Emission and Transmission of Heat 55 



836. In regard to the transmission of heat from one gas to 

 another through a plate of metal, as for equal volume gases have 

 much less capacity for heat than liquids and as their conductivity 

 is very feeble, we may regard the influence of the nature and thick- 

 ness of the plate as absolutely nonexistent, since the quantity of 

 heat which the plate can transmit, even under the most unfavor- 

 able conditions, is incomparably greater than that which actually 

 traverses it, and consequently in no case can the thickness of the 

 metal delay the transmission . The quantity of heat which tra- 

 verses the plate is determined solely by the difference of tempera- 

 tures of the two gases, the absorbing and radiating powers of the 

 two surfaces of the plate, and above all by the movements of the 

 sheets of gas in contact with these surfaces. 



Thus we see that in every case the rapid renewal of the layers 

 of gas or liquid, which bathe the surfaces of the metal plate, has 

 a very great influence on the transmission of heat, but that this 

 circumstance is of much greater importance with gases than with 

 liquids. 



837. One should then in designs seek the disposition 

 most favorable to this renewal, taking advantage of the motions 

 which the fluids must have in order to pass through the appara- 

 tus, and those resulting from heating and cooling. But with gases, 

 one may also artificially produce motions throughout their masses 

 which would cause rapid renewals of the layers in contact with 

 the metallic surfaces, either in some direct way requiring only a 

 small amount of power, or by employing a part of the force re- 

 sulting from the passage of the gas. 



838. Consider, for example, air heated to an extremely high 

 temperature, escaping through a horizontal cylindrical conduit 

 surrounded by water which it is intended to heat. The layers of 

 air in contact with the metal cool very quickly; but as all the 

 little elementary veins possess motion only in the direction of the 

 axis of the conduit, the layers will change place but slowly, for 

 the sole cause of change lies in the increase of density resulting 

 from cooling ; it only exists for the upper half of the conduit, 

 and it only operates slowly. This same state of affairs will con- 

 tinue no matter what the direction of the conduit. We may 

 imagine according to this, that if the area of the conduit were 



