434 HENRY A. EOWLAXD 



of inertia, the work done on it is, as we have seen, TtNWD (1 -f- c 2 ); 

 but if it had no inertia, it is evident that the work would be only 

 TiNWD. If, therefore, the calorimeter is made partially stationary, 

 either by its moment of inertia or by friction, the work will be some- 

 where between these two, and the work spent in friction will be only 

 so much taken from the error. Hence in the latter experiments the 

 inertia bar was taken off, and then the calorimeter constantly vibrated 

 through about half a millimeter on the torsion scale. 



Besides this quick vibration, the calorimeter is constantly moving to 

 the extent of a few millimetres back and forth, according to the vary- 

 ing velocity of the engine. As frequent readings were taken, these 

 changes were eliminated. In very rare cases the weights had to be 

 changed during the experiment; but this was very seldom. 



The vibration and irregular motion of the calorimeter back and forth 

 served a very useful purpose, inasmuch as it caused the friction of the 

 torsion apparatus to act first in one direction and then in the other, so 

 that it was finally eliminated. The torsion apparatus moved very 

 freely when the calorimeter was not in position, and would keep 

 vibrating for some minutes by itself, but with the calorimeter there 

 was necessarily some binding. But the vibration made it so free that 

 it would return quickly to its exact position of equilibrium when drawn 

 aside, and would also quickly show any small addition to the weights. 

 This was tried in each experiment. 



To measure the heat generated, we require to know the calorific 

 capacity of the whole calorimeter, and the rise of temperature which 

 would have taken place provided no heat had been lost by radiation. 

 The capacity of the calorimeter alone I have discussed elsewhere, find- 

 ing the total amount equal to -347 k. of water at ordinary tempera- 

 tures. The total capacity of the calorimeter is then A -f- -347, where 

 A is the weight of water. Hence Joule's equivalent in absolute meas- 

 ure is 



T _ 



~ ( 



where n is the number of revolutions of the chronograph, it making 

 one revolution to 102 of the paddles. 



The corrections needed are as follows : 



1st. Correction for weighing in air. This must be made to W, the 

 cast-iron weights, and to A -f- -347, the water and copper of the calori- 

 meter. If / is the density of the air under the given conditions, the 

 correction is -835 A. 



