THE EUTECTIC RELATION 166 



certain case of this most formal type of crystallization has yet been estab- 

 lished with certainty among the minerals. As in the case of a single 

 mineral cooling down after melting, there is always some nndercooling^^^ 

 from one or other of the causes above outlined, and crystallization is more 

 or less delayed. It is, however, altogether probable that the basic min- 

 erals which are more thinly fluid will furnish numerous illustrations 

 of it. 



It is not necessary that the reader infer from the absence of explicit 

 cases of normal behavior in the crystallization of mixtures of pure min- 

 erals so far studied that we are on the wrong track, and that the existing 

 theory of solutions and the phase rule can not be properly applied to 

 mineral solution, for the converse case of a mineral mixture melting nor- 

 mally has been frequently observed and is unmistakable. The eutectic 

 usually melts promptly for all compositions in which it is present in suffi- 

 cient quantity for observation, while the excess component enters the 

 solution gradually, and disappears at the temperature appropriate to its 

 concentration. It is a matter of great good fortune to the investigator 

 endeavoring to establish quantitative relations between the minerals, that 

 although . individual physical properties frequently intervene to prevent 

 the normal phenomena from appearing promptly on the cooling curve, 

 the heating curve is rarely disturbed by them. 



In this connection, attention may again appropriately be called to a 

 phenomenon first mentioned in a publication from this laboratory on the 

 crystallization of feldspars,^^ in which it was pointed out that in viscous 

 solutions which are not stirred, rapid heating, and even more conspic- 

 uously heating but a short distance above the melting temperature, fre- 

 quently affects the behavior of the solution during subsequent cooling. 

 When crystals are melted, their molecules frequently remain in the same 

 relative positions in the liquid, "as motionless as a school of minnows in 

 a brook,^'-^ for some time, and for a considerable temperature interval (a 

 hundred degrees or more) above the point where melting is complete. In 

 time and at high temperatures the schools are scattered, but recrystalliza- 

 tion on cooling is much assisted if the cooling takes place before these 

 schools have become dispersed. 



Of course, the number of such individual mineral characteristics or 

 properties which enter into the quantitative determination of its relations 

 to other minerals in the laboratory is greater for a two-component system 



i^o See footnote ^a, page 148. 



19 Arthur L. Day and E. T. Allen : Isomorphism and thermal properties of feldspars. 

 Publication No. 31, Carnegie Institution of Washington, p. 54, 

 ** As Doctor Becker first suggested to me. 



