with a diminution, instantly, of about one-fifth of the gas produced, and with a slight explosion. I was now sure, from this phenomenon, and from the unequal augmentation of the bulk of the gas at given times during the process, that combustion had been excited several times before, not only in the present experiment, but perhaps in the former ones, without observing it. I conceived that a gradual combustion also very probably took place in this process by the kindling of bubbles of gas in their ascent through the water. I now perceived that the discharges ought to be produced more slowly, or the tubes to be wider, to allow the bubbles to pass quite through the water, in order to avoid the ascension of gas during the process. My calculation, also, that 35,000 to 40,000 discharges were requisite to produce one cubical inch of gas from water containing its usual quantity of common air, was rendered much more vague by this ascension, so often liable to be occasioned.

To the gas which remained in the tube in this experiment was added an equal bulk of nitrous gas; the mixture diminished to 1.5; and on adding to the residue half its bulk of oxygen gas, and passing through it the electrical spark, no accession or diminution of bulk was produced. Hence all the hydrogen gas and oxygen gas, produced by the decomposition of the water, had been burnt during the process; the oxygen gas thus detected being considered to be only that expelled from the water.

Exp. 4. By means of electrical discharges with the apparatus used in the preceding experiment, I obtained gas from New River water, letting it up into a reservoir, as soon as about 1-20th of a cubic inch was produced, till I had collected 1-8th of a cubic inch. To this was added an equal bulk of nitrous gas on which the mixture diminished to 1.2; and on the addition of a little more nitrous gas, no further diminution took place. To this residue half its bulk of oxygen was added; and this mixture of gases being well dried, by standing over lime and boiled quicksilver, an electric spark was passed through it, by which a diminution of 1-6th of its bulk took place. A little dew was then seen upon the sides of the tube where the quicksilver had risen; and with the aid of a lens the same appearance was perceived on the part of the tube containing the residue of gas.

It may now be expected, that I should have made the experiment with this apparatus on distilled water, freed from its air, not only by long boiling or the air pump, but by sending through it several hundred electrical discharges. It would also have been, to some persons, more satisfactory if the experiment had been made upon a larger scale, so as to have produced the combustion of a much larger quantity of gas, and consequently have produced a greater quantity of water. As, however, I apprehend, the experiments contained in this paper, when well considered by competent judges, will be found to explain the nature of the gas procured from water by electric discharges, and as another very important subject demands my

attention, the honour of more splendid and convincing experiments must be reserved for other enquirers. If the same sacrifices be made by them, as have been made in performing the present experiments, I think it is scarcely possible but that still further light concerning the composition of water should be procured, as well as concerning oils, alcohol, acids, &c. : to the investigation of the composition of which, the mode of analysis and synthesis here indicated may be applied.

SECTION III.

On the Mode of Action of Electric Discharges.

The mere concussion by the electric discharges, appears to extricate not only the air dissolved in water, which can be separated from it by boiling and the air pump, but also that which remains in water, notwithstanding these means of extricating it have been employed. The quantity of this air varies in the same, and different waters, according to circumstances. New River water from the cistern yielded 1-5th of its bulk of air, when placed by Mr. Cuthbertson under the receiver of his most powerful air pump; but in the same same situation, New River water taken from a tub exposed to the atmosphere for some time yielded its own bulk of air. Hence the gas procured by the first one, two, or even three hundred explosions in water, containing its natural quantity of air, is diminished very little by an electric spark.

The gas thus separable from water, like atmospherical air, consists of oxygen and nitrogen, or azotic gas; which may be in exactly the same proportions as in atmospherical air; for the water may retain one kind of gas more tenaciously than the other; and on this account the air separated may be better or worse than atmospherical air at different peiods of the process for extricating it.

With regard to the gas which instantly disappears on passing through it an electric spark, its nature is shewn by (a) this very property of thus diminishing, and by the following properties:

(b) A certain quantity of nitrous gas instantly disappeared, apparently composing nitrous acid by being added to the gas (a) Exp. IV.

