cent of sodium; but the study of the ethereal filtrate from the salt threw a great deal of doubt on this conclusion, since it contained a large amount of phenol. We think it more probable, therefore, that the sodic ethylate decomposed the phenoquinone into quinone and phenol, and that the green salt was then formed by the action of the sodic ethylate upon the quinone, the action being the same as that described in the last section. This inference is strengthened by the marked resemblance in appearance between the salts obtained in these two cases, and also by the study of the properties of the salt made from the phenoquinone, since it dissolved completely in water, forming a dirty green solution, and undoubtedly suffering partial decomposition, as ether extracted from this solution colorless crystals of hydroquinone recognized by their melting point of 169°. The addition of an acid to this solution gave no precipitate, and upon shaking out the acidified liquid with ether, nothing was extracted but hydroquinone. If the substance had been the desired salt of phenoquinone, phenol should have been obtained from this filtrate. Although these experiments tell against the formation of salts of phenoquinone, we do not consider them absolutely final, but the study of the action of alkalies on phenoquinone under other conditions will be continued in this Laboratory during the coming college year.

As we had not succeeded in making a salt of phenoquinone we next turned our attention to the action of sodic phenylate on quinone, as, if the phenoquinone is a hemiacetal, this should act as well as free phenol, whereas according to the other theories of the constitution of phenoquinone, it is hard to see how there should be any action in this case. Sodic phenylate was made by warming the proper amount of metallic sodium with a solution of phenol in benzol until the sodium had entirely disappeared. As the benzol cooled, the white crystalline sodic phenylate separated abundantly, and, after filtering, any free phenol was removed by washing with cold benzol. Upon adding the solid sodic phenylate to a solution of quinone in absolute ether a dark red crystalline substance looking like phenoquinone was formed. After evaporating off the ether slowly, the residue dissolved easily in water with a slight green color, and acids precipitated from this aqueous solution a small amount of a dark reddish solid. The ethereal solution showed a tendency to turn green round the edges during the evaporation which may perhaps have been due to the action of the moisture in the air. If benzol was used to dissolve the quinone instead of ether, a pink substance was formed, which changes to a dark green body when warmed. With ligroine as the solvent, a dark green

precipitate was obtained at first. Unfortunately, this work was undertaken at the very end of the term, so that a more careful study of these products must be postponed until next year, but these preliminary experiments show that sodic phenylate does combine with quinone, and therefore lend a certain amount of countenance to our suggestion that phenoquinone is a hemiacetal.

XIX.

CONTRIBUTIONS FROM THE CHEMICAL LABORATORY OF HARVARD COLLEGE.

ON THE CUPRIAMMONIUM DOUBLE SALTS.

SECOND PAPER.

BY THEODORE WILLIAM RICHARDS AND ANDREW HENDERSON WHITRIDGE.

Presented May 9, 1894.

THE continuation of the study of the cupriammonium double salts, begun in 1891,* has led to the preparation of the following new compounds :

(1.) Cu(NH3)2CICHO.

(2.) Cu(NH),BгC2H2O2. H2O. (3.) Cu(NH3)2 BгC3H5O2.

(4.) Cu(NH), BгC3H5O3. (5.) Cu(NH3)2CIC3H5O3.

(1.) CUPRIAMMONIUM FORMIOCHLORIDE, Cu(NH3)2CICHO

The bromide corresponding to this chloride has been described already by Richards and Shaw. The chloride itself was obtained at the same time by these experimenters, but only in an impure state, and the present problem was to determine the conditions necessary for the preparation of the substance in a state of purity.

If any considerable amount of water is present in the materials, basic salts of copper are certain to be precipitated, and to contaminate the preparation. On the other hand, the slight solubility of cupric formiate and cupriammonium chloride in alcohol makes it difficult to avoid the admixture of these substances with the desired compound, if alcohol is used in the anhydrous condition.

*Theo. W. Richards, Berichte d. d. ch. Gesell., XXV. 1492; T. W. Richards and H. G. Shaw, These Proceedings, XXVIII. 247.

The following procedure was found to be the most successful, but great care was needed to carry it out. Three grams of crystallized cupric formiate were dissolved in just enough warm alcohol to effect solution, and two grams of ammonic chloride were added to the mixture. The whole was then heated to boiling, and dry ammonia was passed in until a very slight excess was present. Upon cooling and evaporation in the air, fine blue prismatic crystals separated, which were fairly pure, as the analyses show. The salt resembles in its properties the formiobromide, being permanent in dry air, but at once decomposed by water. In color it is a purer blue than the formibromide, having much less of the greenish tinge.

In the analysis of the compound the copper was determined electrolytically after the substance had been evaporated with sulphuric and nitric acids. The chlorine from a new portion was weighed as argentic chloride, and the ammonia was distilled after the addition of potash. The formic acid was determined by combustion.

Analyses of Cu(NH3)2CICHO2.

I. 0.0862 gram of the substance gave on electrolysis 0.0305 gram

of copper.

II. 0.0997 gram of the substance gave on electrolysis 0.0356 gram

of copper.

III. 0.1258 gram of the substance yielded 0.0994 gram of argentic

chloride.

IV. 0.1285 gram of the substance yielded 0.1008 gram of argentic

chloride.

V. 0.1060 gram of the substance distilled with caustic potash required 11.75 cubic centimeters of a decinormal solution for neutralization.

VI. 0.0984 gram of the substance yielded on combustion 0.0241 gram of carbon dioxide.

Analyses III. and IV. were made from different samples; hence they prove the definiteness of the compound.

(2.) AMMON-CUPRIAMMONIUM ACETOBROMIDE,

Cu(NHg) BrC,H,O,. H,O.

3

This compound is formed readily when cupric bromide is dissolved in a mixture of alcohol and glacial acetic acid, and an excess of dry ammonia gas is passed into the solution. It is essential to have the solutions concentrated. For example, 2.5 grams of cupric bromide were shaken with 13 cubic centimeters of glacial acetic acid and 25 cubic centimeters of alcohol. Upon cooling after the addition of the ammonia, which raised the temperature of the solution, the desired substance separated out. If when passing in the gas a black precipitate (Cu, Br. (NH3) 10) falls after the solution has become dark purple,* the supernatant liquid should be decanted before it is allowed to crystallize. The precipitate shows the presence of an excess of cupric bromide in proportion to the acetic acid.

Ammon-cupriammonium acetobromide had already been made by Richards and Shaw; but the analyses of the compound were so unsatisfactory that no account of the substance was given in their paper. This unsatisfactoriness was due, not to any difficulty in preparing the

* Richards and Shaw, loc. cit.