Physical and Chemical Behavior of Solids. 211 



gunpowder to high compression without causing it to explode. 

 Now uniform pressure acting equally upon both the solid 

 phase and the liquid phase (produced on fusion) causes a low- 

 ering of melting-point only in a very few exceptional cases* 

 and hence its action is incompetent to account for the facts. 



It can be shown, however, that if the pressure (P) acts on 

 the solid phase but not, or not to the same extent, on the liquid 

 phase, the melting-point of the solid is always lowered, and by 

 an amount which is many times as great as when the same 

 pressure (P) acts on both phases simultaneously. f For pres- 

 sure of this type, we shall, for convenience in what follows, use 

 the term u unequal " pressure. Such unequal pressure may be 

 pictured as occurring when an aggregate of fine crystals is 

 compressed, the liquid formed by the melting flowing into the 

 interstitial spaces, and in this way being subject to a smaller 

 pressure than the adjacent solid particles.^ Moreover, if the 

 possibility of a partial melting under stress be admitted, the 

 force can be imagined only as acting on the solid to a much 

 greater extent than on the liquid phase. 



In formulating the assumption made use of throughout this 

 paper, we have stated " that every permanent deformation of a 

 crystalline aggregate is conditioned by, and consequent upon, 

 a real melting," using the word aggregate advisedly. Firstly, 

 because the experimental work which we are here considering 

 deals exclusively with systems made up of a large number 

 of crystals not regularly oriented ; secondly, because systems 

 composed of a single crystal, or possibly also of a number of 

 similar ciwstals similarly oriented, exhibit phenomena which 

 can be accounted for without assuming that a partial melting 

 has taken place. The phenomena especially referred to are 

 those connected with the presence in crystals of what are 

 termed cleavage and gliding planes ; such properties appear to 

 be somewhat analogous to the quality of brittleness in metals, 

 though of course metals, being agglomerates, do not show a 

 perfectly regular fracture. The above analogy holds moreover 

 in this, that, while the viewpoint adopted in this paper serves 

 to correlate all those properties of metals for which quanti- 

 tative measurements have been made, it does not aid us in 



* Water (though only with pressures less than about 2300 atm.) and 

 bismuth are the best known examples. 



fCf. Ostwald, " Lehrbuch der Allgemeinen Chemie," 2 Aufl., ii, II, 374- 

 9 ; Roozeboom, " Heterogene Gleichgewichte," i, 213-7. (The main part of 

 both of these is quoted by Doelter, Physik. Chem. Mineralogie, Leipzig, 

 1905, pp. 160-2). Cf. also equation IV, p. 214, postea. 



Jit will of course in general freeze again immediately ; (cf. postea, p. 215, 

 where other examples are given) but failure to do so is not necessarily 

 prejudicial to the success of the hypothesis — so long, at least, as the amount 

 melted is a very small fraction of the whole, a condition which in practice 

 would seem to be always satisfied (see also p. 250). 



