226 PROCEEDINGS OP THE AMERICAN ACADEMY 



The lesser readings indicate greater luminosity of the melting wire. 



In another experiment, the wire in its ordinary condition melted 

 when the reading (mean of 5 measurements) was 37.8 in. After 

 annealing in a Bunsen flame, the mean reading at melting (for five 

 experiments) was 37.2 in. After heating to a yellow heat by passing 

 a current of electricity through it for half an hour, the reading at 

 melting (mean of 23 measurements) was 36.9 in. 



If further experiment should show conclusively that the light emitted 

 by melting platinum is variable beyond narrow limits, it is clear that 

 any photometric standard based on the luminosity of melting platinum 

 will present no advantage over those standards now in ordinary use, 

 unless specially prepared platinum freed from gases and consoli- 

 dated by Edison's process of heating in vacuo is found to be available. 

 Probably solidifying platinum, provided its point of solidification is 

 sufficiently definite, would be free from this objection, although this 

 can hardly be accepted without experimental proof. 



KoGEKS Laboratory of Physics, 

 June, 1886. 



