170 



PROCEEDINGS OF THE AMERICAN ACADEMY 



nicla 1-plated, and as seventy-five per cent of the so-called radiation is due 

 to convection by the air, the coefficients of radiation were found to be 

 very constant under similar conditions, even after long intervals of time. 



The experiments were divided into two groups ; one when the 

 temperature of the jacket was about 5° C, and the other when it 

 averaged about 20° C. 



The results were then plotted, and the mean curve drawn through 

 them, from which the following coefficients were obtained. These 

 coefficients are the loss of temperature per minute, and per degree 

 difference of temperature. 



TABLE XXXV. — Coefficients of Radiation. 



As the quantity of water in the calorimeter sometimes varied 

 slightly, the numbers should be modified to suit, they being true when 

 the total capacity of the calorimeter was 8.75 kil. The total surface 

 of the calorimeter was about 2350 sq. cm., and the unit of time 

 one minute. To compare my results with those of McFarlane and 

 of Nichol given in the Proc. R. S. and Proc. R. S. E., I will reduce 

 my results so that they can be compared with the tables given by 

 Professor Everett in his " Illustrations of the Centimeter-Gramme- 

 Second System of Units," pp. 50, 51. 



The reducing factor is .0621, and hence the last results for the 

 jacket at 20° C. become : — 



TABLE XXXVI. 



