8 burgess: a micropyrometer 



lamp is taken as a measure of the temperature of the strip as in 

 the Morse or Holborn-Kurlbaum pyrometer. In practice, the 

 observer with one hand raises the temperature of the platinum 

 strip by increasing the electric current thru it by means of a fine- 

 step rheostat and with the other hand adjusts the rheostat in the 

 pyrometer circuit so as to continuously match in brightness the 

 lamp filament and platinum strip. The eye-piece is furnished 

 with a piece of monochromatic glass such as Jena red filter no. 

 F 4512. For temperatures above which the lamp should not be 

 burned, say 1400°C., an absorption glass is placed between the 

 microscope objective and furnace window. The inside of the 

 metallic container or furnace should be blackened to prevent 

 undesirable reflections of light from the walls. 



The calibration of the pyrometer as sighted upon the platinum 

 strip in the furnace may be made in two ways. The first, which 

 was the only method available when the earlier form of this 

 apparatus was brought out, due to the dearth of well known 

 fixed points in the temperature range studied, consists in cali- 

 brating the pyrometer in the customary manner and then apply- 

 ing the corrections at the temperatures of melting, for the emis- 

 sivity of platinum, furnace atmosphere and window, and for the 

 surface tension of the melting specimens when necessary. 



The second method, which appears to be the more accurate and 

 also more convenient, consists in observing the lamp currents at 

 the known melting points of two or more pure substances such 

 as gold, nickel, cobalt and palladium, and from the equation 

 expressing the relation between temperature and current in lamp, 

 the temperature of melting of any specimen may be computed. 

 For not too great teniperature intervals, the equation log c = 

 a -\- b log t may be used which permits calibration with two fixed 

 points only. 



This second method of calibration has the further advantage 

 that, for materials of approximately the same general properties, 

 the error of method is eliminated. Thus in the case of metals 

 which alloy with platinum, the effects of alloying conductivity 

 and of surface tension enter into the calibration as well as into the 

 determination of the unknown melting point, so that any out- 



