68 THE YEAH. 



below 39^; the instant that it gets colder than that, it 

 becomes lighter, and has of course no tendency to sink ; 

 and thus the heat of all which is below the surface is 

 retained ; and the ice, when once formed, as it freezes 

 slowly downward, actually warms the unfrozen water 

 below, by the quantity of heat that it gives out during 

 the process of congelation. Those who have not re- 

 flected upon the subject are apt to look upon this as a 

 paradox ; but it is a very simple, and far from an un- 

 important truth. When the state of a body is changed 

 from liquid to solid there is a much greater change of 

 heat than the thermometer indicates from the body 

 itself. Water at 32, and ice, have the same tempe- 

 rature, and yet 140 of temperature are required to 

 thaw the ice into water ; and that must be given out 

 before the water can become ice. If water be kept 

 shaded and steady, it may be cooled in a liquid state, 

 considerably below the freezing point; but the moment 

 that it is agitated or otherwise placed in circumstances 

 that would dispose it to freeze, it gets warmer without 

 the abstraction of heat from any thing else, by the 

 separation of a part of the heat that kept it in the 

 liquid state, and which was latent, or had no effect 

 upon the thermometer. Rather it had a negative effect; 

 for water that shows 22 by the thermometer when 

 liquid, will mount up to 32 as it begins to freeze, and 

 thus actually produce 1 of heat ; while in the conge- 

 lation, it will give out just as much as is required to 

 melt the ice into water at 32. Now, when the process 

 of freezing and giving out heat, takes place downwards 

 under water, the heat given out must be communicated 

 to the water, because the ice itself is an imperfect con- 



