OP ARTS AND SCIENCES. 499 



It seems possible, therefore, to assume as a standard of light au iu- 

 candesceut strip which radiates a defiuite amount of energy, this 

 energy being measured at a fixed distance which will best agree 

 numerically with the absolute system of measures now universally 

 adopted in heat and electricity. The method of Draper and Schwendler 

 could be combined with the methods I have described above. For a 

 practical standard, a carbon loop in an exhausted vessel raised to such 

 a point of incandescence that it will radiate a definite amount of 

 energy, — this energy being measured by a bolometer strip or the 

 thermopile at a definite distance from the carbon loop, and also being 

 measured by the formula J H ■=^ C^Rt, — would have a greater range 

 than an incandescent strip of platinum placed in free air. The latter 

 method, however, for the incandescence which produces a light similar 

 in color to that of a sperm candle, is extremely sensitive, and can be 

 made, I think, more exact than present photometric tests. Both 

 methods have the great advantage of substituting a measure of energy 

 for a relative indication by the eye, which is not connected with any 

 absolute measurement. 



These remarks apply to the question of a standard of light for prac- 

 tical purposes, which shall also be scientific in so far that more refined 

 scientific investigation can connect this standard at any time with more 

 precise methods of measuring the exact amount of heat given by 

 radiations of definite wave-length. By means of a Rowland concave 

 grating and with a bolometer strip, one can at present measure the 

 energy of definite radiations. We can say that our scientific standards 

 for light of different colors shall be based upon the energy received 

 upon a definite surface at definite points in the diffraction spectrum. 



Jefferson Physical Laboratory. 



