Latent Heat of Evaporation of Benzene. 9 



the tap-water became so warm that when turned on by the 

 regulator it was unable to keep down the temperature suffi- 

 ciently. The extreme range in the temperature of the cell-tank 

 during these experiments was from 14°*8 to 16°*3 0. As, how- 

 ever, the movement was always extremely slow, it is probable 

 that the tank-temperature closely corresponded to the effective 

 temperature of the cells, and hence the correction 



e= 1-4342{1 + (15-0) x -00077} 



applied by us gave the value of e with sufficient accuracy. 



The time t E was determined by means of a chronograph 

 controlled by an electric clock, whose gaining rate is now less 

 than 1 in 20,000. Any movement of the keys by which the 

 current was switched on to the calorimeter-coil was auto- 

 matically recorded on the tape ; and thus personal errors were 

 eliminated. The times could have been read to -j--J- - of a 

 second, but it was considered unnecessary to read to nearer 

 than y 1 ^, i. e. about 1 in 7000 of t E . 



The value of R^ is expressed in terms of the " true ohm " as 

 given by the B.A. standards, with which (by kind permission 

 of Mr. Grlazebrook) the coils used by us have been directly 

 compared (Paper J, pp. 407-410). 



The increase in Rx due to the rise of temperature caused by 

 the current was determined in the manner described in Paper J, 

 pp. 404-407 (see also Paper W, p. 296). 



The value of J assumed by us was 4*199 (Paper W, p. 314). 

 We would emphasize the fact to which attention has been 

 previously directed (section 2, p. 3) , namely, that even if, in 

 consequence of errors in the standards &c, this value of J is 

 incorrect, it is still the right value to use for the reduction of 

 these observations, for, provided that no change has taken place 

 during the past three years in the standards used (and direct 

 comparisons show no signs of any change which would affect 

 the results), the values of L obtained by us are independent of 

 the numerical values of J and R when expressed in terms of 

 the same units as those assumed during the determination of 

 the mechanical equivalent. 



The value of Q E t % is thus expressed in terms of a " thermal 

 unit at 15° C." 



7. Qs^ 8 . 



The method adopted for finding Q 8 (the thermal units per 

 second due to the work done by stirring) has been fully 

 described in Paper W, pp. 290-293. It was there shown that 

 the value of Q g "varied approximately as r 4 (where r was the 



