44S Iiiquines into llie Laivs of Dilatation of Solids, ^c* 



but we made the experiment at the boiling point of water, iit 

 order to assure ourselves of the degree of exactness which we 

 might expect from our process. By employing in the calculation 

 the dilatation of iron given by Messrs. Laplace and Lavoisier, 

 we found the dilatation of the mercury -5-1^^-5- for one degree. 

 We must add an unit to the third cypher significative of the 

 dilatation of the iron which we have employed, in order to make 

 our result coincide exactly with that of the same experimentalists 

 on the expansion of mercury. 



Now by exposing the vessel at the temperature of 300^ we 

 find that the quantity of mercury which issues from the iron 

 thermometer is far superior to that which ought to issue from it, 

 if the iron and the glass preserved at high temperatures the 

 ratio of expansibility which has been assigned to them below 

 100". — This experiment has been repeated several times, with 

 results very little different ; so that we may conclude that the 

 dilatation of the glass does not remain constant at all tempera- 

 tures, and that it increases more rapidly than that of iron. 



It is extremely probable that the dilatation on the increment 

 of elastic force of the gases remains constantly proportional to the 

 temperatures. (This is what we shall prove in a subsequent pa- 

 per.) By admitting this principle, we see that the indications 

 of the mercurial thermometer will be always superior to the real 

 temperatures, and the more so in proportion as we rise in the 

 scale; but experimentalists will doubtless see with astonishment, 

 the slowness with which these differences increase. This does 

 not seem to arise from the dilatations of the mercury remaining 

 nearly proportional to the real increments of temperature, but 

 from the law of dilatation of the glass being combined with that 

 of the mercury, and an almost exact compensation being the re- 

 sult. Besides, we shall be able to verify this completely. 



It results also from our researches, that the metallic pyro- 

 meters, to which we ascribe a regular march, indicate tempera- 

 tures much too high, when we suppose, as has always been done 

 hitherto, that the dilatation of the metals remains constantly pro- 

 portional to the temperatures. 



Note at the conclusion of the aloue Paper by the Editors of the 

 Annates de Chimie et de Physique. 

 This memoir is printed precisely as it was presented to the 

 Institute in 1S15. We then expected to have been able to 

 publish the continuation speedily afterwards. — Particular circum- 

 stances having interrupted our labours, and not perniiting us to 

 fix the period at which it will be terminated, we thought it best 

 to give to the public the results which have been already ob- 

 tained. 



XCII. New 



