22 Mr. E. H. Griffiths and Miss Marshall on the 



ment was required and, as the temperature of the calorimeter- 

 wire remained constant, the electrical balance required little 

 attention, for the potential- difference rarely altered by as much 



as -. nr> nnn throughout an experiment. As the benzene 



vapour passed into condenser A the pressure in the flask 

 increased, and thus the loss of heat by evaporation diminished. 

 By a rapid movement of the tap between condensers A and B, 

 the pressure could again be diminished and the cooling effect 

 increased. 



It was found impossible to control the rate of evaporation 

 with the same perfection when using benzene as was the 

 case in the experiments with water. The taps had to be 

 constantly manipulated, and a moment's inattention on the 

 part of Observer I. was immediately followed by a sudden 

 rise or fall in X . This was more especially the case during 

 the experiments at low temperatures. From about experi- 

 ment V. or VJ . onwards, however, the swings rarely amounted 

 to as much as 200 or so, except during the first minute when 

 the thermal balance was being obtained, at which time a 

 swing of +400 or 500 was generally experienced. Through- 

 out the whole of an experiment care was taken that any 

 positive swing should be succeeded by a corresponding 

 negative one. Although the announcement of swings of 

 " + 200 }> appeared alarming at the time, the extreme attention 

 devoted to the keeping down of these oscillations was really 

 unnecessary. A swing of + 200 indicated that O was lower 

 than 6 1 (at 40°) by about 0°'02 C, and the radiation &c. 

 coefficient of the calorimeter being about *00009 (in degrees 

 per second per difference of 1°, Paper W, p. 289), this swing, 

 even if maintained throughout the whole time that evaporation 

 was proceeding (on an average less than about 7 minutes), 

 would only have resulted in a total loss of about '00009 x 318 

 x '0216 x 7 x 60 = *25 thermal gram *. In no case, however, 

 was a swing of such magnitude allowed to remain unaltered 

 for more than a few seconds. 



The chief difficulty was experienced near the close of an 

 experiment. When working with water, there were always 

 some indications that the end was near, for the pressure had to 

 be diminished in order to maintain the thermal balance if only 

 a drop or so remained. In the case of benzene, however, 

 there were rarely any such indications ; for the galvanometer- 



* About ^y\ro of the total " heat-supply.'' 



