3G4 PROF. C. FREWEN JENKIN AND MR. D. R. PYE ON THE 



the- new curve and the dotted line the old curve. The maximum shift, at -50 C. is 

 0'0042 carnots. The three pressure curves, 300, 200 and 150 were then redrawn to 

 start on the new limit curve, and the throttle experiments for these curves replotted. 

 The calculations, made as explained in Appendix IV., for this second plotting are set 

 out in Table F, which gives the co-ordinates 9 and < for the points on each line. 

 The co-ordinates are arranged in columns ; each column contains the results for one 

 curve, the value of the pressure being stated at the top of the column. The second 

 line in each column gives the value of I on the limit curve for that pressure. 



The 400-pressure curve was assumed to be correct in plotting the results of all 

 lines which had readings on that curve, i.e., H, J, K, L, M, N, Q and R The O 

 line was plotted from the 300-pressure curve and S, T, U, V, from the 200 curve. P 

 was plotted from the 500 curve. The results, with the three bottom curves replotted in 

 this way, should all fall on the pressure lines : the agreement is remarkably good, the 

 errors being too small to show in fig. G. The alteration in the two limit curves which 

 are shown to be necessary by the new experiments leave the width of the diagram 

 between the limit curves practically unaltered, so that the values previously obtained 

 for the latent heat are hardly changed. 



The reasons for accepting the modification of the bottom of the gas-limit curve are 

 as follows : the extreme simplicity of the throttling experiments make errors improb- 

 able. The close agreement between all the specific heat experiments and all the 

 throttle experiments over wide ranges of temperature makes any systematic error in 

 the throttle experiments improbable. The close confirmation of the accuracy of the 

 upper part of the limit curve also supports the accuracy of the throttle experiment. 

 The temperatures readied in the throttle experiments were lower than ^hose given by 

 the uncorrected chart. Had there been errors in the throttle experiments due to 

 conduction or radiation the difference would have been in the other direction. 



Construction of tlie I^> Chart. 



The chart (Plate 6) is plotted on skew co-ordinates, the angle between the axes being 

 arc-tan ( 0'3 degrees). The vertical I scale is 1 inch = 5 Th.U. ; the horizontal scale 

 is 1 inch = O'OOS carnots. The whole chart is divided into a skew graticule, the distance 

 between horizontal lines being 1 Th.U. (J, inch), and the horizontal distance between 

 the sloping lines being O'Ol carnots (2 inch). 



The part of the chart from 50 C. to -I- 23 G. (and the super-heated area up to + 30 C. 

 for pressures below 800) represents the results of the authors' experiments. The rest 

 of the chart has been completed by using MOLLIEB'S (2) limit curve (approximately) 

 and AMAGAT'S (3) results for the relation between pressure, volume and temperature. 

 These data alone are not sufficient some data on heat quantities are necessary ; we 

 have therefore used JOLT'S (4) experiments on the specific heat at constant volume, 

 which cover a considerable range (between the lines J^ and J 4 on the chart), and for 



