76 CORRELATION OF PHYSICAL FORCES. 



pressure increased as the sensible heat. M. Despretz, in 

 1832, made some experiments, which led him to the con- 

 clusion that the increase was not in the same ratio as the 

 sensible heat, but that yet there was an increase ; a result 

 confirmed and verified with great accuracy by M. Regnault, 

 in some recent and elaborate researches. What seems to 

 have occasioned the error in Watt and Clement Desormes' 

 experiments was, the idea involved in the term latent heat ; 

 by which, supposing the phenomenon of the disappearance of 

 sensible heat to be due to the absorption of a material sub- 

 stance, that substance, ' caloric,' was thought to be restored 

 when the vapour was condensed by water, even though the 

 water was not subjected to pressure ; but to estimate the 

 total heat of vapour under pressure the vapour should be 

 condensed while subjected to the same pressure as that under 

 which it is generated, as was done in M. Despretz and M. 

 Regnault's experiments. 



M. Seguin, in 1839, controverted the position that derived 

 power could be got by the mere transfer of heat, and by 

 calculation from certain known data, such as the law of Mar- 

 iotte, viz. that the elastic force of gases and vapours increas- 

 ed directly with the pressure ; and assuming that for vapour 

 between 100 and 150 centigrade, each degree of elevation 

 of temperature was produced by a thermal unit, he deduced 

 the equivalent of mechanical work capable of being perform- 

 ed by a given decrement of heat ; and thus concluded that, 

 for ordinary pressures, about one gramme of water losing 

 one degree centigrade would produce a force capable of rais- 

 ing a weight of 500 grammes through a space of one metre : 

 this estimate is a little beyond that given by the converse ex- 

 periments of Mr. Joule, already stated, in which the heat 

 produced by a given amount of mechanical action is estimat- 

 ed. I am not aware that the amount of mechanical work 

 which is produced by a given quantity of heat has been di- 

 rectly established by experiment, though some approximative 



