Mecifurement of a Bafe on Hounflow-ITeatli. 47- jt 



roJ to contract to the true freezing temperature, as- Was eafily 

 known by the wires becoming perfedly fixed and flationary 

 with regard to the marks ; the ice was then removed from the 

 copper trough ; and tlie fame being filled with water nearly on 

 the boil, the ebullition was completed, and kept up, by means 

 of the lamps now lighted for the purpole, and flipped in 

 underneath. 



The expanfion, anAvering to the 180^ between freezing and 

 boiling, was now meafured by working with the micrometer 

 fcrew until the bifedion * of its wire with thofe of the mark 

 was again complete ; the obferver at the fixed microfcope taking 

 alfo efpecial care all the while to keep his bifedion perfectly 

 accurate. The number of revolutions, regiflered by the num- 

 ber of entire divifions that the zero of the circular fcale had 

 departed from its dart or index, and alfo the value of any frac- 

 tional revolution, regiftered by the divifions on the head in- 

 tercepted between zero and its proper dart, were then noted, as 

 exprefTed in the firft column of the fubjolned table of experi- 

 ments ; which requires no other explanation than what is 

 therein inferted, and which has been extended purpofely to 

 fhew at one view, from infpedion only, how much the length 

 of our bafe would have been afIe(Sted, if meafured by thefe 

 metals refpe£tively, in temperatures between 32' and 61°, 



All the experiments were repeated at leaft twice, and fome 

 ©f them three times, except the ftandard fcale and glafs pen- 



* This bifei^ion of the wires may always be made to a great degree of preci- 

 fion, by one with a tolerably good eye, and accullomed to obfervations cf this fort.- 

 I have myfelf repeatedly adjufted the wires eight or ten times running, allowing 

 another perfon to read off and unadjuft each time, without the mean difference 

 exceeding one-fourth of a divifion of the head, which is only ^4^o^th part of an- 

 inch. 



duluiia? 



