242 /. E. L. VOGT 



And the individuals of rhodonite in melts with about 80-95 P^r cent 

 rhodonite may reach a length of 37-40 mm. 



In a melt of 100 per cent CaMgSiOe, diopside of a length (=■=<;) 

 of 30 mm. crystallized in the course of thirty-one minutes. This 

 gives a rate of crystaUization of about 1 mm. {or 0.97 mm) per 

 minute. 



For the largest crystals of pseudowollastonite, melilite, §,kerma- 

 nite, fayaUte (Fe2Si04), manganese-fayahte [(Mg,Fe)2Si04], and 

 rhodonite [(Mg,Fe) SiOs], commonly crystalHzed from solutions of 

 75, 80, or 90 per cent (or in one single case of even 100 per cent), 

 we get a corresponding rate of crystallization, viz., about i mm. 

 per minute. In fayalite, and almost certainly in manganese- 

 fayalite, the rate is somewhat more than i mm. per minute, measur- 

 ing the greatest length of the crystals. 



All the minerals mentioned above that reach lengths of 20, 30, 

 exceptionally even of 40 or 50 mm., crystallized (in the course of 

 about half an hour) from very thin fluid melts. If, as a contrast 

 to this, we take extremely viscous siHcate melts as, e.g., KAlSijOs, 

 NaAlSijOg, and SiOa, there are formed, even if they are kept for 

 one or more days just below the melting-point, crystals of only 

 extremely small dimensions. 



The rate of crystallization, no doubt, is a function of several 

 factors, and one of the most important of these is the degree of 

 viscosity. 



All the foregoing minerals examined by me, which in the melted 

 phase are very thin, are characterized by so great a tendency to crystal- 

 lization that, from a solution of 80-100 per cent, they crystallize 

 even under a very quick coohng. If, thus, we pour, e.g., melted 

 Fe^ Si04, (Fe,Mn)2Si04, CaMgSi^Oe, or melted akermanite into a 

 calorimeter pipe, where the coohng from the melting temperature 

 to usual room temperature requires only eight to ten minutes, and 

 where the period of crystallization has a duration of only about 

 one minute, we as a rule at least get some mineral segregated. 



In very viscid molten masses it is quite different. To cause 

 crystallization here, we must keep the molten mass heated just 

 below the melting-point (or below the curve of crystallization) for 



