spectrum on a very bright cloudless day (conditions as before), except that more battery-power was used (10 cells Daniell).

The diffused daylight could not be cut off; so the trial was made in the most shaded part of an ordinary room, the spectrum being superimposed on the ordinary diffused daylight.

The resistances were very carefully balanced in each case, with the following results:

The indications were very clear; and the bar of selenium was so sensitive to the action of the spectrum, that a slight movement of the prism produced a corresponding movement in the spot of light on the galvanometer-scale.

It is to be noted that in this experiment the reflecting galvanometer was placed on a heavy masonry pillar insulated from the floor for observatory purposes, and that the battery, bar of selenium, and resistance-coils were in another room, being connected by long and carefully insulated leads with the galvanometer.

Experiment 4.-The diffused light was cut off as much as possible by screens; and the resistances were again balanced in the solar spectrum. Conditions as in the last case.

Light cut off by lid of box (resistance rising) 310,000

Experiment 5.-The selenium was also exposed to the spectrum of the electric light in a darkened room.

The effect was feeble; but by using more battery-power in balancing, it was possible to measure the swing of the spot of light when the selenium was suddenly exposed to the action of the light of the spectrum.

The maximum effect was obtained in, or just at the edge of the red, the violet and blue rays producing scarcely any effect.

Experiment 6.--The selenium was exposed to the full sunlight; the resistance fell enormously and instantaneously, and on balan

cing it was found to be little more than half what it was in the darkness.

The following were the general results of the experiments: Results.-(1) That the resistance of selenium is largely affected by exposure to light.

(2) That this effect is not produced by the actinic rays, but is at a maximum at, or just outside the red rays, at a place nearly coincident with the locus of the maximum of the heat-rays.

(3) That the effect of varying resistances is certainly not due to any change of temperature in the bar of selenium.

(4) That the effect produced on exposure to light is sensibly instantaneous, but that, on cutting off the light, the return to the normal resistance is not so rapid.

It would seem that there exists a power in rays nearly coincident with the heat-rays of high intensity, of altering instantaneously and without change of temperature the molecular condition of this particular element.

May 15.-William Spottiswoode, M.A., Treasurer and VicePresident, in the Chair.

The following communications were read :—

"Determination of the Number of Electrostatic Units in the Electromagnetic Unit made in the Physical Laboratory of Glasgow University." By Dugald M'Kichan, M.A.

The object of this paper is to describe experiments made at intervals from 1870 to 1872 in the Physical Laboratory of Glasgow University to determine the relation between the fundamental units in the two systems of absolute electrical measurement, the electromagnetic and the electrostatic. A summary is also given of the results of similar observations made by W. F. King in 1867 and 1868.

The two systems of electrical measurement, or the units which they employ, are founded on the fundamental units of time, mass, and space applied to the observed effects of electricity at rest and electricity in motion. The dimensions of quantity in the two systems are such that the ratio of the electromagnetic and the electrostatic unit of quantity is expressible as a velocity.

This velocity, usually known as v, is not only of great importance in all combinations of electromagnetic and electrostatic action, but it is also of great scientific importance in the theory of the propagation of electromagnetic disturbances through a dielectric medium. It occupies a very important place in the development of the electromagnetic theory of light by Professor Clerk Maxwell, according to whose theory this velocity v is the same as the velocity of light.

The first experimental determination of v was made by Weber from a common electrostatic and electromagnetic measure of capacity. As the result of Weber's experiments, v was found to be 310-74 × 108 centims. per second.

Another determination was made by Prof. Clerk Maxwell in

1868, by means of a direct comparison of electrostatic attraction with electromagnetic repulsion. His experiments gave v=288 x 10° centims. per second.

The value of v given by the experiments here described is 293 x 10 centims. per second. The method employed was that of obtaining an absolute electrostatic and an absolute electromagnetic measurement of the same electromotive force. v is defined as the ratio of the units of quantity in the two systems; but it follows from the definition of electromotive force, that v is also the ratio of the units of electromotive force in the two systems.

The electromotive force, or the difference of potentials between the two poles of a constant Daniell's battery, was measured electrostatically by means of Sir William Thomson's absolute electrometer. The absolute electromagnetic value of this electromotive force was given by the effect of the current which it maintained in the circuit of an electrodynamometer. The determination of this value depended on the resistance of the electrodynamometercircuit, which was reckoned in terms of the absolute value of the British-Association standard unit of resistance. Any correction which may hereafter be found to be applicable to the absolute value of this standard coil, as measured at King's College by Professors Clerk Maxwell, Balfour Stewart, and Fleeming Jenkin, must be applied to the value of v given above.

The comparisons made in 1867 and 1868 by Mr. King gave as the mean value of v, 284.6 × 108 centims. per second.

The experiments made in 1870 with the new absolute electrometer gave as the mean result v=294.5 × 108 centims. per second. The result of later observations made under much more favourable circumstances was v=292·4 × 10 centims. per second.

The latest observations (1872) furnish the most probable value of v, 293 × 108 centims. per second.

"On the Heating of a Disk by rapid Rotation in vacuo." By Balfour Stewart, M.A., F.R.S., and P. G. Tait, M.A.

