246 /. H. L. VOGT 



four hours). Even though these figures cannot be transferred 

 directly to pegmatite dikes formed under great pressure, it is 

 evident, at any rate, that their crystalUzation did not necessarily 

 require an exceptionally long time, speaking geologically. Granite- 

 pegmatite dikes, commonly only 5, 10, 20 m., seldom as much as 

 50-100 m. thick, must have solidified in much shorter time than the 

 rock in the large granite fields. 



UNDERCOOLING (SUPERSATURATIOn) 



The degree of supersaturation necessary to crystallization is a 

 function of time and viscosity, and of other general factors. Further- 

 more there will perhaps also be involved a quality specific to each 

 mineral. 



For the melting of quite small quantities as, e.g., 10-20 gr. in 

 platinum crucibles, the supersaturation necessary to crystallization 

 is as a rule quite considerable. It is otherwise with fairly thin 

 melts when we work with quantities of, e.g., 10-20 kg. 



The explanation may probably be found in the fact that, after 

 a certain degree of supersaturation in the first cooled peripheral 

 parts (the surface or the walls. of the crucible) diminutive crystals 

 are formed, which in fairly thin melts abolish the supersaturation 

 in the inner, somewhat more heated parts of the molten mass. 



We shall illustrate this by an example (see Die Sulfid- 

 Silikatschmelzlosungen, I (1919), 75) concerning the melting in 

 graphite crucibles of different quantities of (Ca,Mg)4Si30io (with 

 0.7 Ca:o.3 Mg) from which mainly akermanite crystallized. 



Two experiments were made, in one case with 19 kg., and in the 

 other with i.i kg. The curve of melting (transition from solid to 

 liquid phase) was found to give 1290° at the beginning and 1310° 

 (or 1312°) at the end. 



The curve of crystallization (transition from liquid to solid phase) 

 gave (a) with 19 kg., period of crystallization twenty-five minutes, 

 from 1310° to 1290°; (6) with i.i kg. period of crystallization six min- 

 utes, from 1280° to 1240°. The crystallization of the 19 kg. melt thus 

 took place at the same temperature as the melting, i.e., the crystalliza- 

 tion took place without noticeable undercooling. In the test with only 



