hydroferrocyanic acid were boiled with potassium sulphate solution in an atmosphere of hydrogen, a double surface condenser being used to keep the concentration constant. The water used had been boiled and also cooled in hydrogen to remove dissolved air.

In one case, a 10 per cent. solution of hydroferrocyanic acid with excess of potassium sulphate gave a precipitate of Everitt's salt on boiling. On warming the precipitate with dilute sulphuric acid at 60° and filtering, no ferrous sulphate was found in the filtrate, hence no ferrous cyanide was formed.

A 0.7 per cent. solution of hydroferrocyanic acid was similarly treated, and a precipitate obtained. It was, however, merely ferrous cyanide, as it rapidly and completely dissolved in dilute sulphuric acid at 60°, forming ferrous sulphate and evolving hydrocyanic acid.

The precipitate obtained from the strong (10 per cent.) solution was collected in absence of air, washed repeatedly, placed in a desiccator in an atmosphere of carbon dioxide, and dried in a vacuum at 100° over phosphoric oxide. The pale bluish-yellow precipitate contained potassium, iron, and cyanogen, and was analysed by (i) evaporating down with strong sulphuric acid and a little ammonium sulphate and igniting; (ii) boiling with mercuric oxide, filtering, igniting, and weighing as ferric oxide.

0-3086 gave 0.1457 Fe2O3. Fe=33.1.

K2Fe(CN), requires Fe 32.4 per cent.

=

Decomposition of Hydroferrocyanic Acid.

A number of experiments were made with the object of studying the decomposition represented by the equation:

H4Fe(CN)6 = Fe(CN)2+4HCN.

Fe(CN)2 + H2SO4 = FeSO4 + 2HCN.

According to Berzelius (loc. cit.) the decomposition with boiling water is represented by the equation:

H1Fe(CN)6=4HCN + Fe(CN)2.

Reemann and Carius (Annalen, 1860, 113, 39), and Étard and Bémont, (Compt. rend., 1884, 99, 1024), on the other hand, express it by the equation:

2H,Fe(CN)=6HCN+H2Fe(CN)6

but, in their papers, give neither analyses nor details.

To examine this question, a quantity of hydroferrocyanic acid was prepared by Possell's method, and purified by repeatedly dissolving in absolute alcohol and reprecipitating by ether until spectroscopically

* Ferrous cyanide dissolves completely and rapi lly at 60' in dilute sulphuric acid

free from potassium. After drying in a vacuum over phosphoric
oxide at 60°, this was analysed.
Determination of the iron gave:

On heating with water in an atmosphere of hydrogen, hydrocyanic acid began to be evolved at 60°, and a pale yellow-green solid separated out, thus proving Berzelius' statement to be correct. The hydrocyanic acid evolved was estimated as above.

0-4970 H4Fe(CN), gave 0.2605 HCN. HCN=52.4. 0-3863 H,Fe(CN)6 0.2035 HCN. HCN=52.7.

3HCN requires 375 per cent. 4HCN requires 50 per cent.

The excess of hydrocyanic acid is undoubtedly due to hydrolysis of the ferrous cyanide, as on further boiling hydrocyanic acid is slowly evolved, and the precipitate, after drying in a vacuum, was found to contain ferrous oxide.

In presence of air, a coppery-blue precipitate was formed, but in too small a quantity for analysis. It was undoubtedly Williamson's blue, KFe(CN), as this is formed from Everitt's salt in presence of air and sulphuric acid (Williamson, loc. cit.). That Everitt's salt in presence of air is converted more or less completely to Williamson's blue, which, in the presence of dilute acid and oxygen, decomposes as fast as it is formed, was confirmed by preparing a quantity of the latter. On boiling this with dilute sulphuric acid in presence of air it dissolved, forming hydrocyanic acid and ferric sulphate.

2KFe(CN)6 + 8H,SO, +0=2KHSO4 + Fe2(SO4)3 + H2O+ 12HCN.

4

Summary.

The preceding results may be epitomised:

(i) Concentrated sulphuric acid, H,SO,, dissolves potassium ferrocyanide and shares the potassium with the hydroferrocyanic acid. The ratio must be primarily determined by the active masses and relative affinities of the acids. The following equation represents the initial change:

K4Fe(CN)¿+ H2SO4 = 4K HSO4 + H1Fe(CN)6⋅

The solution is only very slowly decomposed by rise of temperature. Carbon monoxide is given off, but even at 200° the rate of evolution is low and the decomposition proceeds only when the sulphuric acid can dissociate or decompose into water, sulphur trioxide, and sulphur dioxide.

(ii) With the addition of water, marked decomposition occurs, and large quantities of carbon monoxide are formed. This reaction increases with dilution until the concentration H2SO4,2H2O is reached;

at this strength, the whole of the cyanogen appears as carbon monoxide at 180°. The equation which represents this change is: K ̧Fe(CN) ̧+8(H2SO,2H2O) = 4KHSO4+3(NH4)2SO4 + FeSO4+6CO +10H2O.

In this case, there is evidently hydrolysis, and it seems probable that it may be directly due to the molecules of water which are dissociated by solution in the solvent H2SO4. On the other hand, it is possible that sulphuric acid of the concentration H2SO4,2H2O may really act as orthosulphuric acid, S(OH), in which case the above reaction may be evidence of its existence.

