368 PROF. C. FREWEN JENKIN AND MR. D. R. PYE ON THE 



50 C. and the 900 line from zero up to + 23'2 C. where it meets the limit curve. 

 The starting points for these curves are the points 



I = + 0'64, </> = - 0'0024 for 700-lb. line. 

 I = +0'44, </> = - G'0049 900-lb. 



The limit curve from 50 C. to + 23 '2 C. was then set off from these curves; 

 it passes through the zero temperature point where <p = 0, I = + 0'9. The sloping 

 straight lines corresponding to pressures in the saturated area from 900 Ib. to 150 Ib. 

 were then drawn from the limit curve at slopes equal to their temperatures. The 

 gas-limit curve was then determined by marking points on each of these pressure 

 lines at values of <-/> taken from the 0</> chart. The apex of the limit curve was then 

 plotted by means of equation (iv) Cor. (3). 



The gas-pressure and temperature curves up to + 20 C. were th'en plotted from 

 the throttle experiments ; and starting from the 20" C. line JOLY'S four constant-volume 



lines J 1; Jn, J 3 and J 4 were plotted, ten degrees at a time, by means of the expressions 



/j 

 (51 = C t ,M + Anty> and </> OJog, using JULY'S values for C u . Several more constant- 



PI 



volume lines were then drawn between J 4 and C. in the liquid area, assuming that 

 G v = 0'214. The values of ftp were in all cases calculated from AMAGAT'S data. The 

 points on all these volume lines were then marked where they are cut by the 

 constant-pressure curves again using AMAGAT'S data. The constant-temperature 

 curves for 10 (J. intervals were then drawn through the points already marked on 

 the volume lines. The constant-pressure curves were then drawn by drawing an 

 envelope of tangents passing through the points already marked on the volume lines. 

 Any error in the chart was shown up to this stage by the envelope of tangents missing 

 the marked points on some of the volume lines. This check is a very rigorous one. 

 An error of thermal unit in the position of the liquid-limit curve will throw a 

 pressure curve 1 inch away from the point it has to pass through. 



AMAGAT'S data do not go below zero, so a different procedure was necessary for the 

 liquid area below zero. Constant-temperature lines were drawn at 25 C. and 50 C. ; 



the first is horizontal because ( ) =0 at 25 C. as is shown in fig. 9 of the former 



paper ; the second slopes upwards at an angle which can be calculated from the 

 formula 



(dl\ = _(dl\ (d6\ 

 \dplt \dejf ' \dp)i 



= G p x throttle drop* per 100 Ib. 

 using the data given in Table VIII. of the former paper, extrapolated to 50 C. 



^ The throttle drop per 100 Ib. is given in fig. 9 of the former paper, where it is called the Joule- 

 Thomson effect. 



