94 



DRS. J. A. HARKER AND 1'. CIIAITI IS ON A 



hension that, in spite of the care taken with the filling, it might possibly have 

 retained traces of water. 



After a heating of 26 hours above 400 we found in fact an initial pressure 

 considerably greater than that given above, viz. : 



inilliins. 

 October 4 525'359 



4 . . 525-361 



5 ........ 525-341 



7 . 525-348 



Mean .... 525 '352 



Determinations of the coefficient of expansion gave also a value appreciably greater 

 than the one found previously. 



From the mean of four experiments we found 



- 0-003 674 0. 



Our fears having been justified, the thermometer reservoir was again put into 

 communication with the pump and exhausted, being meanwhile heated to a tempe- 

 rature of about 500. After several successive exhaustions and partial fillings of gas 

 the reservoir was then pumped out very thoroughly, and after remaining vacuous for 

 24 hours was filled to the proper pressure with very well dried gas. It was main- 

 tained all the time at a temperature approaching 500. 



The following measurements of the initial pressure and coefficient of expansion were 

 then made : 







