Chemical Equilibrium of Solids. 35 



examination of mercury and solution of zinc sulphate. Both 

 coefficients in this case were negative, and I was not prepared 

 for the positive coefficients (8R/R . Sp) so marked in the 

 present paper. The data as a whole, therefore, show that it is 

 not permissible to pass from one substance to another ; doubt- 

 less because the number of molecules which must split up per 

 unit of area to discharge the field varies largely from sub- 

 stance to substance. Inferences are to be drawn from one 

 and the same solvent, either electrolytically dissociated by 

 temperature, or modified by the solution of some given sub- 

 stance, or both. / 



39. The remarks of § 38 are substantiated by the incidental 

 results of Table VI., where the number of ions is increased by 

 dissolved dirt. It agrees with the negative pressure-coefficient 

 found for zinc sulphate (I. c). Moreover, by keeping the 

 thick machine-oil at 310° for two or more hours, I found 

 that its resistance continually decreased, while the pressure- 

 coefficient passed from indistinctly positive to pronounced 

 negative values. Dissociation is here to be inferred. 



I have only touched upon this question in passing ; but if 

 it can be proved that the occurrence of negative pressure- 

 coefficients is an index of solution, then metals may also be 

 regarded as holding ions in solution. In general a sequence 

 of changes of state or of molecule corresponding to gas-liquid, 

 liquid-solid, may be regarded as crowded into the solid state, 

 though the external manifestations are here no longer obvious 

 at once. 



40. The present research has been excessively tedious and 

 elusive. Moreover the final values for the pressure-coefficient 

 of glass (Tables VII., VIII.), although consistent, do not 

 show the uniformity of values which 1 had hoped to reach. 

 Nevertheless the results are marked, and unless there be some 

 occult behaviour of polarization, I do not believe that I have 

 been led astray in my inferences. In so far as my results 

 are correct, they contain the first direct and connected evi- 

 dence of the change of chemical equilibrium with strain. 



Regarding the behaviour of liquid insulators like the hydro- 

 carbon oils, the results found present many points of interest, 

 and I have no doubt that a more systematic study of the 

 subject than I have been able to make would throw much 

 light on the details of the mechanism by which electricity is 

 conveyed. 



Physical Laboratory, U.S. Geol. Survey, 

 Washington, D.C. 



D2 



