EFFECTS OF HEAT AND COLD. 



437 



quantity of pure spirits of wine is poured upon 

 the water, and lighted, it will burn for some time, 

 and of course communicate heat to the surface of 

 the water ; yet the thermometer, separated only 

 by a thin layer of water from the burning spirit, 

 will not indicate any, or only the smallest possible 

 rise of temperature. This is because fluids are 

 bad conductors of heat. If the flame were applied 

 to the bottom of the water, instead of its sur- 

 face, the result would now be different, and the 

 liquid would boil in consequence of what has 

 been, before called the convection, or carrying 

 upwards of heat, by the 

 light and expanded par- 

 ticles, as they rise to the 

 surface of the fluid. Let 

 us perform the converse 

 of this experiment, and 

 place, instead of a flaming 

 layer of spirit, a flat piece 

 of ice upon the surface of 

 the water. The effect is 

 immediately perceptible 

 by the falling of the ther- 

 mometer ; the particles of 

 water being cooled be- 

 come heavier, and sink ra- 

 pidly downwards. Hence 

 it is evident that while 



