622 JOHN JOHNSTON AND PAUL NIGGLI 
Growth actually takes place only when the temperature is high 
enough for the rate of growth to be appreciable. 
The occurrence of this phenomenon in precipitates in contact 
with water is frequently used to advantage in analytical operations, 
to coarsen the grain of precipitates and so render them easier to 
filter. It has also been demonstrated on a system containing no 
water by Rinne and Boeke,’ who, by heating up calcite crystals 
of various sizes in gas-tight bombs to high temperatures (1,000°) 
for considerable periods, found that the grains had become nearly 
uniform in size. Of similar import is the well-known fact that a 
glass may be made to crystallize in a temperature region in which 
it is by no means fluid, the optimum effect occurring usually at a 
temperature 20°-—50° below its eutectic temperature. 
When rock fragments are surrounded by liquid magma, it will 
act as a solvent and corrode the fragments in a manner depending 
upon the temperature and chemical composition of both. On 
the other hand, the solid particles act as crystal nuclei toward the 
magmatic solution, at least if its temperature is such that crystal- 
lization is possible. It is to be noted that rock transformations, 
due to partial meltings and recrystallizations, may take place at 
temperatures far below the melting-points of the individual minerals 
present; for such processes the positions of the various possible 
eutectic points are the essential criteria. Thus it is well known 
that mixtures of solid components in eutectic proportions can be 
liquefied by long-continued heating at the eutectic temperature, 
which may be very much lower than the melting-point of any one 
of the components. 
The question as to whether solid fragments sink or float in the 
magma cannot be decided by density measurements made at 
ordinary temperature, but its answer involves a knowledge of 
change of volume with temperature, and requires, in addition, 
that the influence of viscosity and of gas evolution be considered. 
R. Brauns has lately directed attention to a new type of meta- 
morphism, the so-called pyrometamorphism, a term which he has 
used to denote the effects of hot gases in partially melting? some 
™ Tschermak’s Petr. Mitt., X XVII (1908), 203. 
2 This idea has also been made use of by Daly in his paper on ‘‘The Nature of 
Volcanic Action,” Proc. Am. Acad., XLVII (1911), 92-93. 
