afcribtd to Heat. 3?3 



At the expiration of that time, I entered the room, — ufing the utmoft caution not to 

 difturb the balance, — wheVi, to my great furprife, I found that the bottle A very fenfibiy 

 preponderated. 



The water which this bottle contained was completely frozen into one folid body of 

 ice ; but the fpirit of wine, in the bottle B, (howed no figns of freezing. 



I now very cautioufly reftored the equilibrium, by adding fmall pieces of the very 

 jGne wire of which gold lace is made, to the arm of the balance to which the bottle B 

 was fufpended, when I found that the bottle A had augmented its weight by -jTioT P^ft 

 of its whole weight at the beginning of the experiment ; the weight of the bottle with 

 its contents having been 48ii;23 grains Troy, (the bottle weighing f 03, 37 grains, and 

 the water 4107,86 grains,) and it requiring now xVVt parts of a grain, added to the op- 

 pofite arms of the balance, to counterbalance it. 



Having had occafion juft at this time to write to my friend, Sir Charles Blagden, 

 upon another fubjeft, I added a poftfcript to my letter, giving him a fhort account of 

 this experiment, and telling him how " very contrary to my expeSation" the refult of it 

 had turned out ; but I foon after found that I had been too hafty in my communication. 

 Sir Charles, in his anfwer to my letter, expreffed doubts refpefting the faft ; but, 

 before his letter had reached me, I had learned from my .own experience, how very 

 dangerous it is, in philofophical inveftigations, to draw conclufions from fingle expe- 

 riments. 



Having removed the balance, with the two bottles attached to it, from the cold ioto 

 the warm room, (which ftill remained at the temperature of 6ii°), :the ice in the bQttle 

 A gradually thawed ; and, being at length totally reduced to water, and this water 

 having acquired the temperature of the furrounding air, the two W.tles, after being 

 wiped perfeftly clean and dry> were found to weigh as at the beginning of the experi- 

 ment, before the water was frozen. 



This experiment being repeated, gave nearly the fame refult, the water appearing, 

 when frozen, to be heavier than in its fluid ftate ; but, fome irregularity in the manner 

 in which the water loft the additional weight which it had appeared to acqpire upon being 

 frozen, when it was afterwards thawed, as alfo a fenfible difference in the quantities of 

 weight apparently acquired in the different experiments, led me to fufpeft, tliat the ex- 

 periment could not be depended on for deciding the faff in queftion ; I therefore fet 

 about to repeat it, with fome variations and improvements ; but, before I give an, ac- 

 count of my further inveftigations relative to this fubjeft, it may not be amifs to -jja^n- 

 »ion the method I purfued for difcovering whether the appearances mentioned in the 

 •foregoing experiments might not arife from the imperfeSions of my balance j .and it 

 «»ay like wife be proper to give an account, in this place, of an intermediate experiaa^nt 



3 D 2 which 



