402 



N. L. BOWEN 



later stages of crystallization, and the evidence of it would be as 

 obvious and unfailing as the evidence of crystallization itself. 

 The complete collection of all the immiscible liquid as a separate 

 and distinct layer is as unlikely as the complete collection of a kind 

 of crystals whose separation continues until a late stage. 



We may perhaps make clear these facts regarding immiscibility 

 by discussing the simplest possible binary example. 'Figure i 

 presents the temperature-composition relations. When a liquid 



of composition x is 

 cooled to the tempera- 

 ture FK, liquid of com- 

 position K, that is, a 

 liquid rich in B, begins 

 to separate from it, and 

 as cooling proceeds the 

 composition of the one 

 liquid changes along /^E 

 and of the other along 

 KD. The liquid repre- 

 sented by points on FE 

 decreases in amount, 

 and that represented 

 by points on KD in- 

 creases in amount. 

 The first separation of 

 liquid must be repre- 

 sented by the forma- 

 tion of minute nuclei 

 that grow to larger and 

 larger globules as the cooling proceeds, and as a result of the 

 slow diffusion of material to these globules. There is no reason 

 why this process should be accomplished any more rapidly for 

 separated liquid than for separated crystals. If the separated glob- 

 ules were heavier than the general mass of liquid they would sink, 

 and here enters the possibility of the growth of these globules to 

 much larger dimensions than crystals, because two globules encoun- 

 tering each other may coalesce. The formation of very large 



Fig. I. — Illustrating behavior of a binary mixture 

 with partial miscibility. 



