Comparing the fifteen "normal" substances (data for two not yet published), it may be stated as a general law that if, for any substance, the temperature ratios are high, the ratios for the saturated vapour will also be high, and those for the liquid will be low, and vice versa. Hexamethylene forms no exception to this rule.

The most interesting comparison of the ratios of the actual to the theoretical densities of saturated vapour is at the critical point. Hexamethylene has a lower ratio than any of the other seven hydrocarbons, or, indeed, than any other substance in the "normal" group except carbon tetrachloride, and, if Amagat's results be included, carbon dioxide. The difference from the theoretical value 3.77 (Trans., 1897, 71, 452) is, however, less than 2 per cent.

In conclusion, we desire to tender our thanks to the Government Grant Committee of the Royal Society for assistance in carrying out this research.

UNIVERSITY COLLEGE, BRISTOL.

LXXXIV.-The Composition and Tensions of Dissociation of the Ammoniacal Chlorides of Cadmium.

By WILLIAM ROBERT LANG, D.Sc., and ALBERT RIGAUT. THE phenomena of dissociation were first investigated by H. St. Claire Deville and by Debray (Compt. rend., 1867, 64, 603; 1868, 66, 134). Following on this, Isambert (Ann. de l'École Normale, 1868) carried out an exhaustive research on the tensions of dissociation of certain ammoniacal compounds of the metals, and, by applying the principles set forth in Debray's papers, was enabled to arrive at the true composition of these compounds. He shows that a definite chemical compound always has a constant tension of dissociation. Quite recently, Jarry (Thèse de la Faculté des Sciences de Paris, 1899) has applied the laws of dissociation in investigating ammoniacal compounds obtained with a solution of ammonia, and has established the fact that, for the ammoniacal halogen compounds of silver, formation and decomposition stop when the liquid saturating the compound is under a pressure equal to the tension of dissociation of the substance in vacuum. We have studied, from this point of view, the compounds formed between ammonia and cadmium chloride, using ammonia in the gaseous condition, in the liquid form, and in solution.

By the action of gaseous ammonia on dry cadmium chloride, Croft

(Phil. Mag., [iii], 21, 355) obtained a dry powder, much heat being developed at the same time, and the mass increasing in bulk. To this compound he gave the formula CdCl2,6NH. He also employed a solution of ammonia, and prepared a granular, crystalline substance corresponding with the formula CdCl2,2NH,. André (Compt. rend, 1887, 114, 908) also, employing a solution at 0°, formed a compound to which, as the result of one experiment, he gave the formula CdCl2,5NH. We have followed the method of preparation indicated by him, and find that this compound is identical with that obtained by Croft.

Our procedure was as follows: three to four grams of cadmium chloride, rendered anhydrous by heating almost to the point of fusion, were placed in a tube with the neck drawn out, to allow of its being afterwards sealed at the blowpipe, and the whole immersed in solid carbonic anhydride. Dry gaseous ammonia was passed slowly into the tube, when the chloride increased in bulk largely, and, in about 2 hours, a layer of liquid ammonia had formed on the surface of the solid compound; the tube was sealed and left for 20 hours. It was found preferable to use the salt in small lumps, as, when in powder, the liquid failed to penetrate through the mass, even on standing for a month. The tube was then cooled to -70° and opened, the temperature cautiously raised to -30° by means of methylic chloride, and the liquid ammonia expelled. In some of our experiments we allowed the ammonia to evaporate at 0°. After this the tube was sealed at the blowpipe and weighed. Analysis gave:

After reopening and heating at 100° until no more ammonia was evolved, analysis gave, as the result of several determinations:

The formula of this compound is therefore CdCl2, 2NH; it is inodorous, and, like the corresponding salt of zinc, only begins to decompose at about 210°. At 360°, decomposition is still incomplete, and when raised to close on 400° it fuses, giving only a trace of oxychloride and of chloride of ammonium.

In order to determine the tensions of dissociation of this hexammoniacal cadmium chloride, we used an apparatus similar to that of Jarry, and which is shown in Fig. 1.

