On flame spectra.

In order to obtain the spectrum of any substance, it has generally been considered sufficient to introduce a small quantity of it into an already formed flame. In the course of a photometrical investigation of flames which had been coloured by injecting the spray from saline solutions into the gas to be burnt, M. Gouy discovered in the spectra of the flames several new lines belonging to the metal contained in the solution. Instead of appearing throughout the whole flame, as did the previously known lines, these new lines were only emitted in the vicinity of the inner blue cone—the origin of the Swan spectrum. The observations of M. Gouy were limited to the examination of certain lines of the visible portion of the spectrum, and, with the advice of Professor Schuster, and under his direction, I have taken up this study with the object of extending it, by means of photography, to the ultraviolet portion of the flame, and also of detecting lines which are too feeble to be visible to the eye.