nate of amyl. The trace of valerianic acid formed is doubtless due to a secondary reaction. The following equation would show the change:

2CH11NO2+ H2SO4 = 2H2O + SO2 + 2NO + C2H,(CH11)O2.

11

3rd. Action of Hydriodic acid on Nitrite of Amyl.

Concentrated hydriodic acid acts upon nitrite of amyl in the cold, very little rise of temperature taking place, although iodine is liberated abundantly.

Concentrated hydriodic acid and phosphorus were introduced into the apparatus already described, gently warmed, and nitrite of amyl added drop by drop. A colourless gas is evolved in abundance, which produces dense red fumes on coming in contact with the air; it is completely soluble in protochloride of iron, and is, in fact, nitric oxide.

On distilling the contents of the retort, a heavy oily liquid passes over. It was washed with dilute solution of carbonate of soda, dried over chloride of calcium, and distilled. It boiled constantly at 146° C. A weighed quantity of it was digested with alcoholic potash, and the amount of iodine determined; 3624 of the substance gave 4276 of iodide of silver.

.. 63.76 % I

theory

64.13 % I

These numbers can leave no doubt that the substance in hand is iodide of amyl.

Permanganate of potash also oxidizes nitrite of amyl. I have not fully investigated the products, but they appear to be the same as those obtained by the action of chromic acid.

XXXI.-Preliminary Notice on Phthalic Aldehyde.

By H. KOLBE and G. WIRCHIN.

A SHORT time ago* one of us predicted the existence of a hitherto unknown class of chemical compounds, namely, the aldehydes and alcohols of polybasic acids. We have since made numerous experiments with the view of discovering some of the members of

*Chem. Soc. J. [2], iv, 54 (Feb, 1866).

this group, and believe that we have succeeded in preparing phthalic aldehyde.

Phthalic chloride, treated with zinc and hydrochloric acid, is converted into a brownish-yellow unctuous mass, which, after dilution with a large quantity of water, may be extracted with ether; and if the residue left on evaporating the ethereal solution be shaken up with a small quantity of aqueous carbonate of ammonia, to remove any phthalic acid that may be present, the insoluble residue again dissolved in ether after repeated agitation with water, and the clear ethereal filtrate evaporated, after treatment with animal charcoal, there remains a compound having exactly the composition of phthalic aldehyde,

Phthalic aldehyde is a white, crystalline, easily fusible substance having a faint aromatic odour. It melts at 65° C., dissolves easily in alcohol and ether, very sparingly in cold water, more freely in boiling water, and crystallises from the latter solution by slow cooling in small rhombic plates. When carefully heated in small quantities, it sublimes apparently undecomposed; but larger quantities, when heated, turn brown, and suffer partial decomposition.

On mixing the warm aqueous solution with a concentrated solution of acid sulphite of sodium, the liquid solidifies, after a while, to a mass of long, slender, silky needles, probably a compound of sulphurous acid with soda and phthalic aldehyde.

A mixture of potassic chromate and sulphuric acid exerts but little oxidising action on phthalic aldehyde.

Phthalic acid, treated in a slightly acid solution with sodiumamalgam, undergoes a similar transformation. By prolonged action there is formed, besides the aldehyde, another body, probably the alcohol of phthalic acid.

We are continuing the investigation, and extending it to other polybasic acids.

XXXII.-The Nitroprussides.

By EDWARD ASH HADOW, King's College.

