56 Sir John Conroy. On the Amount of Luminous and 



incandescent en a 98 pour 100, et celui de la flamme d'alcool 99 

 ponr 100." 



Tyndall (' Phil. Mag.,' 1864, vol. 28, p. 335 ; and 1 Contributions to 

 Molecular Physics,' p. 260) states that he placed in front of his thermo- 

 pile a rock-salt cell and observed the deflection of the galvanometer 

 when the radiation from (1) an incandescent platinum wire, (2) the 

 brightest part of a gas flame, (3) an electric arc lamp, passed through 

 the cell, filled first with pure bisulphide of carbon, and then with 

 bisulphide of carbon in which iodine had been dissolved ; he found 

 that the percentages of luminous to total radiation from the three 

 sources were — 



From measurements made by Mr. Merritt and Mr. Nakano (" Efficiency 

 of Methods of Artificial Illumination," by Professor E. Nichols ; ' The 

 Electrical Engineer,' May, 1889), the efficiency (i.e., the ratio of 

 luminous radiation to total radiation) of arc and incandescent electric 

 lamps appears to be about 10 and 5 per cent, respectively. 



The experiments of which an account is contained in this paper 

 were commenced with the object of repeating and extending Thomsen's 

 observations ; soon after they had been begun a notice appeared in 

 ' Nature ' (July 4, 1889) of a communication made on June 7, 1889, 

 to the Physical Society of Berlin, by Dr. R. von Helmholtz, on the 

 radiation from flames, and under these circumstances it seemed 

 desirable to finish the experiments which had been already com- 

 menced, but not to go on with the enquiry until Dr. von Helmholtz's 

 paper should have been published in full. 



The experiments were made by allowing the radiation from an 

 Argand gas flame to fall on the face of a thermopile, either directly 

 or after passing through different thicknesses of water, or of a solution 

 of alum in water, contained in cells with glass ends. 



Instead of keeping the source of heat at a fixed distance from the 

 thermopile, and noting the deflections of the galvanometer, which 

 was the method employed by Melloni and those who have followed 

 him, the distance between the source and the face of the pile was 

 made the variable, and the distances noted at which equal deflections 

 were produced ; thus the necessity for calibrating the galvanometer was 

 avoided, and also for assuming that the (corrected) deflections of the 

 needle were strictly proportional to the amount of radiant energy 

 incident upon the face of the pile. 



On the other hand, it was necessary to assume that the amount of 

 energy which reached the thermopile varied inversely as the square 



Incandescent platinum. . 

 Bright part of gas flame 

 Arc lamp c 



4-17 

 4-0 

 10-0 



