EXPANSION. 55 



The greaterpart of this expansion is owing to 

 the latter of the causes above mentioned. The 

 freezing mixture employed is made of two 

 parts snow to one part salt, brought into 

 the cup alternately, in small portions. It 

 is well to wrap the cup in flannel, or other 

 cloth, to prevent loss of heat. From ten to fifteen 

 minutes are required for the experiment. If the water 

 is perfectly frozen, the tube will be cracked by its ex- 

 pansion. 



122. COLD WATER FLOATS ON WARMER 



coldwaterfoat WATER AN PROTECTS IT. It Was shown 



on warmer wa- i n the last paragraph that very cold water 

 (below 39) is in an expanded condition, 

 and occupies more space than warmer water. It fol- 

 lows that it is lighter, arid will float on warmer water. 

 As the weather grows colder each winter, and the time 

 approaches for the formation of ice, in rivers and lakes 

 the cold water does actually float on the warmer, on a 

 grand scale, and protect it from the cold. The body 

 of water being thus protected, ice never forms many 

 feet thick. The case would be very different if water 

 grew constantly heavier by cold. The surface water 

 would then constantly sink, until all were reduced to 

 the freezing point. Cooling does, in fact, proceed in 

 this way until the temperature sinks to 39 ; then the 

 exception comes in play, and the surface water, as 

 before stated, retains its place and exerts its protecting 

 influence. When ice is subsequently formed it has the 

 same effect. 



