44 Mr. P. J. Hartog and Dr. J. A. Harker on 



Regnault and Pfaundler assume that the correction for any- 

 intermediate temperature t xi corresponding to the abscissa 

 OE, is represented by the ordinate EF drawn to meet the 

 straight line BD in F. 



For this assumption to be valid two things are 

 necessary : — 



(i) That the change of temperature of the calorimeter 

 should not exceed 3 or 4 ; this condition is 

 satisfied both in M. Berthelot's experiments and 

 in those performed by us ; and 



(2) That the temperature of the bodies radiating heat 

 to the calorimeter should remain approximately 

 constant. 



It is evident that we cannot possibly assume the correction 

 to be merely a function of the temperature of the calori- 

 meter, if we suddenly bring a hot body into its vicinity 

 during the course of our observations. In M. Berthelot's 

 apparatus, we do the reverse ; we suddenly take away the 

 hot burner which was radiating heat to the calorimeter 

 during the actual observation. It is difficult to see how the 

 observation of the march of the thermometer after this 

 removal can serve as a datum for the calculation of the 

 correction to be applied for the period of actual con- 

 densation, during which the thermal conditions were so 

 different. 



We proceed to describe the modified form of apparatus 

 [Fig. 2] which we have adopted in order to be able to use 

 either of the methods of correction just quoted. 



The boiler consists of a flask A, through which the tube 

 BC passes centrally. The upper end of BC is ground coni- 

 cally to fit into a hollow cap D, which is itself attached by a 

 glass rod to the movable bell E. This bell fits loosely into a 

 rim, which is filled with mercury so as to form a lute. The 

 bell and cup may thus be raised or lowered at will, so as to 

 open or close the valve at C, through which the steam passes 



