of a balance and counterpoised so that the amount of wax burned can be determined at any moment without moving the candles. This arrangement is also useful in keeping the flames nearly in the same position, for as the candles burn down the arm supporting them rises. The balance is to be placed so that the candle-flames are vertically over the zero of the scale of a photometer bench in a dark room.

As a source to be compared with these, we use a gas-flame, the supply of gas being regulated and measured thus:

The gas is passed from a gas-holder, where the pressure can be altered by altering the weights on the cover, through a meter, M, fig. 23, which measures the quantity of gas passed

through. One complete revolution of the needle corresponds toth of a cubic foot of gas, so that the numbers on the dial passed over in one minute give the number of cubic feet of gas which pass through the meter in an hour. The gas enters at the middle of the back of the meter and leaves it at the bottom, passing thence to a governor, G, which consists of an inverted bell, partly sunk in water and counterpoised so that the conical plug attached to its top is very close to the conical opening of the entrance pipe q. Any increase of pressure of the gas in the bell raises the bell, narrows the aperture, and diminishes the supply until the pressure falls again. By this means the pressure of the gas at the burner is maintained constant.

The exit pipe from the bell passes to a tube with two stopcocks s, s'. The stopcock s' is provided with a screw adjustment for regulating the supply of gas with extreme nicety; the stopcock s can then be used, being always either turned on full or quite shut, so as to always reproduce the same flame without the trouble of finely adjusting every time. Between these two stopcocks is a manometer M for measuring the pressure of the gas as it burns.

In stating, therefore, the gas-flame employed, we have to put down (1) the burner employed; (2) the pressure of the gas; (3) the amount of gas passing through the meter per hour.1

The gas passes from the stopcocks to the burner, which is fixed on one of the sliding stands of the photometer bar, so that the plane of the flame corresponds to the fiducial mark on the stand. On another sliding stand between the burner and the candles is placed the photometer disc, which consists of a grease spot upon white paper.

The method consists in sliding the photometer disc along the scale until the spot appears of the same brightness as the rest of the paper; the intensities of the lights are then proportional to the squares of their distances from the disc.

The observations should be made by viewing the disc from either side, as it will often be found that when the spot and the rest of the disc appear to be of the same brightness when viewed from one side, they will differ considerably when viewed from the other. This is due, in part, at any rate, to want of uniformity in the two surfaces of the paper of which the disc is made; if the difference be very marked, that disc must be rejected and another used. In all cases, however, observations should be made from each side and the mean taken.

The sources of light should be screened by blackened

In order to test the lighting power of gas' with a standard argand burner, the flow through the meter must be adjusted to 5 cubic feet per hour by means of the micrometer tap.

screens, and the position of the disc determined by several independent observations, and the mean taken.

The lights must be very nearly of the same colour, otherwise it will be impossible to obtain the appearance of equality of illumination over the whole disc. (This may be tried by interposing a coloured glass between one of the lights and the disc.) Instead of trying to find a position in which the disc presents a uniform appearance on one side, the position in which it appears the same as viewed from two corresponding points, one on each side, may be sought for. For additional details see the 'Gas Analysts' Manual,' p. 40, §§ 61, 84.

Experiment.-Compare the illuminating power of the gasflame with that of the standard candle.

Additional experiments.—(a) Compare the intensities of the candles and standard argand burner

(1) Directly.

(2) With a thin plate of glass interposed between one source and the disc. This will give the amount of light lost by reflection and by the absorption of the glass. By rotating the glass plate the variations in the loss at different angles may be tested.

(3) With a thin plate of glass between one source and the disc, and a thick plate on the other side. This will enable you to determine the amount of light lost by the absorption of a thickness of glass equal to the difference of the thicknesses of the two plates.

(b) Obtain two burners and arrange them in connection with a three-way tube. Cover one up by a screen, and measure the intensity of the other. Then interchange them, and so obtain the intensity of each separately. Then place them together so that the two flames unite, and measure the intensity of the combined flame and its relation to the sum of the intensities of each.

(c) Test the intensity of the light from the same amount of gas used in different burners.

Enter results thus :

Gas burning at the rate of 5 cubic feet per hour.

The apparatus for making the comparison consists simply of a bar, at the end of which a ground glass or paper screen is fixed, and on which a support is made to slide, carrying the gas jet or other source of light.

On the bar, and in front of the screen, is placed a wooden rod, about 3 inches from the screen. The two lights to be compared are placed one on the sliding support and the other on the table at a fixed distance (taking care that both are the same height), the positions being so adjusted that the two shadows of the rod thrown on the screen are just in contact with each other without overlapping. The screen must be turned so that it makes equal angles with the direction of the light from each source. The distance of the sliding light has to be adjusted so that the two shadows are of the same depth.

Consider a unit of area, e.g. a square centimetre, of each shadow A and B; let the distance of the unit of area of A from the two sources of light be x, x, and let the distance of the unit of area of the shadow в from the same sources be y, y respectively. Then the unit of area of A is illuminated only by the one source of light, distant x from it, and therefore its illumination is I/x2, where I is the illumination per unit area at unit distance from the source. The unit of area of B is illuminated only by the source of light at distance y, and the illumination therefore is I'/y2, when I' is the illumination per unit area at unit distance from the second

source

Hence, since the illuminations of the shadowed portions of the screen are equal,

If the two unit areas considered be immediately adjacent to the line of junction of the shadows, then we may measure x and y from the same point. Hence the ratio of the intensities of the two sources is the square of the ratio of the distances of the two sources from the line of contact of the shadows. The method has the advantage that the observations do not need a dark room.

The shadows may be so arranged that the line of contact is on the middle line of the bar on which the one source slides, and accordingly the distance may be measured along the bar. The other distance may be measured by a tape.

The arrangements necessary for determining the rate at which the gas is being burnt or the quantity of wax consumed are described in section 45.

Experiment. Compare the illuminating power of the gasflame and standard candle.

Enter results thus:

Candle burns at the rate of 8.1 gms. per hour.