4 02 On the CONTRACTION 



I have already flated, that M. de Luc alleged, that from the 

 temperature of 41 °, the expanfion occasioned by cold was very 

 nearly equal to that produced by the fame number of degrees of 

 heat ; and confequently that water poffefTes the fame denfity at 

 any given number of degrees of temperature above and below 

 41 °. The firft experiments of Mr Dalton appeared to confirm 

 this opinion, and to enlarge the range to which it applied, by 

 extending it to temperatures as far below 32 °, as water allows 

 itfelf to be cooled before it begins to freeze. From one circum- 

 ilance that conftantly occurred, I am inclined to think, that the 

 amount of the dilatation by cold is inferior to that caufed by heat. 



During the heating or cooling of water below 40 , the dif- 

 ference of temperature between the top and bottom of the fluid 

 was lefs than what occurred during the cooling or heating of the 

 fluid through the fame number of degrees above it ; and I con- 

 ceive that, when other circumftances, but particularly the rate 

 of the change, are alike, the difference of temperature between 

 the upper and lower parts of the fluid, as it depends upon, may 

 prove a meafure of, the difference of denfity. 



Alcohol, when heated or cooled, prefents, by reafon of its 

 greater expanfibility, a greater difference of temperature in thefe 

 fituations than water ; and upon the fame principle I infer, that 

 water from 40 ° is more expanded by an equal number of de- 

 grees of elevation than of deprefhon. 



As the concurrence of the teftimony of the experiments above 

 related with the general opinion, will probably remove every 

 doubt refpecling the matter of fact, it remains a very difficult 

 problem for thofe who are fond of philofophical inveftigation, 

 to explain how heat fliall occafion in the fame fluid, without pro- 

 ducing any alteration of mechanical form or chemical condition, 

 at one time contraction and at another expanfion, and to recon- 

 cile the contractive effect to the conceived notions of the mecha- 

 nifm of the operations of this energetic agent. 



When 



