OF ARTS AND SCIENCES. i>43 



be no real flame, but a brilliant, striated structure, from which poured 

 up clouds of smoke. The flickering did not annoy me as much as I 

 had expected in my measurements, but was most noticeable in the red. 

 The limits, however, varied considerably, so I took their maximum 

 position. 



Mean Scale Readings. 



r y g v 



473 310 229 100 



295 362 222 100 



384 336 225 100 



The second red was clearly wrong ; it was therefore discarded and 

 the first only used. 



Brilliancies. 

 r Y G v 



50 100 223 1,129 



The well-known blueness of the flame is clearly accounted for by 

 the great quantity of violet rays present. 



Limits. 



R V Slit. Distance. 



695 411 .03 mm. 1.0 m. 



715 408 .04 mm. 1.0 m. 



20 3 Advanced. 



Total brilliancy, 215 c. p. Intrinsic, 20.8 st. 



Electric Light. 

 The light was obtained with a Foucault regulator, using 40 pint 

 Grove cells. Six observations were made in each series, instead of 

 three, as in the case of the other lights. The intrinsic brilliancy of 

 both the arc and the carbons was measured. I found the arc to be 

 much fainter, and to vary considerably, while the carbons remained 

 quite constant. If a more powerful current had been used, I think 

 the intrinsic brilliancy of the arc might have increased a little, but 

 the chief difference would have been in its area and that of the 

 ignited carbons. 



Mean Scale Readings. 



R y G v 



192 178 151 68 



238 144 155 57 



215 161 153 62 



The second yellow was here discarded as obviously incorrect. 



