bad conductor, as is the case with the very best conductors. A cubic foot of water is only just capable of receiving, or letting pass through it, a full discharge from a jar of one foot of coated surface; and the quantity of water employed in this experiment not being 100th part of a cubic foot, it is a very imperfect conductor; so that a interrupted discharge only can pass through the tube, without dispersing the whole of the water. But if the discharge be not seemingly as strong as the tube can bear without breaking, the gas is not produced from it; and on this point hinges this extremely delicate process.

To succeed by the method of the complete or uninterrupted explosion, the following apparatus must be used, and rules observed:

1. A tube is employed about four or five inches in length, and 1-5th or 1-6th of an inch in diameter. One end is mounted with a brass cap, and the other end is sealed at the lamp, with a wire about 1-40th of an inch in thickness fixed into it, as above described; which extends into the brass cap, so as to be almost in contact when the explosion is made. If the wire touches the brass cap there will be no explosion. The tube being filled with, and set in, a cup of water, the discharge may be made into it as in the above described process; but here the insulated ball must be placed at a greater distance from the prime conductor, and a Leyden jar with only fifty square inches of coating, will answer the purpose. In this way of making the experiment gas is produced by each discharge in the brass tube, in much greater quantity, with much less frequent accidents, and less trouble, than in the former method with the interrupted discharge. But the gas obtained with this apparatus always contains a larger proportion of atmospherical air, on account of the quantity of water, and more immediate and extensive communication of it with the atmosphere. By repeated discharges there is an impression made in the brass tube, in the part where the discharge passes through it, and at last a small hole is made in that part. On this account the same mounted tube cannot serve for producing a large quantity of gas.

The other sort of apparatus invented by Mr. Cuthbertson consisted of a glass tube half an inch wide and about five inches in length, mounted at one end with a brass funnel, and inverted in a brass dish; but afterwards the tube was blown funnel-wise at the end. The other end must have a wire about 1-40th of an inch thick, sealed into it at the lamp; which wire extends to nearly the bottom of the brass dish in which the tube stands.

The exact distance between the end of the wire and brass dish must be found by trials; that which generally answered in my experiments was about 1-20th of an inch. If it be properly arranged, gas will be produced at each discharge.

The Leyden jar used with this apparatus must contain about 150 square inches of coating.

The distance between the insulated ball and the prime conductor, at which the experiment succeeded, was commonly about half an inch.

If experiments be proposed in which electrical discharges must be passed through water or other fluids, for even a much longer time than was consumed in performing those referred to or related in this paper, it may be an object to employ the wind, or perhaps the power of a horse, to turn the electrical machines, the expense of labourers being considerable.

SECTION II.

From my journal of the numerous experiments made during the course of nearly two years, I shall select those which will serve to explain the nature of the process, and shew the power of the plate electrical machines; and I shall particularly relate those experiments which afforded the most useful results concerning the nature of the gas obtained.

1. With interrupted Discharges.

Experiment A. About 1600 of these discharges by means of a 34-inch single plate electrical machine, in nearly three hours, produced from New River water taken from the cistern, and which had not been freed from air by the air pump or boiling, a column of gas, 2-3rds of an inch in length, and 1-9th of an inch wide. On passing through this gas, between the two wires of the tube in which it was produced, a single electrical spark, its bulk was instantly diminished to 2-3rds. In other experiments the bulk of gas was only diminished to about one half; and the result was the same with distilled water.

B. The experiment A being repeated several times with distilled and New River water, freed from air by the air pump or long boiling, the quantity of gas just mentioned was obtained in about four

hours.

On passing an electric spark through this gas in the situation above mentioned, its bulk was instantly diminished, in some cases 15-16ths, and in others 19-20ths.

C. 1600 interrupted discharges, by means of a 32-inch plate machine, produced from New River water and distilled water, freed from their air by the air pump, a column of gas about 3-4ths of an inch in length, and 1-9th of an inch in diameter, in the space of three hours. It was reduced in bulk 19-20ths by passing through it a single electrical spark.

D. 500 revolutions of the 32-inch plate machine, in three quarters of an hour, produced 600 interrupted discharges in river water freed from air by the air pump, by which a column of gas, half an inch in length and 1-10th of an inch in diameter, was obtained. It was diminished as usual by an electric spark 19-20ths of its bulk.

E. Nearly four days' incessant labour with the 32-inch machine, produced only 56,5488 cubes of gas, of 1-10th of an inch each, on account of the usual accidents during the process. The air had been exhausted by setting the water under the receiver of the air pump.

F. It was found that 6000 interrupted discharges produced about three inches in length of gas, measured in a tube 3-20ths of an inch in width, from water out of which its air had been drawn by the air pump.

G. It appeared from many experiments, that the same unboiled water, or water from which the air had not been exhausted by the air pump, which had repeatedly yielded gas by passing through it electrical discharges, always left a residue of gas which the electrical spark did not diminish; and this residue was in nearly the same quantity after six or seven experiments, each of which afforded a column of gas half an inch in length and 1-9th of an inch in diameter, as was left on passing the electric spark through the gas afforded by the third or fourth experiment.

Hence it seems that water is discompounded by the electric discharge, before the whole of the common or atmospherical air is detached from the water by merely the impulse of each discharge. Yet I think it probable that, after the discharges have been passed through the same water for a certain time, the whole of the air contained in water will be expelled, and no gas be produced, but that compounded by means of the electric fire from water; in which case, supposing the gas so produced to be at least merely hydrogen and oxygen gas, it will totally disappear on passing through it an electrical spark. But I have never been able to determine this point, because the tubes were always broken after obtaining a few products, or long before it could reasonably be supposed the whole of the air of the water was expelled from it.

