52 CORRELATION OF PHYSICAL FOKCES. 



which they solidify, we get the apparent anomaly that the 

 motion or mechanical force generated by heat or change of 

 temperature is reversed in direction when we arrive at the 

 point of change from the solid to the liquid state. Thus a 

 piece of ice at the temperature of Zero, Fahrenheit, would 

 expand by heat, and produce a mechanical force by such ex 

 pansion until it arrives at 32 ; but then by an increment of 

 heat it contracts, and if the first expansion had moved a pis 

 ton upwards, the subsequent contraction would bring it back 

 to a certain extent, or move it downwards, an apparent nega 

 tion of the force of heat. 



Again with water above 40, i. e, above its point of 

 maximum density, a progressive increment of cold or decre 

 ment of heat would produce contraction to a certain point, 

 and then expansion or a mechanical force in an opposite direc 

 tion. Thus not only heat or the expansive force given to 

 other bodies by a body cooling would be given out by water 

 freezing, but also the force due to the converse expansion in the 

 body itself, and force would thus seem to be got out of noth 

 ing : but if water in a confined space be gradually cooled, the 

 expansion attendant on its cooling as it approaches the freez 

 ing point would occasion pressure amongst its particles, and 

 thence tend to antagonise the force of dilatation produced in 

 them by cooling, or to resist their tendency to freeze ; or in 

 other words, the pressure would tend to liquefaction, and con 

 versely to the usual effect of pressure, produce cold instead 

 of heat, and thus neutralise some of the heat yielded by the 

 cooling body. Hence we find that it requires a lower tem 

 perature to freeze water under pressure than when exempt 

 from it, or that the freezing point is lowered as the pressure 

 increases for bodies which expand in freezing an effect first 

 predicted by Mr. J. Thompson, and experimentally verified 

 by Mr. W. Thompson ; while as shown by M. Bunsen, the 

 converse effect takes place with bodies which contract in 

 freezing. Here the pressure cooperates with the effects of 



