Fig. 76.—Method of showing the spectrum of substances in solution. Spectroscopic observations are still further complicated by the fact that one and the same substance gives different spectra at different temperatures. This is especially the case with gases whose spectra are obtained by an electric discharge in tubes. Plücker, Wüllner, Schuster, and others showed that at low temperatures and pressures the spectra of iodine, sulphur, nitrogen, oxygen, &c. are quite different from the spectra of the same elements at high temperatures and pressures. This may either depend on the fact that the elements change their molecular structure with a change of temperature, just as ozone is converted into oxygen (for instance, from N2 molecules are obtained containing only one atom of nitrogen), or else it may be because at low temperature certain rays have a greater relative intensity than those which appear at higher temperatures. If we suppose that the molecules of a gas are in continual motion, with a velocity dependent on the temperature, then it must be admitted that they often strike against each other and rebound, and thus communicate peculiar motions to each other and the supposed ether, which express themselves in luminiferous phenomena.
A rise of the temperature or an increase in the density of a gas must have an influence on the collision of its molecules and luminiferous motions thus produced, and this may be the cause of the difference of the spectra under these circumstances. It has been shown by direct experiment that gases compressed by pressure, when the collision of the molecules must be frequent and varied, exhibit a more complex spectrum on the passage of an electric spark than rarefied gases, and that even a continuous spectrum appears. In order to show the variability of the spectrum according to the circumstances under which it proceeds, it may be mentioned that potassium sulphate fused on a platinum wire gives, on the passage of a series of sparks, a distinct system of lines, 583–578, whilst when a series of sparks is passed through a solution of this salt this system of lines is faint, and when Roscoe and Schuster observed the absorption spectrum of the vapour of metallic potassium (which is green) they remarked a number of lines of the same intensity as the above system in the red, orange, and yellow portions. The spectra of solutions are best observed by means of Lecoq de Boisbaudran's arrangement, shown in fig. 76. A bent capillary tube, D F, inside which a platinum wire, A a (from 0·3 to 0·5 mm.