Radium on this hypothesis must be considered as a transformer borrowing energy from the external medium and returning it in the form of radiation. It is not impossible, even, to admit that the energy which the atom of radium withdraws from the surrounding medium may serve to keep up, not only the heat emitted and its complex radiation, but also the dissociation, supposed to be endothermic, of this atom. Such seems to be the idea of M. Debierne and also of M. Sagnac. It does not seem to accord with the experiments that this borrowed energy can be a part of the heat of the ambient medium; and, indeed, such a phenomenon would be contrary to the principle of Carnot if we wished (though we have seen how disputable is this extension) to extend this principle to the phenomena which are produced in the very bosom of the atom. We may also address ourselves to a more noble form of energy, and ask ourselves whether we are not, for the first time, in presence of a transformation of gravitational energy. It may be singular, but it is not absurd, to suppose that the unit of mass of radium is not attached to the earth with the same intensity as an inert body. M. Sagnac has commenced some experiments, as yet unpublished, in order to study the laws of the fall of a fragment of radium. They are necessarily very delicate, and the energetic and ingenious physicist has not yet succeeded in finishing them.[46] Let us suppose that he succeeds in demonstrating that the intensity of gravity is less for radium than for the platinum or the copper of which the pendulums used to illustrate the law of Newton are generally made; it would then be possible still to think that the laws of universal attraction are perfectly exact as regards the stars, and that ponderability is really a particular case of universal attraction, while in the case of radioactive bodies part of the gravitational energy is transformed in the course of its evolution and appears in the form of active radiation. [Footnote 46: In reality M. Sagnac operated in the converse manner. He took two equal _weights_ of a salt of radium and a salt of barium, which he made oscillate one after the other in a torsion balance. Had the durations of oscillation been different, it might be concluded that the mechanical mass is not the same for radium as for barium.] But for this explanation to be admitted, it would evidently need to be supported by very numerous facts. It might, no doubt, appear still more probable that the energy borrowed from the external medium by radium is one of those still unknown to us, but of which a vague instinct causes us to suspect the existence around us. It is indisputable, moreover, that the atmosphere in all directions is furrowed with active radiations; those of radium may be secondary radiations reflected by a kind of resonance phenomenon.