84 M. E.Wiedemann's Investigations on 



different lines can be recognized as characterizing the band 

 spectra. 



The assumption, however, that the band spectra are gene- 

 rated by rotatory motions *, is opposed by the consideration 

 that, if when the temperature is raised the light emitted by 

 gases always consists of rays of the same wave-length, the 

 molecular diameter must increase proportionally to the abso- 

 lute velocity, or to the root of the absolute temperature, since 

 only then could the number of rotations remain invariable. In 

 all other cases the lines must be displaced towards the red or 

 the violet, according as the molecular diameter increases with 

 rising temperature more or less quickly than the velocity of 

 rotation. That such displacements do not occur we learn from 

 all the spectrum-experiments hitherto made. 



Experiments, subsequently to be communicated more fully, 

 have shown me how wide are the limits within which the bands 

 retain their situation. In those experiments nitrogen was 

 heated by the induction-spark to 4000° and 20000° respectively, 

 and yet there were no alterations of position in the band spec- 

 trum. The temperatures were determined calorimetrically. 

 It is therefore in the highest degree improbable that the above 

 spectra ought to be ascribed to the rotations of the molecules. 

 Perhaps the frequently occurring augmentation of brightness 

 of the background from which the spectral lines stand out may 

 be referred to them. 



Spectra of Compounds. — In order, therefore, to explain the 

 band spectra of the elements, and the spectra (composed quite 

 analogously to them) of chemical compounds, we assume that 

 they are produced by the vibrations of the atoms in the mole- 

 cule or of their aether envelopes. That, at the same time, the 

 spectra of the compounds are so much more complicated than 

 the (line) spectra of the free atoms can cause us no surprise. 



If we further compare the spectra of the undecomposed 

 compounds with the band spectra possessed by the elements 

 which compose them, we find that the position and grouping 

 of the lines, as was to be expected, are essentially different in 

 the former from what they are in the latter ; for indeed the 



* Lecoq, indeed, is inclined to account for both, band and line spectra 

 by rotations of the molecules about centres not precisely determined, the 

 motion following elliptical paths, these centres again rotathig about other 

 centres, and so forth. By admitting several such centres, and by a more 

 precise determination whether the motion in the various ellipses takes 

 place in the same or in opposite directions, he succeeds in explaining the 

 various observed phenomena ; but he gives no physical reason, deduced 

 from other well-known phenomena, for the assumption of that sort of 10- 

 tatorv motions. 



