14-6 On the Contract io?i irh'icJi lakes plfce in Mercury 



this instruynent; and. when again reduced to the temperature 

 o]F50 degrees, another arch was drawn : — the contraction of 

 th^eiar of silver hetween the two points appeared to equal 

 •Birth part of its dimensions. T ,1' 



I he silver wag then heated to 2CjQ*' of temperature of^ 

 Fahrenheit's. ^sc^Tile, by boiling it in \vater, and gradually 

 coplmg; an, arch of a circle was then struck on the surface 

 of _ the silver :, reducing it to 150^ of temperature, another 

 arch was struck with the same unvaried centre pins, which 

 Showed evident contraction in the bar of silver: agaip, at 

 100° of temperature another arch was described ; and a 

 fourth, at 50^ showed that the silver had contracted in the 

 deprivation of 15(^^ of temperature -g-i^jth part of its length. 

 _ Tlien, similar parallelopipedbns being to each other as the 

 ciibes of their homologous sides, the increased specific gra- 

 vity, is shown "by multiplying the specific gravity at 50% 

 namely, 10'362, by the cube of v^41; and dividing it by, the 

 cube of 340^ which equals 10-4537; '0917, therefore, is 

 the difference \p.^ 150^ of heatj or '90061 14 is the contrac- 

 tion of silver in each degree; which, from the way this ex- 

 periment is perfqrnied, I call its visible contraction. From 

 it .the specific, gravities at the several' temperatures below 

 are calculated ; — /,„ 



At 200 _ silver is oi the speciijc gravity ] 0*2702; , , 



V^P/o 1, ■ 10-3908 



135 10-3100 



117 . 10-3210 



100 \ 10-3314 



j^"5p.^^ / , 10-3620 



,' "of' 10-3925 



50 below zero 10-4231 



52 — . 10-4243 



56 10-4268 



I. Then, if 10-4268 be the specific gravity of silver at 

 56^ below zero, and weighed in alcohol at that temperature, 

 by. calculation from its loss of weight at C, it will appear 

 tgi be 9*796; by taking '8141 as the supposed density of al- 

 •cphol,, dividing the quantity 1000 by the loss 83-1, multi- 

 plying]; the quotient l£-033 by '81 41 ^ and dividing by 1-000, 



the 



