[ 373 ] 



LXXVII. Ivqulries into the Laws of Dilatation of Solids, LI- 

 quidi, and elastic Fluids, and on the exact Measurement of 

 Temperatures. Bij Messrs. Dxtlong and Petit. Read to 

 tfie Institute 2dtkofMay 1816*. 



When we endeavour to scrutinize most of the questions re- 

 lative to the theory of heat, our progress is soon arrested by a 

 difficulty which reappears at every step under different forms. 

 What are we to understand by the word temperature? and what 

 is the relation which exists between the indications of" the ther- 

 mometer and the quantities of heat added or subtracted, in or- 

 der to produce determinate variations in the temperature ? Here 

 are two questions which must be resolved ere we can find, for 

 example, the trne laws of cooling : and it was. in fact, in order 

 to solve them, that we embarked in the inquiries, a part of 

 which we now submit to the judgement of the Class. 



The construction of all the instruments destined to measure 

 temperatures rests on the property possessed by bodies of 

 changing their volume by the action of heat; but these instru- 

 ments, in order to be rigorously exact, ought to satisfy two con- 

 ditions ecjually indispensable.' The first is, to be capable of 

 comparison with each other, i. e. to agree always in tbtir indi- 

 cations : the second is, to be comparable with themselves, i. e. 

 to have a course which shall be in a known ratio with the vari- 

 ations of temperature. 



The first of these conditions may be now obtained with great 

 exactitu<le. We know well all the causes which can influence 

 the justice of the indications of these instruments, and by em- 

 ploying the proper precautions we may attain a precision suffi- 

 cient for all observations. 



It is much more difficult to satisfy the second condition, which 

 consists in establishing a graduation in which the equal parts of 

 the scale answer to equal variations of temperature. In fact, in 

 order to fulfil it completely, it is not sufficient to determine the 

 relation of dilatations undergone by the therniometrical substance 

 which we employ, to the quantities of heat which we communi- 

 cate to it: we must also ascertain that the capacity of this sub- 

 stance for heat does not change, or at least keep an account of 

 this change if it takes place. The extreme difficulty of deter- 

 laiuing with precision how the specific heat of a body varies, 

 particularly at high temperatures, may be regarded as one of 

 the greatest obstacles to the direct solution of this (piestion. 



Experimentalists, however, have made some attempts to at- 

 tain it. DeLuc, who was the first, and is perhaps the only one 



♦ Annates dc Cliimie et <Ie Thytlqut, July 1816, tome li. p. CIO. 



A a 3 who 



