134 07h the Ground Ice observed in 



speaking, was at 10° will soon be at 9° ; but at 9° the water will 

 possess more density than at 10° ; then, in consequence of the 

 principle of hydrostatics formerly mentioned, it will sink to the 

 bottom of the mass, and be replaced by a layer not yet cooled, 

 whose temperature is 10°. That, in its turn, will be affected 

 like the first layer, and so on of the rest. In a greater or less 

 time the whole mass will then be at -f- 9°. 



Water at + 9° will become cool in the same way as at 10° by 

 consecutive layers. Each in its turn, on coming to the surface, 

 will lose one degreee of temperature. The same phenomenon 

 will reappear, with similar circumstances, at 8°, 7°, 6°, and 5° ; 

 but, on sinking to 4°, every thing will be changed. 



At + 4° (39°.2 Fahr.) water will actually reach its maximum 

 of density. Should the action of the atmosphere take away a 

 degree of heat from the superficial layer, or descend to 3°, the 

 layer will be less dense than the portion of fluid which it co- 

 vers ; it will never sink into it. An additional diminution in 

 the heat will not cause it to sink more, as water at -f 2° is light- 

 er than at + 3°, &c. 



It is quite obvious, however, that the layer in question, by 

 remaining always on the surface, incessantly exposed to the 

 cooling influence of the atmosphere, will at length lose the first 

 4° of its heat. It will end by falling to zero, and freezing. 



The superficial sheet of ice, however singular the phenomenon 

 may be, is then found resting on a liquid mass, whose tempera- 

 ture, at least at the bottom, is 4° above zero. 



The congelation of stagnant water could not evidently take 

 place in any other manner. I repeat, that no person has ever seen 

 the formation of ice beginning at the bottom of a lake or a pond. 



Let us briefly examine the modifications which the motion of 

 the liquid should produce. 



The effect of this motion, when it is rather rapid, when it 

 forms eddies, and flows over a rocky or unequal channel, is per- 

 petually to mix all the layers. The hydrostatic order on which 

 we have insisted so much is overthrown. The water, then, which 

 is lightest does not always float on the surface. The currents 

 are precipitated into the general mass, which is thereby cooled, 

 and whose temperature soon becomes equal throughout. 



To repeat, in a deep mass of stagnant water, the temperature 



