338 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 51 



lifting, or in case of cooling there is a disturbance of equilibrium 

 by the sinking of the center of gravity, which then brings about a 

 motion of the mass of air. 



If the above-given analysis does not yet suffice to make the sub- 

 ject perfectly clear, then perhaps the following considerations may 

 succeed in doing so. 



Assume that we have three cylindrical vessels, two of them filled 

 with mercury to the height of 76o mm , the third filled to the same 

 height with water. Let the external air pressure also amount to 

 76o mm mercury. Now assume further, that at the base of the 

 first vessel, filled with mercury, there is a piece of iron that is at the 

 beginning held down but by some appropriate arrangement may be 

 suddenly left free. When set free, the iron rises until it swims on 

 the surface of the mercury. But now this surface itself stands some- 

 what lower since the floating iron protrudes partly above it and the 

 center of gravity of the whole system is now somewhat lower. 



No one will imagine that the iron cools by rising, but will rather 

 at once perceive that its rising is at the cost of the sinking of the 

 mercury. 



At the bottom of the second vessel imagine a mass of air enclosed 

 in a small bell glass whose mouth opens downward. This 

 air is therefore under a pressure of two atmospheres. Turn the 

 bell glass over by appropriate mechanism so that its mouth opens 

 upward and the air rises through the mercury to its upper surface. 

 In this process the air expands and consequently cools. Assuming 

 that the ascent proceeds so rapidly that there can be no interchange 

 of heat between air and mercury or that the process is adiabatic, 

 then the amount of this cooling can be easily computed. 



The formula for tie computation of the final temperature t 2 is 



k - 1 

 273+/, (p 2 \ « 



273 + t, \p t 



where t t is the initial temperature; p t the initial pressure; and p 2 

 the final pressure and k the well-known constant 1.4 1. Under the 

 assumptions above made we have 



Px 2 



If now t t = o° then we find t 2 = — 50. 2 or a cooling of about 5o°C. 



If now we repeat the experiment last described in the third vessel 



filled with water up to 76o mm , then the air in the bell glass has the 



