chemical nomenclature; but I have heard of many persons, and some of whom were experienced electricians and chemists, who made the attempt.

Since Mr. Cuthbertson came to reside in Londou, I have learned from him the circumstances requisite to the success of the experiment; and I have received from him also very great assistance in continuing a process with the objects I had in view, the tediousness and even difficulties of which can only be conceived by those who have been engaged in the same pursuit.

I am very sensible that it would be unnecessary for me to explain the importance of a process which may at last afford demonstration of the composition of water, by the fullest and unequivocal evidence of its analysis and synthesis; a demonstration which no other single process but the present promises to afford.

I

propose therefore in this

paper:

1. To give such a description of the experiment of rendering water into gas by electric discharges, as shall enable any person who is versed in pneumatic chemistry, and acquainted with the theory and practice of electricity, to repeat it with success. By this description, also, I apprehend I shall make known more generally the very elegant, and frequently most satisfactory mode of decompounding and compounding bodies, by means of the fire of the electric discharge.

2. It is proposed to relate the additional evidence which I have already obtained from this process, concerning the composition of water. For although it seems most probable that water is decompounded in Mr. Van Troostwyk's experiment, it must be confessed that it does not make appear a single unequivocal and decisive property of hydrogen and oxygen in the gas produced. The disappearance of this gas by combustion, or in some other way, instantly on passing through it an electric spark, it is true, is a property known only to belong to the mixture of oxygen and hydrogen gas; but it is well ascertained, that things of totally different species may agree in one or more properties. And there is at least a possibility, that electric discharges may produce various other kinds of gases, in water, beside hydrogen and oxygen from decompounded water; and which may have the property of instantly disappearing on passing through them an electric spark.

3. I shall attempt to resolve the phenomena of the process into a general law of the action of fire, or of the joint action of caloric and light.

SECTION I.

Of the Manner of Conducting the Process.

Electric discharges may be employed in two different manners to decompound water. One of these is by what has been termed the interrupted explosion; which was the method, although not so explained, of Mr. Van Troostwyk. And the other method is by means

of the uninterrupted or complete explosion; for which there are two different kinds of apparatus. These were invented by Mr. Cuthbertson in the course of my investigation of this subject.

To succeed by the method of the interrupted explosion, the following are the necessary parts of the apparatus to be used, and the circumstances to be attended to :

1. The electrical machine must possess sufficient power.—I do not think any cylindrical machine can be made to answer in this process if a large quantity of gas be required; because they cannot be made to act with due regularity, constancy and force. Inequality of the surface of the cylinder is unavoidable, which causes undulation. A cylindrical machine never continues in full force above five or six minutes, without fresh amalgam. Hence, from the repeated amalgamization, the discharges will be so variable that the tubes must be frequently broken.

I used plate machines of a peculiar construction by Mr. Cuthbertson. These machines do not require fresh amalgamization oftener than once in eight hours, and they possess superior powers of acting, in point of regularity, force, and duration. A plate of twenty-four inches in diameter, with a jar containing one hundred and fifty square inches of coating, afforded an adequate discharge every second or third revolution, for several hours; and for a still longer time every third or fourth revolution, with one application of amalgam. A thirty-one inch plate machine afforded a due discharge at first every revolution, and afterwards every second revolution for many hours, with one application of amalgam. The most useful and expeditious machine was that with two plates, each twenty-four inches in diameter, and similar to that of Teyler. It produced twenty-five discharges every fifteen revolutions for an hour or two; and for four or five hours longer a discharge was produced by less than two revolutions, with one amalgamization.

2. The Leyden jar must have a sufficient quantity of coated surface; without which the discharge will not be sufficiently powerful to produce the gas required. The proper quantity, as found by experience, was about one hundred and fifty or one hundred and sixty square inches, with an usual proportional prime conductor.

3. The distance between the insulated ball, and the prime conductor, must always be less than the distance between the extremities of the wires.-Not the least notice of this circumstance has been taken; yet without attention to it the experiment can never succeed, or only for a very short time. Accordingly as the distance between the extremities of the wires within the tube answered hest when it was 5-8ths or 7-8ths of an inch, the distance between the insulated ball and prime conductor was seldom more, but frequently less, than 5-8ths or 6-8ths of an inch. The eye must be kept upon the sparks within the tube, and by practice a person may become a judge of their force by their vividness; which will direct him to bring the receiving ball nearer to the prime conductor, when there appears

danger of the tube being broken; and on the contrary, to remove them to a greater distance from one another when the sparks do not produce gas duly from the water. When the discharge is of the most productive force, both ends of the wire within the tube will be illuminated by a spark; but, when it is weaker, one end only of the wire will be illuminated; and when this is the case, there is no risk of the tube being fractured, but gas will rise from the end of one wire only instead of two.

4. The extremities of the upper and under wire within the tube must be at a certain distance from one another.-If they be too near one another, the points of them will not be illuminated; and provided the insulated ball be as near to the prime conductor as the two wires are to one another, the tube will be broken, because there will be a complete explosion. But if the wires be at too great a distance from one another, the electric fluid of the discharge will be so diffused through the water that no gas will be produced. If the Leyden jar contain, as above stated, 150 or 160 square inches of coated surface; and the ball of the prime conductor and of the insulated ball be about three inches in diameter, the distance between the wires which generally answered best, was about 5-8ths or 7-8ths of an inch, as above said. The narrower the bore of the tube, the greater may be the distance between the two wires; accordingly, the distance may be one inch with a tube 1-14th of an inch wide.

