Now a wave-length--energy curve may be as easily constructed for absorbed as for emitted radiation by means of a Langley's bolometer. The strip of the bolometer is first coated with lamp-black and the spectrum of the incident radiation is explored in exactly the same way as is described in Chapter III. {41} The strip is then coated with the surface under investigation and the spectrum is again explored. Since the incident radiation is exactly the same in the two experiments, the differences in the quantities of heat absorbed must be due solely to the difference in the absorbing powers of the two surfaces. In Fig. 21 the dotted line represents the wave-length--energy curve for the radiation absorbed by the blackened bolometer strip, and the solid line the curve for the strip coated with the surface under investigation. [Illustration: FIG. 21.] The actual form of the curves may and probably will be quite different from the form in Fig. 20, but it will be found for the same wave-length ON that PN/QN is exactly the same in the two figures. It has already been mentioned that dull, dark-coloured surfaces radiate the most heat, and that polished surfaces radiate the least. A radiator for heating a room should therefore have a dull, dark surface, while a vessel which is designed to keep its contents from losing heat should have a highly polished exterior. A perfectly transparent substance would radiate no energy, whatever the temperature to which it is {42} raised, for its absorbing power is zero and therefore its radiating power is also zero. No perfectly transparent substances exist, but some substances are a very near approach to it. A fused bead of microcosmic salt heated in a small loop of platinum wire in a blowpipe flame may be raised to such a temperature that it is quite painful to look at the platinum wire, yet the bead itself is scarcely visible at all. Any speck of metallic dust on the surface of the bead will at the same time shine out like a bright star. +Gases as Radiators.+--Most gases are an even nearer approach to the perfectly transparent substance, and consequently, with one or two exceptions, the simple heating of gases causes no appreciable radiation from them. Of course, gases do radiate heat and light under some circumstances, but the radiation seems to be produced either by chemical action, as in the flames coloured by metallic vapours, or by electric discharge, as in vacuum tubes, the arc or the electric spark.