BRITISH ASSOCIATION. 47 



light, ■which, if examined by the prism, gives a well-defined and characteristic 

 spectrum. By such a spectrum, by any one of its brilliant lines, whose position 

 has been measured, you may recognize the examined gas. This way of proceeding 

 constitutes what is cai ud spectral analysis, to which we owe, until this day, the dis- 

 covery of three new elementary bodies. In order to give to spectral analysis a true 

 and certain basis, you want the spectrum of each elementary substance. Most re- 

 cently, some eminent philosophers, in examining such spectra, met with unexpected 

 difiiculties, and doubts arose in their minds against the new doctrine. These doubts 

 are unfounded. The fact is, that the molecular constitution of gases is much more 

 complicated than it has been generally admitted till now. The spectra, therefore 

 always indicating the molecular constitution of gases, ought to be more complica- 

 ted also than it was thought at first. By these considerations, a new importance 

 a rather physical one, is given to spectral analysis. You may recognize by the 

 spectrum of a gas, not only the chemical nature of the gas, but you may also ob- 

 tain indications of its more intimate molecular structure — quite a new branch of 

 Bcience. Allow me now to select out of the results already obtained two instances 

 only. Let me try to give what I may call the history of the spectra of two ele- 

 mentary bodies — of sulphur and nitrogen. In order to analyse by the prism the 

 beautiful light produced by the electric current, if it pass through a rarefied gas, I 

 gave to the tube in which the gas is included such a form that its middle part was 

 capillary. Thus I got within this part of the tube a brilliant film of light, ex- 

 tremely fitted to h9 examined by the prism. The date of my first paper on this 

 subject is the 12th of March. 1858. After having provided myself with apparatus 

 more suited to my purposes, I asked, about a year ago, my friend. Prof. Hittorf, of 

 Miinster, to join me in taking up my former researches. The very first results we 

 obtained in operating on gases of a greater density opened to us an immense field 

 of new investigation. We found that the very same elementary substance may 

 have two, even three, absolutely different spectra, which only depend on tempera- 

 ture. In our experiments we made use of RiihmkorfiPs induction coil, whose dis- 

 charge was sent through our spectral tubes. In order to increase at other times 

 the heating power of the discharge, we made use of a Leyden jar. Now, let us 

 suppose a spectral tube, most highly exhausted by Geissler's mercury pump, con- 

 tains a very small quantity of sulphur. The discharge of the coil will not pa?8 

 through the tube if it do not meet with ponderable matter, either taken from the 

 surface of the glass, or, if the discharge be very strong, by the chemical decompo- 

 sition of the glass. In heating slowly the tube by means of a lamp, in order to 

 transform a part of the sulphur into vapour, all accidental spectrum, if there be 

 one, will disappear, and you will get a pure and beautiful spectrum of sulphur. I 

 supposed the Leyden jar not to have been interposed. If you now interpose it, 

 the spectrum just spoken of will suddenly be replaced by a quite different one. 

 We were generally led to distinguish two quite different classes of spectra. Spec- 

 tra of the first class consist in a certain number of bands, variously shadowed by 

 dark transversal lines. Spectra of the second class consist in a great number of 

 most brilliant lines on a dark ground. Accordingly, sulphur has one spectrum of 

 the first class and another one of the second class. You ruay as often as you like 

 obtain each of these two spectra. In operating on a spectral tube, containing ni- 

 trogen at a tension of about 60 millim^tves, you will, without the Leyden jar, get 