Oxygen gas being added to the residue, after saturation with nitrous gas, and an electric spark being applied to the mixture of gases, well dried, a considerable diminution immediately took place, and water was produced.

(c) Combustion from hydrogen and oxygen gas took place when the tube was about three-fourths full of gas, Exp. I; which was confirmed by passing an electric discharge, under the same circumstances, through a mixture of hydrogen and oxygen gas, Exp. II.

(d) Combustion from hydrogen and oxygen gas took place when the points of the compasses were accidentally applied to the part of the tube containing gas; which was confirmed by passing a discharge, under the same circumstances, through a mixture of

hydrogen gas, while the points of the compasses were applied to the tube.

(e) The observations made of the kindling of gas, in small quanties, from time to time, during the process of obtaining it, particularly while it was ascending in chains of bubbles, or was adhering to the funnel of the tube, Exp. III, confirm the evidence in favour of this gas being hydrogen and oxygen gas.

The evidence contained under the above heads (a)—(e) considered singly and conjunctively, I apprehend, must be admitted by the most rigorous reasoner, or severest logician, to be demonstrative that hydrogen and oxygen gas were produced by passing electric discharges through water.

With regard to the origin and mode of production of these two gases, our present observations and experiments do not afford complete demonstrative evidence; but although some hypothesis must be admitted, I conceive that the body of evidence we possess can afford a satisfactory interpretation of the phenomena.

It is demonstrable that the electric discharge and spark contain fire; and very probably they are merely a state of fire. Fire may be considered as consisting of caloric and light; but it is at least as consistent with the phenomena, and it is more philosophical, because it is more simple, to consider light not as a distinct species of matter, but as a state of caloric, which is manifested by its producing the sensation termed vision. It is demonstrable also, that the ponderable parts of oxygen and hydrogen gas constitute water. There is strong evidence that these gases consist of a peculiar species of matter which is ponderable; and of imponderable matter, which is that which is separable from them in the state of fire, or flame. If fire could be applied in a sufficiently dense state and quantity, it is warrantable, from a full induction of facts, to conclude, that it is able to disunite the constituent substances of all the compound subtances in nature.

The fire of the electric discharge, in a very condensed state, passes with inconceivable velocity through the whole length of the upper wire, in the case of the interrupted explosion, and of the single wire in the case of the complete explosion; so that it neither exerts its energy on the wire, nor on the water, till it arrives at the extremity of the wire. There it is momentarily interrupted and accumulated; and, in the moment before its diffusion through the water, it is so dense and in such quantity as to manifest itself by a spark at, or nearly in contact with, the extremity of the wire. In the moment of its diffusion, a small part of this condensed wire interposes betwixt the constituent elements of the ultimate and invisible particles of water, that is, betwixt the hydrogen and oxygen, of which water is compounded, so as to place them beyond the sphere of their chemical attraction for one another; and each ultimate particle of hydrogen and of oxygen uniting with a determinate quantity of fire, new compound ultimate

particles, consisting of hydrogen and caloric, and of oxygen and caloric, that is, hydrogen gas and oxygen gas, are compounded. This mode of action of electric fire on water is confirmed by the effect of electric fire and common fire, of a due degree of density, on oxide of quicksilver, which is by them resolved into oxygen gas, and quicksilver in the vapour state. All calculations must needs be extremely vague; yet, perhaps, some elucidation will be obtained by considering that as it probably requires seventy or eighty thousand discharges to produce a cubic inch of gas from the supposed decompounded water, the gas produced by each discharge cannot amount to one 200,000th of a grain weight. The quantity of condensed fire at the extremity of the wire must be immensely great, comparatively with the quantity of it which enters into the composition of the gases from decompounded water; otherwise it is not easy to conceive how even the minutest particle of water could be decompounded, the electric fire being in contact with a large body of water, and passing through it with a velocity which is incalculable.