H. To the gas obtained in the experiment E was added, over water, an equal bulk of almost pure nitrous gas. Fumes of nitrous acid appeared, and the gas examined was reduced almost 1-3rd of its bulk. A small bubble more of nitrous gas being let up, no further diminution took place. To this residue was added half its bulk of oxygen gas, obtained from oxy-muriate of potash.

This mixture of gases having stood several days over well burnt lime and boiled quicksilver, an electric spark was passed through the mixture over quicksilver, by which its bulk was instantly diminished 1-4th; but no moisture could be perceived upon the sides of the tube or on the quicksilver. The failure of the appearance of moisture was imputed to a bit of lime accidentally left in the tube which was burst by the explosion, and dispersed through the tube; or else the quantity of water produced was so small, comparatively with the residuary gas, that the water was dissolved by it in the moment of its composition. For, supposing water to have been compounded, it could not amount to the 1-100th part of a grain, and the residuary gas was at least 2,000 times this bulk.

That a quantity of water can be compounded under the same circumstances as in this experiment, and be apparently dissolved in air, so as to escape observation, even with a lens, was proved by passing an electric spark through a mixture of hydrogen and oxygen gas, well dried by standing over lime

2. With complete or uninterrupted Discharges.

The gas obtained by the first described kind of apparatus for the uninterrupted discharges, always left a residue of at least one-fourth of its bulk, on passing through it the electric spark, even when water was used which had been freed from air by boiling, or the air pump. Nor will this result appear surprising, when it is considered how liable the water in this apparatus is to mix and absorb air during the experiment. However, this method would have been extremely valuable, if the next other method had not been discovered; for gas may be obtained by it with fewer accidents, and much more rapidly, than with the interrupted discharges. The apparatus is also much more easily fitted-up, and is more simple. But I think it unnecessary to particularly relate any experiments, as they afforded the same results as those already described, and as those next to be related.

The following experiments were made with the apparatus :—

Experiment 1. At Oh. 40' P.M., began to produce discharges with a double plate 24-inch machine, in water taken from the cistern; and at 12h. 6′ P.M., of the same day, there had been written down 10,200 discharges, each of which occasioned air to ascend from the bottom of the wire and brass cup. The quantity of air obtained was now apparently about one-fourth of a cubical inch, and it occupied nearly half of the tube, the water in which was by this time very muddy.

After standing till the day following at noon, when the process was again commenced, it did not appear that any of the gas had been absorbed by the water over which it stood.

At 2h. 35' P.M., began to produce discharges, and at 8h. P. M., had passed 6,636; which, together with those of the preceding day, amounted to 16,836. The tube was now 5-8ths full of gas, and there seemed to be almost half a cubical inch; for it was observed that the gas was this day yielded at double the rate it had been the

day before. This was accounted for from the diminished pressure upon the electric fire, by the tube containing gas instead of water.

At this time, namely, at 8h. P.M., I was surprised on the passing of a discharge by a vivid illumination of the whole tube, and a violent commotion within it, with, at the same time, the rushing up of water, instantly to occupy rather more than 5-8ths of the space which had been occupied by gas.

The residue of gas was not diminished further by an electric spark; and to the test of nitrous gas it appeared to be rather worse than atmospherical air, as it consisted of rather less than one part of oxygen, and three parts of nitrogen or azotic gas.

It seemed as if the electrical discharge had kindled the oxygen and hydrogen gas of the decompounded gas, by flying from the bottom of the wire to the brass funnel, so that the fire returned into the tube where it passed through the gas. Or the combustion

might be occasioned by a chain of bubbles reaching from the brass dish to the surface of the water in the tube, which was set on fire in its ascent, and thus produced combustion of the whole of the gas of decompounded water.

That this phenomena was from the combustion here supposed, was in some degree proved, by finding that the mixture of hydrogen gas and atmospherical air, under the same circumstances, was kindled in the same manner.

Exp. 2. With a double plate electrical machine, 24-inches in diameter, and a similar apparatus to that in the last experiment, 14,600 discharges produced at least one-third of a cubical inch of gas. While I was measuring with a pair of compasses the quantity of gas produced, the points of them being in contact with part of the tube occupied by gas, I was again surprised on the passing of a discharge by an illumination of the whole tube, and the rushing up, with considerable commotion, of water, to occupy about twothirds of the space filled with gas.

The residuary air was found, as in the former experiment, to be rather worse than atmospherical air.

It was concluded that the points of the compasses had attracted electrical fire from the wire to the sides of the glass, and thereby kindled the hydrogen and oxygen gas of decompounded water. But to determine this question, I introduced into the same tube a mixture of one measure of oxygen and two measures of hydrogen gas, to occupy nearly the same space in the tube as the gas had occupied, then passing an electrical discharge through it, no combustion was excited; but on passing a discharge while the compasses were in contact with the tube, as just mentioned, an illumination and violent commotion were produced, with the rushing up of water, to leave only 1-8th of the gas as a residue. On repeating this experiment with two measures of atmospherical air, and one of hydrogen gas, combustion could not be excited; nor with one measure of atmospherical air, and two of hydrogen; but on adding to this last mixture one measure of oxygen gas, the electrical discharge produced the phenomena of combustion just mentioned, with the rushing up of water, to occupy about two-thirds of the space which was occupied by the gases.

Exp. 3. Having passed 12,000 discharges through water, with the apparatus of the preceding experiment, and thereby obtained. only one-fifth of a cubical inch of gas; and having observed that the quantity of gas was not greater than when only 8,000 discharges had been passed, and yet bubbles had been seen to be produced on each discharge, as copiously, or more so, by the last 3,000 or 4,000 discharges, as before, I began to suspect that part of the gas had been destroyed during the process, or had been absorbed. While I was considering how to account for this disappearance of gas, and was at the same time looking at the tube through which the discharges were passing, I observed one of them to be attended