Sources of Artificial Illumination. 



113 



dies, increases very nearly in the ratio 1 : 3*7 ; whereas the 

 corresponding ratio for the neighborhood of the Gr line is 

 approximately 1 : 22 ; intermediate regions giving values lying 

 between these limits. In the yellow, (region represented by 

 \=6000), this ratio has a value which is identical with that 

 which expresses the change of candle-power. 



In all regions toward the red from this particular wave- 

 length, the brightness of the spectrum increases less rapidly, in 

 all regions of shorter wave-length it increases more rapidly 

 than the candle-power. 



The position of this re- 

 gion, in which the bright- 

 ness of the spectrum in- 

 creases in direct proportion 

 to the candle-power, is indi- 

 cated in figure 4, by a ver- 

 tical line. 



It lies slightly toward the 

 red from the position of 

 maximum luminosity in the 

 spectrum of the incandes- 

 cent lamp, as determined 

 for a lamp of the same type 

 by Otto Schumann.* 



The location of this line 

 is of especial interest be- 

 cause of the light it throws 

 upon the' results obtained 

 with the Bunsen photome- 

 ter when that instrument is 

 applied to the comparison 



of Sources of light differing Curves of relative intensities in the spec- 



in composition. The line in tri ™ ° f ,^ n in f fi nd f„ een 9 l s lamp / ^T 



. -r - marked 4c, 10c, 16c, 22c, 28c, refer to the 



(question does notCllVlCle trie standard incandescent lamp at various tem- 

 spectrum equally either in peratures. Curve B refers to a low volt- 



respect to the total energy a s e lam P at normal candle-power. 

 of luminous radiation of the portions lying between it and the 

 red on the one side and between it and the violet on the 

 other, nor in respect to the total luminosity of those portions. 



Suppose the case of four incandescent lamps, so constructed 

 that when they give 16 candles each, their temperatures are 

 respectively those of the Edison lamp upon which our measure- 

 ments were made, when the latter gives 4, 10, 22 and 28 can- 

 dles. 



As measured in the Bunsen photometer these lamps would 

 be found of equal candle-power, but they would differ greatly 



* Otto Schumann, Elektrotechnische Zeitschrift, v, p. 220. 



