10 W. Ferrel — Law of Thermal Radiation. 



ing the laws is, therefore, even more uncertain than that by 

 means of the residuals as in § 4, and it leaves considerable un- 

 certainty with regard to the best form of expression of the 

 law or the values of the constants to be used in the expression. 

 9. From the differences in the rates of cooling of a bare and 

 a silvered cylindrical thermometer from r=75° to r = 137° 

 Stefan obtained from his law the following quotients : 



4848 4588 4621 4624 4641 



These indicate that Stefan's law must hold pretty well for 

 this range of temperature, the mean temperature being 106° ; 

 but the range of temperature being short, the law might be 

 varied considerably, that is the value of a in (15) might be con- 

 siderably greater or less than 4, without affecting much the 

 equality of the quotients. 



Dividing the differences in the rates of cooling by (1*0077 

 — 1) he obtained the following ratios : 



6212 6236 632? 6373 6432 



These numbers do not satisfy so well the condition of equal- 

 ity, but show, allowing for small errors of observation, a regu- 

 lar increase of values with increase of temperature, indicating 

 that Dulong and Petit's law in some measure fails, and is not 

 as correct as Stefan's law for this range of temperature, and 

 that a value of a in (14) considerably greater than 1*0077 is re- 

 quired here. 



Again, Stefan obtained the quotients below corresponding to 

 the values of d in the first line, the temperature of the inclosure 

 being 14-7.° 



48-18° 



55-58 



80-98 



5407 

 7044 



5418 

 7105 



5417 by Stefan's law. 



7277 by Dulong and Petit's law. 



Quotients 



These indicate, so far as can be inferred from so short a 

 range of temperature, that Stefan's law, at these temperatures, 

 is more nearly correct than that of Dulong and Petit, the latter 

 quotients again indicating that Dulong and Petit's law fails 

 here and that a value of a in (14) greater than 1*0077 is re- 

 quired to make the quotients equal. 



It should be considered here that where the law of the radia- 

 tion of glass is deduced from the differences in the rate of 

 cooling of bare and silvered bulbs, it is assumed that the laws 

 of both are the same. This is, most probably, not the case, 

 but the radiation of the silver is so small that it cannot affect 

 the results much. 



10. We come now to the examination of a series of experi- 

 mental observations of a different kind, in which the relative 

 radiativities of the face of a Leslie's cube coated with lamp- 



