1880.] Prof. J. Dewar. Studies on the Electric Arc, 91 



"parabolic curve." Assuming the general accuracy of this law for 

 high temperatures, the total radiation may be taken as nearly propor- 

 tional to the square of the temperature. From this law the hypo- 

 thetical temperature of the sun was "estimated as at least 11,000 C." 

 Rosetti has recently made a more elaborate investigation on the sub- 

 ject, and has arrived independently at a formula of a parabolic order. 

 Rosetti * represents his results by the equation 



/u=aT 3 (T — 0) — &(T— 0), 



where fx is the total radiation measured by thermo-electric current, 

 T° the absolute temperature of the source, 0° that of the medium 

 surrounding the pile, and a and h constants. However well this 

 formula may represent the complete series of the experiments, 

 it is certain that his results for temperatures above 150° may be 

 expressed within the limits of probable error as proportional to the 

 square of the temperature. To be convinced of this, it is sufficient to- 

 plot the logarithm of the respective values of the radiation and tem- 

 perature, when it will be found the results arrange themselves in a 

 straight line, the tangent of which may be 1*9 or 2 for the observa- 

 tions above 150°. Experiments made with the thermopile, surrounded 

 with an annular vessel, through which a continuous current of water 

 at constant temperature is caused to circulate, as represented in fig. 3, 

 where EF represents the section of the vessel, and CD a large water 

 screen, on the same plan, each having a narrow opening, about half an 

 inch in diameter, through which the radiant heat passed to the pile, 

 have confirmed the earlier results. The vessel for holding the mercury 

 or other substance to be heated to different temperatures has a radia- 

 ting face, which was made of the sheet iron used in the construction of 

 telephone plates, and the thermometer must be placed close to the 

 back of the front surface, and the face guarded with a screen, FGr. 

 The tube, CE, is connected with a condenser, when substances at their 

 boiling point are employed for giving fixed points. The form of the 

 apparatus is shown in fig. 4. 



This arrangement of the apparatus is necessary in order to get 

 anything like comparable results. The two following tables give the 

 records of two series of experiments, without any correction being 

 made in the numbers representing the deviations of the Thomson 

 galvanometer. 



• * " Becherches experimental sur la Temperature du Soleil " (Acad. E. dei Lincei. 

 1877-78). 



