373] The method I used in making all the plates of the second table was this. I first cast a round plate of the substance, three or four times as thick as I intended it should be, and rather thinner near the edges than in the middle, taking care to cast it as free from air bubbles as I could.

I then heated it between two thick flat plates of brass, till it was become soft, and then pressed it out to the proper thickness by squeezing the plates together with screws. In order to prevent its sticking to the brass plates, I put a piece of thin tinfoil between it and each plate, and I found the tinfoil did not stick to it so fast but what I could get it off without any danger of damaging them.

374] The heat necessary to melt shell lac is so great as to make it froth and boil; which makes it impossible to cast a plate of it free from air bubbles. The plate mentioned in the preceding table was as free from them as I could make it. It contained, however, a great quantity of minute bubbles, but no large ones.

375] Bees wax melts with a heat of about 145°. If it is then heated to a degree rather greater than that of boiling water, it froths very much, and seems to lose a good deal of watery matter, and if it is kept at this heat till it has ceased frothing, it will then bear being heated to a much higher degree without frothing or boiling. Bees wax thus prepared I call dephlegmated.

In order that the plates of dephlegmated bees wax should all be equally so, I dephlegmated some bees wax with a pretty considerable heat, and suffered it to cool and harden, and out of this lump I made all three plates, taking care in casting them not to heat them more than necessary.

I used the same precautions also in casting the plates of a mixture of rosin and bees wax, the proportion of the rosin to the bees wax was forgot to be set down.

What are called in the table the 4th and 5th plate of rosin and bees wax are in reality the same plate as the 3rd, only with a smaller coating.

376] It appears from these experiments, first, that there is a very sensible difference in the charge of plates of the same

dimensions according to the different sort of glass they consist of, the charge of the plates O and Q, which consisted of the greenish foreign plate glass mentioned in [Art. 301] being the greatest in proportion to their computed charge of any, next to them the crown glass, and the flint glass being the least of all.

Secondly. The charge of the Lac plate is much less in proportion to its computed charge than that of any glass plate, and that of a plate of bees wax, or of the mixture of rosin and bees wax still less.

But it must be observed that there is a very considerable difference between the three different plates of dephlegmated bees wax in that respect. The same thing, too, obtains in the mixture of rosin and bees wax*.

377] As the proportion of the real charge to the computed is greater in the thick plates than the thin ones, one might be inclined to think that this was owing to the electricity being not spread uniformly. But as the difference seems to be greater than could well proceed from that cause, I am inclined to think that it must have been partly owing to some difference in the nature of the plates. Perhaps it may have been owing to some of the plates having been less heated, and consequently having suffered a greater degree of compression in pressing out than the others.

378] The piece of ground crown glass mentioned in the first of the foregoing tables was made out of a piece of crown glass about † of an inch thick, and ground down to the thickness mentioned in the table, care being taken by the workman to take away as much from one side as the other, so that the plate consisted only of the middle part of the glass.

My reason for making it was that as there appears to be a considerable difference in the charge of different sorts of glass, it was suspected that there might possibly be a difference between the inside of the piece and the outside, and if there had, it would have affected the justness of the experiments with the ten pieces of glass ground out of the same piece.

But by comparing the charges of the plates of crown glass with those of the two other pieces of crown glass in the table, * [Note 27.]

There are pieces of that thickness sometimes blown for the use of the Opticians.

there does not seem to be any difference which can be depended on with certainty.

The experiment indeed would have been more satisfactory if the piece of ground glass and the pieces with which it was compared had been all made out of the same pot. But as it would have been difficult procuring such pieces, and as I have found very little difference in the specific gravity of different pieces of crown glass, and as I am informed it is all made at the same glass house, I did not take that precaution.

Fig. 29.

