LIGHT AND ITS ARTIFICIAL PRODUCTION. 297 



uess, for the melting point of platinum is about 2,000°, while the redness 

 corresponds to 800° on the absolute scale. If we assume Wien's law 

 as expressing the increase in intensity with the temperature, the energy 

 emitted by platinum at the absolute temperature of 800° is to that at 

 2,000° as 800'^ is to 2,000% i. e., roughly in the ratio of 1 to 100. The 

 mantle of an incandescent gas light is probably at about the same tem- 

 perature as that of melting iilatinum if we assume that it attains 

 the highest temperature of the flame. If we assume, in addition, that 

 thorium oxide radiates as much energy as jjlatinum at the same tem- 

 perature, an assumption not altogether justifiable, the amount of light 

 emitted by a given area of the incandescent substance would have to 

 be equal in both cases. In reality thorium oxide corresponds more 

 nearly to a black body, in Kirchhoff's sense of that word, and therefore 

 the former would radiate, ceteris paribus, more than the latter. The 

 experiment I will now show will teach us how much more light ferric 

 oxide radiates than polished platinum. I have written with ink a few 

 words on the piece of platinum, and I now heat it to incipient white- 

 ness by passing an electric current through it. I will project on the 

 screen the side of the strip on which I have written and you can plainly 

 see the words appear bright on a darker background. This shows that 

 the iron oxide left on heating the ink radiates much more light than 

 does the polished surface of the platinum at the same temperature. 



By the assistance of our knowledge of the radiating power of lumi- 

 nous substances and of the temperature at which they become luminous 

 I can compare theoretically the intensities of different sources of light. 

 Unfortunately, considerable difficulties are met with in the measure- 

 ment of high temperatures. According to the latest measurement of 

 the radiation of the sun by Paschen its temperature is about 5,400° C, 

 and according to Yiolle's measurements the temperature of the electric 

 arc is about 3,600° C. Assuming that the material of the sun radiates 

 light as well as carbon, the quantity of light emitted by a given area 

 in the two sources is in the ratio of 3 Ho 2^, or about as 8 is to 1. Kow 

 the sun subtends at the earth an angle of 32 minutes of arc, and a sur- 

 face of a square centimeter in area would subtend about the same angle 

 at a distance of 1 meter. On the assumptions made we obtain the 

 result that the sun will illuminate a surface eight times as brightly as 

 an incandescent carbon surface 1 square centimeter in area at a tem- 

 perature of 3,600° (that of the electric arc) at a distance of 1 meter, or, 

 cseteris paribus, just as brightly as a surface of 8 square centimeters at 

 the same distance. 



The results obtained by comparing sources of light in which the same 

 substance, for instance carbon, is brought to incandescence are more 

 reliable. To this class belong, besides all free burning flames, the 

 incandescent electric light and the arc light. Since in all of these 

 the incandescent substance approximates a black body, their relative 

 temperatures can be determined by the color of the light emitted. The 



