MELLOXI ON THE NOCTURNAL COOLING OF BODIES. 



'161 



tioo limits, 0°'03 must be subtracted from the thermometer B, 

 and 0°-062 from the thermometer C, in order to obtain the tem- 

 perature of thermometer A. After having determined the cor- 

 rections to be made in the actual circumstances of the experiment, 

 the blackened armatures of the thermometers A and B were re- 

 moved and replaced by well-polished armatures of silver. The 

 recipients of these two thermometers were then closed, as well 

 as the third recipient containing the blackened thermometer C. 



The weather continued calm and serene, and the temperature 

 of the air had not altered much on the terrace where the appa- 

 ratus was placed. At ten o'clock A indicated 17°*65, B 17°"69, 

 C 17°'70. Applying to the two latter their respective corrections, 

 we have l7°-66 for B, and 17°'638 for C, so that the three ther- 

 mometers were at the same temperature. The covers of B and 

 C were then removed, leaving A enclosed, and at half-past ten 

 o'clock the observations were resumed, of which the following 

 Table contains the results : — 



Subtracting from the mean of A the corrected means of B and 

 C, we obtain the differences 0°-108 for B and 3--5S1 for C ; and 

 consequently if we make 3-581 = 100, the value of the emissive 

 or radiating power of the silver forming the armature of the 



thermometer B will be i^, 100 = 3-026, a value which may be 

 3581 



regarded as exact within a limit of error of fifteen or twenty 



