228 Johnston and Adams — High Pressures on the 



welding (p. 225), if we look upon the net effect as a breaking 

 down or crumbling of the crystals by the compressive force, 

 accompanied by a simultaneous (and, perhaps, also a sub- 

 sequent) welding together again of the grains produced in this 

 way. This way of regarding the phenomenon is supported by 

 the following observation : a small quantity (about 0*7 grams) 

 of finely powdered dry salt (NaCl) was placed in a steel bomb 

 with a hole, 1/2 inch in diameter, into which a hardened steel 

 plunger fitted tightly, and was by this means subjected to a 

 pressure of some thousands of atmospheres ; after being com- 

 pressed it appeared as a button, 13 mm in diameter and 3 min thick, 

 which was sufficiently transparent to enable one to read print 

 through it.* Thus this observation shows that, starting with a 

 substance in the form of fine powder, we can by applying suffi- 

 cient compressive force obtain a block of material altogether 

 similar to that produced by the deformation of a single crystal 

 of the same substance. 



None of these observations on the effect of differential stress 

 on crystals and crystalline aggregates conflict with the idea that 

 permanent deformation is conditioned by a real local melting. 

 Indeed many of the observations directly support this idea: 

 thus, for instance, the production of twinning on deformation, 

 as noted by Adams, would seem to be more easily accounted 

 for by considering that an actual partial melting had occurred 

 than by ascribing it to the operation of purely mechanical 

 factors ; and again, substances exhibit greater plasticity the 

 higher the temperature. Moreover we note again that there is 

 a general parallelism between relative ease of flow and melting 

 point, a parallelism which is precisely what we should expect 

 if an actual local melting takes place during the process of flow 

 of a solid. 



Correlation of the Relative Ease of Flow with Physical 

 Constants of the Material. — The conclusion that there should 

 be a parallelism between relative ease of flow and melting 

 point, although almost obvious from general principles, can be 

 derived readily from equations III or V. For, according to 

 the views advocated in this paper, the ease of welding is 

 parallel to the <£ value, which in turn depends upon the den- 

 sity, heat of melting, and temperature of melting. Now of 

 these three factors, the first lies within comparatively narrow 

 limits for a series of substances, and especially for a series of 

 minerals ; the second is not known, but probably lies also 

 between comparatively narrow limits in general *\ the limits 

 within which the melting point lies are comparatively much 



* A similar observation has been made previously by some one else (prob- 

 ably by Spring), but we have been unable to locate the reference. 



f The exceptional cases would be those in which the heat of melting is 

 very small. 



