upon the same piece of wood, made strong enough to support them without bending; and we arrange below an equal number of wooden, porcelain, or glass troughs, filled with the conducting liquid. If we wish to make use of the apparatus, we lower the Fig. 68. wooden bar, and each pair is immersed in the corresponding trough. When our experiments are completed, we raise the bar, and the troughs filled with liquid remain prepared for another experiment; but if we think that the liquid needs to be renewed, we empty the troughs and fill them again. Experience has proved, that of all the arrangements at present in use, this is the most simple, convenient, and efficient.†

157. Voltaic piles have been constructed of wires of a single metal, bent in the form of an arc and immersed in vessels filled with a single liquid, under which are placed lamps alternately, by which they are heated. It is said that the mere inequality of temperature thus established between the two branches of the same wire is sufficient to cause an unequal electric charge. I have not had an opportunity of verifying this fact; but it accords fully with the general theory of Volta, if we adopt the explanation which has been given of his experiments.

158. I shall conclude this section with an account of some remarkable phenomena, which are very easily explained upon the principles we have established relative to the influence which the conducting power has on the effects of the voltaic pile. We take two wires, A, Z, one of silver, the other of zinc, and im- Fig. 69. merse them both in a very weak solution of sulphuric or muriatic acid. As long as the two wires do not touch in any point, the zinc dissolves in the acid and disengages hydrogen, while no gas bubbles appear upon the silver wire. But if we bring the two wires. into contact, at their dry extremities, then gas escapes from each of them. This is a very simple phenomenon. Until the contact of the two wires takes place, no derangement is produced in the equilibrium of their natural electricities; but the contact being established, the derangement takes place, and an electric current passes from one to the other through the conducting liquid. Thus far there is nothing irreconcilable with the other phenomena. But what follows seems more extraordinary. If we bend

See note on Hare's Deflagrator and Calorimotor.

the silver wire so that its immersed extremity shall also touch the zinc, or even if it be soldered to it so as to form a continued ring, half silver, and half zinc, the same effects still take place. Nevertheless the zinc wire is then in contact with the silver by its two ends; and according to the theory of Volta, the electromotive actions, exerted at these two ends should counteract each other, and hence it ought to remain in its natural state, which is contrary to observation. But this contradiction disappears, if we give, as before, to the fundamental experiments, their true interpretation, independently of any hypothesis. We have seen that these experiments indicate simply a state of electric equilibrium, which must take place in the contact of the metals with each other, in virtue of which, the silver in contact with the zinc. ought to have, for instance, an excesse of resinous electricity, and the zinc an excess+e of vitreous electricity. But this condition must be satisfied in the ring at the two points of junction of the silver wire with the zinc wire; therefore, the same electric state extends to each wire, on account of the free circulation of the electricity through their substance. Now, if we immerse the ring in a conducting liquid below the points of junction of the two wires, so that a portion of each wire shall be immersed, the two opposite electricities which these portions possess will unite through the conducting liquid; and as they are incessantly renewed at the points of contact of the two metals, there results a continued circulation which ought to produce all the phenomena of a voltaic pair. This case is absoFig. 71. lutely the same as that of a complete plate, of which the upper half is of zinc, the lower of silver, and which is immersed in a conducting liquid below the point of junction. In such a plate, however, the inequality of the electric charge obtains for all the points of the line of contact AB; and it is hence communicated by the conducting power to the whole of each plate; whereas, in the ring, this inequality originally exists only in two points, which are the points of junction of the wires.

159. Soon after the voltaic troughs were constructed the two opposite sides of each trough were formed of the metallic plates themselves, which gave the arrangement represented in fig. 72, in which the letter Z indicates the zinc plates, C the copper, and L the liquid interposed. Now it often happens, that after

having poured the liquid into the troughs, the disengagement of the gases causes it to overflow their upper edges, which are thus moistened so that each zinc plate communicates with the contiguous copper by a moist stratum. Nevertheless, the effects of a general current are still produced, although with less energy than when the edges of the plates are kept dry. This is because the communication thus established by each moist stratum is far from being sufficient to transmit all the electricity developed by the contact of the entire surfaces of the plates. The remainder passes, therefore, through the liquid of the troughs, and, being incessantly renewed by their contact, causes in the usual way a continued electrical current.

