398 On the CONTRACTION 



ing from the bottom, and thefe cold currents cannot move up- 

 wards, were not the water of them fpeclfically Hghter than that 

 of the incumbent warmer fluid. 



The water, therefore, which at the bottom is cooled by the 

 contiguous frigcrific mixture, muft be expanded by the lofs of ca- 

 loric. 



This experiment fecures full force to the laft, as it obviates 

 the objedlion already noticed, and alfo precludes another. I 

 have already ftated, that it may perhaps be alleged, that the fluid 

 at the top, in experiment third, though cooled to 32° did not 

 defcend, becaufe below 40° the contradllon is fo trifling, that it 

 does not occafion a difference of fpecific gravity fuflaciently great 

 to caufe the particles to defcend, when oppofed by the inertia and 

 tenacity of the fluid through which they have to fall ; or it may 

 be conceived, that the defcent is fo tardy, that time is given to 

 the ambient air or fubjacent fluid to furnifli heat enough to raife 

 the temperature of the defcend ing flream, and by that arreft it 

 in its downward courfe. 



But from the particulars above recorded, it is manifeft, that 

 the change of denfity between the temperature of 32° and 40° 

 is quite fufficient to put into motion the particles, and to enable 

 them to overcome the obftacle arifing from inertia and tenacity, 

 and to withftand the arrefl:ing effecfls of atmofpheric heat. 



Though thefe experiments, and fome others of a fimilar na* 

 ture, carried convidlion to my mind, and perfed:ly fatisfied me 

 refpefling the i-eality of the anomaly of water, I determined to 

 vary fomewhat the mode of making the experiment, fo as to ob- 

 tain ftill more fl;riking refults. 



For the fifth experiment, 1 ufed an apparatus which confifted 

 of a ftill taller jar. It was 2 1 inches high, and 4 in diameter. 

 I adjufted at the middle of its height a perforated bafon of tin- 

 ned iron, 2 inches, in depth, and 1 o in diameter. As this bafon 



embraced 



