COMPARISON OF PLATIMM AND GAS THERMOMETERS. 



91 



[For the dilatation of mercury, which enters into the calculations, the formula found 

 by one of us (' Proces-verhaux des Stances du Coinit6 International,' 1891, p. 37) 



V, = V [l + (182008* - 11 -380 4/* + 0-169 21/*) lO' 9 ] 

 was adopted.] 



Treated by the method of least squares, these olwervations give for the cubic 

 dilatation of porcelain the following formula: 



V, = V [l + O'OOO 007 593 OGt + O'OOO 000 013 750J 2 ]. 



The observations between and 100 which determine the value of the term in 

 t~ not being numerous, we can consider only the mean dilatation between the extreme 

 points and 100 as having been determined with sufficient ticcuracy. As, however, 

 it was of importance to know the second term more exactly, as its influence increases 

 at high temperatures, we made a second determination of the dilatation by means of 

 the Fizeau apparatus. The specimen which served for this determination was pre- 

 pared from a fragment of the capillary tube of a precisely similar reservoir made at 

 the same time at the ImjMjrial Porcelain Factory at Berlin. 



The results of this determination, which comprised 37 ol>ser vat ions lietween the 

 temperatures 2 and 82, are for the linear exj>ansion 



and for the cubical expansion 



a, = O'OOO 002 687 62, 

 /8, = 0-000 000 002 987 3 ; 



a, = O'OOO 008 062 8, 

 & = O'OOO 000 008 983. 



The two methods give practically the same result for the mean dilatation between 

 and 100; by the weight thermometer we have 



a, + 100ft = O'OOO 008 968 71, 

 N 2 



