from its surface; nor will the action of the latter be much ac celerated, though the charged air be removed from around it by a pair of bellows. The action of the flame and the point are different; the former never accumulates electricity enough to overcome the pressure of the atmosphere, but radiates the fluid either to or between the molecules of air; the latter does accumulate electricity enough to push back the molecules immediately opposed to its point, and the surrounding air rushing into the minute vacuum thus formed, creates that current which is always felt to proceed from such a point. As the electricity gets weaker, this vacuum contracts, and no points being perfectly sharp, there will be an intensity of electricity, which is balanced by the pressure of the air even when accumulated at the point of the finest needle. The action of a point depends chiefly on this vacuum, through which the electricity radiates, as through a crack in the vase that surrounds it: but the lamp is surrounded by an atmosphere of electrified particles of air, and its action will be impeded and stopped, unless these are removed; when abstracting electricity from a distant body, the effect of a current towards this last will be still more powerful, since it then forms a medium of communication between the two, and this circumstance forms one of the greatest difficulties which we have to encounter in such experiments.

I shall conclude this part of the subject with some remarks upon the analogous cases which occur in phenomena connected with electricity. It has been long remarked, that all bodies when nearly red-hot become good conductors; those which soften before this temperature become so from their change of state; but glass, a very perfect non-conductor, becomes a conductor at a lower temperature than that at which it softens, and the same is probably true of other bodies: if such is the case, we must attribute the effect either to bodies at that temperature losing their attraction for electricity, or, which is more probable, to the radiating force impressed by the heat. The magnetic experiments which I before mentioned are still more striking; the magnetic fluid, whether in the form which we meet it in the loadstone, or under that modification which forms the electro-magnetic fluid, is possessed of so little mobility, or meets such obstruction in flowing through bodies, that

Iron was

there are very few that will serve as conductors to it. long supposed to be the only material in which the magnetic fluid was developed; and the electro-magnetic fluid is so feeble in its motive powers, that when generated by a battery of a few large plates, it cannot be made to flow through water, and a single thickness of silk is sufficient to arrest its progress. Yet when bars of iron are heated white-hot, the magnetism induced by the earth seems to pervade it with such rapidity, that it ceases to act on the compass. It is true, that we know very little of the real cause of this phenomenon; but when we find exactly the same temperature to produce the radiation of electricity, and of the intermediate galvanic fluid, it seems probable that the effects must be the result of similar causes. The radiation of the galvanic fluid may also be doubted, but the circumstances are so precisely similar to those which occur in electricity, that I consider them as quite satisfactory. If the wires of a battery of large surface be furnished with charcoal points, and brought within the tenth of an inch asunder, no effect is produced; but when they are separated by an interval which does not exceed one fiftieth of an inch, a bright spark appears, the charcoal points become intensely ignited, and a stream of fire commences flowing from one point to the other. The points may now be separated many inches asunder, and the fluid will continue to flow in a stream of the brightest light. From the positive point it parts in a stream resembling that which issues from a pointed wire on a conductor electrified positively, whilst the negative fluid presents that appearance of a star, so well known as characteristic of electricity flowing into a negative conductor. These phenomena are precisely such as we might expect: whilst the charcoal is not ignited, the galvanic fluid is incapable of penetrating a plate of air more than the fiftieth of an inch in thickness. Brought within this limit, it overcomes the obstacle opposed to its passage, and the immense quantity of fluid which then begins to circulate, raises the temperature of the charcoal till it ignites, and the fluid obtains from this elevation of temperature, a power of radiation which suffices to penetrate a plate of air several inches thick*. If this view of the subject be correct, the phenomena should take

* The effect ought, theoretically, to be increased in a vacuum; and we are assured by BIOT, that such is the case. Precis. I. 648.

place more readily when the charcoal is previously ignited.no have not made any satisfactory experiments upon this subject. The fact that light is thrown off abundantly at the radiating temperature is a singular coincidence, and when united with the known fact that few of the solar phosphori act till they are warmed, and that some of them which give out blue or yellow light when nearly cold, give off the red rays when heated, would seem to point out an analogy between the radiation of light and that of electricity.

With respect to the latter, the preceding pages contain, I believe, the first inquiry into the subject; and much must be done before we can consider the effect of caloric upon elec+ tricity as fully developed. My own views will be answered, if I succeed in attracting the attention of electricians to the subject, which promises to afford a favourable method of inquiring into the laws which regulate the motions of the electric fluids, a branch of the science far more imperfect than that which treats of their equilibrium. In concluding, I shall make a few remarks upon this last, as it is exhibited in the works of PoISSON and BIOT; the former of whom has treated the subject more profoundly, and the latter more methodically, than any preceding writers.

