emitted by a source having a constant temperature. This condition could be complied with by means only of certain flames and boiling liquids. I was therefore unable to vary the experiments so much as I should have desired. The sources however which I have employed present the most remarkable phases of the hating and combustion of bodies. They are four in number; namely, the flame of oil without the interposition of glass, incandescent platina, copper heated to 390°, and boiling water. Thus I had two luminous and two non-luminous sources. The first is furnished by a Locatelli lamp[1]; the second is a spiral of platina wire kept in a state of incandescence by means of a lamp fed with spirit of wine; the third is obtained by covering a flame of alcohol with a plate of copper, which soon acquires a fixed temperature whose mean value, as found by the method of immersion, is 390° Cent.(732° Fahr.); and the last source is merely a vessel of thin copper, blackened on the outside and filled with boiling water.
The intensities of the radiations have been always ascertained by the thermomultiplier. The means necessary to be adopted in order to obtain with this instrument the measure of the immediate transmission having been stated in the Memoir already quoted, I think it needless to enter here into further detail as to the arrangement of the apparatus and the nature of the galvanometric indications. I shall only remind the reader that as this method requires that the operation should be performed under the influence of a radiation equivalent to 30° of my thermomultiplier, the diaphanous substances, if placed at a suitable distance between the thermoelectric pile and the source of heat, cannot acquire a temperature sufficient to produce in the instrument any perceptible action. This is proved in three ways: first, by placing the screens on their stand after having exposed them to a calorific radiation of the same intensity as that to which they are exposed during the experiment; secondly, by substituting for the diaphanous body plates of blackened glass or metal, flakes of wood or stone, or sheets of paper; thirdly, by varying the nature and thickness of the medium (more or less transparent) through which the rays are to pass, from the thinnest plate of mica to pieces of rock crystal, glass, or Iceland spar several inches in thickness. In the first case the index of the galvanometer remains unmoved, notwithstand-
- ↑ The Locatelli lamp is merely a common lamp with one current of air and fed with oil. It has a wick of the shape of a quadrangular prism, which exactly fills the beak, but has no funnel. It gives a fine flame of constant temperature. The Argand lamp produces a flame of much greater intensity.
In the first series of experiments the main object was to determine the difference between the calorific and the luminous transparency. We therefore preferred the source which was least favourable to the establishment of the principal fact which it was then our purpose to verify. In the present experiments we proposed more particularly to examine the calorific transparency by itself. It was therefore necessary to operate upon rays that had not been forced to undergo a transmission previously to their being employed in the experiments.
