LAWS OF HEAT. AIR. 107 



even the most violent tempests and torrents of 

 rain, maybe considered as oscillations about the 

 mean or average condition belonging to each 

 place. All these oscillations are limited and 

 transient ; the storm spends its fury, the inunda- 

 tion passes off, the sky clears, the calmer course 

 of nature succeeds. In the forces which produce 

 this derangement, there is a provision for making 

 it short and moderate. The oscillation stops of 

 itself, like the rolling of a ship, when no longer 

 impelled by the wind. Now, why should this be 

 so ? Why should the oscillations, produced by 

 the conflict of so many laws, seemingly quite un- 

 connected with each other, be of this converging 

 and subsiding character? Would it be so under 

 all arrangements ? Is it a matter of mechanical 

 necessity that disturbance must end in the res- 

 toration of the medium condition ? By no means. 

 There may be an utter subversion of the equi- 

 librium. The ship may roll too far, and may 

 capsize. The oscillations may go on, becoming 

 larger and larger, till all trace of the original con- 

 dition is lost; till new forces of inequality and 

 disturbance are brought into play ; and disorder 

 and irregularity may succeed, without apparent 

 limit or check in its own nature, like the spread 

 of a conflagration in a city. This is a possibility 

 in any combination of mechanical forces ; why 

 does it not happen in the one now before us ? By 

 what good fortune are the powers of heat, of 



