J. Trowbridge — A Standard of Light. 131 



light emitted. This phenomenon had been noticed early by 

 Dr. J. W. Draper. One Leclanche cell with five ohms in the 

 circuit beside the resistance of the strip was sufficient to raise 

 the latter to a red heat, and precautions were then necessary to 

 prevent a change of resistance from the heating effect of the 

 battery employed with the Wheatstone's bridge. Being desir- 

 ous of ascertaining whether the resistance of the platinum wire 

 changed after it had been heated to a red heat and had been 

 allowed to cool, I arranged the resistance of the battery circuit 

 outside the bridge, so that the wire could be raised to a red 

 heat, and then, having quickly weakened the battery circuit, 

 remeasured the resistance of the strip. No difference could be 

 perceived in the resistance of the strip. This illustrated the 

 fact discovered by Professor Langley, that thin strips of metal 

 arranged as bolometer strips give up heat very quickly. 



The results of this experiment led me to think that a bolo- 

 meter strip of definite surface could be placed at a fixed 

 distance from a carbon loop of definite dimensions inside an 

 exhausted glass vessel. The amount of radiation which the 

 bolometer strip receives could be calculated; and we might 

 base our standard of light upon the point of incandescence 

 which would give a definite radiation at a fixed distance. We 

 could not distinguish by this method the energy produced by 

 rays of different refrangibility. It seems desirable, however, 

 to substitute for the uncertain estimation of colored lights by 

 the eye an instrument which will measure the energy produced 

 by the radiating source at a certain distance. Within certain 

 limits I found that the bolometer strip would indicate an 

 increase or decrease of the amount of radiant energy received 

 while the difference in color of the incandescent lamp made 

 the observer at the photometer entirely uncertain of his meas- 

 urements. 



Owing to the difficulty of obtaining the proper apparatus for 

 the prosecution of the study of this method, I then studied the 

 question of the practicability of employing a thermopile to 

 measure the amount of radiation from an incandescent strip of 

 platinum at a fixed distance. Within a long photometer box 

 ■was placed a thin brass vessel containing water. Steam was 

 passed by means of a rubber hose into the water of this vessel 

 which was thus maintained at a constant temperature of about 

 94° C. The outside of the vessel was about 92° C. This was 

 ascertained by making the side of the vessel constitute one 

 metal of a thermal junction. Between this vessel and the pla- 

 tinum strip, which was made incandescent by a current of 

 from 8 to 9 webers, was placed a thermopile. The face of the 

 thermopile was thus exposed to the radiation from a given 

 amount of heated surface at a constant temperature, while the 



