Hajplodioritic and Related Magmas. 181 



had the composition H complete crystallization would not then 

 take place until the temperature had fallen to 1125° and the 

 final liquid has then the composition Y, exceedingly rich in 

 albite. 



If this separation of early crystals from liquid is a continuous 

 process accomplished through zoning of the crystals, it is clear 

 that this continual lowering of the temperature of final con- 

 solidation and offsetting in the composition of the liquid is 

 limited only by the eutectic albite-diopside, 1085° and 97 per 

 cent albite. There is a certain theoretical rate of cooling 

 which will give maximal zoning, in which case this limiting 

 temperature and composition of the liquid will be actually 

 attained. An increased rate of cooling will bring about under- 

 cooling, and crystallization will be completed when the liquid 

 has the composition Y or X or H, or with very rapid cooling, 

 the liquid M may be cooled below 1170° before crystallization 

 begins, in which case it will crystallize in toto without any off- 

 setting of composition. A rate of cooling slower than that 

 which gives maximal zoning will also limit the offsetting in 

 the composition of the liquid on account of a certain amount 

 of adjustment between liquid and crystals. The result is that 

 the final liquid may be Y or X, or with exceedingly slow cool- 

 ing (perfect equilibrium) H. 



However, even when the cooling is slow enough to permit 

 perfect equilibrium other factors may intervene to produce the 

 same offsetting in the composition of the liquid as does zoning. 

 These are the separation of crystals from the liquid, or from 

 some of the liquid, by their sinking or by a squeezing-out or 

 draining-off of residual liquid. Clearly the opportunity for 

 both of these, especially the former, increases with slow cool- 

 ing. The rate of cooling of any liquid of the present system 

 is, therefore, of fundamental importance in determining the 

 range of composition which will be covered by the liquid as it 

 changes during crystallization, and likewise the range of com- 

 position of the crystalline products. 



The Significance of the Results in Petrologic Problems. 



The aim of experimental investigations of silicate melts is, 

 of course, the explanation of some of the multitudinous, more 

 or less disconnected facts concerning igneous rocks that have 

 accumulated with the advance of descriptive petrography. It 

 still is, and possibly always will be, a considerable extrapola- 

 tion from the simple systems that can be investigated quantita- 

 tively to the more complex systems, or perhaps rather system, 

 represented by magmas. Nevertheless, considerable progress 

 is being made in the shortening of this extrapolation. It is 