26. In two previous communications (Proc. Roy. Soc. June 15, 1865, and No. 88, 1866) to this Society, we gave an account of some experiments which we had made upon the heating of a disk through rotation in vacuo. In these experiments the increase of radiation of the heated disk was observed by means of a delicate thermopile and galvanometer. Three aluminium disks of various thicknesses and one ebonite disk were used; and the results derived from the experiments were as follows:

(1) The heating effect observed appeared to be independent of the density, and of the chemical constitution, of the residual air and vapour surrounding the disks.

(2) The quantity of heat developed under similar circumstances of rotation in three aluminium disks 05, 0375, 025 of an inch in thickness respectively appeared to be the same, inasmuch as the relative thermometric effect for these disks varied inversely as their thickness.

(3) Besides the heating effect alluded to in (1) and (2), there was found to be, when the vacuum had been recently made, a strictly temporary effect, sometimes in the direction of heat, sometimes in that of cold, owing probably to the condensation or evaporation of small quantities of aqueous vapour; but this effect was only noticeable during rotation, disappearing the moment the motion was stopped.

27. The experiments described in these communications were resumed in 1870. In the interval an addition had been made to the apparatus, in virtue of which an ordinary carbonic-acid vacuum might be subjected to the influence of a vessel containing potash allowed to open in it, and thus to absorb as much as possible of the

remaining gas.

On May 4 a carbonic-acid vacuum was obtained by this means (pressure 0.05 in.). A disk made of cartridge-paper, when made to rotate in this vacuum, gave a very perceptible result. Carbonic acid was then allowed to enter the vacuum until the pressure became 0.65 in. The consequence of this increased pressure was in this instance an increase in the effect, which was probably of a permanent nature, inasmuch as it remained after three days. Unfortunately the exact increase was not noted; but it is believed that the heat-indication became about three times as great in consequence of the additional pressure.

28. On May 11 another carbonic-acid vacuum was made (pressure 0.12 in.), and at the suggestion of Professor Maxwell a sulphuric-acid gauge was placed in the receiver. A rotation was then made; and the result of the rotation was a hardly perceptible rise in the sulphuric-acid gauge. We may therefore imagine that the residual air was not greatly heated.

29. On May 16 another carbonic-acid vacuum was obtained (pressure 0.08 in.), and with an ebonite disk of about inch thickness, the heat-indication was 16. Carbonic acid was introduced until the pressure was 1.6 in.; but the indication was only 18. This result is in conformity with our previous experiments, in which an increase of pressure of the residual air produced little or no effect.

30. The ebonite disk was likewise tried in a carbonic-acid vacuum, pressure 0.04 in.; and also in one, pressure 0·02 in., which was the lowest we could obtain. The result of these experiments appeared to show that in all probability the cartridge-paper disk radiated more than twice as much as the thin ebonite disk. The experiments were put a stop to by a collapse of the glass receiver during rotation, fortunately without injuring any one present.

31. In June 1871 the experiments were resumed. In the mean time Mr. Beckley had fitted the apparatus with an arrangement working through a barometer-tube, by means of which, instead of trusting to radiation, the disk itself might, after rotation, be tapped by means of the pile, which could be brought up to it and then withdrawn. By this means a much larger effect might be obtained; and it became possible, by varying the adjustment, to find according to what law the heat effect varies with the distance from the centre.

32. These experiments were conducted in the following manner:-The disk was first of all tapped before rotation several times; at each tapping the momentary swing of the needle was recorded, and the mean of the readings was regarded as indicating the state of the disk with respect to heat.

The disk was next tapped after rotation, and the difference between the readings before and after was taken as indicating the change in the state of the disk produced by rotation. In the later experiments the disk before rotation was kept in slight motion in order to equalize any tendency to unequal heating of its various parts; but this was not done in the experiments of June 1871. In these experiments (June 1871) the disk was of brown paper, and the results obtained were as follows:-For a carbonic-acid vacuum (pressure 0·065 in.) a swing of 307 divisions was recorded, while for a hydrogen vacuum (pressure 0-150 in.) a swing of 281 divisions was recorded. Each result was the mean of three rotations. 33. The next experiments were made in January 1872. The galvanometer was one of Thomson's, but more adapted for battery currents, and hence not in a very delicate state for these experiments; the time of vibration of the needle was 3 seconds. The disk used was of ebonite (thickness about in.). The ebonite of this disk was completely black; and in this respect, as well as in being thinner, it differed from the ebonite disk first used in the radiation-experiments (art. 17). In the present experiments the centre of the pile was made to tap the disk at a distance of 1.5 in. from the rim. The amount and velocity of rotation were represented by 30 turns of the handle, or 3750 turns of the disk, in about 40 seconds. The following results were obtained, a result representing on an average somewhat more than four rotations :Heat-indication for

Pressure of residual gas, in inches.

15.0

It would thus appear that the results derived by tapping are very different from the radiation-results, inasmuch as in the former the effect of the pressure and quality of the residual air is very apparent, while in the radiation-results it is hardly perceptible. A probable explanation of this will be given afterwards (art. 46); but in the mean time, in view of these results, it will be expedient to discuss them quite independently and by themselves, with the view of ascertaining whether they can best be explained by a gas-effect alone, or whether they likewise indicate a residual effect independent of gas.

34. With this object let us take (A)+(E

2

as representing the whole effect at a pressure of in., due to whatever cause or causes. We thus obtain

20

Dry hydrogen.

Whole effect at .

9.5

Dry air.
25.0

Dry carbonic acid. 24.0