in Fluids. 317 



ance tends to retard the cooling of water when it is ex- 

 posed in a cold atmosphere. 



It is well known that there is no communication of 

 Heat between two bodies as long as they are both at the 

 same temperature ; and it is likewise known that the 

 tendency of Heat to pass from a hot body into one 

 which is colder, with which it is in contact,, is greater, 

 as the difference is greater in the temperature of the two 

 bodies. 



Suppose now that a mass of very cold air reposes on 

 the quiet surface of a large lake of fresh water at the 

 temperature of 55 of Fahrenheit's thermometer. The 

 particles of water at the surface, on giving off a part of 

 their Heat to the cold air with which they are in contact, 

 and in consequence of this loss of Heat becoming spe- 

 cifically heavier than those hotter particles on which they 

 repose, must of course descend. This descent of the 

 particles which have been cooled necessarily forces other 

 hotter particles to the surface, and these being cooled in 

 their turns bend their course downwards ; and the whole 

 mass of water is put into motion, and continues in mo- 

 tion as long as the process of cooling goes on. 



Before I proceed to trace this operation through all 

 its various stages, I must endeavour to remove an ob- 

 jection which may perhaps be made to my explanation 

 of this phenomenon. As I have supposed the mass of 

 air which rests on the surface of the water to be very 

 cold, and as I have taken it for granted that there is no 

 communication whatever of Heat between the particles 

 of water in contact with this very cold air and the neigh- 

 bouring warmer particles of water, it may be asked how 

 it happens that these particles at the surface are not so 

 much cooled as to be immediately changed to ice. To 



