UNCERTAINTY OF SOLIDIFICATION 149 



teristic of the substance, )3iit at random, the temperatures being dependent 

 on the rate of cooling and various extraneous conditions which need not 

 be recounted here. It is even possible that such a substance, when cooled 

 rapidly, may be so inert with respect to changes of molecular arrange- 

 ment, or so viscous, that it may cool far below the melting temperature, 

 becoming more viscous all the while, until finally it reaches the tempera- 

 ture of the room without having crystallized at all. Such substances are 

 not uncommon; they are known, both to physics and to geology, as 

 glasses, and they possess no molecular arrangement characteristic of the 

 quasi-rigid condition which they appear to have reached. Strictly speak- 

 ing, therefore, from the physico-chemical standpoint, such substances are 

 undercooled liquids, and are still in an unstable condition, albeit oppos- 

 ing with overwhelming inertia (viscosity) any molecular rearrangement 

 whatever. Proof of this lies in the fact that the latent heat of fusion 

 which was absorbed during melting, and which is an absolute concomitant 

 of the change of state, is not released on cooling to glass. In this condi- 

 tion days, or years, or even geologic time, may conceivably elapse before 

 the characteristic crystal forms appear. In the extreme case where the 

 crystallizing force encounters resistance of overwhelming magnitude, 

 crystal formation is no doubt permanently stopped until new forces or 

 conditions intervene to bring relief. Ordinary window glasses, or the 

 obsidian cliffs of the Yellowstone Park, offer familiar illustrations of 

 such permanently interrupted though still incomplete operations. 



Some Difficulties of Interpretation where Glass is Present 



The physicist, therefore, soon learns that there is much individuality 

 among the characteristic properties of different minerals, and that deter- 

 minations of their physical constants must proceed with unusual caution 

 lest the result be misleading. In particular, having established the fact 

 that many minerals do not solidify (crj^stallize) regularly at definite 

 temperatures, he must, not always, perhaps, but usually, depend on melt- 

 ing-point determinations if he would learn the normal temperature of 

 their change of state. He is also forced to take into account the fact 

 that the phenomenon of melting does not have the same appearance with 

 different minerals — that some change more conspicuously in one property, 

 some in another — and that his observation of the moment of melting 

 must be made with proper respect for the individual characteristics of 

 the particular mineral under investigation. If a mineral has been found 

 to melt to a viscous liquid, he may not set up a thin sliver, as was the 

 common practice a few. years ago, and observe with a telescope the first 



I 



