jga On the CONTRACTION 



Bottom, Top. 



Ini hour 44 min, 39.5 



{A crufl of ice of feme thick- 

 nefs now lined the glafs, 

 and air had fallen to 40". 

 — 5i hours, 39 



~ "at^SightJ 39 Cruft of ice complete. 



— 19 hours, ?. £' \ 

 next morningjj" 



39 Air 40' 



— 26 hours, 40. 



Air 40°. So much ice had 

 I melted that the cake was de- 

 1 tached from the fide of the vef- 



— 32 40 (^fei^ and floated. 



— 41 — — 40 Air 41 ". Ice not all melted. 



— 50 41 Air 42**. Ice not entirely gone. 



This long protraded experiment prefents fome flriking fads, 

 and its general import, with regard to the fubjedt of inveftiga- 

 tion, agrees with the preceding. In it we fee, that when the 

 frigorific mixture abftradled caloric from the upper extremity of 

 a cylinder of water, nearly 18 inches long, and at 50°, the re- 

 dudion of temperature appeared fooner, and advanced quicker, 

 at its lower extremity than in the axis at the top, not two and a 

 half inches diftant from the cooling power. No one can entertain 

 a doubt that this is owing to a current of cooled and condenfed 

 fluid defcending, and a correfponding one of a warmer tempera- 

 ture afcending. Now, if water obferved the fame law that other 

 bodies do, and had no peculiarity of conflitution, the fame pro- 

 grefs of cooling fhould continue. This, however, the experi- 

 ment teaches us, is not the cafe : as foon as the fluid at the bot- 

 tom exhibits a temperature of 40°, it ceafes. The colder fluid 

 remains at top, and quickly lofing temperature, ere long begins 

 to freeze- The continuance of the colder fluid at the furface 

 furely denotes, that it is not more denfe than the fubjacent 



warmer 



