Pp with the prime conductor, a communication being made between the outside of the vial A and the ground, and the vial a being made to communicate with the ground by the wire xyz, which rests on Gg, and is suspended from the wire bt by silk strings represented in the figure by dotted lines. When the vials are sufficiently charged, the wire bt is lifted up till xy bears against the bottom of Cc, xy being still suffered to communicate with the ground as before, and the communication between the outside of the vial A and the ground being still preserved. At the same time the wires Ss and Nn are let fall upon Ee and Ff. For the sake of doing this more commodiously I make use of the silk strings represented in the figure by dotted lines and passing over the pulley H. A weight is fastened to the string at w, which is supported while the vials are charging in such manner that the wires Ss and Nn are lifted up so as not to touch Ee and Ff, and the wire bt is suffered to rest on Ee and Cc, and the wire xy on Gg; and when the vials are sufficiently charged the weight is let down, by which means Ss and Nn are suffered to fall down upon Ee and Ff, and the wire bt is lifted up till xy bears against the bottom of Cc.

242] From what has been said it appears that whilst the vials were charging, the outsides of each of them communicated with the ground, and consequently the inside of each vial is overcharged and the outside undercharged. As soon as the vials are charged the communication of each of them with the prime conductor is taken away, and at the same time the communication between the outside of the vial a and the ground is taken away, so that it is intirely insulated, and, immediately after, a communication is made between its inside and the ground, and at the same time the body B is made to communicate with the inside of the vial A, and the trial plate with the outside of the vial a; consequently the body B will be overcharged as it communicates with the overcharged part of the vial A, while the undercharged side communicates with the ground; and the trial plate will be undercharged, as it communicates with the undercharged side of the vial a, while the overcharged side communicates with the ground.

Immediately after this operation is performed the wires Rr and Mm are lifted up, so as to cut off the communications of the

bodies B and T with the vials, and, instantly after, the wires Dd and DS are let down, so as to make a communication between the body B and the trial plate. For the sake of expedition this operation was performed nearly in the same manner as the former, by means of the silk strings passing over the pullies L and l, and represented in the figure by dotted lines. I also employed an assistant to turn the electrical machine and to manage the silk strings passing over the pulley H, while I stood ready near D to perform the last mentioned operation as soon as the wires Ss and Nn were let down, and also to see whether the pith balls separated

or not.

243] From the manner of performing the last mentioned operation it appears that the communication is not made between B and T till after their communication with the vials and all other bodies is cut off; consequently, if the quantity of redundant fluid communicated to B is more than sufficient to saturate the redundant matter in T, they will be overcharged after the communication is made between them, and the pith balls at D will separate positively, but if the redundant fluid in B is not sufficient to saturate the redundant matter in T they will be undercharged, and the pith balls will separate negatively.

244] The balls were made of pith of elder, turned round in a lathe, about one-fifth of an inch in diameter, and were suspended by the finest linen threads that could be procured, about 9 inches long.

245] In making these experiments I did not open the trial plate to such a surface that the pith balls should not separate at all on making the communication between B and T, and assume that for the size which must be given to the trial plate in order that the deficience of fluid in it should be equal to the redundance in B (or for the required surface of the trial plate, as I shall call it for shortness); but I first made the surface of the trial plate such that the deficient fluid therein should exceed the redundant in B, and that the pith balls should separate negatively, just enough for me to be sure they separated: I then diminished the surface of the trial plate till I found, on repeating the experiment, that the pith balls separated positively as much as they before separated negatively, and the mean between these I concluded to be the required surface of the trial plate.

246. This way of making the experiment I found much more accurate than the other, for supposing the required surface of the trial plate to be expressed by the number 16, I found that its surface must be increased to about 20 before I could be certain that the pith balls would separate negatively, and that it must be diminished to about 12 before they would separate positively; whereas I found that increasing its surface from 20 to 21 would make the balls separate sensibly further, and that diminishing its surface from 12 to 11 would have the same effect; so that I could determine the required surface of the trial plate at least four times more exactly by the latter method than by the former. 247] It will be shewn hereafter that the quantity of deficient fluid in the trial plate is in proportion to the square root of its surface; consequently the redundant fluid in B must exceed, or fall short of, the deficient fluid in the trial plate by about 4th part, in order that the balls should separate, and moreover the increasing or diminishing the deficience of fluid in the trial plate by about part will make a sensible difference in the separa

tion of the balls.

*

248] It is plain that this way of finding the required surface of the trial plate is not just, unless the vials are charged equally in both trials, namely, that in which the balls separate positively and that in which they separate negatively; I therefore fastened an electrometer to the wire Pp, at a sufficient distance from the vials, consisting of two paper cylinders about three-quarters of an inch in diameter and one inch in height, suspended by linen threads about eight inches long, and in changing the vials took care always to turn the globet till these cylinders just began to separate.

249] In all the later experiments, however, I made use of a more exact kind of electrometer, consisting of two wheaten straws, Aa and Bb (Fig. 30), eleven inches long, with cork balls A and B at the bottom, each one-third of an inch in diameter, and supported at a and b by fine steel pins bearing on notches in the brass plate C, and turning on these pins as centers. This electrometer was suspended by the piece of brass C from the prime conductor, and a piece of pasteboard, with two black lines drawn upon it, was placed six inches behind the electrometer on a level with the balls, in order to judge of the distance to which the balls separated, the eye being placed before the electrometer at thirty inches distance [Arts. 284, 479, 682.]

+ [Of Nairne's electrical machine, see Art. 563.]

from them (a guide for the eye being placed for that purpose and the electrical machine was turned till the balls appeared even

with those lines. By these means I could judge of the strength of the electricity to a considerable degree of exactness. In order to make the straws conduct the better they were gilt over.

250] In order to estimate what error may arise from the vials being not equally charged in both trials, let the required surface of the trial plate be called 16; then must the surface which must be given to it in order that the balls may separate negatively be 20, or 16 +4, supposing the vials to be charged with the usual degree of strength. Suppose now that in the next trial, in which the balls are to separate positively, the vials are charged stronger than before, in the ratio of a to 1, so that the quantity of redundant fluid in B shall be greater than before, in the ratio of x to 1, and that the deficience in the trial plate should be greater than before in the same ratio, provided its surface remained unaltered; then must the surface which must be given to the trial plate, in

* It is necessary that the eye should always be placed nearly at the same distance from the electrometer, as it is evident that the nearer the eye is placed the further the balls will appear to separate. But as the distance of the balls from the eye is so much greater than their distance from the pasteboard, a small alteration in the distance of the balls either from the eye or the pasteboard will make no sensible alteration in the distance to which the balls appear to separate.