400 DISTRIBUTION OF HEAT IN THE SPECTRUM. [MEMOIR XXVIII. 



be traced to errors in placing the screen i with exactness 

 on the centre of the spectrum and on the line A. 



For the sake of more convenient comparison, I have re- 

 duced all the different sets of experiments to the stand- 

 ard of 100 for the whole visible spectrum. 



I have made use of four prisms: (1) rock-salt; (2) 

 flint-glass ; (3) bisulphide of carbon ; (4) quartz, cut out 

 of the crystal so as to give a single image. 



All the observations here recorded were made on days 

 when there was a cloudless sky. 



TABLE I. Distribution of Heat by Hock-salt. 



Series I. Series II. 



(1) Heat of the whole visible spectrum 100 100 



(2) " more refrangible region 53 51 



(3) " less " " 47 49 



In this table the column marked Series I. gives the 

 mean of four sets of measures, and that marked II. of 

 three. At the beginning of each set the rock-salt was 

 repolished. 



TABLE II. Distribution of Heat by Flint-glass. 



Series I. Series II. 



(1 ) Heat of the whole visible spectrum 100 1 00 



(2) " more refrangible region 49 52 



(3) " less " " 51 48 



Series I. gives the mean of ten sets of measures, Series 

 II. of eight. 



TABLE III. Distribution of Heat by Bisulphide of Carbon. 



Series I. Series II. 



(1) Heat of the whole visible spectrum, 100 100 



(2) more refrangible region 52 49 



(3) " less " " 48 51 



The sulphide employed was devoid of any yellowish 

 tinge ; it was quite clear. Series I. is the mean of eight 

 experiments, Series II. of ten. 



