* ན།

This hypothesis agrees with the experiments already detailed, but it remains for us to inquire concerning the mode in which flame impresses on electricity a velocity of radiation: which will connect our subject with the laws that regulate the equilibrium of electricity when developed on conducting surfaces. CAVENDISH established these laws nearly as they are now received, but his mathematical acquirements were scarcely adequate to encounter a subject of such extraordinary difficulty, whilst by adopting the hypothesis of FRANKLIN, he was led to a theory less perspicuous and simple than would have resulted from that of DUFAY. POISSON aided by LAPLACE, and conversant with the most refined inventions of the modern analysis, has advanced further in this intricate subject, and has shewn the exact coincidence of theory and experiment,,, The same coincidence results from either hypothesis, but not with equal simplicity and clearness. The advocates for a single fluid have relied chiefly on the direction of the spark; asserting that it always moves from positive to negative. Were not this assertion supported by names of the first respectability, and nearly universally admitted, I should not hesitate to say that it is utterly without foundation; but against such a weight of testimony even the evidence of the senses should be received with caution and distrust. From my own experiments I am led to believe that by varying the mode in which sparks are drawn from the prime conductor, they may be made to move in either direction. If we present a ball to the extremity of the prime conductor where the electricity is strong, and to the

"

1

insulated body, or screwed to it, and after being adjusted to the magnetic meridian, and the electricity communicated, the deviations are measured by the sight K, which traverses on an arc that is independent of the instrument. The support of the balance should be as slender as possible compared with the surface of the body electrified, otherwise the quantity of the fluid which is at first rendered inactive on the surface of the support, and which afterwards returns as the electricity becomes faint, interferes materially with the indications of the instrument.

Other, and very serious difficulties present themselves in the experiments which I have described: the least motion of the air alters the rate of abstraction, whilst the radiation of electricity from the light would rapidly charge the sides of any cover of moderate dimensions, and by saturating the inclosed air with electricity would affect the velocity of of radiation. These difficulties have hitherto prevented me from determining whether any difference exists in the radiation of positive and negative electricity,

-MARCH 1823.

JANM

centre, where it is feeble, the direction of the spark will generally vary. The expelling force, as LAPLACE has shewn, is proportional to the thickness of the shell of fluid at the place where the spark takes place; hence to produce a spark from negative to positive, we must choose a part of the positive conductor where the electricity is weak, but where the action of the whole conductor would be much increased at an inch or two above the surface. In a cylindrical conductor, this may generally be effected by making a ball project some distance from the middle of the conductor, and bringing the negative ball, which should be connected by a chain with the negative conductor, over the former.

But it occurred to me that a much more decisive experiment might be made by disposing the apparatus in such a manner that the negative fluid should be strongly impelled to quit the surface, whilst a body at a short distance from the surface should be slightly acted on. The phenomena of charged plates will readily suggest such a disposition. I have before mentioned that a plate intensely charged will not affect a gold leaf electrometer on the neutralized side; the action of the two surfaces neutralizing each other at nearly all distances—whilst on the charged side the action is incomparably less than that by which the two fluids attract each other. If therefore we can obtain such an arrangement of plates, that instead of an attraction for the negative fluid there shall be a repulsion, we may expect to have the spark from negative to positive. Such n arrangement is found in the electrophorus. When this instrument is placed on its insulated stand, and the cover lifted, the electricity distributes itself as follows. The upper part, a b, of the resin, fig. 7, will be negative, the upper sur

face, mn, of the sole will be positive, and the lower surface p q will be negative. The quantity of fluid p q is nearly equal to that at a b, and consequently the force with which it endeavours to escape, and which we have before remarked is proportional to the thickness of the stratum of fluid, will be considerable. But at the distance of the ball k, the attraction of the plate p q will be nearly neutralized by that of mn; and we shall have a much greater thickness of stratum at p q than on the ball, and accordingly the spark will be found to move to the ball. That the fluid at the lower surface of the sole is negative may be shewn by removing a small part of it on a proof plate, and testing it by a gold leaf electrometer.

It has also been urged by those who suppose the electric fluid to circulate from positive to negative, that when a piece of tin-foil is placed on Lane's discharger, the burr always appears on the negative side. I have frequently made the experiment, and when the foil lay loosely on the negative ball, have always found the reverse to be true. Even the different appearance of the light thrown off by a pointed wire when positively or negatively electrified may be imitated with one fluid by arming the wire alternately with a ball of pith gilt and not gilt; the brush of rays seems produced by a resistance experienced in leaving the pith ball.

