298 Shepherd, Rankin, Wright — Binary Systems of 



Fig. 2, 



EEFfl 



at the melting point the retardation of the temperature of the 

 charge, communicated along the shaft of the tack, should cause 

 the head to flash, i. e., become visible. Although the condition 

 of even approximate blackness is seldom attained in such 

 furnaces, we were much assisted by this method of procedure. 

 At the present time observations of melting temperatures 

 above 1600° are only determinable for the case of compounds 

 at the maximum, and for eutectic compositions. The course 

 of the liquidus cannot be established by any optical method yet 

 devised. For that reason we have sketched all curves above 

 1600° in dotted lines to indicate that only the maxima and 

 minima are determined. Of course the phases present along 

 the liquidus can be determined microscopically, and this has 

 been done in every case. The method of making these tem- 

 perature observations is similar to that used on the orthosilicate 

 of lime. A (fig. 2) is the tube of the iridium furnace ; B, a 



magnesia cylinder on which 

 rests the iridium crucible 0. 

 D is the cylinder whose 

 melting temperature is de- 

 sired, with the iridium tack 

 in positi on . E is a magnesia 

 > G lid with a small hole through 

 which to observe the charge; 

 F an exhaust tube to carry 

 away the iridium vapors and 

 heated air which would 

 affect the prism H of the 

 pyrometer. Gr is an asbestos 

 shield to further protect the 

 pyrometer. The bottom of 

 the furnace is closed by the 

 circular cup, I, thus prevent- 

 ing air currents from pass- 

 ing up through the furnace. 

 If .the furnace acted as a 

 perfect black body, neither 

 the metal tack nor the 

 charge w r ould be visible. 

 When such blackness does 

 not occur, the scale of a 

 theoretically black body does not apply, and the pyrometer 

 must be calibrated arbitrarily in terms of the conditions in 

 which it is nsed. The difference between the theoretical and 

 actual scales may reach 100° C. The necessity for frequent 

 recalibration is also obvious. 



^ 



^ 



^ 



