notwithstanding the heat acquired by the screens; in the second case also it remains unmoved, although in this case the plates (blackened or opake) are submitted to the actual radiation of the source itself. In the third case the index of the galvanometer leaves its position of equilibrium and describes arcs of greater or less extent according to the quality and thickness of the screen. But the time which it takes to reach the extremity of these arcs is invariable, and equal to that which it takes to describe a deviation of 30° when there is no screen interposed.
This third proof, though indirect, is nevertheless the most convincing, and possesses the additional advantage of showing, as it were, palpably, that the manner in which radiant heat is transmitted in the interior of diathermanous substances is altogether analogous to that in which light is propagated through transparent media whether solid or fluid. For in respect to the latter we perceive no appretiable difference between the times which the luminous rays take to pass through layers of any quality and thickness whatsoever.
The analogy between the transmission of light and that of radiant heat is rendered still more striking if, by shaking or otherwise, a motion is produced in the mass of the screen submitted to the experiment. I have passed the different parts of a large square of glass rapidly before the narrow aperture of the metallic plate through which the calorific rays that strike the surface of the pile are transmitted. By means of a bow I made it vibrate; it emitted sounds more or less acute: the index of the galvanometer pointed invariably to the same degree of its scale. I found the deviation of the magnetic needle equally invariable when I measured the intensity of the calorific radiation through a layer of acidulated water, at first still, but afterwards set in motion by agitators or traversed by a strong electric current.
Here then, though under different forms, the fact observed in the experiments of Pictet and Saussure when we agitate the mass of air interposed betwen the reflectors is reproduced; namely, the impossibility of altering by these means the direction or the intensity of the luminous or the calorific rays passing through atmospheric air or any other diaphanous medium.
These different considerations seem to me well calculated to dispel every shade of doubt that may yet be entertained as to the immediate transmission of radiant heat by diathermanous bodies, whether solid or liquid. But (to return to the four sources) we have already observed that in our method of proceeding it is necessary to operate uniformly under the influence of a radiation equal to 30° of my thermomultiplier. Now to effect this with sources of various temperatures they must be brought more or less close to the thermoelectric pile until we have obtained the galvanometric indication required, and such is the way in which we have proceeded in all our experiments of transmission. The same screen
