ON THE EFFICIENCY OF EDISON'S ELECTRIC LIGHT 



201 



radiation, but the error was avoided as much as possible by having the 

 mean temperature of the calorimeter as near that of the air as possible, 

 and the rise of temperature small. The error would then be much less 

 than one per cent. A small portion of the light escaped through the 

 apertures in the cover, but the amount of energy must have been very 

 minute. 



In order to obtain the amount of light and eliminate all changes of 

 the engine and machine, two lamps of nearly equal power were gener- 

 ally used, one being in the calorimeter while the other was being 

 measured. They were then reversed and the mean of the results taken. 

 The apparatus for measuring the light was one of the ordinary Bunsen 

 instruments used for determining gas-lights, with a single candle at 

 ten inches distance. The candles used were the ordinary standards, 

 burning 120 grains per hour. They were weighed before and after 

 each experiment, but as the amount burned did not vary more than 

 one per cent from 120 grains per hour, no correction was made. 



As the strips of carbonized paper were flat, very much more light 

 was given out in a direction perpendicular to the surface than in the 

 plane of the edge. Two observations were taken of the photometric 

 power, one in a direction perpendicular to the paper, and the other 

 in the direction of the edge, and we are required to obtain the average 

 light from these. If L is the photometric power perpendicular to the 

 paper, and I that of the edge, then the average, I, will evidently be 

 very nearly 



Xo 



COS a sin a d a + I I Sin 2 a d a, 



/ 



I 



Ft 



A = J L + p. 



In the paper lamps we found l = 

 The lamps used were as follows: 



nearly; hence x =|L nearly 



The capacity of the calorimeter was obtained by adding to the capac- 

 ity of the water, the copper of the calorimeter and the glass of the 



