267 



apparatus is discharged, is a dark blue, in the capillary part of the 

 same tube the finest gold-colour, while the light surrounding the 

 negative wire is of a fawn-colour. [The finest appearance is obtained 

 with a larger tube containing residual traces of the vapour of this 

 salt, put on the iron-pieces of a powerful electro-magnet ; within the 

 blue light of the discharge numerous golden flashes are produced, 

 and variously directed by the magnetic force.] 



A piece of metallic sodium within an atmosphere of rarefied hy- 

 drogen does not alter the spectrum of this gas when at the ordinary 

 temperature ; but when it is heated, a single brilliant yellow band is 

 added to the three original bands of hydrogen. The middle of the 

 new band exactly coincides with Fraunhofer's dark line D. The 

 vapours of the metal are condensed in the cooler parts of the 

 apparatus. 



Phosphorus, when treated in the same way, instead of adding new 

 bands to those of hydrogen, at a certain temperature even extin- 

 guished the spectrum of the gas. 



In the case of mercury, I gave to the tube the shape of 

 the annexed drawing. The electrodes merely entered the 

 two flask-shaped ends of the tube, where they were covered 

 about half an inch with mercury. Peculiar bands were 

 obtained without heating the mercury ; when it was 

 heated, the spectrum became most brilliant. 



When traces of two gases, not acting on each other, 

 are mixed within a spectrum -tube, the spectra of both 

 are simultaneously obtained. 



A result of some importance, following from the 

 researches of which I have here given an abstract, 

 is, that in all such dioptrical researches, where Fraun- 

 hofer's dark lines were used in order to get exact measures, these dark 

 lines may with great advantage be replaced by the middle lines of the 

 new brilliant bands of the gas spectra. To these bands the most 

 convenient breadth may be given in each particular case by regulating 

 the aperture of the apparatus. A spectrum-tube of hydrogen, 

 exhibiting three well-defined bright bands, is well suited for this 

 purpose. During a whole year I made use of such a tube, which 

 remained absolutely unaltered. 



Every gas being characterized by its spectrum (even by one of the 



VOL. X. 



