frangibility of a ray of light depends upon its wave-length. Hence, then, we see that glowing hydrogen emits a number of rays possessing certain definite wave-lengths and no others. The rays of intermediate wave-length are entirely wanting. In this respect, of course, a fundamental contrast is presented between such a spectrum as that we are considering and the spectrum of an incandescent solid in which light of every wave-length between certain limits is manifested. Each wave-length corresponds, of course, to a certain system of undulation in the ether. We are hence assured that the radiation from hydrogen translates itself into a certain system of waves of ether, each with its own particular period.
We must, therefore, expect to find that the hydrogen gas possesses the means of imparting to quiescent ether those particular vibrations that the spectroscope reveals. It will be obvious that the movements of the molecule of hydrogen as a whole are not what will answer the purpose. Such movements are not in the nature of vibrations. It is indeed known that the speed with which a molecule of hydrogen is animated undergoes frequent changes. If the positions of the lines in the spectrum were directly dependent on the velocity of translation of the molecules, the spectrum could not be expected to exhibit the characters which we actually find it to possess. It is, therefore, impossible for the origin of the spectral lines to be attributed to any other source than those internal agitations which each molecule itself possesses. The oscillations of the several parts of the molecule impart vibrations to the surrounding ether which are from thence propagated as radiant light. The vibrations of an elastic body are isochronous, that is to say, per-