The reason of the metal wire not uniting to the oxygen of the decompounded water, as in the experiment of passing water through red hot iron tubes, might be assigned from the intensity of the fire only; but it is also on account of the rapid motion of the discharge, as well as partly from the great quantity of light. In a very low temperature, light decompounds oxide of silver, and of several other metals; also oxy-muriatic acid, nitric acid, &c. Hence light both decompounds bodies, and prevents oxygen from coming within the sphere of chemical attraction of the metal.

It is supposed that but a very minute proportion indeed, of the electric discharge, is consumed at the end of the wire in the composition of gases, during a momentary interruption, as above said; for it diffuses itself through the water, between the wires, yielding a volume of vivid light till it arrives at the extremity of the under wire or point of the convexity of the metal cup, where it is again condensed by the superior conducting power of the metal to that of water, and where, if it be in due quantity, and of sufficient density, it manifests itself by a spark, and infallibly again decompounds water hence bubbles are seen to rise from the end of the lower wire, and from the metal cup, as well as from the point of the upper wire.

From these interpretations it will not be difficult to explain the reason of a spark appearing in some cases at the points of both the upper and under wire; why in other cases it appears at the point of the upper wire only; why in others it appears at the point of the lower wire only; namely, according to the density of the fire of the discharge there accumulated.

Concerning the agency of electrical fire in causing the hydrogen and oxygen gas of the supposed decompounded water to undergo combustion and produce water, it is well known that the smallest visible spark, or particle of flame, or fire, can kindle as rapidly a

very large quantity of hydrogen and oxygen gas, as the greatest quantity of flame, sparks, or fire, can kindle the smallest quantity of these two gases; while, on the other hand, the largest mass of matter, heated most intensely, but short of ignition, cannot produce combustion of oxygen and hydrogen gas.

Although, as hath been above explained, caloric and light, in a sufficiently dense state, may decompound every compound substance in nature; it is also well ascertained, that caloric, in certain states, universally promotes chemical combinations. The mode of agency, in this latter case, I apprehend to be in a different way from that commonly accepted; for I think it is unnecessary to suppose that it operates by increasing the power of chemical attraction, and I conceive its agency to be merely diminishing or destroying the powers which counteract chemical union; especially diminishing cohesive attraction, and exciting motion among the particles of the different substances: hence these substances are applied to one another, within their spheres of chemical attraction; or the chemical attraction acts between a greater number of points of the different bodies, as when caloric renders solids and inelastic fluids into the elastic fluid state.

Accordingly, when an electric spark, or the smallest particle of flame, or of an ignited substance, is applied to the gas produced in the above process, or to the mixture of hydrogen and oxygen gas, the ultimate particles of these gases nearest to the flame are driven from it in all directions, as from a centre, by the interposition of fire, or of caloric and light; so that they are brought within the sphere of their chemical attraction for the ultimate particles of the gases at a certain distance from the centre of application of fire; which therefore unite, and the caloric and light, disengaged by that union, act in a similar manner in producing union among the next set in order of proximity of the ultimate particles of the gases; the disengaged caloric and light of which act in producing union of the next set of ultimate particles in order; and so on successively, but with incalculable velocity, the greatest bulk of ultimate particles of the two gases unite with one another; the known products of which union are fire and water; or light, caloric, and water. According to this hypothesis, if caloric or fire, or merely caloric of sufficient intensity and quantity, be applied to a given bulk of hydrogen and oxygen gas, no combustion should be produced; as the caloric will be interposed in such quantity that all the ultimate particles must be at the same instant driven from one another in all directions, so that they are beyond their spheres of chemical attraction for one another; and in various instances common experience verifies this hypothesis.

If light be considered as a different species of substance from caloric, then the theory of its agency, lately published by Dr. Parr, M.D., may be applied very happily to explain the explosion from the combustion of oxygen with hydrogen gas. From a very large induction of facts, Dr. Parr infers that, although light and caloric