464 REPORT — 1873. 



mass at near the temperature required. This latter fact was ascertained by 

 finding the amount of heat the platinum emitted when thrown into a calori- 

 meter containing a known quantity of water. As the amount of heat emitted 

 was very small, special precautions had to be taken in guarding the calorimeter 

 and in getting the mass of platinum transferred. The calorimeter, containing 

 about 100 grammes of water, was floated in a cistern (having been pre- 

 viously placed in the middle of a tin cylinder, leaving an annular space 

 between), and so loaded that the water in the calorimeter was sunk to the 

 level of the water in the cistern. The Buusen burner was placed in a tin 

 vessel loaded with shot, so as to give a flame the upper half of which was 

 above the level of the water in the cistern. By this means constancy of 

 temperature was maintained, and the results agreed closely together. It is 

 easy to be convinced that a mass of platinum like that employed, radiating 

 freely, is rarely heated above a temperature of 1100° or 1200° C. Compa- 

 risons were made between platinum in the Bunsen burner and lime in the 

 oxyhydi'ogen flame, and also between lime in both. 



The photometer employed for comparing the lights was on the principle of 

 that recommended by Bunsen, A wooden box, about 8 inches long, 4 inches 

 broad, and 3 inches deep, containing several diaphragms with circular aper- 

 tures, thoroughly blackened in the interior, and ha\'ing the aperture of the middle 

 diaphragm covered with a piece of Swedish filter-paper, marked with one or 

 two circular spots of paraffin, was employed to exclude extraneous light and 

 to obtain good definition. By this means it is possible to obliterate com- 

 pletely the spot of paraffin, and thus gain greater confidence in the results. 



From the mean of a great number of experiments made in this way, the 

 luminous intensity at about 2000° C. is from 500 to 550 times that at 1040° C. 

 The calculated amount given by the above formula for the exact temperature 

 of 2000° C. is 484 times that at the lower temperature. According to the 

 formula of Becquerel, it would be about 24,000,000 times that at the lower 

 temperature. This empirical law, therefore, gives with considerable approxi- 

 mation the luminous intensity up to a temperature of 2000° C. 



Total Radiation. — If the law of Dulong and Petit for the velocity of cooling 

 was true for temperatures above the range of the actual observations made in 

 support of the law, the amount of heat radiated per unit of time would be 

 found by multiplying the velocity of cooling at the temperature considered 

 into the specific heat at that temperature and into the weight of the substance. 

 From this may also be calculated the amount radiated per unit of surface. 

 In fact, for the same substance the relative quantities of heat evolved at two 

 difl'erent temperatures would be to each other as the velocities of cooling if 

 the specific heat and the emissive power remained constant. This would give 

 an extraordinarily rapid rate to the growth of total radiation. For instance, 

 taking the temperatures of 2000° C. and 700° C, we find, according to Dulong 

 and Petit's law, 



Q^--for-« -^l,54o, 



where a is the constant 1'0077. 



Thus a substance radiates at a temperature of 2000° C. 21,000 times as 

 much heat per unit of time as it does at a temperature of 700° C. 



In order to compare the total radiation as given from the law of Dulong 

 and Petit with that of actual experiment, a series of observations were made, 

 and the total heat evolved registered by the use of Pouillet's pyrheHometer. 

 For this purpose, a spherical ball of lime, 8 millims. in diameter, was formed 



