HEATING BY HOT WATER, HOT AIR, AND STEAM. 173 



bodies tend to rise vertically, in a dense medium, it follows, that 

 when heated air enters a house by an aperture at one part of it, 

 a very large portion of the heated air thus entering, must rise 

 immediately towards the roof; and in practice we find this to be 

 exactly the case. For, let any person examine the roof of 

 a hot-house in a frosty night, heated by a hot-air stove, and he 

 will perceive the part immediately above the entrance of the 

 air quite warm by the ascending heat, while all the rest of the 

 roof may be covered with ice or snow. 



But the atmosphere of a house heated artifically, by whatever 

 means, is always in motion ; with hot-water pipes it may be 

 less perceptible, for the reasons already stated, but it is not 

 the less real. The motion given to the atmosphere of a house 

 depends upon the difference of temperature between the two 

 bodies .of air, externally and internally; therefore a motion must 

 continue in the air of a house artificially warmed, so long as the 

 house requires warming, — that is, as long as any difference ex- 

 ists between the internal and external atmospheres. 



Some advocates of hot-air heating found their arguments upon 

 the fact that air can be raised to a higher temperature, in a 

 given time, by a given amount of caloric, than water. This is 

 probably true, if we calculate according to the bulk, without re- 

 gard to the density, of the respective bodies ; but, supposing it to 

 be true, then we know that, by the law already referred to, its 

 rapidity in warming will just be in exact proportion to its rapid- 

 ity in cooling, and vice versa. It is, therefore, manifest, that this 

 property militates against it as an agent in heating horticultural 

 buildings, as it is well known to be an all-important point, in 

 warming these structures, to obtain an equilibrium of heat for 

 the greatest length of time, and with the least possible amount 

 of attention, and experience has fully concluded that this is most 

 effectually and most easily obtained by the circulation of hot 

 water through wooden and metallic radiators and conductors. 



Suppose, for instance, that a house, containing 4000 cubic 

 feet of air, is required to be heated, from 32° to 60°, and 

 suppose the external thermometer to remain stationary at this 

 point ; then, by calculation, we find that it requires double the 

 amount of fuel to heat the atmosphere through the 28 degrees 



