On the Chemical Equilibrium of Solids. 9 



This gives E c for fluorine = O88, and, again, a scarcely 

 appreciable negative dispersion. 



As the six metallic fluosilicates are all doubly refractive, 

 and we do not know how to deal with that phenomenon, we 

 have not endeavoured to make any precise calculation, espe- 

 cially as the dispersion of the metallic element in each case 

 has not yet been determined. Mr. George Gladstone has 

 already shown that they give a probable refraction-equivalent 

 for fluorine varying from 0*62 to 0*84; and it is quite evident 

 that the dispersion is excessively small, even if it be not a 

 negative quantity. 



These results all tell the same tale as fluorbenzene in regard 

 to the optical properties of fluorine. It is true the figures do 

 not appear very accordant ; but it must be remembered that 

 all the errors of experiment as well as those arising from 

 impurity of substance fall upon these residual numbers, and 

 where the numbers are so very small discrepancies appear 

 important which would otherwise scarcely be noticed. The 

 values taken for the atomic refraction and dispersion of silicon 

 and the metals must be looked upon as merely approximate ; 

 but, whatever rectifications may be made in them hereafter, 

 it is hardly possible that they should affect our general con- 

 clusions, viz. that the atomic refraction of fluorine in combi- 

 nation is extremely small, in fact less than 1*0, and that it is 

 smaller for the violet than for the red end of the spectrum. 



II. The Chemical Equilibrium of Solids, in its Relation to 

 Pressure and to Temperature. By Cakl Bakus*. 



1. FN a recent number of the ' American Journal/ f I com- 

 J- municated certain data on the effect of dilatational 

 strain on the electric resistance of glass. These experiments, 

 though decisive as to results, were somewhat complex in cha- 

 racter ; and from their important bearing on the chemical 

 equilibrium of solids, I resolved to attack the subject from a 

 new point of view. The present work is direct, and I hoped 

 that results could thus be reached without necessitating auxi- 

 liary hypotheses or experiments. In this expectation 1 de- 

 ceived myself : the insulating oil, which transmits pressure to 

 glass, shows a pressure-coefficient usually much larger than 

 the glass itself. Both must therefore be studied. Neverthe- 

 less, the confirmation of my earlier results which the present 



* Communicated by the Author. 



t Amer. Journ. xxxvii. pp. 339 to 351 (1889). 



