The loss of water in analysis V. amounted to 44 per cent. ; the theory CH,, 2 U, O, PO1+ Aq requires 4:35 per cent.

2

4

Arsenious Ethylphosphate.-The replacement of the basic hydrogen of ethylphosphoric acid by such an element as arsenic appeared to present some features of interest. The experiment was thus made. To a weighed quantity of pure arsenious chloride in a small flask, an equivalent quantity of anhydrous ethylphosphate of lead was added (in one experiment ethylphosphate of silver). The mixture became warm, and after moderate heating solidified. It was extracted with warm water, and the filtered extract evaporated. Beautiful feathery crystals separated in considerable quantity. Once crystallized from a solution, they appeared to dissolve less readily a second time in water. The cause of this phenomenon was soon discovered. Water gradually decomposes this salt, giving arsenious anhydride and ethylphosphoric acid. Although the analysis of the first crop of crystals was tolerably satisfactory, the original method of preparing the compound was abandoned, and another plan adopted. It was found that ethylphosphoric acid readily dissolves arsenious acid at the boiling-point, and that on heating and evaporating the solution, beautiful crystals of the arsenious ethylphosphate separate. In order to study this reaction more closely, the experiment was repeated, substituting, however, common orthophosphoric for the ethylphosphoric. The arsenious anhydride readily dissolved in considerable quantity on ebullition; and after filtration and cooling, an abundant crop of brilliant crystals was deposited from the filtrate. These crystals were not perceptibly affected by washing with cold water, and proved to be completely volatile when heated in a test-tube over a spirit-lamp. In fact they were nothing but octahedra of arsenious anhydride. Further experiments showed that it was not possible in this manner to form an arsenious phosphate; so that the statement in Gmelin's Handbook, referring to this salt as probably obtainable by the process above given, would seem to require correction. The normal arsenious phosphate, As PO,, remains to be discovered; a peculiar interest consequently attaches to the salt now under review, as the only arsenious phosphate known. Prepared by either of the processes above given, pressed between folds of filtered paper, and dried in vacuo, it gave on analysis numbers very nearly agreeing with the expression

(CH) 3 PO..

The arsenic in the arsenious ethylphosphate was determined as sulphide, the precipitation being effected according to the directions given by Fresenius. The salt was dried by pressure between folds of filter-paper, it was then placed in vacuo over sulphuric acid, and finally heated for a short time in the water-oven. The arsenious sulphide obtained on its analysis

The formula (C, H,), As, 3 PO, requires 28.74 per cent. of arsenic : analysis I. gave 28.95, while II. gave 29.58, the specimen of salt submitted to analysis in the latter case having been partially decomposed by washing, ethylphosphoric acid being thus removed, and consequently an excess of arsenious anhydride remaining in the residual salt.

The ease with which the arsenious ethylphosphate is formed induced me to hope that similar success would attend experiments made with another triatomic element, phosphorus. When an action is established between terchloride of phosphorus and ethylphosphate of silver, an oily product may be removed from the mass by means of anhydrous ether, but it yields on analysis results less definite than could be wished. Yet the reaction is a promising one: I hope to recur to it shortly, and to experiment in a similar manner with antimony and bismuth compounds.

Ditetrethyliac Ethylphosphate.-The ordinary ethylphosphate of ammonium is very readily made; its aqueous solution becomes acid on evaporation, but the salt may be obtained in a semicrystalline form by drying its saturated solution in vacuo over sulphuric acid. The salt heated carefully in an oil-bath for some time loses ammonia as well as water, but yet appears to yield, among other products, ethylphosphamic acid. A different and much more definite kind of decomposition takes place with the compound next to be described.

59

When a hot solution of argentic ethylphosphate is mixed with a solution of iodide of tetrethylium, a change occurs expressed by the equation C2 H ̧, Ag2, PO1+2 [(C2 H2), NI]=C, H„, [(C, H ̧), N]2, PO1+2Ag I. If the two salts be employed in the exact proportions indicated by this equation, it will be found that after boiling them together the new compound is contained in the filtrate. On evaporating this liquid first of all at 100°, and then in vacuo over sulphuric acid, a syrup, and finally a mass of confused crystals will be obtained; by long drying, these crystals lose their transparency, most probably because they have thus parted with some of their water of crystallization. The salt is intensely soluble in cold water, and deliquescent. The analyses of this salt were not exact, but corresponded sufficiently with the formula of an ethylphosphoric acid in which two atoms of hydrogen had been replaced by two atoms of tetrethylium. This view of its constitution is amply confirmed, not only by the mode in which the salt is prepared, but also by a singular decomposition

which it undergoes when heated. It begins indeed to decompose, though very slightly, at 100°, even when water is present, giving off a distinct odour of triethylamine. But on heating the salt itself to a temperature exceeding 100°, decomposition becomes more rapid, and the substance is finally resolved into triethylamine and triethylic phosphate,

C2 H,, [(C2 H ̧), N]2, PO1=(C2 H ̧) ̧ PO ̧+2 (С2 H ̧), N.

2

The triethylamine was analytically identified by a platinum-determination in the double chloride made from it by addition of hydrochloric acid and platinic chloride. The metamorphosis of this ethylphosphate is perhaps more easily seen by means of the following arrangement of its formula :—

Several ethylphosphates have been prepared besides those described in the present paper; most of these salts, however, presented no marked features of interest. The ammonium, nickel, chromic, mercurous, and platinic compounds were investigated more particularly. The mercurous ethylphosphate is somewhat difficult to prepare; it is best made by adding a few drops of mercurous nitrate to a strong solution of potassic ethylphosphate, filtering off the grey precipitate first formed, and then adding a further quantity of the mercurous nitrate in solution; if the solutions are not too concentrated the salt gradually separates in pearly plates. Hot water partially dissolves this salt, the residue becoming yellow, and the solution acquiring a distinct acid reaction. It is slightly soluble in cold water, though not altogether without decomposition; it is insoluble in alcohol. Dissolved in dilute nitric acid and precipitated by chloride of sodium, the air-dried crystals of this salt, gave the following result :

236 grm. gave 191 grm. of Hg" Cl.

