ON DIFFUSION IN SOLIDS. 349 
Diffusion in Glasses, and Devitrification. 
The production of coloured glass by the application of a layer of 
strongly coloured glass to a thicker colourless mass involves diffusion, 
but the temperature employed is so high that the glass is soft, and 
cannot be regarded as solid. The colouring matter does not appreciably 
diffuse at atmospheric temperatures. Thus, a sheet of fourteenth- 
century ruby glass, 5 mm. thick, was found on microscopical examina- 
tion to be built up of 55 coloured and 88 colourless layers, the 
boundaries between which were yet distinct, although under a higher 
magnification zones of intermediate shade, which may have arisen 
during the process of manufacture, could be detected.* The observa- 
tions on diffusion in glass at moderate temperatures are almost entirely 
limited to experiments depending on electrolysis, which are described 
in the following section. The process of devitrification, however, 
involves diffusion, although only through comparatively small distances, 
and the evidence on this point may be briefly reviewed. 
The formation of crystals in a glassy material (devitrification) is 
strictly comparable with the process of crystallisation from a liquid, 
and its course depends on two factors—the number of centres appear- 
ing in a given time—and the linear velocity of crystallisation.‘ 
Quantitative measurements have been made in the case of glassy 
borates,° and numerous observations of the devitrification of amorphous 
organic compounds have been recorded.* In silicate glasses the 
crystals formed are usually very minute. They may become so 
numerous, when the glass is maintained for a long time at a suitable 
temperature, as to result in complete opacity (Réaumur’s porcelain).? 
The formation of much larger crystals during the slow cooling of glass 
has been observed in a few cases. Thus, a quantity of 400 tons of 
glass having escaped from a furnace, the slow cooling of the mass led 
to the formation of spherulites.* In this case the glass was an almost 
pure mixture of silica, lime, and soda, and the crystals were entirely 
composed of wollastonite. In a similar accident, in which the glass 
also contained much magnesia, very large spherulites of diopside, up 
to 4 cm. in diameter, were formed, together with wollastonite.® It is 
to be noted that the spherulitic mode of crystallisation is most, liable 
to occur when the crystals separate at a temperature below the 
metastable limit.?° 
Similar conditions are observable in blast-furnace slags, which may 
be of any texture from that of a pure glass to that of a stony, minutely 
crystalline mass. A large number of minerals have been identified in 
3 J. Fowler, Archeologia, 1880, 46, 65. 
4 G. Tammann, Zettsch. Elektrochem., 1904, 10, 532 ; ‘ Krystallisieren und Schmel- 
zen,’ p. 148. 
5 W. Guertler, Zeitsch. Anorg. Chem., 1904, 40, 268. 
6 QO. Lehmann, ‘ Molekularphysik ’ (Leipzig, 1888), i., 707. 
7 Experiments conducted 1727-29. 
8 F, Fouqué, Compt. rend., 1899, 109, 1. 
9 L, V. Pirsson, Amer. Jour. Sci., 1910 [iv.], 30, 97. 
10 J, Chevalier, Min. Mag., 1909, 15, 224. 
