y -i-j') ) 



Uy iiieaiis of (1") and (3') — IIiü lasl (Mjiialioii for llic calciilaliuii 



of ril, after <l' lias heen cak'ulated from the tirsr — or of (4) and 



(5), wheji ??? <^ 0,4, we now easily ealenhite the |>rc('eding table. 



I - m 

 In the eoliunn with the heading J -|- — , we have ealcnlated 



the value of the abscissae of the straujltt luw which connects the ci-itical 



point with the point - (</ -|- ;/) = 1,5. The deviations are so slight 



Li 



[the greatest deviation at about //?= 0,77 amounts onl}' to 1 7o ) 



that in a graphical representation made at the ordinary scale the 



1 



deviations of the curve [d -\- d') z^ J- [ni] from the straight line 



Nvould be hardly perceptible. Hence the above rei)resentation is 



greatly exaggerated . 



As we shall immediately see, the coefliciejit of direction of thé cui-ve 



1 

 in the critical point is =0,4, while that for //a = 0, - (r/-j- (/)== 1,'^ 



Li 



4 

 iiets the valne ot or 0,44. The coefticient of direction of the straight 



line being =: 0,5, the real "straight" diameter will have moved 

 somewhat to the side of the vnpour at the critical point, and some- 

 what to the side of the liquid at iii = 0. About at m =z 0,35 it 

 passes the straight line. As is easily derived from the above table, 

 the coefficient of direction of the curve in this latter point is = 0,5B, 

 Near m = we derive from the table 0,46, which will pass into 

 0,444 exactly at in = 0. It follows namely from (4j (see the note), that 



IS' — VJcZ+ti') V/, — '/.,d 



= ^ approaches to 7,, at y/i = 0. If we wish 



w m 



also to derive the dii-ection iiear the critical point from the table, 



we cannot make use of the first values, because the calculation of 



(/' from d by means of (3) is not accurate enough. But from the 



values for f/z=l,3 and 1,5 we calculate the value 0,405 for r/ = J, 4, 



which comes already very near to the limiting value 0,4. Finally 



we point out that for d=2 the coefticient of direction of the curve 



amounts h) 0,41), so that there (at the place of the maximum deviation 



from the straight line) the curve runs almost parallel with the 



straight line. 



So il foHow's fi'oni the above that the fact thai Ihe curve 72 (^^~l~'^') =^ 



= f{i/i) is almost strdijht, must not be ascribed to a very special 



form of the e([uation of state, as Cardoso does among others, who puts 



