High Temperatures. 177 



2. Influence of Pressure on the Gas Thermometer Bulb. 



At the highest temperatures measured the inside pressure in 

 the platin-iridium bulb exceeded that outside by nearly an 

 atmosphere. Although no permanent change in the volume 

 of the bulb after heating could be detected, it seemed, never- 

 theless, not impossible that a temporary yielding of the 

 glowing walls under the comparatively high pressure might take 

 place. We have investigated this question by redetermining 

 the higher temperatures with a much diminished pressure. 



This was accomplished by allowing a portion of the expand- 

 ing gas to pass over into the closed tube of the manometer, 

 thus varying pressure and volume simultaneously. The expan- 

 sion-coefficient of the gas suffers a small change thereby which 

 we estimated to be about 0*1 per cent and therefore neglected it. 



The closed manometer tube had already been provided with 

 two points (see first paper — loc. cit.) at which readings could 

 be made for volumetric purposes, so that by bringing the mer- 

 cury tangent to the lower point a definite volume could be 

 added to that of the bulb. To enable a more accurate meas- 

 urement of the temperature of this supplementary volume to 

 be obtained than was possible with the small thermometer in 

 the cap, a second and more sensitive thermometer was fastened 

 to the tube and the whole carefully packed in a thick layer of 

 cotton batting containing only the two small openings neces- 

 sary to make the readings above and below. 



As a rule the observations were made in sets of three, the 

 first and last with the mercury adjusted to the lower, the 

 second to the upper point. Since it was necessary to wait 

 some time between observations until the temperature condi- 

 tions in the manometer became uniform, the temperature in the 

 oven had opportunity to change a few degrees. These changes 

 (8) were measured with the thermo-elements. The supplemen- 

 tary volume between the two points was measured by weigh- 

 ing the mercury required to fill it, the results being as follows : 



50-630 cm3 

 50-609 

 50641 

 50-658 



Mean 50-635 cm3 



The differences between the determinations are largely due 

 to an imperfectly formed meniscus at the lower point, the 

 adjustment being necessarily made with a falling column as 

 the mercury was drawn off through a fine capillary tube. The 

 adjustment with a rising column as used for the temperature 

 •measurements gives much more exact results. Also the deter- 