5. The upper wire fixed into the closed extremity of the tubes must be of a proper length and thickness.-If this wire be too long, either the discharge will not be carried through to the end of the lower wire in sufficient quantity to produce gas; or, if it be in sufficient quantity to produce gas, the tube will be fractured. The smaller the diameter of the tube, the longer may be the upper wire for a reason to be given under the next head. I generally found that the discharge requisite to produce gas fractured the tube if the upper wire was more than 6-8ths or 7-8ths of an inch in length, within a tube of more than 1-8th of an inch in diameter. But with very narrow tubes, such as those of 1-16th of an inch in diameter, I frequently succeeded when the upper wire was 14 of an inch in length. It is obvious, that the shorter the upper wire the more readily will gas be produced: the process, however, will be rendered still more tedious in those cases in which a quantity of gas is to be collected in a reservoir for examination; on account of the time consumed in transferring such small parcels of gas by each experiment.

The diameter of the upper wire cannot, perhaps, be too small; for the greater its superficies, the more electric fluid will be parted with to the surrounding water. Hence platina wire of the finest sort, as that 1-240th of an inch in diameter, may be used with superior advantages. This sort of wire also cannot be melted while it is soldering to the glass, which can hardly be prevented with fine wire of other metals. However, I found that copper, brass, or gilt wire, of about 1-80 or 1-100th of an inch in diameter, could be soldered to the glass, and answered perfectly. I did not find that any of the

metal wires were affected by the discharge; but iron or steel ones are not proper, on account of their being so soon oxidized by the water, and consequently extricating from it hydrogen gas. I did not find any advantage from using several small wires twisted together, but separated at the end within the tube; for gas was extricated generally at the point of one of them only, namely the undermost. I think care should be taken to fix the upper wire so that it be in the middle of the tube, as in that case the tube will be less liable to be broken. As to the under wire, the diameter of it seems to be of little importance; for, if it be a thick one, as much gas will be extricated as if it were a small one; because the electric discharge will take the first point of the surface of the wire where the gas is produced.

6. The tubes must be of a proper length and diameter.-If they be shorter than seven inches, the discharge will be liable to pass over the outside; and if they be longer than twelve inches, they will be of an inconvenient length. I found the most convenient length to be from nine to ten inches, exclusive of the curved part. The curved part was found very useful in preventing air ascending, which was accidentally let into the tube, by which the product of gas from the experiment would have been contaminated. Such curved extremities were however less convenient than straight ones, on account of the greater difficulty of transferring from the former than from the latter.

The diameter of the tubes should not be more than 1-8th or less than 1-12th of an inch. At least in my experiments these tubes answered best. If they were wider, the discharge requisite to produce gas broke the tube. If the tubes be narrower than just mentioned, the experimenter will find it difficult and tedious to transfer the gas through the curved part into a reservoir, in those cases in which a large quantity is wanted for examination. Where however the object is merely to show the production of gas by means of the electrical discharge passed through water between two wires, and the instant disappearance of gas so produced by passing through it an electric spark, the narrowest tubes are most eligible, as with them the experiment can be made in a shorter time. But there was one inconvenience experienced from very narrow tubes, namely, the bubbles of gas were very apt to hang near the end of the upper wire instead of ascending; and they were apt to form large bubbles with water between each; in which situation the discharge frequently fractured the tube, or made gas to disappear by combustion in the course of the experiment. To possess at once the advantages of the narrow and wide tubes, about three to six inches of narrow tube were joined by fusion to a bottom part of a wider tube; into the curve of which tube the gas produced was let up from time to time; so that this part frequently contained the products of ten or more experiments before the gas was transferred into the reservoir. The vast number of tubes which were broken in this experiment induced me to try various different kinds of them. I experienced no advan

tages from annealing tubes; but their being seemingly rendered more brittle and harder by this treatment was an effect which I least of all expected. Tubes with thin sides answered just as well as with thick ones. Bohemian green glass tubes were found excellent for this experiment, as they were much less apt to be cracked by the discharges than any kind of English glass. To save time and trouble in so frequently letting out gas produced in very narrow tubes, a small bulb was made at the sealed end. In this case, however, as the upper wire must be shorter than other narrow tubes, it was more difficult to regulate the explosion.

Although common atmospherical air is an electric, and water is a conductor of electricity, it appears that the discharge passes with more resistance from wire to wire through water in the above experiment, than through common air under otherwise the same circumstances. The reason of which is this: air being an elastic and very rare fluid, it more readily gives way to the electric discharge than water; and it can therefore pass through a longer and thicker column of air between two wires, without breaking the glass tubes, than it can through water. For although water is a conductor, yet in a very small quantity it is a very indifferent one; so that its density and defect of elasticity more than compensate for its conducting power. Hence also, and on account of the conducting power of water, the reason of the upper wire in this experiment being shorter in proportion as the tube is wider; and on the same account will be seen the reason of the advantages of a small upper wire over thicker ones.

It will be necessary to add, that the tubes with curved extremities can only be filled by setting them in water under a receiver, and exhausting the air from the receiver, tubes, and water; then, by letting in the air again, the water will be forced up into the tubes. Sometimes, however, I have filled the tubes by setting them in Papin's digester.

These are the directions for making the experiment; but the rationale of it cannot be understood unless the nature of the interrupted explosion be explained; because I believe books on electricity do not contain the necessary information. It must be considered, in the above experiment, that if in place of water the tubes be filled with air, the whole of the charge of the Leyden jar will pass, at each explosion, from the upper to the under wire, and no interruption in the discharge will happen; but if they are filled with water, then an interrupted discharge may be caused; by which is meant that a part of the charge only passes at each explosion through the water, from wire to wire, and with much diminished velocity. The residuary electricity in the Leyden jar is nearly one half; as may be accurately demonstrated from the difference in point of density, elasticity, and conducting power of the medium of water and air, as already observed. It must be added, that although water in large quantity is a good conductor, and air is not, yet water being here in very small quantity, it proves a