of the total amount of light acting behind each color screen. 



. ie true ueeauue lEhe difference between hent and light is only a 



-r- ^^ 



ter of wave length. -i^. a truo thtct (the thermopile measures jj 7 in 



terms of the electric current produced by a difference in temperat- 

 ure of the exposed and unexposed junctions; but it does so by vir- 

 tue of the fact that the energy of whatever vibrations fall upon 

 it be thtfy long and therefore heating in their physiological ef- 

 fect or short and therefore only perceived ad light, is converted 

 into heat energy upon being absorbed by the exposed junction of the 

 thermopile. In other words, the ligh- of the blue end of the spec 

 1rum produces an electromotive force much less than that of the 

 infra- red but no less measurable. 



The instrumetfnta ur td in the enejjy calibration of these 

 screens were Si ariffin^T 1 + " table thermopile with junctions 

 of bismuth and Hwrt^SSv and a moderately sensitive galvanometer 

 (d'Arsonval) .^BsdaP electric arc similar to the one later used in * 

 the experiments themselves nas the source of light* The thermopile -** 

 with the open end screened by on,e of ;the lighjt f J.^exp, was / 



exposed to the light until the galvanometer indicator reached 

 a maximum deflection which took, ordinarily^sbout thirty second^. 

 The number of divisions through which the spot of light re- 



^^j-a^a JU^a-frJL^-ca-fQ *x*/ ~X-i^e, *>-^<*-&e. 



fleeted from the galvanometer mirror.was noted. This was repea't" 

 ed six times and the average deflection recorded. The other 

 filte^rs were then used in turn to screen the thermopile t 

 and thermopile and screen' moved to such ^ distance 'rrom the 

 arc that the displacement^ of the galvanometer &3&& indi- ^ 



cator opt*'* approximately equal. This -rdi3tan.c'was ^.al.so j^^^y 

 for the ui'icnjroGne.d thermopile^ which rerres en red 



the removal of the control from the source. For the experiment 

 the quantity of light used can be varied for the whole set ^ 

 of screens by multiplying or dividing these distances by 

 the same numbervandthe intensity infall the boxes will remain 

 equal. The actual amount o lisht in candlepowers can be 



