384 Dr. A. Schuster on the Spectra of Metalloids. [May 16, 



for each gas; we have to, study the effect of various impurities, the 

 influence of the electrodes, and that of the glass which in the tubes 

 generally used is considerably heated up by the spark. I have chosen 

 oxygen as a first subject of investigation. Though Pliicker and Wiillner 

 have, as far as their experiments went, accurately described the phe- 

 nomena seen in oxygen tubes, the following contains much that is 

 new, and will put some of the older facts on a firmer basis. 



As some of the facts brought to light by the investigation bear 

 directly on the question of double spectra, our knowledge on that 

 point must be briefly referred to. We divide all known spectra into 

 three orders. Continuous spectra, channelled space spectra, and line 

 spectra. With regard to continuous spectra, it is shown that the 

 older statement which limited them to -liquid and solid bodies is no 

 longer tenable. Most gases give continuous spectra long before they 

 condense. Two theories of continuous spectra are noticed. The one 

 considers that the vibrations of a molecule always tend to take place 

 in a fixed period, but that the impacts of other molecules may, when 

 the pressure is great or in liquid and solid bodies, prevent complete 

 oscillations taking place, and thus produce a continuous spectrum. 

 The other theory considers that, when a gas condenses, molecular 

 combinations take place, which make the molecular structure more 

 complicated, and may produce channelled space spectra or continuous 

 spectra. According to the latter theory such molecular combinations 

 are possible before the gas condenses, and thus the state of aggrega- 

 tion of the gas only indirectly affects the spectrum. The latter theory 

 seems to be more consistent with experiment than the former one. 

 For instance;, it is shown that oxygen gives a continuous spectrum at 

 the lowest temperature at which it is luminous. If the temperature 

 be raised, the continuous spectrum is replaced by a line spectrum. 

 This seems to be inexplicable by theory of molecular impacts. 



With regard to channelled space spectra attention is drawn to a 

 new kind, which finds a representative in the spectrum of the negative 

 pole in oxygen. The ordinary channelled space spectra show bands, 

 which when seen under small dispersion appear to have a sharp 

 boundary on one side, and fade gradually away towards the other. 

 When seen under large dispersion these bands appear to be made up 

 of a number of lines which get nearer and nearer to each other as 

 they approach the sharp boundary. The spectrum of the negative 

 pole in oxygen, when looked at with small powers, seems to be made 

 up of a number of bands which are uniformly illuminated throughout, 

 but with high powers each band is found to be made up of a number 

 of lines at about equal intervals. 



With regard to line spectra it is shown that a body may at a low 

 temperature show a different set of lines altogether from what it 

 shows at a high temperature. It is difficult to decide whether the 



