condition. Four years of labor on this subject had not failed to suggest many other researches.
A detailed study of the distribution of the heat of the solar surface was begun about this time, by means of the thermopile, and was quickly rewarded by the discovery of an unknown thermo-chroic action in the sun's atmosphere, such that it transmits the light less readily than the heat, owing to the difference in wave-length. An interesting consequence of this action is that, if, at any time, the sun's atmosphere should grow thicker, the color of the sun would tend toward red; if thinner, then toward blue. These changes, which are quite possible, suggest interesting explanations of some of the phenomena of the variable stars. The glacial epochs on the earth may be connected with changes in the solar atmosphere.
In 1877 we find another outcome of the series of measures of the heat from various parts of the sun's disk, and especially from the umbræ, etc., of sun-spots. The periodic changes in the spotted area of the solar disk, which had long been known, induced the inquiry whether changes in the amount of spotted surface bore any relation to changes of temperature on the earth's surface.
The result of the extremely delicate measures of Professor Langley led plainly to the conclusion that the direct effect of sun-spots on terrestrial temperature is sensible; that, when the spotted area is a maximum, the temperature is on that account lower, and the converse; but that the total direct effects of the periodic changes in the spotted area on the earth's mean temperature are extremely small, not more than a change of three tenths of 1° C. in eleven years, and not less than one twentieth of 1° C. The indirect effects are not here considered.
A thermopile used in connection with the most sensitive galvanometers is an extremely delicate instrument; and Allegheny Observatory now possessed a most complete outfit of this sort.
But the most important and pressing questions in solar physics demanded a means of measurement of heat still more delicate. When it was a question to measure the heat radiation from the different parts of the sun's disk, the thermopile was adequate. But if the heat from one of these parts is spread out into a heat-spectrum several feet or even yards long, it becomes necessary to devise new means of measuring the minute differences between the various parts. Such a device is the bolometer, which consists of two systems of extremely thin steel or platinum strips. Through these two systems an electric current passes. A sensitive galvanometer connected with both systems keeps its needle steady when the currents are equal.
If one system is now exposed to heat radiations while the other is protected from them, the temperature of the first is raised, its electric resistance is increased, and the battery-currents through the two systems and the galvanometer no longer balance. The galvanometer--
