34 Mr. J. Joly. On the [Nov. 22, 



The first method of procedure adopted was to compress by means of 

 a pump a certain quantity of dry air into a thin copper sphere, the 

 sphere being then closed by a screw valve. The quantity of gas in 

 the sphere is ascertained by weighing. 



The sphere is now hung in the calorimeter, suspended from a 

 delicate balance, reading to one-tenth of a milligram, and its thermal 

 capacity determined in a certain number of experiments. The gas is 

 then released, and the sphere further exhausted by means of an air- 

 pump, sealed, and its thermal capacity again determined in a number 

 of experiments. This allows of a computation of the thermal capacity 

 of the contained gas, 



This method I at first used, but as in dealing with the effect on the 

 weighings, due to the transference of so bulky a body from air to 

 steam, much troublesome calculation and risk of error was involved, I 

 modified it in the following manner : 



Two spheres are prepared, alike with respect to external volume, 

 and approximately of the same weight. The thermal capacities of 

 these are compared in a double calorimeter, being suspended one 

 from each arm of a short beam balance. If their thermal capacities 

 are not alike a calculated weight of copper wire is introduced into 

 that of least thermal capacity. They are in this way brought to have 

 the same thermal capacity, so that in an experiment on the empty 

 spheres there is no effect on the balance. 



One of these is now pumped full of air, and the specific heats of the 

 spheres again compared. The weight of condensation now obtained is 

 that due to the gas alone. It is evident that many sources of error 

 obtaining in the former method are removed in the latter. The 

 results obtained are also far more consistent one with another. In this 

 case the specific heat is calculated directly on the formula given in 

 my paper on the steam calorimeter 



where X is the latent heat of steam, 10 the weight of steam condensed 

 by the gas, W the weight of gas, and 2 ^ the extremes of tempera- 

 ture obtaining. S so calculated may be subject to some slight correc- 

 tions, which I will not here enter into. 



Up to the present I have only dealt with air, but I have made pre- 

 parations for resuming shortly my work, dealing with other gases, 

 over critical temperatures if possible in some cases, and making con- 

 firmatory experiments on air and also in extension of those given 

 below. 



The spheres used are about 6' 7 cm. internal diameter; volume 

 158*5 c.c. They weigh about 92'2 grams. That containing the air is 

 tested hydraulically to 1000 Ibs. per square inch. 



