326 Prof. Spencer U. Pickering on Experimerital 



platinum vessel is altered, and consequently this vessel may 

 not be in such intimate contact with the disk at one time as 

 at another, and the amount of heat absorbed by the latter 

 will therefore not always be the same. The error thus intro- 

 duced, however, cannot be very large, since the disk was never 

 found to be equivalent to more than one hundredth of the 

 contents of the calorimeter. 



It was thought possible that an error might also arise in 

 the following manner. Before the introduction of any salt 

 into the calorimeter, all the parts of the apparatus are at the 

 same temperature; but as soon as a reaction takes place, in- 

 volving, say, the evolution of heat, the radiation from the 

 calorimeter will raise the temperature of the vessel which is 

 interposed between it and the water-jacket (owing to its great 

 bulk the water-jacket itself remains practically unaltered in 

 temperature). During the rest of the time allowed for the 

 experiment this vessel will therefore be at a higher tempera- 

 ture than it was initially; and consequently the rate of cooling 

 during the last intervals will be less than during the first one, 

 the correction to be applied for cooling will be under-estimated, 

 the intermediate vessel having in reality acted as so much 

 calorimetric water. In order to ascertain whether the error 

 from this source attained appreciable dimensions, four experi- 

 ments with potassium sulphate were performed ; in two of 

 these the intermediate vessel w r as present as usual, while in 

 the other two it was dispensed with. The results obtained 

 were as follows : — 



A. With the intermediate vessel : — 

 1 —7215 calories 



ories 1 

 B. Without the intermediate vessel 



rMean —7198 cal. 



I: :::::: i?iS c tT ies } Me ^ - 7mcai - 



Showing that the use of the intermediate vessel had no appre- 

 ciable effect on the results obtained *. Where, however, large 

 heat-disturbances take place in the calorimeter, it is safer to 

 dispense with it entirely; and indeed the employment of it is 

 in no case attended by any advantage. 



The time required to complete the dissolution of the salt 

 varied, of course, in different cases, but rarely exceeded 5 

 minutes. The loss (or gain) of heat by radiation &c. during 

 this interval was determined separately in each experiment. 



* If otherwise, the results of A would be smaller than those of B. 



