272 Day and Shepherd — Lime-Silica Series of Minerals. 



careful determination of it lias ever been made. Since this 

 oxide melts to an extremely viscous liquid, attempts to deter- 

 mine the melting temperature by observing the softening of 

 the charge are wholly misleading. The molecular deorienta- 

 tion proceeds very slowly, extending over a considerable range 

 of temperature, as albite and orthoclase have been found to 

 do,* but with the disadvantage that this temperature region is 

 too high to be reached with a thermoelement, and no other 

 method of temperature measurement possesses sufficient sensi- 

 tiveness in this region to locate the melting temperature by 

 the heat absorbed during slow fusion. Determinations of the 

 freezing-point are out of the question, owing to the inertness 

 of the viscous melt. 



An approximate determination of the melting temperature 

 was made in this way : A gram or two of finely powdered 

 quartz was placed in a small iridium crucible and heated in an 

 iridium tube furnace (see p. 286). (Experience has shown that 

 melting and inversion phenomena in very viscous substances 

 take place much more readily if the material is finely divided.) 

 A tiny fragment of platinum foil was then laid on the top of 

 the charge and the furnace slowly heated until the foil was 

 observed to melt. Upon removing the charge from the fur- 

 nace and examining it microscopically, evidence of fusion was 

 found throughout the mass. The crystal grains had inverted 

 to tridymite and the superficial liquefaction had caused them 

 to sinter tightly together, but no displacement of the grains 

 had taken place. At the temperature of melting platinum, 

 therefore (1720°), silica shows positive evidence of fusion. 

 Other similar charges were then prepared and the operation 

 repeated with longer exposures and temperatures slightly below 

 the melting point of platinum, the temperatures being meas- 

 ured with a Holborn-Kurlbaum optical pyrometerf focused on 

 the platinum fragment. By repeating this process at short 

 temperature intervals and with about 20 minutes exposure, 

 melting was definitely established as low as 1625° C. 



The iridium furnace is unfortunately not adapted for long- 

 continued heating, and the platinum coil furnace will not 

 reach this temperature, so that an effort to discover a 

 definite temperature below which the solid is stable and above 

 which it will melt if given time enough, was abandoned. If 

 the heating is moderately rapid, the crystalline solid will per- 

 sist far above the melting point of platinum ; if slow enough, 

 it liquefies completely at 1625° or even lower. It is probably 

 a fair assumption, that pure silica begins to melt at about 1600,° 

 and will continue to complete fusion if given time enough, — 



*Day and Allen, loc. cit. 



f Holborn and Knrlbaum : Ann. d. Phys. x, p. 225, 1903. 