This corresponds to 71.09 per cent. of mercury; the formula C, H, Hg", PO +2 Aq requires 71-45 per cent. The more probable formula, C, H, Hg", PO +Aq, requires 73.82 per cent.

In offering the foregoing results to the Society, I do not wish it to be supposed that I consider them conclusive so far as regards the theoretical considerations introduced into the present paper. It is possible that the various aluminium and iron salts described may be mixtures only, in spite of their apparent constancy of composition; or, again, it may be that their formulæ ought to be doubled or quadrupled. Moreover the constitution of ethylphosphoric acid itself has not been made out: I trust that the study of diethyl-, ethylpyro-, and ethylmeta-phosphoric acids, and of the pro

ucts of the action of heat on the ethylphosphates, may aid in determining this question. Meanwhile the observation, already recorded, as to the stability of ordinary ethylphosphoric acid and its salts in the presence of permanganate of potassium requires a word or two of further comment. When argentic diethylphosphate is acted upon with iodine, the silver and one atom of ethyl may be removed, and after treatment with finely divided silver and a little oxide of silver to remove any iodine and hydriodic acid, and then with excess of carbonate of barium, an ethylphosphate of barium is obtained, which, unlike the ordinary salt, immediately reduces permanganate of potassium; perhaps the ethyl in this salt exists in a different and less intimate form of combination. I am inclined to think that the permanganates will afford, in some cases, criteria for the detection of slight differences in isomeric compounds, although it would be premature at present to hazard an exact interpretation of the phenomena to which they give rise. I may add that treatment of an ethylphosphate with strong nitric acid fails to decompose the ethylphosphoric acid; so that phosphoric acid cannot thus be separated from this remarkably stable body.

e

II. "A Dynamical Theory of the Electromagnetic Field." By Professor J. CLERK MAXWELL, F.R.S. Received October 27, 1864.

(Abstract.)

The proposed Theory seeks for the origin of electromagnetic effects in the medium surrounding the electric or magnetic bodies, and assumes that they act on each other not immediately at a distance, but through the E intervention of this medium.

The existence of the medium is assumed as probable, since the investigations of Optics have led philosophers to believe that in such a medium the propagation of light takes place.

The properties attributed to the medium in order to explain the propagation of light are

1st. That the motion of one part communicates motion to the parts in its neighbourhood.

2nd. That this communication is not instantaneous but progressive, and depends on the elasticity of the medium as compared with its density. The kind of motion attributed to the medium when transmitting light is that called transverse vibration.

An elastic medium capable of such motions must be also capable of a vast variety of other motions, and its elasticity may be called into play in te other ways, some of which may be discoverable by their effects.

One phenomenon which seems to indicate the existence of other motions than those of light in the medium, is that discovered by Faraday, in which the plane of polarization of a ray of light is caused to rotate by the action

of magnetic force. Professor W. Thomson has shown that this phenomenon cannot be explained without admitting that there is motion of the luminiferous medium in the neighbourhood of magnets and currents.

The phenomena of electromotive force seem also to indicate the elasticity →→ or tenacity of the medium. When the state of the field is being altered by the introduction or motion of currents or magnets, every part of the field experiences a force, which, if the medium in that part of the field is a conductor, produces a current. If the medium is an electrolyte, and the electromotive force is strong enough, the components of the electrolyte are separated in spite of their chemical affinity, and carried in opposite directions. If the medium is a dielectric, all its parts are put into a state of electric polarization, a state in which the opposite sides of every such part are oppositely electrified, and this to an extent proportioned to the intensity of the electromotive force which causes the polarization. If the intensity of this polarization is increased beyond a certain limit, the electric tenacity of the medium gives way, and there is a spark or "disruptive discharge."

Thus the action of electromotive force on a dielectric produces an electric displacement within it, and in this way stores up energy which will reappear when the dielectric is relieved from this state of constraint.

A dynamical theory of the Electromagnetic Field must therefore assume that, wherever magnetic effects occur, there is matter in motion, and that, wherever electromotive force is exerted, there is a medium in a state of constraint; so that the medium must be regarded as the recipient of two kinds of energy—the actual energy of the magnetic motion, and the potential energy of the electric displacement. According to this theory we look for the explanation of electric and magnetic phenomena to the mutual actions between the medium and the electrified or magnetic bodies, and not to any direct action between those bodies themselves.

In the case of an electric current flowing in a circuit A, we know that the magnetic action at every point of the field depends on its position relative to A, and is proportional to the strength of the current. If there is another circuit B in the field, the magnetic effects due to B are simply added to those due to A, according to the well-known law of composition of forces, velocities, &c. According to our theory, the motion of every part of the medium depends partly on the strength of the current in A, and partly on that in B, and when these are given the whole is determined. The mechanical conditions therefore are those of a system of bodies connected with two driving-points A and B, in which we may determine the relation between the motions of A and B, and the forces acting on them, by purely dynamical principles. It is shown that in this case we may find two quantities, namely, the "reduced momentum" of the system referred to A and to B, each of which is a linear function of the velocities of A and B. The effect of the force on A is to increase the momentum of the system * Proceedings of the Royal Society June 1856 and June 1861.