92 DRS. J. A. MARKER AND P. CHAPPUIS ON A 



and by the Fizean method we have 



a, + 100& = 0-000 008 961 2. 



Admitting the coefficient &, deduced from the observations by the Fizeau method, 

 we have calculated the coefficient 2 from the relation 



which gives 



2 + 100& = O'OOO 008 968 71, 



a,, = 0-000 008 070 35. 

 We have thus adopted as our final formula for the cubic expansion of Berlin 



V, = V (1 + O'OOO 008 070 35* + O'OOO 000 008 983*). 



porcelai 



in 



(c) Pressure Coefficient of the Porcelain Reservoir, 



The measurement of the pressure coefficient of the porcelain reservoir was made in 

 precisely the same way as that of the glass reservoir previously described on p. 75. 

 We determined by three series of observations the variation of volume Av corre- 

 sponding to a variation of pressure of 1 millim., obtaining the following results : 



(1) 



(3) 



microlitres. 

 Av = 0-003 803 5 



Av = 0-003 701 7 

 Aw = 0-003 746 6 



We have adopted the mean of these three determinations, viz. : 



Ai> = 0*003 750 microlitre per millim. 



(d) Determination of the " Dead Space." 



For this determination we followed exactly the method already described on 

 p. 75. The nine weighings made gave divergences from the mean of four parts 

 per thousand. After all reductions we found for the whole volume of the " dead 

 space " 



v = 709 "5 microlitres. 



The effective capacity of the thermometric reservoir being 



V = 164*805 cub. centims., 

 we have 



v/V = 0*004 305. 



This result, which is appreciably higher than the corresponding one for the 



