148 A. L. DAY MINERAL RELATIONS FROM LABORATORY VIEWPOINT 



point can best be found by locating the temperature at which the latent 

 heat of fusion is absorbed or released. 



Uncertainty of Solidification — Undercooling 



Let us carr}^ the examination of the properties of different minerals 

 somewhat further. We may undertake to determine the melting temper- 

 ature of a particular mineral, and may obtain, as we suppose, a competent 

 measurement, but on cooling again we note that the solidifying tempera- 

 ture falls at some distance below the melting temperature. From the 

 petrological standpoint, the solidifying temperature will perhaps appear 

 the more important of the two, and proceeding from this viewpoint alone, 

 if these two temperatures should differ in the same substance, we might 

 be tempted to reject the former and adopt the latter without special en- 

 deavor to obtain corroborative evidence of the significance of the differ- 

 ence. But here again we may mislead ourselves by an over-hasty conclu- 

 sion. Laboratory experience has also shoAvn us that minerals in which 

 the solidifying point falls at a different temperature from the melting 

 point, also display considerable differences between successive determina- 

 tions of the solidifying point if the rate of cooling is changed or if the 

 mineral, while melted, happened to be heated high above its melting tem- 

 j)erature before cooling. In the explanation of this phenomenon there is 

 valuable information for the student of mineral crystallization, if he will 

 take the trouble to seek it; for it appears that minerals, much more than 

 metals, possess the kind of molecular inertia by virtue of which changes 

 of molecular arrangement, such as melting or solidifying, take place with 

 some difficulty and slowly, instead of promptly and regularly, as metals 

 do. Accordingly, as the liquid mineral cools down to its melting tem- 

 perature, it may not crystallize at once, and, delaying, may cool for a 

 considerable distance into the unstable region below the melting tempera- 

 ture before crystallization begins. Even after it has begun to crystallize 

 the process may proceed so slowly, while the temperature continues to 

 fall, that the visible evidence of a crystalline condition may not appear 

 for some time longer. In such a mineral, obviously the crystallizing tem- 

 perature is a property which can be determined only by the most carelul 

 observation, if at all. In all the pure minerals so far studied in the Geo- 

 physical Laboratory it happens that it can not be determined at alF^ — 

 that is, minerals, when cooling under atmospheric pressure, do not crys- 

 tallize promptly at a particular temperature which is constant and charac- 



''a Since the text of this article was written, accurate measurements of chemically 

 pure pyrrhotite (by E. T. Allen) show the temperature of crystallization to occur 

 promptly and to be identical with the melting point. 



