From these circumstances, I am inclined to think that the charge of two plates together is to that of one plate alone as 21.96 to 10:34, and that the charge of the four plates together is to that of one alone as 42:06 to 10:34, and consequently that the charges of the tin circles of 93 inches, 18 inches and 36 inches are to each other as 9-3, 2019 and 43.75*.

338] Though I do not know how to calculate how much the charge of the circles ought to be increased by the attraction of the undercharged ground, yet I think there can be little doubt but that if the charge of the plate of 18 inches is increased in any ratio whatever as that of x to x-18, the charge of the plate of 36 inches will be increased in the ratio of x to x 36, and that of the plate of 9.3 inches in the ratio of x to x-93; therefore if we suppose that the charge of the 18 inch plate is increased in the ratio of 9 to 8, or of 166 to 166-18, the charges of the three plates should be to each other as

which agrees very nearly with experiment, and nearer so than it would have done if we had supposed the charge of the 18 inch plate to have been increased in any other proportion which can be expressed in small numbers.

339] I think we may conclude therefore that the charge of the 12.1 inch globe was increased by the attraction of the undercharged ground nearly in the proportion of 9 to 8, for I think there can be little doubt but that the charge of the globe must be increased thereby in nearly the same ratio as that of the 18 inch plate, and therefore we may conclude that the charge of the plate D is to the charge which the 121 inch globe would receive, if it was placed at a great distance from any over or under-charged matter, nearly in the proportion of 263 to 121, or, in other words, the charge of the plate D is 26-3, which is rather more than eight times greater than it ought to be if the electric fluid did not penetrate into the glass. I shall speak further as to the cause of this in [Art. 349].

340] In order to try the charge of what Æpinus‡ calls a plate of air, I took two flat circular plates of brass, 8 inches in diameter * [Art. 649.] + [Art. 652, and Note 24.] + [Mém. Berl. 1756, p. 119.]

and thick, and placed them on the bars Nn and Pp of the machine (Fig. 20), the two plates being placed one over the other, and kept at a proper distance from each other by three small supports of sealing-wax placed between them, the supports being all of the same height, so that the plates were exactly parallel to each other. Care was also taken to place the plates perpendicularly over each other, or so that the line joining their centers should be perpendicular to their planes.

The lowermost plate communicated with the ground by the wire RS, and the uppermost communicated with Mm by the wire V, just as was done in trying the Leyden vials.

I then found its charge, or the quantity of redundant fluid in the uppermost plate, in the usual manner, by comparing it with the plate D, and found it to be to that of D as*...

341] As I was desirous of trying larger plates than these, and was unwilling to be at the trouble of getting brass plates made, I took two pieces of plate-glass† 11 inches square, and coated each of them on one side with a circular plate of tinfoil 11.5 inches in - diameter, and placed them on the machine as I did the brass plates in the former experiment, with the tinfoil coatings turned towards each other, and kept at the proper distance by supports of sealing-wax as before, care being taken that the tinfoil coatings should be perpendicularly over each other.

For the more easy making a communication between the circular coating of the lower plate and the ground, and between that of the upper plate and the wire Mm, I stuck a piece of tinfoil on the back of each plate, communicating by a narrow slip of the same metal with the circular coatings on the other side.

I then tried the charge as before, the lower plate communicating with the ground and the upper with the wire Mm.

As glass does not conduct electricity, it is plain that the quantity of electric fluid in the pieces of tinfoil will be just the same that it would be if the glass was taken away, and the pieces of tinfoil kept at the same distance as before.

The memoranda I took of that experiment are lost, but to the best of my remembrance the result agreed very well with the following experiment.

+ [Art. 517.]

The distance of the two circular coatings of tinfoil was measured by the same instrument with which I measured the thickness of the plates of glass, and may be depended on to the 1000th or at least to the 500th part of an inch*.

342] In this manner I made the experiment with the plates at four different distances, namely 910, 420, 288 and 256, and when I had made a sufficient number of trials with the plates at each distance, I took off these circular coatings and put on smaller, namely of 6:35 inches diameter, and tried the experiment as before with the plates at 259 inches distance. The result of the experiments is given in the following table:

It is plain that some allowance ought to be made in these trials for the spreading of the electricity on the surface of the glass. In the above table I have supposed it to spread '05 of an inch, but the effect is so small that it is of very little signification whether that allowance is made or not.

