CunnincHam— Crystallization of Minerals in Igneous Rocks. 389 
temperature, and then begin to diminish again, also, perhaps, 
more or Jess gradually. We assume that a number of other pro- 
perties also change more or less gradually at the same time. Thus, 
if we could measure the viscosity of such a substance as we gradu- 
ally raised its temperature, we should probably find that its 
viscosity began to diminish at first slowly, then more rapidly 
until its rate of decrease of viscosity with rise of temperature (=) 
attained a maximum at the melting temperature. Its viscosity 
would then begin to diminish less rapidly with rise of tempera- 
ture, as it passed into the liquid state. A perceptible “ softening”’ 
may thus be taken to indicate the first upward bending of the 
corner B, figs. 2 and 38, which may now be looked upon as 
curves showing the relation between general properties (ordinates) 
and temperature (abscissee) to a suitable scale.t It must be 
remembered that a gradual passage from solid to liquid (fig. 3) 
is highly characteristic of amorphous substances,’ whereas a 
sudden change from solid to liquid (figs. 1 and 2) is charac- 
teristic of crystalline bodies.* Indeed, we know that quartz does 
undergo such a sudden volume change on melting as indicated by 
the observation of Moissan,* and by the density of quartz. 
Hence, I infer that the other physical properties of quartz also 
change suddenly on melting. Further, the fact that molten silica 
may be quenched in cold water without developing a single 
erack,’ and will cool down as the amorphous substance, with 
1 Of course certain properties will diminish with rise of temperature, but then we 
can think of these curves (figs. 1 to 4) as having the reciprocal of such properties for 
ordinates. It is the rate of increase or decrease of the physical properties which 
concerns us here. 
2 The gradual expansion of amorphous silcates is illustrated by Dr. Joly’s experi- 
ments on mineral glass beads: (‘‘ On the volume Change of Rocks and Minerals 
attending Fusion,”’ Trans. Roy. Dublin Soc., Ser. 11., vol. vi. (1897), p. 283). The 
expansion curves show very much rounded corners corresponding to B, fig. 3., but 
they lack the upper portion C. The volume changes obtained by Dr. Joly’s method do 
not, of course, exhibit the volume changes that go on in rocks when crystallizing, but 
probably only a small part of them. 
3 Cf. Preston’s ‘‘ Theory of Heat,” pp. 270 and 286. london, 1894. 
£M. H. Moissan, ‘‘Le Four Electrique,” 7), GOs Iara, Wee, Il, Sis I 
volatilisation de la silice,’? &c. Comptes Rendus, 116 (1893), p. 1222. 
5 Of. Shenstone and Lacell, ‘‘ Working Silica in the Oxy-gas Blowpipe Flame,”’ 
Nature, vol. lxii., p. 20. (8rd May, 1900); and C. V. Boys, “Quartz Fibres,’’ _ 
Nature, vol. xlii. (1890), p. 605. 
