concerning Heat. 151 



then, that the experiment with boiling water at the one 

 end and freezing at the other be made when the tem- 

 perature of the air is at the freezing point, or 32 of 

 Fahrenheit. 



In this case it is evident that the surface of the cyl- 

 inder must everywhere be cooled by the influence of the 

 surrounding atmosphere. The question, then, is to 

 determine the comparative effects, or the relative quan- 

 tities of refrigeration or loss of heat, that must take 

 place in the different farts of the cylinder ; and, in the first 

 place, it is clear that the hotter a given part of the cyl- 

 inder is, the more heat it must lose in a given time, by 

 the influence of the surrounding cold bodies; whence 

 we may conclude that the refrigeration of the surface of 

 the cylinder by the influence of the air and other sur- 

 rounding cold bodies must necessarily diminish from 

 the extremity of the cylinder, A, which is in contact with 

 the hot water, to its extremity, E, which is in contact 

 with the cold. 



From reasoning which appears incontrovertible, and 

 which the results of a great number of experiments 

 appear to confirm, it has been concluded that the celer- 

 ity with which a hot body placed in a cold medium is 

 cooled is always proportional to the difference between 

 the temperature of the hot body and that of the me- 

 dium. Considering this conclusion as established, we 

 may determine a priori what ought to be the gradation 

 of temperatures in the interior of a given solid cylinder 

 surrounded by air, one extremity of which is in contact 

 with a considerable body of boiling water, while the 

 other is similarly in contact with cold. 



We have seen that, if the surface of the cylinder were 

 perfectly isolated, the decrease of temperature from the 



