262 A r . L. Bowen — The Ternary System : 



In summing up, then, it may be stated that any igneous 

 magma or mineral melt may best be regarded from the view- 

 point of the kinetic theory : all possible combinations and dis- 

 sociations are going on simultaneously. At any temperature 

 all the reactions have a definite rate and at equilibrium there are 

 definite relative amounts of all the constituents. On cooling 

 such a magma the first substance to crystallize will be that one 

 of all those present wmich happens to reach first its point of 

 saturation with respect to the liquid under the prevailing con- 

 ditions. Under different conditions there may crystallize first 

 an entirely different substance. This is true of the substance 

 which crystallizes second and so on till all is solid. For this 

 reason the same magma might give rise to different assem- 

 blages of minerals under different conditions. The. whole 

 reduces itself to a question of relative solubilities (relative 

 degree of saturation) considered in conjunction with equili- 

 brium between the molecules in the liquid phase. 



Freezing-Point Lowering in Silicate Mixtures. 



Many attempts have been made to calculate freezing-point 

 lowering in silicate melts and therefrom to calculate eutectic 

 points of various pairs of minerals. Usually the law of freez- 

 ing-point lowering in dilute solutions has been used. It is 

 well known that this law cannot apply to concentrated solutions 

 and that even the form of equation used for concentrated 

 solutions need correction factors when there is association or 

 dissociation in the liquid phase. The binary system, diopside- 

 forsterite, exhibits an extreme case of variation from the 

 simple rules. The addition of forsterite to diopside depresses 

 the freezing-point an almost inappreciable amount — the eutec- 

 tic point at the composition 12 per cent forsterite is only 

 about 4°-5° below the melting-point of diopside. The micro- 

 scope shows that there is no solid solution, so the explanation 

 of this slight depression must be sought in a different direc- 

 tion. It has already been stated as probable that diopside is 

 dissociated in the liquid state and a possible equation for this 

 dissociation was given. 



2CaMgSi 2 6 ^=± 2CaSi0 3 +Mg 9 Si0 4 + SiO s . 



If forsterite is added to a liquid containing all the above mole- 

 cules, a necessary result is the driving of the reaction from 

 right to left and, therefore, the actual concentration of diop- 

 side molecules may be changed very little. At the freezing- 

 point of diopside the complex liquid is just saturated with 

 diopside molecules. The addition of forsterite (Mg. 2 Si0 4 ), if it 

 affects very little the concentration of diopside molecules, 

 should therefore change the freezing-point very little. The very 



