constitutions

SbBr3.3KCl or SbCl,.3KBr,

and on heating either all the bromine or all the chlorine would be driven off, not one-half of each.

In making the above experiments I obtained combinations in other proportions, which have not been examined crystallographically, owing to the difficulty of getting sufficiently bright surfaces.

1. Sb2Cl¿BrзK3 + 2Aq.-This compound was deposited from a solution of the crystals previously described, containing a small excess of antimonious chloride. The crystals apparently belong to the rhombic system, and are much lighter in colour than the former. The following gives their composition:

Theory for

2. SbCl,BrK + H2O.-From a mixture of two molecules of antimonious chloride with three of potassic bromide in saturated solution, successive crops of crystals were obtained, the earlier ones being con-taminated with crystals of potassic bromide, and only towards the last were they found to be free from that salt. The earlier crystallisations gave numbers leading to the formula 5SbCl,.6KBr+ 8H2O, but these were probably not homogeneous, though no difference could be detected between the different crystals. The later crystals were pale-yellow octahedra, having the following composition:

In their composition they resemble the salt formed from antimonious iodide and potassic iodide, which has the formula SbIK + H2O.

BISMUTH COMPOUND.

Attempts were made to obtain a bismuth salt similar to the antimony one first described, viz., from BiCl, + 3KBr, and from BiBr, + 3KC1, but hitherto without success. The following new salt was obtained during the experiments, but the crystals have not been examined crystallographically, further than to ascertain that they belong to the rhombic system.

3. A solution of bismuthous bromide in a saturated solution of potassic chloride, after standing in a desiccator over oil of vitriol, deposited thin pale-yellow lozenges. An analysis gave the following

The formula with 1Aq is more probably correct, the crystals doubtless retaining a little water.

A second crop of crystals obtained from the same solution had a darker yellow colour and a stumpy prismatic appearance, but, except that they were slightly contaminated with some potassic bromide, they gave similar numbers :

:

In conclusion I have much pleasure in recording my obligations to Dr. Williamson, For. Sec. R.S., for his kindness in permitting me to carry out these experiments in the Laboratory at University College, and also for much valuable advice during the course of the investigation.

293

XXXV.-Crystallographic Examination of the Crystals of Antimoniopotassic Chlorobromide.

By R. H. SOLLY, Esq., Cambridge.

THE crystals of the two salts in the solutions marked A and Ba solution from SbBr, + 3KCI

A
B

....

have been carefully examined. More than 20 crystals were measured; some were measured twice after an interval of a fortnight. The bright faces gave the same angles, but the dull and rough ones varied. Only a very few of the crystals gave distinct reflections of the bright signal, the best angles range from 62.25-62.34°; 62-29° was found three times on the A crystals, and five times on B crystals; 62.29° is the angle used to calculate the element, as it was obtained thrice from very good faces. The face (001) was found well developed on two crystals in B solution. These crystals are therefore crystallographically similar, both crystallising in the tetragonal system, and exhibiting similar development. There is no apparent cleavage.

System tetragonal. Common form (111), but sometimes in combination with (001), Fig. 1.