Supposing the existence of two fluids established, and it should be observed that one clear instance of the spark moving from negative to positive is in this case equivalent to such a proof, there can be no objection to admit Poisson's theory, which supposes either fluid, when accumulated in a conductor, to move by the mutual repulsion of its particles to the surface of the body, and to be retained there by the pressure of the air; which PorsSON regards as a vase encompassing the electric fluid. But if the air is the immediate cause that retains the electricity, it can only do so by repulsion; whereas all the phenomena of radiation imply an attraction between either of the electric fluids and the particles of atmospheric air. We must therefore conceive the particles of the electric fluid on their first arrival at the surface of a conductor not to be retained there, but to pass on into the surrounding atmosphere; where, attracted by the molecules of air, they quickly form an electric

atmosphere of the same kind as the electricity in the conductor; the repulsion of these electrified molecules will, after a time, equal the expanding power of the fluid in the latter, and will then prevent the further escape of the electricity. We shall examine this hypothesis further presently; but we may remark, that the atmosphere of electricity which is supposed to surround all conducting bodies, must be extremely limited, as is sufficiently evinced by an experiment of Franklin, who electrified a small ball, and after repeatedly whirling it round his head by a silk string, found the ball still electrified. The experiment may be readily repeated with the small plate we used at the commencement of these experiments, and which, when tested by the gold leaf electrometer, shews scarcely any loss of electricity after being swung round many times.

So far we have a clear idea of the process by which electricity is retained on the surface of conductors; and if we admit that the apparent conducting power of a vacuum is, in fact, a radiation, we shall be enabled to trace the successive steps which take place in the action of flame upon electricity. POISSON and BIOT seem to consider the fact, that a perfect vacuum is a conductor as settled; though the well-known experiments of MORGAN are opposed to this conclusion *; but these experiments are inconsistent with nearly all others that have been made on the subject, especially with those detailed by CAVALLO. Indeed, it is difficult to conclude anything decided from MORGAN's experiments; for if the hypothesis which he suggested, and which BIOT has rendered highly probable, be true, that the light observed in electrical experiments is produced by the compression of the air, light ought not to have been seen in a perfect vacuum. And with respect to the charging of the tube, it is to be observed, that the electricity would not flow so rapidly through a vacuum, where it acts by radiation, as it would through a conductor, where it acts by induction, and where the charge on one surface empties itself at once upon the other. Unless, therefore, MORGAN drew off the charge into a condenser, I do not know how he was to prove that the tube was not charged. He does not inform us on

* Treatise on Electricity, vol. ii. p. 90."

what he grounded this conclusion, unless on some vague and imperfectly described facts respecting the breaking of the tubes.

Since electricity radiates in a vacuum, part of the radiation produced at the temperature of combustion may arise from the dilation of the surrounding air by the heat; but this cause does not appear sufficient to produce the whole of the velocity impressed. At the highest temperature which we can assign for the combustion of oil, the contiguous atmosphere cannot be more than four or five times less dense than in its natural state; and although electricity of any intensity would radiate freely in such an atmosphere, we cannot imagine the very weak electricity which would be induced at two feet distance by a ball of an inch and a half diameter, moderately charged, to radiate under such a pressure, at least with sufficient force to penetrate two feet of air. We must, therefore, either conclude that a force of radiation is impressed by the heat on the electric fluid, and that this latter, aided by the rare state of the immediately surrounding medium, penetrates it to some depth; or, that flame is of itself a perfect conductor, and that the electric fluid, in expanding through it, obtains from the continued repulsion of its particles the requisite velocity. In either case the surrounding atmosphere would speedily become charged with electricity, whose repulsion would prevent the radiation from going on. Of this effect we have abundant proof; a lamp placed on an electrified conductor reduces its charge almost instantaneously to a certain strength, after which it acts very slowly, unless aided by a current of air. Thus an electrified body, when heated to the temperature of combustion, differs from other conductors merely in being surrounded by a much more extended atmosphere of electricity. The extent, in the former case, apparently amounting to several feet, whilst in the latter, it cannot exceed a very minute fraction of an inch, as we see by the experiment of FRANKLIN already adduced; and as is more decidedly shewn by the effect of points in discharging electricity. It seems commonly supposed, that a very sharp point will discharge an electrified body immediately, and whilst we confine ourselves to strong charges the fact is true; but if an insulated conductor be furnished with a delicate electrometer, it will be found to retain a small charge for nearly an hour, although a fine sewing needle projects an inch or two