39© On the CONTRACTION 



This experiment is the counterpart of the foregoing, and from 

 the teftimony of the fame inflruments, it appears, that when a 

 cylinder of water of 5$° is cooled by circumfluent iced fluid, 

 the colder part of the water takes poffeflion of the bottom of the' 

 vefTel, fo as to eftablifh a difference of temperature from the fur- 

 face, amounting fometimes to 8°. And that as foon as the fluid 

 at the bottom arrives at the 40th degree, the temperature of the 

 fluid in that fituation is ftationary till the furface reaches the 

 fame point. 



During the fubfequent refrigeration, the progrefs of the 

 cooling undergoes a total change. The thermometers tell that 

 the colder fluid rifes to the furface ; fo that the top gets the ftart 

 of the bottom foon by 4% and attains the loweft temperature of 

 34 very long before the other falls to the fame degree. 



These circumftances, I think, lead to the conclufion, that by 

 the lofs of caloric, water at 53° is contracted and rendered fpe- 

 cifically heavier, and that this continues to happen till the water 

 come to the temperature of 40 °, at which period an oppofite ef- 

 fect is produced ; for now the water, as it cools, becomes fpecifi- 

 cally lighter, or is expanded. 



In this, as well as the former experiment, the complete change 

 in the fituation, which the warmer and colder parts of the fluid 

 affected, in the progrefs both of the heating and cooling, while 

 every external circumftance of the procefs continued unaltered, 

 is particularly worthy of remark. 



Experiment III. 



I took a glafs jar, 17.8 inches deep, and 4.5 in diameter, in- 

 ternal meafure, having a neck and tubulature very near the bot- 

 tom. I provided alfo a cylindrical bafon of tinned iron, 4.8 



inches 



