M. A. J. Angstrom's Optical Researches. 329 



according to Stokes, produce the luminous effects discovered by 

 Herschel. 



3. Now, as according to the fundamental principle of Euler, 

 a body absorbs all the series of oscillations which it can itself 

 assume, it follows from this that tbe same body, when heated so 

 as to become luminous, must emit the precise rays which, at its 

 ordinary temperature, is absorbed. The proof of the correctness 

 of this proposition is, however, surrounded with great difficulties ; 

 for the condition of the heated body, as regards elasticity, is 

 altogether different from the state in which the light is supposed 

 to be absorbed. An indirect proof of the truth of the proposi- 

 tion is furnished by the connexion, discovered by M. Niepce de 

 Saint Victor, between the colour imparted by a body to the flame 

 of alcohol, and that developed by light upon a disc of silver 

 which has been chlorinized by the body under consideration. As 

 the disc of silver, treated with chlorine alone, assumes all the 

 tints of the solar spectrum, and, when treated at the same time 

 with a colouring body, exhibits almost exclusively the colour of 

 the latter, this cannot occur otherwise than by the exclusive ab- 

 sorption on the part of the so-prepared silver disc of the precise 

 tint which belongs to the colouring body. 



4. One of the most convenient and most practicable means of 

 studying a glowing body is presented by the electric spark, 

 although the body in this case is probably in the gaseous con- 

 dition. I have therefore believed that an investigation of the 

 spectra obtained from e*lectric sparks drawn from different metals 

 would not be without interest for the theory of light. The sub- 

 ject has been already treated by Wollaston, Fraunhofer, Wheat- 

 stone, and lastly by Masson, in a memoir in the Annates de 

 Chimie et de Physique, 1851. Masson has also measured the 

 electric spectra of various metals, and has drawn them by means 

 of the Camera lucida. Nevertheless I venture to hope that the 

 present investigation will not be deemed superfluous, when the 

 results obtained by me are compared with those obtained by my 

 predecessors. 



I have found that the spectrum of the electric spark must 

 really be regarded as consisting of two distinct spectra ; one of 

 which belongs to the gas through which the spark passes, 

 and the other to the metal or the body which forms the con- 

 ductor. 



The electric spectrum is traversed, — 1. By a multitude of 

 luminous lines, comparable in number and distribution with the 

 lines of Fraunhofer in the solar spectrum. These lines arc the 

 same for all metals, but vary in intensity with the strength of 

 tin- condenser, and the greater or less humidity of the air. 

 The nature of the conductor does not appear to exercise any 



