386 On the CONTRACTION 



Unwilling to remain in uncertainty, and confidering it as a 

 point of much curiofity and intereft, I have endeavoured to in- 

 vefligate the fubject by experiments conducted in a totally dif- 

 ferent manner, equally calculated to exhibit the fingular truth, 

 but free from the objections to which the others are liable. In 

 them, it was my object: to provide, that neither the changes of 

 the actual volume of the water, nor the alterations in the dimen- 

 fions of the inft.rument, fhould have any influence whatever. 



I have already taken occalion to date, that the purpofe of this 

 paper is to prove, by experiments on the principle now mention- 

 ed, that in the conftitution of water there really exifts the angu- 

 larity often noticed. 



I shall nrft flate the plan of the experiments, and then detail 

 the particulars of the mod remarkable of them. 



When any body is dilated, whether by heat or cold, it necef- 

 farily becomes lefs denfe, or fpecifically lighter j and the oppofite 

 effects refult from contraction. This is the circumflance, as eve- 

 ry one knows, which caufes various movements among the parti- 

 cles of fluids, when any inequality of temperature prevails in 

 the mafs ; hence thefe particles are little acquainted with a flate 

 of reft. 



If a partial application or fubtraction of heat produce an in- 

 equality of denfity in a mafs of fluid, the lighter parts rife to the 

 furface, or the denfer fall to the bottom. 



It readily occurred, that 1 might avail myfelf of thefe move- 

 ments, and upon ftatical principles determine the queflion in dif- 

 pute. 



I had only to examine attentively water, as it was heated or 

 cooled in ajar, and to obferve, by means of thermometers, what 

 fituation the warmer, and what the cooler parts of this fluid af- 

 feaed. 



If I fhould find that ice-cold water, in acquiring temperature, 

 fhowed, in its whole progrefs, the warmer parts near the top, it 



would 



