Alumina with Silica, Lime and Magnesia. 297 



homogeneous. A conspicuous visible demonstration of the 

 need for such careful preliminary mixing occurs with the 

 composition A1 2 3 8 per cent, CaO 69 per cent, Si0 2 23 per 

 cent. As long as the mass is incompletely combined, quantities 

 of calcium orthosilicate will disintegrate (dust), yielding a 

 mixture of fine powdered orthosilicate and cinders of the more 

 aluminous material. After three or four fusions and grindings, 

 the charge remains solid, i. e., practical homogeneity has been 

 secured. This shows that diffusion in silicate melts is extremely 

 slow and emphasizes the necessity for the somewhat tedious 

 care in preparation. 



The effect of insufficient mixing on the observed melting 

 points may be very great. Even in a sharply melting compound 

 like anorthite, the cones prepared from the most carefully 

 mixed oxides often melt 20° or more below those made from 

 previously combined oxides. Here the heat of combination 

 is evidently dominant. Obviously, quite the reverse might 

 happen with other mixtures. One immediate consequence of 

 this is that experiments in which the Seger cone method is 

 employed for the measurement of temperature must always be 

 made with carefully mixed and combined charges. It is also 

 clear that comparison of the bending of a Seger cone with the 

 bending of a cone of wholly different mixtures may mean very 

 little, for the viscosities of the different mixtures bear no 

 definite relation to each other, and viscosity is the chief factor 

 in the bending of these cones.* Thus minerals like quartz or 

 albite will retain their stiffness far above their melting points, 

 i. e., after their crystalline structure is entirely gone. 



Apparatus. — Those portions of the series which melt below 

 1600° were determined by the thermoelectric method described 

 in the various publications from this laboratory. Above 1600° 

 only the optical method described in the lime-silica paper is 

 available, i. e., an optical pyrometer (Holborn and Kurlbaum 

 type) sighted on a strip of platinum or iridium. We have 

 found it useful to make a small tack of platinum, as figured 

 (fig. 1, 1>). This tack has a polished upper surface upon which to 

 sight. The prepared charge is moulded into a cylinder (A) of 

 the shape shown; the shaft of the tack 

 fits snugly into the hole (C) in the 

 cylinder ; the head of the tack (B) rests 

 firmly on the flat surface of the charge, 

 while the raised shoulder (D) diminishes 

 the reflection from the walls of the 

 furnace. If ideal conditions obtained, 

 the tack would be invisible, i. e., we ^: 

 would have an absolute black body, but 



* See The Lime-Silica Series of Minerals, this Journal (4), xxii, 2( 



TfiT 



-7 



