F. J. Rogers — Magnesium as a Source of Light. 307 



The method employed for the determination of this ratio 

 was to note the galvanometer deflections produced by the 

 direct radiation of the source upon the face of a thermopile in 

 circuit with a galvanometer, and then the effect of radiation 

 from the same source when the " dark " heat is cut off by a 

 glass cell containing a saturated solution of alum. Two alum 

 cells were used in all cases. The second cell eliminated the 

 effect of the warming up of the first one due to the absorp- 

 tion of radiant heat. The total thickness of the two alum 

 cells was 72 mm . The E. M. F., produced in a thermopile by 

 radiation on its face, varies directly as the intensity of the 

 radiation. Therefore the ratio of the galvanometer deflection 

 obtained when the alum cells are interposed, to the deflection 

 produced by the unobstructed radiation, properly corrected if 

 there has been a change in the resistance of the galvanometer 

 circuit, gives the ratio of the radiation through the alum cells 

 to the total radiation. When this ratio is corrected for the 

 light absorbed and reflected by the alum cells and for the 

 " dark " heat which passes through them, we have the radiant 

 efficiency of the source of light. 



The amount of light absorbed and reflected by the alum 

 cells was determined photometrically. The candle power of a 

 constant source of light was measured both when the two alum 

 cells were and were not interposed between it and the photo- 

 meter, and it was found that "611 of the light incident upon 

 the face of the first alum cell passed through both. The 

 " dark " heat transmitted by the alum cells was measured with 

 the aid of a cell containing an opaque solution of iodine in 

 bisulphide of carbon. This transmitted the " dark " heat but 

 intercepted the light. In the case of a gas flame '004 of the 

 incident "dark" heat was transmitted through both alum cells. 



For the sake of comparison the radiant efficiency of candle 

 light and gaslight was determined. The final results of a num- 

 ber of measurements are as given below. 



Table III. 

 Radiant efficiency of candle-light, '0153 



u u u o-ociio-ht j Bat's wing burner, -0128 



. gaslight, | Argand « . 0161 



Julius Thomsen* found the efficiency of petroleum flame to 

 be about "02. His method was similar to the above except 

 that he used distilled water instead of a solution of alum and 

 disregarded the heat transmitted. S. P. Langleyf by a very 

 different method, namely by measuring the intensity of the 



* Poggendorff's Annalen, cxxv, p. 348. f Science, vol. i, p. 483. 



