On the Contraction of Water hj Heat. 1 01 



riment equally shows, that the warmer fluid ascends an^ 

 occupies the summit, by its route announcing it* dimi- 

 nished density, and proving that water is now expand^ 

 ed by heat. 



Experiment II. 



I filled the same jar with water of temperatyre 53°; an<i 

 that I might observe the phcenomena of cooling, I placed 

 it in the axis of a much larger cylindrical vessel, nearly full 

 of water, of temperature 41°, and, by an earthen-ware 

 support, raised it about three inches from the bottom, 

 taking care that the water ;jhould be on the same level in 

 both vessels. As soon as I had adjusted the two thermoc 

 meters, as in the former experiment, I observed that the 

 top of the fluid was still at 53", but the bottom had fallen 

 to 49°. 



Top. Bottom. 



In 9 minutes 52° - 45 



— 15 52 - 44 



Now, to accelerate the cooling, I withdrew by a syphon 

 the water from the large cylinder, and supplied its place by 

 ice-cold water, mixed with fragments of ice, which by 

 repeated cautious agitation was kept uniformly at the tem- 

 perature of 32°. 



In 23 minutes 48° - 42 + 



— 38 44 - 40 



— 43 42 - 40 



— 46 40 - 40 



— 52 36 - 40 



— 58 35— - 39 



— Q^ 31 _ 37 



— 75 34 - 36 



— 103 34 - 34 



This experiment is the counterpart of the foregoing, and 

 from (he testimony of the same instruments, it appears, 

 that when a cylinder of water of 53° is cooled by circum- 

 fluent iced fluid, the colder part of the water takes posses- 

 sion of the bottom of the vessel, so as to establish a dif- 

 ference of temperature from the surface, amounting some- 

 times to 8". And that as soon as the fluid at the bottom 

 arrives at ihe 4()th degree, the tc;Tiperature of the fluid in 

 that situation is stationary till the surface reaches the same 

 point. 



During the subsequent refrigeration, the progress of the 

 co(jling unclLrgoes a total change. The thermometers tell 

 that the colder fluid rises to the surface ; so that the top 



Vol. 23. No. 90. Nov. 1805. L gets 



