GEOPHYSICAIv REISEARCH — DAY. 1 83 



pyroxene, melting at 1,521°. Into this form the other three modifications 

 change more or less readily with the application of heat. If a melt of this 

 composition be cooled rapidly, one of the amphiboles results; if somewhat 

 less rapidly, an orthorhombic pyroxene identical with enstatite ; if slowly, the 

 stable pyroxene always reappears. Curiously enough, the stablest form of 

 this composition has never been identified in natural rocks, although we are 

 confident that intergrowths of it with enstatite must have been found and 

 overlooked. It does sometimes appear in meteorites. On the other hand, 

 the less stable pyroxene, enstatite, is a very familiar natural mineral. 



The situation seems somewhat anomalous, and, indeed, no complete expla- 

 nation for it has yet been found, but a curious side-light is thrown upon it 

 by experiments which we made in the laboratory to the following effect: If 

 magnesium metasilicate in any one of its crystal forms is mixed with a small 

 quantity of a thinly fluid flux (calcium vanadate, magnesium vanadate, mag- 

 nesium tellurite, magnesium chloride in the presence of dry hydrochloric 

 acid) and heated, the stable pyroxene will always reappear. If, on the other 

 hand, a solution be formed between the magnesium silicate and albite, ortho- 

 clase, or some other viscous silicate, enstatite, and even the orthorhombic 

 amphibole crystallizes out quite readily. It thus appears that the retarding 

 influence of viscosity is sometimes sufficient to prevent or postpone the crys- 

 tallization of that form of the mineral which is relatively stablest in favor of 

 some one of the unstable forms which may or may not go over into the stable 

 form afterward. This final transition probably also depends upon tempera- 

 ture and viscosity. 



The fourth form of this mineral is a monoclinic amphibole which we were 

 able to form in aqueous solutions at low temperatures, but only occasionally 

 and in small quantities from the melt. It appears to be unknown among 

 natural rocks. 



The final proof of the relation between these minerals was established by 

 showing that each of the three lower forms passes into the monoclinic pyrox- 

 ene with evolution of heat. 



This paper, although restricted to a study of the different possible forms 

 of a single mineral, was immensely instructive in explaining the appearance 

 of relatively unstable components among natural rocks, and in showing that 

 contiguous minerals, both in natural rocks and in pure mineral melts, are by 

 no means necessarily in equilibrium. 



In all branches of quantitative research, especially in its earlier stages, 

 nearly as much attention requires to be paid to the methods of measurement 

 as to the results. It therefore frequently happens that investigations which 

 appear to bear only indirectly upon the ultimate purpose of the research are 

 among the most important which the laboratory is called upon to undertake. 

 During the past winter the methods of temperature measurement which we 



