Intelligence and Miscellaneous Articles, 387 



rounding temperature on account of radiation, contact of the air, 

 and of the supports. 



Into this first body I introduce a second, which I suppose also 

 homothermous. Let y be its temperature, C its thermal capacity, 

 e its loss of heat per second for 1° of excess at all the points where 

 it is not in contact with the first body, X its loss at all points where 

 it is in contact with the first body (this will be the external con- 

 ductivity). 



The immersion of the second body in the first will cause the 

 temperatures x and y to vary according to the following differential 

 equations : — 



Integration gives for x and y the functions : — 



x=Me- mi +'Ne- nt ; 



y—~p e - mt -\-Qe- nt . 



Observation of the course of the temperatures will give the quan- 

 tities M, N, P, Q, w, n — with which we obtain 



E PQ ' (1) 



x _ -( B - W p 



F MQ-NP ' y) 



p = mM(Q-N)-n(P-M) 



P """ MQ-NP ' w 



e_ mP(Q-N)-nQ(P-M) 



C MQ-NP • K) 



A second experiment, in which, P remaining constant, will 

 have been augmented by a known quantity Jc, will give 



C + fc_-M,N, 



~r p^7 (5 -' 



and, with (1), permit the calculation of P and C. 



If the immersed body is not homothermous, I suppose it (for 

 simplicity) of a cylindrical form, and replaced by two concentric 

 bodies : — the one exterior, having in all its points the temperature 

 z of the surface of the real body ; the other interior, its tempera- 

 ture y suitably chosen in order that the sum of these bodies may 

 replace the real body in its effects upon the exterior body. This 

 second case, which I will call dithermy, will give three equations 

 instead of two. The calculations will be longer, but not more com- 

 plicated; and they will furnish, besides the physical unknown 

 quantities of the preceding problem, the measure of the internal 

 conductivity of the real body. 



In order to approximate to the conditions assumed in this method, 

 certain experimental arrangements are necessary : — 



