THE 



AMERICANJOURNALOFSCIENCE 



[FOURTH SERIES.] 



Art. I. — The Determination of Mineral and Rock Densities 

 at High Temperatures y by Arthur L. Day, R. B. Sosman, 

 and J. C. Hostetter. 



1. Existing Data. 



Before discussing briefly the existing data on the change of 

 specific volume with temperature, we shall mention several 

 misinterpretations and misunderstandings which have occasion- 

 ally appeared. 



In the first place, the sharp distinction between a glassy and 

 a crystalline substance has not always been kept in mind. A 

 glass is essentially a liquid which has cooled without crystalli- 

 zation to a point where its viscosity has become so great as to 

 give it the mechanical rigidity of a solid without the addition 

 of any new property. The cooling has resulted in no release 

 of the heat of fusion, nor has it developed a discontinuity in 

 any physical or chemical property. There has therefore been 

 no change of state (freezing point) and no one temperature 

 which possessed more physical significance than another in the 

 entire process of cooling to the temperature of the surround- 

 ings. The same is true of the reverse operation ; inasmuch as 

 the glass acquired no new property on cooling, it has none to 

 give up on reheating, and there is no melting and no melting 

 temperature. We must, therefore, associate glasses with 

 "liquids," not with "solids." 



By definition, the temperature at which crystals of a sub- 

 stance (the "solid" phase) will exist indefinitely in contact with 

 their "liquid," unless heat is added or withdrawn, is the true 

 melting temperature of the substance. Above it the liquid 

 alone is stable, and below it only the crystals are stable. But 

 if the substance be very inert it may exist for very long periods 

 in the temperature region in which it is not stable (as, for ex- 



Am. Jour. Sci.— Fourth Series, Vol. XXXVII, No. 217. — January, 1914. 

 1 