(iii) With further dilution to the concentration H2SO4,4H2O there is another definite change in the reaction, since, in addition to carbon monoxide, Everitt's salt, K,Fe(CN), and hydrocyanic acid make their appearance. That this is a definite change is shown by the fact that warm acid of the strength H2SO4, 2H2O immediately decomposes Everitt's salt.

The Everitt's salt appears to be formed from hydroferrocyanic acid by the action of potassium sulphate solution, thus:

2H4Fe(CN)¿Aq+2K2SO4Aq = K2Fe(CN)6 + 2KHSO1Aq+6HCNAq. Some hydroferrocyanic acid is decomposed at the same time, forming hydrogen cyanide and ferrous cyanide. With increasing dilution, this becomes the more important and eventually, with acid of the concentration H2SO4,10H2O, the sole reaction. The equation which may represent this stage is:

H1Fe(CN)6Aq=4HCNAq + Fe(CN)2

In presence of the sulphuric acid, the ferrous cyanide dissolves with the formation of ferrous sulphate and hydrocyanic acid, and the change as a whole may be considered as due to the molecules of sulphuric acid dissociated by solution in the solvent water.

The final equation now becomes :

K4Fe(CN)6 +5(H2SO4,10H2O) = 4KHSO4 + FeSO4 + 6HCN +50H2O.

(iv) The final decomposition is hindered by increasing the mass of the salt, but helped by increase of temperature and the presence of air.

This last condition is important, and assists in the rapid decomposition of the salt. It also explains the use of porous brick or the passage of an air current to assist in the preparation of hydrocyanic acid on the manufacturing scale.

When air is present, Williamson's blue, KFe(CN), is formed in small quantity, and the solution contains ferric salts; it is probably formed from Everitt's salt, and then decomposed by the action of the oxygen of the air, as explained above.

These results, therefore, confirm the oldest account of the hydrolysis of hydroferrocyanic acid, namely, that due to Berzelius (loc. cit.).

The authors are engaged in further investigations of changes of the character of that described under (ii) produced by sulphuric acid, to see if they admit of further elucidation.

ST. JOHN'S COLLEGE, LABORATORY,

CAMBRIDGE.

XVII.-Action of Alkyl Iodides on the Mercuric Iodide Sulphides of the Fatty Series.

By SAMUEL SMILES, B.Sc.

SOME years ago, Krüger (J. pr. Chem., 1876, [ii], 14, 207) investigated the action of methyl iodide on methyl ethyl sulphide, and of ethyl iodide on dimethyl sulphide, and found that in each case a different dimethylethylsulphine iodide was produced. This work was repeated by Nasini and Scala (Gazzetta, 1888, 18, 67) a few years later and confirmed. Klinger and Maassen (Annalen, 1888, 243, 193), however, in a careful series of experiments, obtained by the two methods. identical sulphine iodides which gave rise to identical series of double salts. At the same time, they pointed out that Krüger's method of preparation was at fault and that his substances were impure.

From the facts that conflicting results have been obtained, and that the bulk of the evidence in these researches is based upon the behaviour of certain double salts, it seemed possible that an investigation of these substances might throw some light upon the matter. The following research has therefore been undertaken, firstly, to attempt to determine the constitution of the double salts of the sulphine bases, and, secondly, to ascertain whether, by varying the methods of preparation, stereoisomeric compounds could be obtained.

In acetone or alcoholic solution, the sulphides unite with mercuric iodide to form compounds of the general formula R,SHgI. These are analogous to such substances as RSI, RSBr2, R,SRI, &c., and their constitution must therefore be represented in a corresponding manner, that is, as containing quadrivalent sulphur : Rs.

R

HgI I

Mercuric iodide and the alkyl iodides might be expected to react in a similar way with the mercaptans forming substances such as and R>S<. These, however, do not appear as pro

R>s<HgI

H

ducts of the reaction, and may be assumed to be unstable com

pounds, formed at an intermediate stage, and readily splitting off hydrogen iodide as follows:

RSH + HgI,

R HgI
>S<

1

H

Regarded from this point of view, the acid nature of the alcohols and mercaptans appears to be dependent on the tendency of oxygen and sulphur to become quadrivalent, and since oxygen shows this tendency to a less degree than sulphur, it follows that the alcohols show weaker acid properties than the mercaptans. That this is the case is illustrated by the fact that ethyl mercaptan, dissolved in an alcoholic solution of sodium ethoxide, shows all the reactions of sodium mercaptan ; for example, on treating with alkyl iodides, the sulphides are formed. The alkyl iodides react with the mercuric iodide compounds of the sulphides to form stable yellow substances, which are found to be identical with those produced from the corresponding sulphine iodides and mercuric iodide. Accordingly, their formation must be represented as follows:

Dobbin and Masson (Trans., 1885, 47, 56) have already investigated the products of addition of the halogens to the sulphine halides, and conclude that they are not molecular compounds, but owe their formation to either sexavalent sulphur or tervalent iodine. From their stability and from the fact that the sulphine sulphates give similar compounds, these authors incline to the former hypothesis.

In support of this view, I have found that dimethyl sulphide bromide, (CH),SBr2, reacts with methyl iodide to give the same substance, (CH),SBr,I, as is produced from trimethylsulphine iodide and bromine in molecular proportion.

The following series of reactions were carried out with the object of investigating the stereochemical properties of sexavalent sulphur. As yet, only negative evidence has resulted. It has been found that the same compound of trimethylsulphine iodide with mercuric iodide is produced in the following three different ways.

From dimethyl sulphide-mercuric iodide and methyl iodide: ]