Hence, it is evident, that this current would still exist if each piece of zinc were brought in contact with each piece of copper by a better conductor than a simple moist stratum; except that its effects would thus be still more weakened, the quantity of electricity being less. For example, take a single pair of plates Fig. 73. having a large surface like those already described, and immerse

it in an acid mixture; it will cause a red heat in the platina wire 145. which communicates from the copper to the zinc. This being done, interpose somewhere between the two plates of zinc and copper, another small wire, as f', bent in such a way, as to sustain itself between the two plates by its elasticity. It will then be seen, that notwithstanding this communication the platina wire ƒ is still red hot, although in a less degree; or if we choose, we can cause another wire of less diameter to become entirely red. This is because the communication established by the second wire is not sufficient to transmit all the electricity developed by the contract of the whole surface of the plates; and the remainder, passing through the first wire is still sufficient to cause its ignition; in the same way as in traversing the liquid conductor, this remainder produces a disengagement of the disengagement of the gas; if any doubt existed of the division which thus takes place in the electric current between the two wires f, f', we may assure ourselves of it by this circumstance, that the wire f', the simple pressure of which causes a less perfect communication, becomes itself sensibly warm, while the wire ƒ is red hot. These curious experiments were communicated to me by M. Gay-Lussac, to show that the theory of Volta required to be modified, and 1

have been led to refer them to simple conditions of electric equi 131. librium as heretofore explained. Upon the same principles it may be shown also, why the action of a pile, connected by a liquid of great conducting power, does not cease to act when it is immersed entirely in water. It is because the electricity circulates through the water less rapidly than in the interior of the pile; whence it follows, that the communication established by the water cannot entirely discharge it, so long as the interposed liquid remains.

139.

Of Secondary Piles.

160. While all sorts of combinations were tried for the purpose of forming voltaic piles entirely of dry, and consequently unalterable substances, Ritter discovered one, which although incapable of developing electricity by its own action, is nevertheless susceptible of being charged by the voltaic pile in such a way as to acquire, for a moment, all its properties. This is called the secondary pile of Ritter.

To form a just idea of this pile, it is necessary to call to mind an observation of Volta, already mentioned, and which proves the imperfect conducting power of vegetable substances saturated with water. If we insulate an electrical column, of which the upper pole is vitreous and the lower pole resinous, and if we make these two poles communicate by an imperfect conductor, as a strip of paper, for instance, moistened with pure water, each half of this strip will take the electricity of the pole with which it communicates. The upper part will be vitreous, and the lower, resinous. We have remarked that this phenomenon is an evident consequence of the laws by which the electric principle is governed, when distributed over bodies which transmit it imperfectly.

Let us now suppose that we remove this imperfect conductor, with a nonconducting body, as a glass rod; the equilibrium will not be instantaneously established between its two extremities; they will remain for some time, the one vitreous, the other resinous, as when they communicated with the two poles of the pile.

These differences will gradually diminish, according as the opposite electricities re-combine, and their neutralized actions will soon become altogether insensible.

It is to this, precisely, that the fundamental experiment of Ritter refers itself; except that he substitutes for the moist strip of paper a column composed of copper discs and moistened pasteboard alternately. This column is incapable by itself of putting the electricity in motion, at least if we suppose its elements of each kind to be homogeneous in themselves; but it becomes charged by communication with the pile, like the band of moist paper of which we have spoken. Nevertheless there is an essential difference in the two results. It appears that the electricity, when it is feeble, meets with some difficulty in passing from one surface to the other. This seems at least to result from the experiments of Ritter, and perhaps the resistance is produced by the imperceptible stratum of nonconducting air, which adheres to the surfaces of all bodies. The electricity introduced into the column, composed of a single metal, meets therefore with a similar difficulty in passing from the metal to the moist pasteboard; this obstacle increases according as the alternations are more numerous. Thus a pile once charged must lose its electricity very gradually when there is no direct communication between its two poles. But if we establish this communication by a good conductor, the passage of the two electricities, and their combination, taking place rapidly, will cause a discharge, as in the Leyden jar. A new state of equilibrium will follow this effect, in which the repulsive forces of the different plates will be diminished in the ratio of the quantity of electricity which is instantly neutralized. The discharges must therefore be repeated with diminished effects, according to the number of contacts; but they soon cease to be sensible on account of the equal charge which they tend to establish between all parts of the apparatus. In a word, the action of the column depends on this, that it becomes a better or worse conductor, according as its two extremities do or do not communicate with each other.

As to the manner in which the electricity arranges itself in this case, it must be such, that its repulsive force at the surface of each plate, combining with the resistance of the contiguous E. & M.

24