250 Johnston and Adams — High Pressures on the 



Now imagine our molecular complex exposed to uniform 

 pressure: it would merely be compressed, and would regain 

 its original state when freed from compression. Following out 

 this conception, too, we should expect that uniform pressure 

 would increase the stability of the complex ; and this is what 

 we find, since uniform pressure raises the melting point of 

 practically all substances. 



With unequal pressure, on the other hand, it is otherwise. 

 For if, the temperature being kept constant, a gradually 

 increasing one-sided pressure — that is, a shearing stress — be 

 applied to the complex, it will ultimately overcome the resist- 

 ance offered by the complex, causing a dissociation of the latter 

 into one or more simpler molecular species ; and 4f this molec- 

 ular species is that characteristic of the liquid phase of the 

 substance, we have a melting — or something exactly equivalent 

 to it — which takes place at a perfectly definite pressure, the 

 magnitude of which depends on the temperature. Now, if 

 our particle is within a crystal — that is, if it is surrounded by 

 a configuration of similar particles, arranged, as it would seem 

 they must be if we are to account for compressibility, in such 

 a way that there are spaces of some kind between the particles 

 — the melted parts will flow away, thus relieving the stress, 

 and quickly recombine to form the complex which is stable 

 under the particular conditions ; for no conditions could be 

 imagined more favorable to recrystallization, especially if there 

 be any reality corresponding to the idea that a particle exerts 

 influence within the space surrounding it. According to this 

 view, then, we need scarcely be surprised if the particle in 

 recrystallizing regained its original orientation. 



It may be urged that the general hypothesis is inapplicable 

 in the case of substances — for instance, quartz — which form 

 highly viscous melts or which dissociate on melting ; but these 

 objections lose weight if we remember (1) that viscosity of the 

 melt will result only in an increase in the apparent melting 

 pressure by an amount sufficient to overcome the viscosity ; 

 (2) that the condition of the melt as obtained under ordinary 

 pressure is absolutely no criterion of its condition under high 

 pressure ; indeed, it appears to be a fact in harmony with the 

 views outlined above that pressure tends to inhibit dissocia- 

 tion, or to favor recombination. Moreover, it is in no wise 

 essential to the success of the hypothesis that the substance 

 should in every case completely recrystallize ; for, since the 

 total fraction which melts during the deformation is in all 

 probability very small indeed, it seems to us doubtful if even 

 the most refined optical methods now in use are sensitive 

 enough to enable us to detect with certainty the presence of 

 the small quantity of sub-cooled liquid resulting from the 



