248 Johnston and Adams — High Pressures on the 



These results are in agreement with the geological evidence 

 and with what one would expect from general principles ; they 

 required to be demonstrated practically only because state- 

 ments to the opposite effect had been made and, apparently, 

 accepted. 



The Interdiffusion of Solids. 



Closely related to the subjects discussed above is the question 

 of the interdiff usion of solids. Various early isolated observa- 

 tions had been made,* but the first thorough investigation was 

 made by Roberts-Austen,f on systems composed of two 

 metals. Spring^ also made some experiments along this line, 

 and since that time a few isolated observations have been 

 recorded, amongst them those already referred to on the sys- 

 tem BaC0 3 - Na 2 S0 4 , and those of Masing on alloys. 



The general conclusion from this work is that diffusion 

 between solids is appreciable, but only in those systems in 

 which solid solutions can form, as indeed one might expect a 

 priori; and that the rate of diffusion increases enormously 

 with rise of temperature. As regards the effect of uniform 

 pressure on the rate of diffusion, some experiments made in 

 this laboratory indicate that subjecting a system to uniform 

 pressure is accompanied by an increased rate of diffusion ; but 

 it is premature to consider this as established, and it would be 

 still more hazardous to generalize from this that uniform pres- 

 sure should always have such an effect. 



General Remarks. 



To some the hypothesis made use of in the foregoing dis- 

 cussion — namely, that every permanent deformation of a crys- 

 talline aggregate is conditioned by a real local melting — may 

 appear to be of rather daring character, despite the fact that 

 it has been shown to reconcile successfully some divergences 

 hitherto outstanding. In reply to such persons, we would 

 observe that the criterion by which a hypothesis is to be 

 judged is not so much its inherent fitness to take a place in 

 the group of our preconceived notions of the behavior of 

 matter, as its usefulness in correlating known facts and in sug- 

 gesting new inquiries. And it is precisely on this account 

 that we deem the hypothesis here advocated to be satisfactory, 

 since it does provide a definite basis for new experimental 

 work ; whereas the older hypotheses tend to be mere post 

 facto descriptions of the phenomena, and to lead nowhither. 



* See Winkelmann's Handbuch der Physik, 2 Aufl., vol. i, 1505. 

 fPhil. Trans. Eoy. Soc. London, clxxxvii, 383, 1896. 



JZs. physik. Chem., xv, 65-78, 1894; Bull. Acad. Eoy. Belg. (3), xxviii, 

 23-46, 1894. 



