ATMOSPHERIC PRESSURE BETWEEN 104 C. AND 115 C. 419 



There are two series of experiments in this group : one series (a) in which the 

 vacuum-jackets were exhausted before their insertion in the double-walled brass 

 jacket, and a later series (b) with the silica calorimeter, in which case the vacuum 

 was formed when the calorimeter was in situ. In this case the space between the 

 double walls of the calorimeter was connected with a Geissler tube, a charcoal tube, 

 and a pump. Three sets of experiments were made with different pressures in this 

 jacket. 



Calculation of the Results. 



The values of CE/Q dd, obtained with three different flows, but with the same rise 

 of temperature dd, were calculated from the reduced tables of observations, and then, 

 by means of three equations of the form 



the values of S m , h, and k were obtained. S. m is the mean specific heat over the 

 range 104 '5 C. to (I04'5 + c0) C. Similar calculations were then made from the 

 results obtained under the same experimental conditions but with other values of dO, 

 and it was found that, although the values of h and k were constant within the limits 

 of experimental error, the values obtained for S m decreased witli an increase in the 

 value of d9. This relation is represented by the equation 



S c = S^ + O'OOlOdfl. 



S e = specific heat at the cold temperature, i.e., 104 '5 C. 



This means that, for any given flow with a given calorimetric arrangement, each 

 quantity CE/Q d6 depends on the magnitude of d9, hence by applying this tempe- 

 rature correction to the observed values of CE/Q d9 the latter are brought to the 

 values they would have had if the rise of temperature had been zero. Then, by 

 taking the mean of all the maximum, all the medium, and all the minimum values of 

 CE/Qdf0, thus corrected, and substituting these in the three above equations, the 

 most probable values of S c , h, and k can be deduced. The values of S c thus obtained 

 show a distinct variation with the barometric height. This variation of the specific 

 heat at constant pressure with pressure can be calculated by means of equation (7), 

 p. 387. In the case of steam, at a temperature of 104 '5 C., if S is expressed in 

 joules/gr. deg., and 1 cm. of mercury is taken as the unit of pressure dS/dp = O'OOIS, 

 this means an increase in the value of S of 0'065 per cent, for an increase of pressure 

 of 1 cm. of mercury. Experimentally, the value of dS/dp is found to be about twice 

 as great as this, but, taking into consideration both the smallness of the correction 

 and the fact that the barometer only varied by about 2 cm. during any one series of 

 experiments under identical conditions, the agreement is very satisfactory. A pressure 



