of Gases at Constant Volume. 



391 



initiations of specific heat have been carried to densities at which the 

 substance was partly in the liquid state at the lower limit of tem- 

 [perature of the experiments. Observations dealing with true specific 

 heat, uncomplicated by the presence of thermal effects due to the 

 presence of liquid, are limited by the density 0'1444. At this density 

 |the mean specific heat over the range, 12 C. to 100 C., is 0'2035. 



The following table contains a summary of the mean results of 

 the experiments in which no liquid is present at the initial tempera- 

 ture. The range of these experiments lies between air temperature 

 (12 to 16) and steam temperature 



These observations, combined with those contained in Part I (loc. 

 cif.), afford a well defined line, which rises slowly at the higher 

 densities, turning away from the axis of density. 



According to an empirical equation to this line, the specific heat 

 of carbon dioxide at constant volume is given in terms of its variation 

 with density p, as follows : 



C v = 0-1650 +0'2125 / > + 0-340(V 2 . 



This is in fair agreement with the linear equation deduced in 

 Part I from the limited number of experiments at low densities 

 therein contained : 



C v = 0-16577 + 0-2064 P . 



In these experiments a spherical vessel of copper was used to hold 

 the gas, having a voluminal capacity of about 86 c.c., a mass of 

 137 grams, and an estimated resistance to bursting of 300 atmos. 

 This, as in the former experiments, was equilibrated against a similar 

 vessel in a differential steam calorimeter. These vessels produced 

 each a precipitation, due to its own calorific capacity of 2'1 grams of 

 steam. It was found, however, that closely agreeing results (to 1 per 

 cent, about) were obtained when the precipitation due to the gas fell 

 to as little as 0'15 gram. 



