EXPERIMENTAL STUDY OF HEAT LEAKAGE. 795 



h. Existence of a pressure coefficient of fx. 



In 1909, Davis ^^ made a very complete study of all experimental 

 results on the Joule-Thomson effect in steam which had at that time 

 been published, and of some which had not been published. His 

 purpose was to test the applicability of the law of corresponding 

 states to the Joule-Thomson effect in steam and in CO2. He states 

 that a careful scrutiny of all the observations with which his paper 

 concerns itself failed to reveal any systematic variation of fx with 

 pressure, and concludes that if such a variation exists at all, it is 

 within the limits of error of the experimental work. These limits 

 are not narrow, but the wide range of pressure covered by the four 

 experimenters of whose results Davis made use certainly justifies the 

 inference that the pressure-coefficient of n must be small to have 

 escaped detection. 



In the present research, a few experiments have been made at mean 

 pressures of 83.0 Ibs./in.- (5.84 kgm./cm.-) and 32.7 lbs./in.-(2.30 

 kgm./'cm.-), these pressures being such that the mean pressure of the 

 other experiments is approximately their arithmetic mean. The 

 experiments referred to were made with the VI plug, without inter- 

 change of thermometers. The ju' vs. , plots are shown in Fig. 17, to- 

 gether with the curve for plug VI already shown in Fig. 12. Owing 

 to unusually large accidental errors (or possibly, genuine curvature 

 of the plot) in the case of the highest pressure, and to the fact that the 

 flow through the plug was comparatively small for the lowest, the 



rectilinear extrapolation of the ^l' vs. . plot is unreliable in both cases. 



Moreo^•er, the presence of the regeneration effect in this plug dimin- 

 ishes one's confidence in employing the results to arrive at a value of 

 the pressure coefficient of /x, particularly as this effect seems more 

 prominent in the 5.84 kgm./cm.- runs than in the others. There is 

 also the possibility of wet steam in the 5.84 kgm./cm.- runs, which 

 were fairly close to the saturation line. A rough estimate of the 

 magnitude of {djx/dp)t can, however, be obtained by considering those 



parts of the three curves on the /u' vs. .diagram which overlap, ascrib- 



11 Proc. Am. Acad., 45, 243-264 (1910). 



