METEOROLOGICAL PARADOXES HUMPHREYS. 187 



case 1° C. per 120 meters change of elevation. If, then, under these 

 conditions, a mass of air having the temperature and elevation indi- 

 cated by C, say, of the figure, be heated 1° C, or shifted in the figure 

 to W, it will correspondingly expand and consequently be forced 

 up by the surrounding denser air — will ascend, as we say. As it rises, 

 it will cool, by expansion, along the adiabatic gradient A'B', a ad, 

 therefore, will come into equilibrium with the surrounding atmos- 

 phere where this gradient intersects the actual gradient EE, or at 

 the level and temperature indicated by W. Clearly, then, under the 

 assumed conditions, such as are very common in nature, a mass of 

 air heated 1° C. rises 600 meters, and in so doing cools 6° C, or to a 

 temperature 5° C. lower than it had before it was heated. Of course, 

 the warm air does not rise strictly adiabatically, though probably 

 very nearly so ; but in so far as it actually does lone heat it comes to 

 equilibrium at a correspondingly lower level and warmer tempera- 

 ture. 



It is precisely this paradoxical process of cooling by heating, the 

 heating being mainly at the surface, however, that leads to the for- 

 mation of cumulus clouds and generates the familiar " heat " thun- 

 derstorm. In fact, it is quite possible to produce a cumulus cloud, 

 and even a local shower, through the action of a large surface fire. 

 It should be noted in this connection that though combustion adds 

 much water vapor to the air, five-ninths the weight of the fuel con- 

 sumed even in the case of absolutely dry cellulose, nevertheless, the 

 cumulus cloud over the fire is due essentially to the expansional or 

 dynamical cooling of the ascending air. 



TO WARM AIR, COOL IT. 



This paradox is the converse of the one just discussed, and is readily 

 explained in much the same way. Referring again to figure 2, let a 

 mass of free air having the altitude and temperature indicated by 

 W in the figure, be cooled 1° C, or its position shifted to C. It 

 will at once become denser than it was, follow the adiabatic gradient 

 AB as it falls to lower levels, and, therefore, come to rest at the level 

 and temperature indicated by C, or at the intersection of the adia- 

 batic gradient followed and the existing gradient. That is, as a result 

 of the initial cooling of 1° C, the given mass of air will fall 600 meters 

 and become 5° C. warmer than it was before it was first cooled. In 

 so far, however, as the falling air gains heat from the surrounding 

 warmer atmosphere, it will come to rest at a correspondingly greater 

 elevation and lower temperature. 



This paradoxical phenomenon of warming by cooling is very fre- 

 quently and very prettily illustrated by the evening disappearance of 

 small detached clouds, such as alto-cumuli, fraeto-stratus, etc. As 



