ON DIFFUSION IN SOLIDS. 367 
films and threads of metallic silver make their appearance when a 
current is passed through a crystal under the microscope. Great 
weight is to be attached to microscopical. investigations of this kind. 
Thus, it was found that-solid silver chloride remained unchanged -in 
appearance, and no polarisation could be observed. when a current 
was passed through it, although the resistance diminished in a remark- 
able manner, '*4 but subsequent-examination !*° proved-that. the increased 
conductivity was due to microscopic filaments of metallic silver. 
Metallic alloys do not conduct electrolytically. A series of experi- 
ments in which two cylinders of different metals were placed in close 
contact, a current being passed, showed that no transport of matter 
took place, even when considerable. .currents were used.°* This. is 
entirely in accordance with older observations. An. isolated result, 
that carbon could be caused. to travel with the current in steel,*” is 
probably susceptible of another explanation. 
Diffusion. in Minerals. 
The question whether true solid diffusion ever occurs in minerals 
is very difficult to answer. The term ‘ diffusion’ has sometimes been 
employed rather loosely in geological writings, so that a complex rock 
such as granite has been spoken of as diffusing into surrounding schists, 
a comparison being made with Roberts-Austen’s experiments with 
metals.1°® In such a case the action is evidently one of mechanical 
percolation. The most favourable instances might be expected to be 
the secondary replacement of minerals and the formation of pseudo- 
morphs. Attention has been drawn to the fact that whilst certain 
alterations, such as the conversion ‘of olivine into serpentine, obviously 
proceed along cracks in the original crystal, other changes are of a 
much finer character, and must be assumed to depend on a molecular 
porosity.15° The writer has not, however, been able to find, from the 
examination of photo-micrographs and of rock-sections, that any cases 
of alteration can be attributed unhesitatingly to diffusion, as even when 
cracks are absent the alteration very commonly proceeds along cleavage 
planes. It is possible, but by no means certain, that such reactions 
proceed by true diffusion. 
It has been urged as a fatal objection to the occurrence of solid 
diffusion in crystalline substances that crystals of native minerals are 
found in which successive zones of minerals known to be isomorphous 
with one another are in‘intimate contact, and yet there is no indication 
of inter-diffusion having taken place, although geological ages have 
elapsed since the crystals have been deposited from a fluid magma. 
Augite and the triclinic felspars often.exhibit zonal structures in great 
perfection. In the case of metals, as described above (p. 353), zonal 
structures occur when the rate of cooling of a liquid depositing crystals 
F 134 M. Le Blanc and F. Kerschbaum, Zeitsch. Blektrochem., 1910, 16. 242, 
» 135 Jbid., 680. 
86 J, Kinsky, Zettsch. Elektrochem., 1908, 4, 406. 
87 J, Garnier, Compt. rend., 1893, 116, 1449. 
88 HB, Greenly, Geol. Mag., 1903 [iv.], 10, 207. 
189 C. A. McMahon, Brit. Assoc. Reports, 1902, 589 
