CRYSTALLIZATION OF IGNEOUS ROCKS. 
97 
would make it necessary for the temperature to be lowered 
before the proper degree of saturation was reached. 
The effects of temperature and pressure on such solutions 
near their point of crystallization are consequently the re¬ 
verse of one another. An increase of temperature hinders 
crystallization; a decrease of temperature induces it. An 
increase of pressure induces crystallization; a decrease of 
pressure hinders it. 
It is well known that salts differ greatly in their solubility 
and in the tendency they possess to saturate a solution; or, 
in other words, that a solvent may take up different amounts 
of various substances before becoming saturated. The amount 
is generally greater as the temperature is higher. A solution 
that is saturated at a certain high temperature becomes 
supersaturated as the temperature is lowered and parts with 
the excess of the dissolved substance, which crystallizes out. 
It is possible to render a solution highly supersaturated by 
cooling it very gradually and quietly. 
The relative tendency of various chemical compounds that 
occur in rocks to saturate siliceous magmas has been investi¬ 
gated by Pelouze * and others, who experimented on artificial 
glass to discover what amount of these compounds could be 
introduced into the glass without producing devitrification 
or crystallization. The results of these experiments are 
summed up by Lagorio,f in the paper already cited, as fol¬ 
lows : “ Potash silicates saturate molten silicate solutions 
with great difficulty, the same is true of alumina; soda 
silicates more easily, then follows CaO, MgO, as well as 
the oxides of the heavier metals (FeO more difficultly than 
Fe 2 0 3 ), as well as titanium and zirconium. Si0 2 saturates 
silicate solutions with difficulty.” Hence the order in which 
these compounds tend to saturate silicate solutions, com¬ 
mencing with that having the strongest tendency, is as fol¬ 
lows : Ti and Zr, Fe 2 0 3 , FeO, MgO, CaO, Na 2 0, Si0 2 , Al 2 0 3 , 
K 2 0, the position of silica not being definitely stated. 
* Comptes rendus Acad. Sci., Paris, vol. 64, 1867, p. 53. 
t Loc. cit., p. 501. 
11-Bull. Phil. Soc., Wash., Vol. 11. 
