placed sixteen or eighteen inches from a screen of tin foil: the a scre latter was one foot square, and hung by silk threads as in fig. 4. The lower part of the foil was in contact with a gold

leaf electrometer. A chain of ten or twelve feet long led from the wire which supported the candle to the electrophorus. On raising the plate of the latter, the leaves of the electrometer diverged strongly with the same species of electricity, and remained permanently divergent. Some degree of divergence was produced when the candle was not employed; but in this case the leaves collapsed immediately the electrophorus was closed, and the nature of the electricity possessed by the leaves depended on the position of the electrometer, that is, whether the leaves were before or behind the sheet of foil." It could not be doubted, therefore, that a real transfer of electricity took place in the first instance, and that merely induced eléctricity was present in the second.

L

Ex. 11. The last experiment was repeated; but in place of the sheet of foil, a disc of silk of three inches diameter, stretched upon wire, and insulated by a pillar of glass with a

97851

ball and socket, 1102 & 13%

Fig. 5.

On Radiating Electricity.

SAT 199E9 10 (19912

141

(fig. 5) was employed. By turning the socket, this disc could be placed in any position with respect to the candle, and several experiments were made with it in front, behind, on the side, and above the flame. The electricity stopped by the silk was examined by taking the disc to the electrometer; and was found in all the instances to be positive, and nearly of the same intensity; rather less electricity, however, seemed to radiate downward than horizontally.

Ex. 12. As the radiation in these cases was undoubtedly from the light, it seemed fair to infer that such is the case when the combustible is not itself electrified, but is exerting its abstracting influence on other bodies. Were this the case, it seemed to follow that a screen interposed between the two ought to receive electricity from the body in combustion only, and consequently would be in an opposite state to the charged body. An arrangement was accordingly made similar to that in the 9th Ex. (fig. 2), but a screen of glass one foot square, and a quarter of an inch thick was interposed between the charged plate and the condenser: the result was remarkable; immediately the light was placed on the electrometer, the leaves of the latter began, to diverge, and continued to rise and discharge themselves against the guard at intervals of nearly a minute; whilst the ball E fell as slowly as if no light had been employed. When this was actually the case, the same phenomenon was observed; but it was found that the leaves whilst diverging were now charged with negative electricity, whilst, when the light was used, the electricity with which they diverged was positive. This difference was readily explained, and was consistent with the theory adopted. When the electrometer was exposed to the influence of the charged plate, with or without an interposed screen, but without employing a candle, the electricities in it soon became neutralized, and the leaves hung parallel: after a minute or two, the plate having lost positive electricity, the negative electricity in the electrometer was no

longer neutralized; the excess in the guard readily escaped, but the leaves being insulated, the excess was confined in them, and caused them to diverge with negative electricity. That such is the true explanation was rendered manifest by suddenly discharging the plate, when the leaves instantly flashed and discharged themselves. With the candle the effect was reversed, because, in this case, negative electricity radiated from the electrometer; and this radiation produced little or no effect on the charged plate, because it was intercepted by the glass screen. When the latter is very large, or the plate is inclosed in a glass case, I have found both the dissipation, and the effect on the electrometer, to be reduced to little or nothing; and this too is consistent with the hypothesis for the screen will itself become charged with induced electricity, and will neutralize the effect which the charged plate could otherwise have had on the lamp; as may be satisfactorily shewn by charging a plate of glass coated on both sides: so completely does the negative side'neutralize the positive, that if this side be held over a gold leaf electrometer, it would produce no effect on the leaves, although the plate may contain electricity enough for a strong shock. The same action of the induced electricity in the screen which rendered that on the light feeble, interfered with the result which had been anticipated respecting the charge which the screen would take, both by causing the light to radiate less electricity, and by the induced electricity, which does not immediately recover its equilibrium in a non-conductor, mingling its action with that of the radiated electricity. I have since, however, removed this difficulty by interposing the small silk screen, fig. 5, between a lamp, and the prime conductor of an electric machine, whilst the latter was in action; the screen, as anticipated, became charged with negative electricity.

