MR. J. H. BRTNKWORTH ON SPECIFIC HEAT OF STEAM. 397 



consisting of 20 small-bore tubes in parallel, immersed in a water bath to keep the 

 temperature steady. The current was measured and adjusted by observing with an 

 oil gauge the difference of pressure between the ends of the resistance. The same 

 apparatus was subsequently employed by SWANN (' Phil. Trans.' A, 210, pp. 199-238, 

 1910) in a more accurate and extended investigation of the specific heats of air and 

 CO 2 . SWANN found the value 0'2414 for air at C., which corroborated the 

 conclusion that REGNAULT'S value was appreciably too low. 



A more direct verification of the result for steam was obtained by measuring the 

 adiabatic index n+l in the equation P/T B+1 = constant, at 108 C. in the neighbour- 

 hood of atmospheric pressure, by means of a compensated platinum thermometer of 

 '001 -in. wire, similar to those employed in the experiments with Prof. NIOOLSON on the 

 law of condensation of steam in 1895. The resulting values of the index lay between 

 4'26 and 4'30, giving, according to the foregoing theory, values of S at atmospheric 

 pressure and 108 C. lying between 0'488 and 0'493 calories per gr. deg. These 

 experiments were performed by W. MAKOWER at University College under my 

 direction in 1901-2, and were, published in the ' Phil. Mag.' for January, 1903. 



Jt<wtltn of otlcr CM.srnvr.s-. 



The experiments of HOLUORN and HENNTNO (' Ann. Phys.,' 18, p. 739, 1905), 

 extending over the range 110 C. to 820 C. at atmospheric pressure, indicated that 

 the specific heat of steam was constant within the limits of error of their observations, 

 but threw no light on the variation with pressure in the neighbourhood of saturation, 

 as their Icnvest range extended from 110 C. to 270 (J. Their results gave most 

 accurately the ratio of the specific heat of steam to that of air, which was found to 

 be 1'940 over their lowest range. Their value for air being 0'2315 over the same 

 range, that for steam would be 0'449, which is evidently too low, on account of the 

 error of the value 0'2315 for air. SWANN'S value (loc. cit.) for air over this range, 

 namely, 0'2443,* with the same ratio T940, would give the value 0'474 for steam, 

 agreeing closely with REGNAULT'S value from 125 C. to 225 C. If equation (7) is 

 adopted for the variation of the specific heat, this result would correspond to a value 

 0'482 for the specific heat at 108 C. and 1 atmosphere. The agreement with the 

 value 0'485 found for dry steam at 108 C. is fairly close, but rests on the assumption 

 that the specific heat at zero pressure is accurately constant, which may be to some 

 extent in error. The results of explosion experiments indicate a considerable increase 

 in the specific heat at high temperatures, which, if represented by a linear formula, 

 according to the common practice, would require an increase 0'00034 in the specific 

 heat per 1 C., or 0'027 between 110 C. and 190 C., which is more than sufficient to 

 compensate for the diminution O'OOS over this range represented by the term 



* This value has quite recently been confirmed by HOLBORN and HENNING. 



