AT THE earth's SURFACE. 419 



This is true, at least, so far as relates to any power of the at- 

 mosphere to convey heat. There is still to be considered, 

 howevei", another mode in which caloric may be supposed to 

 be discharged from our planet, — that by radiation. Rays pos- 

 sessed of heating power are thrown off from a body at a high 

 temperature, and by the discharge of these its temperature is 

 reduced. But various considerations shew, that this would be 

 a very inadequate source of the escape of heat from the inte- 

 rior part of the earth. 



Thus the radiation of caloric is inconsiderable, except from 

 a body which is heated ; and the quantity radiated increases 

 as the temperature rises in a much higher ratio than the in- 

 crease of temperature itself. At low temperatures, therefore, 

 it must be extremely small. At 100° it is scarcely apparent, 

 from experiment ; and at 50° is not sensible. Not only, too, 

 does the quantity diminish rapidly with the temperature, but 

 the projectile force of the rays emitted becomes less, so that 

 those which are discharged at low temperatures, are incapable of 

 penetrating media such as glass, which those at high tempera- 

 tures penetrate with facility. And as even the latter are, to a 

 certain extent, intercepted by the atmospheric air, the former 

 must be more completely arrested. At 50", therefore, the me- 

 dium temperature of the globe, and which prevails over so 

 large a portion of its surface, we have no reason to believe 

 that a discharge of caloric by radiation, into remote space 

 can take place to any extent. 



At 



ascent and descent of tlie aerial mass, and the principal effect will be, to accumu- 

 late temperature at the surface. The condition is one, however, which need not 

 to be taken into account, as it is one which has never existed with regard to our 

 planet ; and the argument remains as is stated above, while there is inequality in 

 the distribution of heat over the circumference of the globe. 



■!'<' 



