258 JV. L. Bowen — The Ternary System : 



separated should be redissolved and that mix-crystals already 

 separated should change their composition, whereas in the 

 second type no such changes take place. Between the extremes 

 of perfect adjustment of equilibrium between the liquid and 

 the early-formed crystals and complete lack of adjustment 

 there is every stage of partial adjustment. If the mechanism 

 of crystallization is considered, it is apparent that perfect 

 adjustment must be exceedingly difficult to obtain. In order 

 to change the composition of a mix-crystal there must be a 

 diffusion of material in the solid crystal itself and an inter- 

 change of material between it and the liquid. In the mean- 

 time crystallization of a further quantity of mix-crystal of 

 somewhat different composition is in progress and is almost 

 certain to take place as a layer or zone about the early crystal 

 as a nucleus, rather than to form a new nucleus. Unless dif- 

 fusion in the solid can take place fast enough to bring about a 

 change, the composition of this outer layer w T ill remain, dif- 

 ferent from that of the layer it surrounds, while still other 

 layers are added. Cooling at an exceedingly slow rate might, 

 perhaps, bring about perfect adjustment, but usually zoned 

 crystals will result. 



It is well known that zoned crystals are common in natural 

 rocks. The significance of the formation of zoned crystals is 

 great because, as has been shown in the discussion of crystalli- 

 zation of the second type, the temperature of final solidifica- 

 tion is thereby very much lowered — the temperature and time 

 interval of crystallization, therefore, much increased.* 



Even when the cooling is extremely slow, a factor other 

 than zoning may intervene to bring about the same result of 

 preventing the early crystals from participating in any inter- 

 change with the liquid necessitated by equilibrium at a lower 

 temperature. This factor is the rising or sinking of crystals 

 under the action of gravity as they grow in the liquid. 



Consequences of the Possible Sinking of Early Crystals in a 



Fluid Magma. 



The idea that the sinking of crystals in a fluid magma has 

 some importance in the differentiation of igneous rocks has 

 recently been gaining ground. That such action must take 

 place seems hardly debatable, only its relative importance is at 

 issue. 



In the melts here described, from our precise knowledge of 

 equilibrium, we are able to determine definitely the effect of a 

 possible settling of crystals. In discussing the crystallization 



* See also N. L. Bowen : The Melting Phenomena of the Plagioclase Feld- 

 spars, this Journal (4), xxxv, 597, 1913. 



