1895.] The Latent Heat of Evaporation of Water. 213- 



singularly limpid oil, consisting of hydrocarbons only. This oil was- 

 stirred by paddles revolving about 320 times per minute. Within 

 the oil was a platinum-silver coil whose ends were kept at a known 

 potential difference when an experiment was proceeding. The supply 

 of heat was balanced by the loss due to evaporation in the flask, and 

 thus the temperature of the calorimeter (0j) was kept practically 

 constant and equal to 9 Q throughout an experiment. True d l oscillated 

 slightly about , but &i rarely attained to 0*05 C. in experiments- 

 lasting 70 or 80 minutes, and the oscillations could be so controlled 

 that #100 was alternately positive and negative. The advantages 

 of the above method are that the results are not appreciably affected 

 by- 



(1.) Errors in thermometry. 



(2.) Changes in the specific heat of water. 



(3.) The capacity for heat of the calorimeter. 



(4.) Loss or gain by convection, &c. 



Had the calorimeter been filled with water, its capacity for heat 

 would have been so great that a considerable loss or gain of heat 

 might have caused but small change in O l9 hence the employment of 

 the oil, above referred to, whose specific heat and specific gravity 

 were both small, and thus its " volume heat " was only about four- 

 tenths that of water. It was necessary to detect, and in some cases- 

 to measure, very small differences between and lt and this was 

 done by the use of differential platinum thermometers. A full 

 description of these thermometers has recently been published. * 

 Extremely minute differences of temperature could not only be 

 detected, but also be measured by means of the arrangement adopted. 

 In Paper L I describe the experimental proofs of the following state- 

 ment : " It follows that the differences of temperature could be 

 determined to 0'0004 C., and I am confident that differences of 

 0-0001 C. could be detected." 



Had it been possible to secure (a) that should remain absolutely 

 unchanged ; (6) that the value of 0\. should be zero at the com- 

 mencement and also at the end of an experiment, it would have been 

 unnecessary to measure. differences of temperature, since the variations 

 during an experiment would, if the above conditions were observed r 

 have no effect on the result. Let 0i and #/' be the initial and final 

 temperatures of the calorimeter, and C^ the capacity for heat of the- 

 calorimeter and contents at the temperature lt then C^ (#/ #1") 

 gives the loss of heat of the calorimeter, and as I have shown that 

 the loss or gain by radiation, &c., may be neglected, it follows that 

 this heat was abstracted by the evaporation of some water. Also, if 

 varied, an additional correction had to be made. I have indicated 



* " The Specific Heat of Aniline." ' Phil. Mag.,' January 1895. 



