Of WATER by HEAT. 401 



This experiment fpeaks in as decided language as the prece- 

 ding. It mows that when a portion, in the middle of a column 

 of water at temperature 39.5 is cooled, the colder fluid rifes, and 

 does not defcend through the warmer mafs, and prefents the 

 unequivocal demonftration, that water of temperature 39t° is 

 actually expanded by lofing heat. 



The different experiments which I have in detail recorded, 

 agree perfectly with each other in the evidence they give relative 

 to the fubject of inquiry. The general import of them is, that 

 water which is ice-cold, or a few degrees warmer, when heated, 

 becomes fpecifically heavier, — that water of 40 ° when heated 

 becomes fpecifically lighter, — that water above 40 % by the lofs 

 of heat, or by cold, is rendered fpecifically heavier ; and that 

 water below 40 ° is, by the fame caufe, rendered fpecifically 

 lighter. 



Such being the general import, the conclufion is irrefiftible, 

 that heat, in low temperatures, caufes water to contract, and at 

 fuperior temperatures to expand. The opinion, therefore, is 

 founded in truth, that water pofTeiTes a peculiarity of conftitu- 

 tion in relation to the effects of caloric, and that it is, within a 

 fhort range of temperature, an exception to the general law of 

 " expanfion by heat." 



So far as I can judge from thefe experiments, I am difpofed to 

 believe that the point at which the change in the conftitution of 

 this fluid in relation to heat takes place, lies between the 39^ y 

 and the 40th degree. 



1 am not at prefent aware of any objection to the method I 

 have followed in eftablifhing this fingular anomaly, and in re- 

 moving any doubts which may have arifen from the unavoid- 

 able influence which the inftrument muft have in the mode of 

 conducting the invefligation that had previoufly been adopted. 



The plan of operation above defcribed, however, only af- 

 certains the fact ; it gives no data for afcertaining the amount of 

 the anomalous effect of heat. 



3F 2 I 



