6 Allen, etc. — Diopside and its Relations 



of the determinations is shown by comparing the measure- 

 ments among themselves.' Their interpretation in terms of 

 an absolute scale depends upon an agreement among observers, 

 and will vary from time to time whenever more accurate funda- 

 mental observations are available. At the present time the 

 Reichsanstalt Scale is generally accepted. 



The melting points measured, like those obtained in this lab- 

 oratory with the feldspars, and by workers in high tempera- 

 tures generally, are not entirely sharp, even with substances 

 which theoretically should melt at a strictly constant temper- 

 ature. Indications that melting has already begun invariably 

 appear on the thermal curve 20° or 30° below the melting point 

 proper, and the region of strongest absorption of heat is distri- 

 buted over an interval of from 2° to 3°. The main cause of 

 this phenomenon in the present case probably lies neither 

 in any unusual molecular viscosity attending the change 

 of state* nor in experimental error of the temperature obser- 

 vations. It may be due to the slight impurity ('1 to *2 per 

 cent) which chemical analysis shows to be present, even in the 

 most carefully prepared artificial mixtures. Similar curves are 

 obtained with ice to which 2 per cent or 3 per cent of salt has 

 been added, and at an absolute temperature five or six times 

 as high the same effect should be produced by about one-thirtieth 

 as much impurity. f The highest portion, that is, the end, 

 of the melting interval is taken as the melting point, since this 

 is probably the nearest attainable approach to what would be 

 the sharp melting point of the substance unaffected by traces 

 of impurity. 



The Melting Point Curve from jper cent— lfi'3 -per cent 

 MgSi0 2 . — The temperature-time curves were taken on a series 

 of mixtures of the metasilicates 5 per cent to 10 per cent apart 

 except in critical parts of the curve, where shorter intervals were 

 chosen. The temperatures at which the heat absorptions took 

 place are plotted in fig. 1. The results in the case of about 

 one-third of the mixtures were controlled by repeating the 

 observations on several different preparations, and since the 

 thermal phenomena in certain parts of the curve w T ere rather 

 complicated, a considerable number of the observations were 

 many times repeated, making in all nearly four hundred inde- 

 pendent determinations. 



By reference to the diagram (fig. 1) it will be seen that the 

 addition of magnesium silicate to pure calcium silicate (pseudo- 

 wollastonite) lowers the melting point rapidly, the curve fol- 



* This, of course, is not true of the feldspars, quartz, etc. 

 f In accordance with the Van't Hoff-Eaoult formula, where A, the 

 depression of the freezing point due to impurity, equals '02 T- 