A glass vessel, A a, containing the substances under consideration,

and in which the compound was prepared, is fixed by means of sealing wax to one of the extremities of a length of narrow lead tubing. The use of rubber should be avoided when working with gaseous ammonia, as the latter is readily dissolved by the rubber, and is given off again when the pressure is diminished. This tube is attached at b to the apparatus, which is mounted on a solid stand. The glass tube b c d e allows of communication being made with a mercury pump by means

of the stopcock R; the small U-tube c d e communicates by means of the vertical tube dƒ with B, which contains mercury, and which can be lowered or raised at will. This allows of the tube c d e being filled with mercury, and so shutting off the vessel 4, when a vacuum is produced, from the stopcock R. This is a very necessary precaution, as the gaseous ammonia dissolves in the grease with which this stopcock is lubricated. When a vacuum is maintained in the tube m, the

difference of levels at m and m' measures directly the pressure in A. As the tube m m' is only some 30 cm. high, higher pressure can be read on the tube M, which is open to the air, and stands about 2 metres high.

The readings were made in the following manner: The temperature of the water-bath in which the vessel A was immersed was very gradually raised, and the lamp removed; the thermometer fell slowly, and the mercury in the manometer rose, and, when it appeared to have attained a maximum, the readings were taken, at the same instant the temperature in the water-bath being noted.

It will be seen that above 62° the substance CdCl2,6NH ̧ can no longer exist, whilst we have at 100° a very stable compound, CdCl2,2NHg. It must therefore be at temperatures between 62° and 100° that the intermediate products are obtained.

We prepared this ammoniacal compound in the wet way also. By moistening hydrated cadmium chloride with a 20 per cent. solution of ammonia until thoroughly saturated, the compound is formed, increase in bulk taking place and much heat being developed. The crystals so obtained were dried over lime, and found to correspond with the formula CdCl2,2NH. If a solution of cadmium chloride in aqueous ammonia is evaporated over lime, or if the crystals formed when this solution is heated to 50° are collected, the same compound is in every case obtained.

8

In order to obtain a compound containing more than 2 molecules of NH by means of aqueous ammonia, it is necessary to pass a current of gaseous ammonia into a solution of cadmium chloride in a 20 per cent. solution of ammonia, keeping the temperature at 0°. An abundant precipitate is thus got of small, transparent crystals, which, when dried and pressed between filter paper, become opaque, losing ammonia. André (loc. cit.) also prepared those crystals, and, as mentioned elsewhere, gave them the formula CdCl2,5NH. From an analysis of them, we are led to the conclusion that the formula is not CdCl2,5NH, but that it is as indicated by

Croft (loc. cit.), namely, CdCl2,6NH. This confirms Jarry's conclusions regarding the ammoniacal compounds of silver (loc. cit.). We have, besides, determined the tension of dissociation of this compound. At 22°, we found it to have a tension of 147 mm. instead of 135 mm. The excess obtained from this product is due to a small quantity of vapour of water, difficult to completely remove by pressing with bibulous paper.

From these experiments, it appears to us that it is possible to obtain the same ammoniacal compounds, either by the use of gaseous ammonia or from a sufficiently concentrated solution.

A short note, containing a portion of the results here set forth, has been communicated to the Académie des Sciences, Paris, by Prof. Troost, of The Sorbonne, to whom we desire to tender our best thanks for his kind assistance and advice.

CHEMISTRY DEpartment,

UNIVERSITY OF GLASGOW,

August, 1899.

LXXXV.-The Aluminium-Mercury Couple. Part I. Action of Sulphur Chloride on some Hydrocarbons in presence of the Couple.

By JULIUS BEREND COHEN and FREDERICK WILLIAM SKIRROW,

The Yorkshire College.

IT has been shown by Hirst and Cohen (Trans., 1895, 67, 826) that the action of the aluminium-mercury couple closely resembles that of aluminium chloride, investigated by Friedel and Crafts, in bringing about combinations between certain halogen compounds and the aromatic hydrocarbons.

We have continued the investigation into the action of the couple in other directions, and find that, like aluminium chloride, it is capable of very general application as a condensing agent and also as a halogen carrier. The advantage of using the couple consists in the ease with which it is prepared and in the fact that, the action being apparently a contact action, a very minute quantity of the couple suffices for the preparation of a considerable quantity of material.

If a few drops of sulphur chloride (SCI) are added to a little benzene and a small fragment of the couple dropped in, a violent reaction ensues. The liquid boils and the colour rapidly darkens whilst large volumes of hydrogen chloride gas are evolved.

As the same vigorous action occurs with the majority of aromatic hydrocarbons investigated, it is necessary to moderate it by using