296 
PROFESSOR J. A. EWING AND MR. W. ROSENHAIN 
cease to be visible, but the presence of the impurity makes itself felt by the formation 
of deep grooves or channels along the inter-crystalline boundaries on etching. Even 
a very minute amount of impurity would suffice to form a thin but practically 
continuous film of eutectic in the crystalline boundaries. 
The close analogy between alloys and salt solutions has been so fully worked out 
that it is barely an assumption to say that a eutectic alloy in contact with crystals of 
one of its constituents will behave in much the same way as a saturated solution of a 
salt in contact with crystals of that salt would do. Our view, then, is that there is 
constant diffusion from the surface of the crystal into the eutectic film and equally 
constant re-deposition of metal upon the crystal from the eutectic film. If there are 
several crystals in contact with the same eutectic, then there will be, under some 
conditions, a state of dynamic equilibrium between them, the amount dissolved from 
each being exactly counterbalanced by the amount deposited upon it; if, however, 
there is any difference in the solubility of various crystals, in other words, if there 
is any difference in their “ solution pressure ” in respect to the eutectic, then the 
less soluble will grow at the exj^ense of the more soluble. In the case of salt crystals 
in an ordinary solution it is well known that the large crystals gradually absorb the 
small ones, a transformation generally explained on the ground that a system tends 
to assume a position of minimum potential energy. In the case of metallic crystals 
in a solid metal we have, however, this special condition, that the eutectic (or solu¬ 
tion) exists as a mere thin film in contact with only one face of one crystal on either 
side. If, therefore, these two crystal faces differ in solubility or “ solution pressure ” 
in the eutectic, gradual transfer of the metal by diffusion through the eutectic film 
from one crystal to the other would result, and the one crystal would grow at the 
expense of the other. This action might go on while the eutectic was solid—Sir W. 
C. Roberts-Austen has demonstrated that diffusion does occur in solid metals,* but 
at extremely slow rates, unless the metals are near their melting-points. The metal 
constituting the eutectic films, being much nearer its melting-point than the rest of 
the mass, would thus be favourable to comparatively rapid diffusion, but the rate of 
such diffusion and, consequently, the rate of growth of crystals, would be enormously 
increased by heating the metal to a temperature above the melting-point of the 
eutectic in question. 
The theoryf which we suggest to explain crystal growth in a solid metal depends 
upon the existence of a difference in the solubility of the two crystal faces in 
contact with a eutectic film. The only difference between these two faces is, appa¬ 
rently, in the orientation of the crystalline elements; but this very difference is 
sufficient to produce a difference in the rate of solution of such surfaces in an acid. 
A polished surface of metal when attacked by an acid is not attacked at a uniform 
* Bakerian Lecture, “Diffusion of Metals,” by W. C. Roberts-Austen, ‘Phil. Trans.,’A, 1896, vol. 187, 
pp. 383-416; “Diffusion of Gold into Lead,” ‘Proc Roy. Soc.,’ May 5th, 1900. 
t It is proper to say that the credit for this theory belongs to Mr. Rosenhain. —J. A. E. 