THE nitroprussides have allured me for years by the beauty of their reaction with soluble sulphides, and also because there is an evident uncertainty about their composition. They are evidently formed on the type of the ferridcyanides. Thus :

The only doubtful question seems to be what oxide of nitrogen replaces the MCy. A mere glance at Play fair's most laborious and careful ultimate analyses showed the hopelessness of endeavouring to gain light in this way by improving on them. The only conclusive proof, it was felt, would be the actual formation of nitroprusside from ferridcyanide by the displacement of an atom of alkaline cyanide by one of the oxides of nitrogen. Protoxide of nitrogen, NO, being generally rejected in published formulæ, was not tried; but as Gerhardt and almost all other chemists represent it as the binoxide NO2, an apparatus was fitted up as follows, for the purpose of testing the correctness of this view :

The flask C has a solution of pure ferridcyanide acidulated with sulphuric acid and kept hot, so that the binoxide NO2, instead of having to displace MCy, has the far easier work of displacing HCy, if it can. The whole apparatus was first filled with car

bonic acid to expel oxygen, and pure nitric oxide (NO2) from A, washed from any nitrous acid in B, was then passed in for some time. Flask C by degrees became very blue and muddy, and when, after an hour, the contents were examined, no trace of nitroprusside was found.

The operation was then exactly repeated, only that instead of pure nitric oxide the red nitrous acid vapours from starch and nitric acid were passed in. The difference was wonderful; no Prussian blue appeared in flask C. The colour changed rapidly to the red of nitroprusside, and much hydrocyanic acid came off into D; and on opening the flask all ferridcyanide of potassium had disappeared, the conversion to nitroprusside being almost perfect. Since then I have never had a doubt but that the generally received formula Fe2Cy,NO,Na, was wrong, and that the true formula has NO, or NO instead, but which, the last experiment does not determine, for NO, from starch always has NO mixed with it. The presence of NO, or NO4 as the nitrogenoxide essential to nitroprusside, is further shown by a very simple experiment--the reverse of the last. On adding to a solution of pure nitroprusside a solution of potash and some prussic acid, and warming, the red of the nitroprusside changes for an instant to deep yellow, and then to pale yellow. The solution will be found to contain ferrideyanide, with plenty of nitrite, but no trace of nitroprusside. Here NO, or NO4 has been displaced by KCy, and ferridcyanide reproduced.

I felt that nothing could be conclusive but the actual displacement of an alkaline cyanide in ferrideyanide by pure NO3, which can be obtained by the action of acetic acid on an alkaline nitrite, which mixture absolutely excludes NO4. After trying many methods for displacing KCy in ferrideyanide of potassium, the following experiment was tried with perfect success :-Ferridcyanide of potassium solution was mixed with a solution of corrosive sublimate and acetic acid, and a nitrite added, and the whole was left for some hours, when, on examining the mixture, nitroprusside was found in abundance, which experiment conclusively stamped NO, as the oxide. The reaction was as follows:

3

3

Nitroprusside.

Fe2CyK+HgCl + NO2 = Fe2Cу¿NOŽK2 + HgCy + KCl; and the true formula of the sodium-salt is

Fe, Cy,NO,Na + 4HO = 306.

The fact of this being the true formula is further confirmed by the close correspondence between the percentage of each of the elements calculated from this formula and Playfair's own numbers, which is closer than the numbers derived from Gerhardt's formula.

The success of this reaction on the small scale suggested it as a means of manufacturing nitroprusside on the large scale. The present retail price of the sodium-salt, 6s. per oz., or £4 16s. per lb., while the ferrocyanide from which it is derived is worth about 1s. 6d. per lb., sufficiently shows the difficulties attending the present method.

After very many experiments the following method has been arrived at, which, though by no means perfect, will furnish nitroprusside of sodium by the pound without difficulty.

1. A strong solution of caustic soda is prepared, and thoroughly saturated with the nitrous acid vapours from starch and nitric acid (old battery acid does well). The amount of true nitrite of soda, NaONO3, in this solution is determined by permanganate by taking a small measured quantity, diluting largely with water, acidifying with sulphuric acid, and observing the amount of standard permanganate decolorised. Fe NaO.NO, in decolorising power. The amount of real nitrite of soda in the

=

solution being known, is recorded on the bottle.

A mixture of any bulk is then made in the following propor

Pour the cold

adding acetic acid, if necessary, until quite clear.

make up to pint with cold water,

solution B into the hot solution A. The mixture becomes at