RELATIONS TO OTHER SCIENCES 599 



first minerals to crystallize out, even though, in common with the 

 ores, they sometimes show a not inconsiderable period of separation. 

 It is to be doubted if the early solidification of these accessories is really 

 due to their small proportionate amount, as is often believed, for since 

 the solution seems to be diluted with reference to them, they should 

 not crystallize out until quite late. Since it is, also, not permissible 

 to adopt the rather drastic view that the magma strives first to rid 

 itself of these foreign bodies, it is therefore preferable to assume 

 that these minerals are especially difficult of solution in the silicate 

 solution at the lower temperatures. 



The causes of the variable behavior of the characteristic constitu- 

 ents as regards order of crystallization are still in large part unknown. 

 The treatment of the problem is, however, made especially difficult, 

 since both experimental and theoretical considerations have been 

 accustomed to assume only two substances in solution, whereas a 

 silicate-rock magma usually contains more than four substances in 

 solution contemporaneously. 



Attention is drawn to the fact that in certain solutions the range of 

 temperature appropriate for the separation of one compound, e. g., 

 leucite, may be very limited, while, under otherwise the same condi- 

 tions, the range appropriate to the separation of another compound, 

 e. g., augite, may be much greater. Thus, according to the tempera- 

 ture conditions, one and the same magma may yield up its augite 

 now before, now after, its leucite. Furthermore, Meyerhoffer has 

 shown that, according to the labile equilibrium, now a and now 6 

 may first crystallize out of the same slag. 



The order of separation may be affected by yet another factor, viz., 

 pressure. Since, according to the usual view, the rock-forming min- 

 erals contract as they crystallize out of their magma, then, as Sorby 

 and Bunsen have shown, increased pressure must aid this contraction, 

 i. e., must accelerate crystallization. The concomitant shifting of the 

 temperature of solidification (freezing-point) takes place unequally 

 for different substances. Thus two substances which under ordinary 

 atmospheric pressure have different freezing-points, may under some- 

 what greater pressure, and consequently more nearly equal melting- 

 points, freeze at the same temperature, while under the influence of 

 yet higher pressure the normally quicker-freezing substance may be- 

 come the slower-freezing one. On the ground of such changes, then, 

 the order of separation may be altered, as shown for example by the 

 varying relations between the more easily fusible augite and the 

 more difficultly fusible orthoclase. 



According to Doelter, the rate of crystallization may also be of im- 

 portance, in so far as the advantage (start) in separation given to the 

 substance a by its lower solubility may be equaled or overbalanced 

 by the tendency to more rapid crystallization possessed by the more 



