239 



lo write R 



f=N. 



It must be remarked that, in consequence of the more or less relative 

 mobility of the molecules, N must have a different value for different 

 substances. 



16. The elementary substances, according to Boedecker's theory, 

 not entering into mutual combination with the same density and 

 correlation of atoms as they possess in the state of liberty, it may be 

 admitted that they are probably compelled to form double molecules, 

 or semi-molecules. In the second case of combination the specific 

 power of refraction would be the double, in the first only the half 

 of what it was in the elementary substance, when still chemically free. 



17. The indices of refraction of chemical combinations and 

 mixtures may be calculated on the supposition enounced in 16. 

 Absolute exactitude may be obtained for any extensive groups, of 

 which some members, and consequently the law of their chemical 

 combination, are known*. 



The general law here is Substances with nearly equal powers of 

 refraction may combine without alteration in them. If these powers 

 are very different, and far distant from each other, they tend to 

 become homogeneous, so that the greater power of refraction is 

 halved, and the lesser is doubled (see 16). 



18. The coefficient of dispersion of combined substancesf seems 

 also to be derivable from the principles stated in 9 and 16 ; as B 

 being dependent on P, must change proportionally to the square of 

 molecukr condensation (see 16). 



19. In consequence of 9, the lines of Fraunhofer (although the 

 constancy of the lengths of undulation, to which they answer, remains 

 unaltered), when refracting substances are condensed, must advance 

 their centre toward the violet end, in analogy to the phenomena of 

 absorption, observed by Weiss (Poggendorff's ' Annalen,* 1861). For 

 this reason absorption must depend on the length of the undulation. 



* Evidence is afforded by Dulong's, Dale's, and Gladstone's determinations, also 

 by all other, especially organic, combinations, on account of the simplicity of their 

 formulae ; also by Deville's, Grailich's, Handl's, and Weiss's determinations con- 

 cerning mixed substances. 



t Our knowledge of the coefficients of dispersion proper to chemical elements 

 being still very defective, it is impossible to bring this proposition to full evidence. 

 Sulphuret of carbon may serve for demonstration, and similar inferences may be 

 deduced from investigations concerning mixtures. 



s2 



