222 



Mr. J. Joly. 



The procedure adopted was the obvious one of measuring directly 

 the effect of the displacement difference, air to steam, on the weight 

 of the copper sphere ; condensation of the steam upon it being pre- 

 vented by raising it to a temperature above that of the steam before 

 the vapour was admitted into the calorimeter. The calorimeter used 

 was spherical, and 14 cm. in diameter. The sphere was first equili- 

 brated when cold as it hung in the calorimeter, the air in the calori- 

 meter being to a great extent dried by leaving in it a vessel of calcium 

 chloride throughout the previous night. In considering, then, the 

 density of the air in which equilibration was effected, the hygrometric 

 state of the air need not be taken into account. At this point the 

 temperature of the air in the calorimeter and the height of the 

 barometer were observed. Steam was now got up in the boiler 

 attached to the calorimeter, the calorimeter opened, the calcium 

 chloride removed, and the process of heating the sphere begun. 



This consisted in applying to it a spirit flame as it was slowly 

 swung round on the suspending wire. Of course the first effect of 

 the flame is to precipitate moisture on the cold metal, but as this 

 grows hot the moisture drys off. To avoid as far as possible a change 

 of weight during this process, due to oxidation, the precaution had 

 been taken of subjecting the sphere to a prolonged course of similar 

 treatment previously, till further heating over a reasonable interval 

 of time had no appreciable effect on its weight. 



When the temperature of the sphere all over is well above that of 

 the steam, shown by touching it here and there with a stirring-rod 

 wet with water, steam is admitted into the calorimeter. 



It is observable that if now, immediately the calorimeter is filled 

 with steam, the counterpoise be adjusted till equilibrium obtains, this 

 counterpoise will be excessive. The steam is superheated in the 

 vicinity of the copper sphere, and its density diminished below its 

 true density at the prevailing pressure. In a few moments the 

 apparent weight of the sphere diminishes. The change may be as 

 much as a milligram. The vibrations of the balance now become 

 steady, and this state of equilibrium continues for from 10 to 

 14 minutes. The quantity by which the counterpoise has been in- 

 creased to maintain this equilibrium is the true result of the experi- 

 ment. Subsequently a slow and uniform increase in the apparent 

 weight of the sphere takes place, at the rate of 1 milligram in five 

 minutes. This effect, which is considered further on, is apparently 

 due to radiation, and consequent slow continued precipitation of 

 water in the sphere. It does not apparently interfere with the ex- 

 periment. When the experiment is concluded, I found it necessary 

 sometimes to make a second observation of the height of the 

 barometer. 



To reduce the experiment the following values are required : — 



