ON DIFFUSION IN SOLIDS. 365 
In alloys which contain two solid solutions in equilibrium with one 
another, such as the af alloys of copper and zinc, the structure also 
becomes coarser when the alloys are heated at a temperature at which 
diffusion takes place. 
The increase in size of crystals is equally pronounced when only a 
single constituent—a pure metal or a solid solution—is present. The 
growth of the ferrite grains in soft steel at 700°-720° is extremely 
rapid,’’* and the process may be very conveniently studied in such a 
solid solution as 70:30 brass.’"* The growth of crystals may be 
watched in lead even at 60° if the metal has been previously 
strained.'* It is always the larger crystals that absorb the smaller. 
Whether the metal in which such recrystallisation takes place is 
homogeneous or heterogeneous, diffusion must occur in order that the 
rearrangement may come about. The effect has been explained by the 
principle that small crystals have a greater solubility than large, so 
that if small and large crystals of the same substance are both in 
presence of a solvent, solution and re-deposition tend to go on *** until 
only crystals above a certain size are present. This has been verified 
for the case of calcium sulphate in water.”’ A thermo-dynamical 
explanation has also been given’ of the principle that the bounding 
surface between a crystal and its saturated solution tends to become 
a minimum, so that equilibrium is only finally reached when all the 
small crystals have united to form a single crystal. 
The principle of differing solubility is rejected as an explanation by 
G. Tammann,*”® who assumes that the surface tension, which is less 
than the forces producing rigidity in a crystal at the ordinary tempera- 
ture, may become much more considerable with increase of tempera- 
ture. When the surface tension exceeds the opposing forces, two 
crystals unite as two drops of fluid would do. The hypothesis is 
ingeniously applied to explain the recrystallisation of strained metals. 
The Influence of the Amorphous Modification in Metals. 
The view has been taken !*° that the occurrence of diffusion in 
heated metals is not a proof that diffusion may occur in a crystalline 
medium, as Spring’s experiments have rendered the presence of amor- 
phous material in metals extremely probable. Since then the existence . 
of an amorphous modification in worked metals has been proved by 
much experimental work,'?! but it has also been shown that complete 
recrystallisation occurs at very moderate temperatures, such as 250° 
for copper and 300° for silver.'*° 
us J. E. Stead, Jour. Iron and Steel Inst., 1898, i. 145; A. Joisten, Metallurgie, 
1910, 7, 456. 
u4 G. Charpy, ‘ Etude des Alliages’ (Paris, 1901), p. 1. 
us J. C. W. Humfrey, Phil. Trans., 1902, 200A. 225. 
6 W. Ostwald, Zeitsch. Physikal. Chem., 1900, $4, 495. 
7G, A, Hulett, Zeitsch. Physikal. Chem., 1901, 37, 385. 
8 J. Curie, Bull. Soc. Franc. Min., 1885, 8, 145. 
9 Nachr. K. Ges. Wiss. Gottingen, 1912, 1. 
120 G. Bodlinder, N. Jahrb. Min. Beil. Bd., 1899, 12, 52. 
121 G. T. Beilby, Phil. Mag., 1904 [vi.], 8, 258. 
= G. T. Beilby, Jour. Inst. Metals, 191, 6, 5. 
