Changes in Dielectric Constants and in Volumes. 69 



would absorb enough saturated solution on its surface to affect 

 the temperature. Propyl alcohol saturated with acenaphthene, 

 which dissolves very slightly in it, as liquid was used and 

 acenaphthene as solid. The quantities were 3*1 grms. of satu- 

 rated solution and 5*6 grms. of acenaphthene. After 5 mins. 

 the mercury had risen 0"001° while the regular rise per minute 

 before running in the liquid was 0'0001°. This lay within the 

 uncertainty of reading and the conclusion was drawn that the 

 surface action was too slight to be noticed. 



The thermometer was read by a telescope and under favor- 

 able circumstances 0*001° could be estimated, but in general 

 this was not possible and so tlie thermometer readings are to 

 be considered uncertain to drO'001°. 



The correction for radiation was deduced on the assumption 

 that the change in temperature of a body is proportional to 

 the difference between its temperature and that of its surround- 

 ings. Let T be the total change in temperature of the calori- 

 meter from the time of running in the solvent to the time it 

 is saturated, t^ be the change in temperature at the end of the 

 hrst minute after running the solvent in, /^t^ be the average 

 change per minute of the calorimeter for 10 mins. before run- 

 ning in the solvent, and A^ be the average change per minute 

 for 10 mins. after the solvent is saturated. The change T is 

 always a di'op, always — . So for the first minute the correc- 

 tion is 



rT-t^ t^ -1 



and for the second minute 





and for the n^^ minute 



ni:M^ + {t^ + t^-\- .... 0(A^„~A^J 

 T 



n being the number of minutes needed for the solvent to get 

 saturated, beginning from the time the solvent was run into c. 

 This was sho\vn by the mercury beginning to rise after its fall 

 or by A?^., being less than A^^^ if A^^ was negative. Sometimes 

 the external temperature would change during an experiment 

 in such way that Az^^ ^^^^ ^Ix got opposite signs, and sometimes 

 when T was small A4 = A25j, and it became very difficult to fix 

 on a proper value for n. In general, n was less than ttO mins., 

 but once it rose to 102 mins. and once fell to S mins. 



