2^6 Expanjion of Water at Ttmpefalures helotu 4a Degyees. 



inches with mercury from 32° to ai2', and filled U with 

 pure water. I then graduated it by an accurate mercurial 

 thermometer, putting them together into a bafon filled with 

 water of various degrees of heat, and ftining it occafionally. 

 As it is well known that water does not expand in propor- 

 tion to its heat, it does not therefore afford a thermometric 

 fcale of equal parts, like quickfilver. 



From repeated trials agreeing in the refult, I find that the 

 water thermometer is at the loweft point of the fcale it is 

 capable of, that is, water is of the greateft denfity, at 42^" of 



uhe mercurial thermometer. From 41° to 44° inclufively 

 the variation is fo finall as to be juft perceptible on the fcalej 

 but above or below thofe degrees the expanfion has an in- 

 creafing ratio, and at 32' it amounts to T-8th of an inch, or 

 about x^yth part of the whole expanfion from 43^° to 212^, or 



I boiling heat. During the iiiveftigation of this fubjeft, my at- 



. tention was arrelled by the circumltance, that the expanfion of 

 water was the fame for any number of degrees from the point 

 of greateft condenfation, no matter whether above or below 

 it : thus, I found that 32°, which are 10^ ° below the point of 

 greateft denfity, agreed exaftly with 53", which are lof-* 

 above the faid point ; and fo did all the intermediate degrees on 

 both fides. Confcqucntly, when the water thermometer flood 



• at 53°, it was impoflible to fay, without a knowledge of other 

 circumftances, whether its temperature was really 53'^ or 32°. 

 Recollecting fome experiments of Dr. Blagden in the Philo- 

 fophical Tranfadtions, from which it appears that water was 

 cooled down to 2l° or 22° without freezing, I was curious to 

 fee how far this law of expanfion would continue below the 

 freezing point, previoufly to the congelation of the water, 

 and therefore ventured to put the water thermometer in- 

 to a mixture of fnow and fait, about 25° below the freez- 

 ing point, expecting the bulb to be burft when the fuddcn 

 congelation took place. After taking it out of a mixture of 

 fnow and water, where it ftood at 32^ (that is 53° per fcale), 

 [ immerfcd it into the cold mixture ; when it rofe, at firft 

 (lowly, but incrcafing in velocity, it paflTcd 60°, 70°, and was 

 going up towards So*', when I took it out to fee if there was 

 uuy ice in the bulb — but it remained perfe6lly tranfparent. 

 f immerffd it again and raifed it to 75° per fcale, when in 

 an inOant it darted up to 128^; and that moment taking it 

 out, the bulb appeared whife and opaque, the water within 

 being frozen. Fortunately it was not burft; and the liquid 

 which was raifed thus to the top of the fcale was not thrown 

 out, though the tube was unfealed. IJpou ajiplying the 

 hand, theice was melted, and the liquid refumcd its ftation. 

 T This 



