FOOTE, MOHLER, AND FAIRCHILD : PYROMETER GLASS 547 



below the point where the safe operation of the pyrometer lamp 

 is in question, and to use this experimentally determined value 

 of A for extrapolation of the temperature scale without the neces- 

 sity of applying corrections for change of A with 6. 



Such a glass is ideal for other reasons. There will be always 

 a perfect color match between the black-body furnace and the 

 pyrometer lamp. This becomes evident upon considering in- 

 tensities at any monochromatic wave length. We have for a 

 match at wave length X 



J (x, e) T' = J (x, S) (4) 



or 



H) 



Since by the assumed form of T' the quantity X log e T' is con- 

 stant, - — o is constant, and a match for one wave length holds 



io 



for all other wave lengths. Also this condition of color match 

 holds for all other temperatures. 



As a consequence of the above fact, the settings made with 

 this type of absorption glass are absolutely independent of the 

 type of red glass ocular used. The same settings are obtained 

 for a red glass, a blue glass, or no-colored glass ocular at all. 

 The settings are independent of the visibility curve of the ob- 

 server and a color-blind person would obtain the same readings 

 as the normal observer. With the usual type of absorption 

 glass or sectored disk we do not have a color match and settings 

 are therefore difficult to make; different observers obtain differ- 

 ent values of A on account of differences in their visibility 

 curves; and the variation of A with 0, which must be accurately 

 known for accurate temperature measurement, is very difficult 

 to determine experimentally. 



In order to check the constancy of A with 6 luminosity curves 

 were drawn for each system, viz., lamp + red glass and furnace 

 + black glass + red glass, and comparisons made by use of 

 equation (1). The quantity A was found to be constant within 

 the errors of the necessary graphical integrations. Another 