So powerful is the effect of a lamp in abstracting electricity, that a large jar may be readily charged when six inches from the prime conductor of an electric machine, by placing a lamp on the cover of the jar. In this experiment the charge will not increase beyond a certain degree, and if the radiant electricity: be then examined with the small silk screen, it will be found. to be positive on one side of the light, and negative on the other; so that one part of the flame is charging the jar, and

the opposite discharging it. A jar with a foot and a half of coating was charged in this way to ten or fifteen degrees of the quadrant electrometer, when two feet from the prime conductor of a seven-inch cylinder. If the light be inclosed in a metal vessel, leaving only communication enough with the atmosphere to sustain the flame, the jar will still be charged, though to a less intensity; but if the vessel be of glass, the jar will not be charged at all, and the glass envelope will become strongly negative. A jar will not be completely discharged by a light placed on the top of it whilst the air is still, but a current of air facilitates greatly both the discharging and ab. stracting power of combustion.

The phenomena which I have described will be explained by assuming, that a body in a state of combustion has the property of impressing on either of the electric fluids, when flowing through it, a force of radiation which enables them to penetrate many feet of atmospheric air. When a lamp is placed on an electrified body, the electric fluid passes to the light, and is thence radiated into space. But as the effect ceases before the body is completely discharged, we must further assume, that the radiated electricity adheres to the particles of the air, and forms round the lamp an atmosphere of electricity which by its repulsion prevents any further radiation from taking place. Hence the efficacy of a current of air in increasing the dissipating power of the lamp.

When the latter is not placed upon the electrified body, but abstracts electricity from it at a distance, the effect is still produced in a similar way. The electricity on the conductor induces an opposite electricity in the lamp, which, ascending to the flame, is radiated into the surrounding atmosphere, electrifies its particles, and these last, attracted to the charged body, gradually neutralize its electricity. Hence if there is not a free communication through the atmosphere between the lamp and the conductor, the latter will not be deprived of its electricity; but the radiation of induced electricity, from the lamp, will still go on, until the accumulation on the obstacle neutralizes the action of the charged surface. Admitting the truth of this hypothesis, it seems that the abstracting should decrease faster than the fourth power of the distance; for the effect of induction, or the propelling power, we know to decrease as the

square of the distance; and the intensity of the electricity when propelled, were there no resistance from the air, would decrease in the same ratio. I have examined the diminution attentively, and find it to agree with the fourth power of the distance within a certain limit, beyond which it appears to decrease faster. But these experiments require extreme caution when much accuracy is required, and are subject, after remaining regular and consistent for many repetitions, to fluctuations that appear at first altogether arbitrary, but which I have since been led to ascribe to the capacity of the atmosphere for electricity; though it must be confessed that if this be the cause, and I have diligently endeavoured to avoid every other, the electrical state of the atmosphere must vary to an extent and with a rapidity which I am not inclined to admit without further experiments *.

*All the electrometers with which I am acquainted, not excepting COULOMB'S, are imperfect when electricity of much intensity is to be measured; both from their insulating supports taking a charge, and from the glass which is placed over them becoming charged by radiation. COULOMB has in some measure avoided the first by measuring only a small quantity of the electricity to be examined, and the second defect he has likewise attempted to remove by using a very large case; but though his balance is delicate, theoretically elegant, and often the best that can be used in practice, it is cumbrous from its size, tedious in its operation, and only lessens to a certain extent the evils of which I have spoken. For experiments which required less time, and where portability was a convenience, I have made use of the electrometer delineated in Fig. 6, AB

is a delicate magnetic needle two or three inches long, having its ends furnished with small pith balls. M and N are large balls of thin brass attached to the extremities of a wire M m n N, which is fixed to the metal foot of the instrument. The electrometer can be placed on the