concerning Heat. 157 



when with the quill I made a rapid friction against the 

 end of the cylinder immersed in the boiling water, I 

 occasioned all the thermometers to rise several degrees. 



It may perhaps be imagined, at first sight of the re- 

 sults of the experiment, that as the three thermometers, 

 which occupied the parts about the middle of the axis 

 of the cylinder, did not indicate a decrease perfectly 

 agreeing with the theory, the theory itself cannot be 

 true ; but a moment's reflection will show that this in- 

 ference would be too hasty, and that the difference be- 

 tween the theory and the results of our experiments, far 

 from proving anything adverse to the theory, serve on 

 the contrary to render it more probable. 



The results of such experiments can never agree with 

 the theory, except the divisions of our thermometers be 

 perfectly accurate; but it is well known to everyone 

 who has any knowledge of natural philosophy that the 

 divisions of our thermometers are defective. 



One of the objects I had in view in the experiments 

 of which I have just given an account to the Class, and 

 in several others which I intend to make without delay, 

 is to improve the division of the scale of the thermom- 

 eter, in order to render this valuable instrument of 

 greater utility in the delicate investigations of natural 

 philosophy. 



It appears certain that the increase of the elasticity 

 of air by heat is much more nearly proportionate to the 

 increase of temperature than the dilatation of mercury 

 or any known fluid ; consequently, it is the air thermom- 

 eter we ought to endeavour to improve, and which 

 must ultimately afford us the most accurate measure of 

 heat that it is possible for us to procure. 



