E. Merritt — Light from Incandescent Lamps. 173 



alum cell was removed and the deflection corresponding to 

 total radiation was observed. The ratio of the two deflections 

 gave the ratio of the light energy to the total energy. The 

 total energy being determined by electrical measurements, the 

 energy of the light could be calculated. 



Fig. 2 shows the arrangement of the apparatus. The lamp, 

 L, was placed close behind the screen W. This was a metallic 

 vessel of the form shown, 

 and was filled with water. 

 A cylindrical opening 

 "through the center allowed 

 part of the light from the 

 lamp to pass through. This 

 screen prevented any heat 

 from the supports of the 

 lamp from reaching the 

 pile. Beyond this water- 

 screen was the cell C con- 

 taining the solution of alum, and beyond C the large screen 

 S of asbestus paper. There was a small opening in this screen, 

 directly in line with the opening in W, and covered by a mov- 

 able piece of asbestus paper A. The pile P was placed with its 

 funnel-shaped tube about 5 cm from A. Close to the end of 

 the other funnel of the pile was the asbestus screen S', which 

 served to protect the right-hand face from sudden changes of 

 temperature. The total distance from the lamp to the pile 

 was about two feet. 



The pile contained 56 pairs with a total surface of about 

 2-5 sq cm_ Wires led from it to a Thomson tripod galva- 

 nometer of about - 3 ohms resistance. The galvanometer 

 and scale were placed at some little distance from the pile, and 

 the screen S was between the pile and the observer. The gal- 

 vanometer was so delicate, that it was impossible to keep its 

 zero point from drifting. As it took fully four minutes for 

 the needle to reach its final deflection when the light was al- 

 lowed to fall on the pile, and as the lamp current was not at 

 all steady, this movement of the zero point made it difficult 

 to obtain reliable deflections, but a peculiarity in the manner 

 in which the needle reached its final position, simplified the 

 matter greatly. The screen A could be suddenly drawn aside 

 by means of a cord reaching to the observer, so as to allow the 

 lamp to shine upon the pile. When this was done the spot 

 of light which indicated the motion of the galvanometer nee- 

 dle, moved quickly to one side. But in two or three seconds 

 it began to move more slowly, and in about five seconds 

 stopped, moved backward a short distance, and then on again. 

 Several of these maxima and minima could be noticed before 