A

B

C

D

379] Let two or more flat plates of different non-conducting substances, as Aab B, Bbc C and Ccd D, (Fig. 29) be placed close together and coated in the manner of a single plate with the coatings Ee and Ff. Let the charge of the plate AabB, supposing it placed by itself and coated in the usual manner, be equal to that of a plate of glass whose thickness is A and whose coatings are of the same size as those of AabB.

In like manner let the charge of Bbc C be equal to that of a plate of the same glass whose thickness is equal to B, and let that of CcdD equal that of one whose thickness is C.

E

a

с

Then whichever of the three ways of accounting for the excess of the real charge of glass plates above the computed we prefer, it is a necessary consequence of our theory that the charge of this compound plate AadD should be equal to that of a single plate of glass whose thickness equals A+B+C, and whose coatings are of the same size as Ee and Ff.

380] In like manner if two or more plates of the same kind of glass are placed together and coated as above, the charge of this compound plate should be equal to that of a single plate of the same glass whose thickness is equal to that of all the plates together. This appears from the following experiments to be the case, for

1st. I took the three plates of glass A, B and C*, and laid them on one another, having first taken off their old coatings and coated the outside surfaces as in fig. 29 with circles of tinfoil 6·6 inches in diameter. The charge of this compound plate was found to be to that of the three plates D, E and F together as 944 to 1. The sum of the thicknesses of A, B and C together is 6309, and the computed charge of a plate of that thickness with coatings 6-6 in. diameter is to that of D, E and F together, allowing in the same manner as in [Art. 328] for the instantaneous spreading of the electricity, as 94 to one. So that the charge of this compound plate is exactly the same that it ought to be according to the foregoing rule.

381] 2ndly, I made a plate of a mixture of rosin and bees waxt, about 8 inches square and somewhat more than 12 thick, and coated it with circles 6:61 in. diameter. Its charge was found to be to that of the plates K, D and E together as 56 to 55, and therefore should be equal to that of a plate of glass of the same kind as K whose thickness is 345 and the diameter of whose coatings is the same as those of the rosin plate, namely 661 inches.

This plate was then inclosed between the glass plates B and II, the coatings being first taken off, and the outside surfaces of B and H coated with circles 66 inches in diameter. Its charge was found to be to that of K as 7·56 to 8.

According to the foregoing rule, its charge should be the same as that of a plate of glass of the same kind as B 634 of an inch thick with coatings 66 inches in diameter, and should therefore be to that of K as 7:34 to 8, which is very nearly the same that it was actually found to be.

382] On the charges of such Leyden vials as do not consist of flat plates of glass.

These experiments were made with hollow cylindrical pieces of glass, open at both ends, and coated both within and without with pieces of tinfoil surrounding the cylinder in the form of a ring, the breadth of the ring being everywhere the same, and the inside and outside coatings being of the same breadth, and placed exactly opposite to each other. Only as the inside diameter of the two +[Arts. 548, 678.]

* [Arts. 534, 544, 546, 677.]

[Arts. 552, 679.]

thermometer tubes was too small to admit of being coated in this manner, they were filled with mercury by way of inside coating.

The thickness of the glass was found by suspending the cylinder by one end from a pair of scales with its axis in a vertical position, and the lower part immersed in a vessel of water, and finding the alteration of the weight of the cylinder according as a greater or less portion of it was under water*.

383] The result of the experiments is contained in the following tablet.

The lengths of the coating here set down are the real lengths. But in computing the charges of the white jar and cylinder and the three green cylinders, these lengths were increased on account of the spreading of the electricity according to the same supposition as was used in computing the charges of the flat plates.

But in computing the charges of the thermometer tubes no correction was made, as I was uncertain how much to allow, but as the length of their coatings is so great, this can hardly make any sensible error.

384] It should seem from these experiments as if the proportion of the real to the computed charge was rather less in a cylinder in which the thickness of the glass is of the semidiameter than in one in which it is only, and most likely rather less in that than in a flat plate, but then it seems to be not much less in a cylinder in which the inside diameter is many times less than

* [Art. 594.]

+ [See Art. 676, and Note 28.]