966 
DR. J. JOLY ON THE SPECIFIC HEATS 
evidently unaffected by any buoyancy effect due to a small quantity of the jacketting 
vapour leaking in—as already indicated—and so facilitate calculations lia^dng reference 
to displacement errors. 
To correct for the unequal thermal capacities of the spheres over the several ranges, 
it was thought best to calculate the correction upon the basis of experiments over the 
widest range of temperature. Table I. contains the experiments upon the spheres 
between air temperature and steam temperature. 
Table I.—Correction for Unequal Thermal Capacity of Spheres. 
fi- 
b- 
^2 ~ h- 
00 . 
11-89 
100-16 
88-27 
0-0245 
11-61 
100-24 
88-63 
0-0243 
12-07 
100-18 
88-11 
0-0251 
13-30 
99-88 
87-58 
0-0251 
Means. 
88-15 
-0-02476 
The mean result is a deductive correction of 0'02475 over 88'15 degrees. To 
determine the proper deduction to be applied to other intervals of temperature it is 
necessary first to make a correction by deducting a constant effect due to unequal 
displacement, and thus arrive at the true calorific efiect due to heating through 
88'15 degrees. The external volume of the active sphere is 105’595 cub. centims., 
that of the idle sphere but 99'810 cub. centims. (see Part II., pp. 944 and 945), this 
gives a buoyancy correction of — 0'0035 grm. ; leaving a calorific effect of 0’0212. 
Assuming Bede’s results for the specific heat of copper, we have the formula 
C = 0'0892 + ’000065 f, by which to calculate the deduction proper to other ranges 
of temperature. To this the displacement effect proper to each case must be added. 
The “ correction for spheres ” at base of each table is got in this manner. 
Regarding other corrections, the remarks made in Parts I. and II., and the system 
of tabulation adopted in Part II., are adhered to. The initial temperatures are, of 
course, the boiling-points of the liquids used in jacketting, and P^ is the pressure at 
that temperature, due to W grammes of gas. The jrressure for the experiments of 
Tables II. to XVI. inclusive are obtained from a plotting of Amagat’s results against 
density. The mean piessure, P, is obtained by adding one atmosphere to the value 
read dmectly from the curve, as explained in Part II. The pressures obtaining in 
the remaining experiments. Tables XVII. to XX., are obtained from Andrews’ 
results. 
The purity of the gas used was tested on a sanqrle released from the sphere when 
reducing the mass under experinrent. 66 cub. centims. were absorbed b}" caustic 
