of water, carefully weighed, was placed on a light wooden support, touching it at only three points. This was put inside of a considerably larger cylinder, also of tinned iron; this outer cylinder having a double cover with a hole in it—the cover large enough to shade the sides of the vessel, and the hole a little less than three inches in diameter. A delicate thermometer was immersed in the water, with a sort of dasher of mica for the purpose of stirring it and keeping the temperature uniform throughout the mass. The apparatus was so placed and adjusted that the whole of the light and heat passing through the hole in the cover would fall upon the surface of the water, the sun at that time (December 31st) being within 12° of the zenith at noon.
This apparatus was placed in the sunshine and allowed to stand for ten minutes, shaded by an umbrella, and the slight rise in the temperature of the water was noted. Then the umbrella was removed and. the solar rays were allowed to fall upon the water for the same length of time, and the much larger rise of temperature was noted; finally, the apparatus was again shaded and the change for ten minutes again observed. The mean between the effects in the first and last ten-minute intervals can be taken as the measure of the influence of other causes besides the sun, and, deducting this from the rise during the ten minutes' insolation, we have the effect of the simple sunshine.
Herschel's figures for his first experiment run as follows:
| Rise of temperature in first ten minutes | 0·25° |
| Rise„ of„ temperature„ in„ second ten minutes (sun) | 3·90 |
| Rise„ of„ temperature„ in„ third ten minutes | 0·10 |
The mean of the first and third is 0·17°, and this deducted from the second gives 3·73° Fahr. as the rise of temperature produced by a sunbeam three inches in diameter, absorbed by a mass of matter equivalent to 4,638 grains of water. (We do not indicate the minutiae of the process by which the weight of the tin vessel, thermometer, stirrer, etc., are allowed for.) Nothing more is now necessary to enable us to compute just how much heat is received by the earth in a day or a year, except, indeed, the determination of the very troublesome and somewhat uncertain correction for the absorption of heat by the earth's atmosphere; a correction deduced by means of observations made at varying heights of the sun above the horizon.
Herschel preferred to express his results in terms of melting ice, and put it in this way:
The amount of heat received on the earth's surface, with the sun in the zenith, would melt an inch thickness of ice in two hours thirteen minutes, nearly.
Since there is every reason to believe that the sun's radiation is equal in all directions, it follows that, if the sun were surrounded by a
