the former, and consequently the diameters of their coatings nearly the same. The computed charges of A, B and C were intended to be three times as great as those of K, L and M, and consequently the diameter of their coatings about three times as great, and the computed charge of H was intended to be three times as great as that of A. By some mistake, however, the coatings of K, L and M were made rather too small, but the error is very trifling.

317] My first trials with these plates were to examine whether the charge of the three plates D, E, and F together was sensibly less when they were placed close together than when they were placed at 6 inches distance from each other, that is at as great a distance as my machine would allow of. I could not perceive any difference. This is conformable to the theory, as is shown in [Art. 185]. I chose to make the experiment with these three plates, as the difference should be more sensible with them than with any of the others.

318] Secondly. I compared together each of the plates D, E and F. I could not perceive any sensible difference in their charges*.

Thirdly. The charge of the plate K was found to exceed that of the three plates D, E and F together in the proportion of 1·016 to 1. The charge of L was not sensibly different from that of K, and that of M very little different.

Fourth. The charge of each of the plates A, B and C was to that of the three plates L, K and M together, as 0·905 to 1.

Fifth. The charge of H was equal to that of the three plates A, B and C together.

Therefore the charges of D, K, A and H were to each other as 1, 305, 8.28 and 249+.

319] It appears, therefore, that the proportion which the charge of K bears to that of D, and which H bears to that of A, is very nearly the same as that of their computed charges, but the proportion which the charge of A hears to that of K is near less than it ought to be.

part

This in all probability proceeds from the effect of the instantaneous spreading of the electricity bearing a greater proportion to

*

[See Art. 489, Feb. 4, 1772.]

[See also Arts. 656 to 658.]

the whole in the plate K than it does in A, the diameter of whose coating is near three times as great.

320] In order to form some judgment, if possible, how great the effect of this instantaneous spreading of the electricity was, I took off the coatings from the plates A and B*, and put on others of just the same area in the form of a rectangular parallelogram (that of A was 6414 long and 5310 broad, and that of C 6.398 long and 5-201 broad), and compared their charges with that of the plate B, whose charge, as was before said, was just the same as those of these two plates before their coatings were altered.

321] I then took off these coatings†, and on A I put a square coating 6.388 each way with slits cut in it, as in Fig. 23, each broad, so as to divide it into 9 smaller squares, each 1.863 inches

each way. The narrow communications marked in the figure between these squares were each of an inch broad.

On C I put an oblong coating 6:377 long and 6:343 broad, with four parallel slits cut in it, as in Fig. 24, each broad, the narrow space left between these slits and the outside being broad. Having done this, I compared their charges with that of the plate B as before.

It must be observed that the area of these slit coatings was somewhat less than that of the circular or oblong ones, but their whole circumference, including the circumference of the slits, is more than three times as great as that of the circular or oblong ones, so that the surface of glass charged by means of the instantaneous spreading of the electricity was more than three times as great in these coatings as the former, and consequently the quantity of that surface may be determined thereby, supposing that, if it was not for the spreading of the electricity on the surface, the charge of a coated plate would be the same whatever shape its coating is of, provided the area of the coated surface is given.

322. In order to find whether the electricity spread to the same distance upon thin glass as thick, I also took off the coatings from the plate H, and in its room put on first a square coating 6.03 inches each way, and then an oblong one 6708 long and 6.514 broad, with four slits in it, as in Fig. 24, each broad, and ascertained the proportion which its charge with each of these coatings bore to that with the circular coating by comparing it with another plate, the proportion of whose charge to that of the circular coating I had before ascertained*.

323] It appeared from these experiments that if we suppose the electricity to spread instantaneously about 07 of an inch on the thick glass plates such as A and C, and about 09 on the thin ones, not only the charges of A, C and H with the three different coatings, but also the charges of all the plates will agree very well with the theory, as will appear by the following table; whereas, if we suppose that the electricity does not spread sensibly on the surface of the glass, the charge of the plate H with the slit coating would be greater in proportion to its charge with the circular or oblong coating than it ought to be in the ratio of 7 to 6, and the error in the plates A and C would not be much less.

* [Arts. 659–663.]

By the observed charge in the foregoing table, I mean only the proportion which the observed charges bore to each other, not the real observed charges. [See Art. 671.]

326] From the circumstance of the light mentioned in [Art. 307], it appears plainly that the electricity does actually spread instantaneously to a small distance on the surface, and from the rings of dirt taken notice of in Art. 308 it seems likely that the distance to which it spreads is not very different from what we have here supposed; moreover, if the distance to which the electricity spreads is such as we have supposed, the charges of all these plates bear very exactly the same proportion to each other that they ought to do by theory, whereas if the distance to which the electricity spreads is different from that here assigned, and consequently the proportion of the charges of different plates to each other different from that furnished by theory, it seems very strange that their charges should all have happened to agree with computation, notwithstanding that their thickness and the size and shape

of their coatings are so very different. I think therefore that we may fairly infer both that the distance here assigned to the spreading of the electricity is right, and that, if it was not for this spreading of the electricity, the charge of any plate of glass would be as the square of the radius of the circle equal in area to the coated surface divided by twice the thickness of the glass, that is, that the actual charges are in proportion to the computed ones.

327] Though it seems likely from these experiments that the electricity spreads further on the surface of thin glass than it does on thick, yet I can not be sure that it does, as the difference observed is not greater than what might proceed from the error of the experiment. However, as there seems nothing improbable in the supposition, I shall suppose in the following pages that it does really do so.

of an inch

328] When I say that the electricity spreads on the surface of the glass, I mean that the quantity of electricity thereby spread on the uncoated part of the glass is the same that it would be if it actually spread to that distance, and if all that part of the glass which it spread over was charged in the same degree as the coated part, and consequently that the charge of the plate is the same as if the size of the coating was increased by a ring drawn round it 07 of an inch broad, and that the electricity was prevented from spreading any further. But I would by no means be understood to mean that no part of the electricity spreads to a greater distance than that, as it seems very likely that it does so, but that the part furthest from the coating is less charged with electricity than that nearest to it.

329] What is said above must be understood of the distance to which the electricity spreads with that degree of strength which I commonly made use of in my experiments, but I also made some trials with the plates A and C to determine to what distance it would spread with two other degrees of electricity.

If a jar with Lane's electrometer fixed to it was charged to the higher degree, it would discharge itself when the knobs of the electrometer were at 053 inches distance; when it was charged to the lower degree, it discharged itself when they were at about half that distance, or at 027 of an inch; and when it was charged to

[Art. 540, Feb. 16, 1773.]