186 ANNUAL REPORT SMITHSONIAN INSTITUTION, 1920. 



sidering both velocity of vertical movement and volume occupied, 

 or velocity times volume, the atmosphere as a whole is always ascend- 

 ing, a fact not only interesting itself, but also of some importance to 

 both the aeronaut and the aviator. 



Measured in terras of mass. — Whatever the volume relations 

 between ascending and descending air may be, it would seem that 

 at least the mass that goes up and the mass that eventually returns 

 must certainly be the same. But, on the contrary, they indeed are 

 far from it, for one of the important constituents of the atmosphere, 

 water vapor, often amounting, in places, to 1 per cent, and occa- 

 sionally to more than 2 per cent of the whole, invariably ascends 

 as a gas, as a distinct part and parcel of the air; but descends, in 

 great measure, not as a gas at all, not as any part whatever of the 



air, but as a liquid in 

 the form of rain, or a 

 solid, such as snow and 

 hail. 



Paradoxical, there- 

 fore, as it may be, a 

 greater mass of air ac- 

 tually does go up — 

 more by at least 20,- 

 000,000 tons per second, 

 the measure of world- 

 w i d e precipitation — 

 than ever comes down. 



TO COOL AIR, HEAT IT. 



The air referred to in 

 this seemingly absurd 

 statement is not that 

 topsy-turvy kind Alice 

 might have found in 

 Wonderland, but just 



Fig. 2. — Vertical temperature gradients of free air. 



that ordinary kind in which we have always lived ; and the phenom- 

 enon itself, however contrary to experience it may seem, one of great 

 importance and almost continuous occurrence. 



This paradoxical result is easy to explain with a diagram. To 

 this end let AB and A'B', figure 2. be two adiabatic gradients of 

 the free air; that is, let each indicate a temperature change of 1° C. 

 for every 100 meters change in elevation — the relation between the 

 temperature and elevation of a rising or falling mass of air that 

 during its travel neither gains heat from, nor loses it to, any outside 

 object, such as the surrounding atmosphere. Let EE be any actual 

 temperature gradient (nearly always less than the adiabatic), in this 