344] In my former paper [Art. 134] I expressed a doubt whether the air contained between the two plates in this experiment is overcharged on one side and undercharged on the other, as is the case with the plate of glass in the Leyden vial, or whether the redundant and deficient fluid is lodged only in the plates, and that the air between them serves only to prevent the electricity from running from one plate to the other, but the following experiment shows that the latter opinion is true.

I placed the two brass plates on the machine (Fig. 20), and tried their charge as before, except that, after having charged the plates‡, I immediately lifted up the upper plate by a silk string so as to separate it two or three inches from the lower one, and let it * [See Art. 459, "Bird's instrument," and "dividing machine," Art. 517. Also 594, 595.]

[See Arts. 669, 519.]

[Arts. 511, 516, Dec. 18, 26, 1772.]

down again in its place before I found its charge by making the communication between Bb and Dd and between Aa and Ee.

The way I did this was that as soon as I had let down the wire Cc on Aa and Bb, and thereby charged the plates, I lifted it up again half way so as to take away the communication between Cc and the upper plate &c., but did not lift it quite up, so as to make the communication between Bb and Dd, and between Aa and Ee, till after I had separated the upper plate from the lower, and put it back in its place.

I could not perceive any sensible difference in the charge, whether I lifted up the upper plate in the above-mentioned manner, or whether I tried its charge without lifting it up.

345] It is plain that in lifting up the upper plate from the lower and letting it down again, the greatest part of the air contained between the two plates must be dissipated and mixed with the other air of the room, so that if the air contained between the two plates was overcharged on one side and undercharged on the other, the charge must have been very much diminished by lifting up the upper plate and letting it down again, whereas, as I said before, it was not sensibly diminished.

I think we may conclude, therefore, that redundant and deficient fluid is lodged only in the plates, and that the air between them serves only to prevent the electricity from running from one plate to the other.

346] As this is the case, the charge of these plates ought, according to the theory, to be equal to that of a globe whose diameter equals the square of the radius of the plate or circular coating divided by twice their distance, that is, to their computed charge, provided the electricity is spread uniformly on the surface of the plates, and therefore in reality the numbers in the last column but one ought to be rather greater than in the last but two, and moreover the less the distance of the plates is in proportion to the diameter of the coating, the less should be the proportion in which those numbers differed, and if the distance is infinitely small in proportion to the diameter, the proportion in which those numbers differ, should also be infinitely small.

347] This will appear by inspecting the table to be the case, only it seems from the manner in which the numbers decrease,

that they would never become equal to unity though the distance of the plates was ever so small in respect of their diameter, and I should think, or rather I imagine, would never be less than 1·1, so that it seems as if the charge of a plate of air was rather greater in proportion to that of the globe than it ought to be, and I believe nearly in the proportion of 11 to 10*.

348] The reason of this, I imagine, is as follows. It seems reasonable to conclude from the theory that when a globe or any other shaped body is connected by a wire to a charged Leyden vial, and thereby electrified, the quantity of redundant fluid in the globe will bear a less proportion to that on the positive side of the jar than it would do if they could be connected by a canal of incompressible fluid†, but in all probability when a plate of air is connected in like manner to the Leyden vial, the quantity of redundant fluid on its positive side will bear nearly the same proportion to that in the vial that it would do if they were connected by a canal of incompressible fluid, and consequently the charge of the plate of air in these experiments ought to bear a greater proportion to that of the globe than if they had been connected to the vial by which they were electrified by canals of incompressible fluid.

349] It was said in Art. 339 that the charges of the glass plates were rather more than eight times greater than they ought to be by the theory, if the electric fluid did not penetrate to any sensible depth into the glass. Though this is what I did not expect before I made the experiment, yet it will agree very well with the theory if we suppose that the electricity, instead of entering into the glass to an extremely small depth, as I thought most likely when I wrote the second part of this work‡, is in reality able to enter into the glass to the depth of of the whole thickness of the glass, that is, to such a depth that the space into which it can not penetrate is only of the thickness of the glass, as in that case it is evident that the charge should be as great as it would be if the thickness of the glass was only of its real thickness, and the electricity was unable to penetrate into it at all.

350] There is also a way of accounting for it without suppos

* [Art. 670.]

+ This seems likely from Appendix, Coroll. 5 [Art. 184].

[Refers to Art. 132.]