These philosophers set out from the law which COULOMB demonstrates, that the particles of the electric fluids repel each other with a force that varies inversely as the square of the distance; and assuming that conducting bodies have no action on electricity, they conclude that the latter will be spread over their surface only, where, if the body be a sphere, it will form a shell of uniform thickness; pressed on its external surface by the atmosphere which prevents the dissipation of the fluid

But this conclusion involves several hypotheses: in the first place it is necessary that the law of repulsion between the particles of the electric fluid should differ from the inverse square of the distance when the particles are brought very near together; for were that law to hold, neither would a spherical shell of fluid be in equilibrium, nor would it be retained by the pressure of the air. With such an arrangement, the interior particles of the shell could not be acted on, and consequently they would continue to approach those which were without them until they actually touched; and, strange as this result may ap

pear, it was adopted by Cavendish in his theory.

BIOT, in the

Traite de Physique, has not remarked this difficulty, and consequently, the explanation which he has given after Poisson, of the equilibrium of the electric fluids, is not sufficient. In a subsequent work, however, he has observed, that the absence of electricity in the interior of bodies leads to the important result, that the electric fluid is incompressible, and in other instances speaks of it as flowing after the manner of liquids; such comparisons are probably merely inaccurate modes of expression. If the phenomena of electricity be explained by the intervention of peculiar fluids, these must be assumed as elastic, but when compressed within certain limits, increasing the law of their repulsion to a high power of the inverse distance. The chief difference between fluids so constituted, and such elastic fluids as the gases, consists in the latter diffusing themselves uniformly, whilst the former everywhere press to their boundaries, and if not free, almost immediately acquire a uniform velocity of radiation.

A second hypothesis, which is required to explain the equilibrium of the electric fluids at the surface of conductors is, that the atmosphere immediately surrounding the conductor, becomes highly charged with the same species of electricity, which, by its repulsion, retains that on the surface for the air of itself cannot retain it there, having no repulsion, but rather an attraction for each of the electricities; as is shewn by many experiments: it is impossible to work an electric machine for any length of time without charging the air of the apartment with electricity; if a screen of tin foil connected with an electrometer (fig.4), be carried into the room, it will immediately render sensible the electricity which is diffused there, and which must be adhering to the particles of air, and forming minute atmospheres of electricity round them. Thus, when a conductor is electrified, the electricity radiates to the surface, and a small portion of it passing on, attaches itself to the immediately contiguous particles of air, where it is retained by the attraction of the latter, and by its own repulsion prevents the further escape of the fluid. The quantity so accumulated in the adjoining particles of air need be very small to preserve the equilibrium. If we assume the particles of air and of the electric

fluids to be at the same distance under equal pressures, it would be sufficient for one particle of the former to attach itself to one of the latter; hence there is no proof from the experiment of FRANKLIN before mentioned, that such electrified atmospheres do not exist.

The thickness of the shell of fluid which can be retained by a given pressure, affords an excellent means of comparing differ ent electrometers, or rather furnishes a standard by which to compare them. Ift is the thickness of the shell, m a constant co-efficient, and h the height of the barometer, we have

t = m√h

or, assuming t = 1 when h = 30 inches

t = 182 √h

An electric machine in good order will readily charge a conductor high enough for the thickness of the shell of fluid on a ball an inch in diameter, and projecting some inches from the conductor, to exceed this limit, and consequently to continue discharging itself into the air.

With the suppositions which I have mentioned, and which there appears every reason to believe correct, the theory of the equilibrium of electricity as improved by POISSON from the prior theories of ÆPINUS and CAVENDISH, leaves little to be wished for on this part of the subject, and may readily be made to embrace both the equilibrium of the magnetic fluids and that of the electricity disturbed by the action of the electromotive apparatus. The first of these applications I made in 1820, in a couple of papers published in the 55th volume of the "Philosophical Magazine;" where, proceeding upon the theory which Potsson had applied with so much success to the distribution of electricity on spheres and ellipsoids, I deduced, by a very easy investigation, the distribution of the magnetic fluids, on similar solids when subject to the action of the earth. The result, which was wholly theoretic, agreed with all the experiments of Mr. BARLOW, and seemed to leave no doubt respecting the truth of the principles. Examining more nearly however, we found an error of one-sixteenth in the constant co-efficient by which the several expressions for the deviation were multiplied. This minute error arises from the resistance which iron offers to the motion of the magnetic fluids; Mr. BARLOW,