( 161 ) 



The boundary oui-ve for af^on has been lenj-thened on llic li(|ni(l 

 side by means of Bait and Donnan's ol)servalions in Ihe noi.uiiboni-- 

 hood of (he boiling point. 



We mentioned in our former paper (hat the in(ro(bjction of (he 



quantity ;.o = -^ v ^) enabled us to leave entii-elv on( of accouiU (he 



value of the critical volume deduced for isopejitane from the diameter 



law '\ and for argon fron> the equation ( . 1 =1 )■ The same 



principle has been also utilised in drawing the whole of the accom- 

 panying diagram with the exception of the critical points themselves 

 (marked k on the diagram) which here appear as points of contact 

 between the experimental boundary curves and the cri(ical (empe- 

 rature axis. 



From the diagram it is quite evident that (he various differences 

 between argon and ordinary normal substances, in pai'ticulai- isopentane 

 are systematically connected over the whole i-egion, and (hat we 

 must ascribe the appearance of this difference graph to a definite 

 cause common to the various deviations. We may obviously look 

 for this cause in the hypothesis that (he mona(omic argon molecule 

 is less compressible (han the poly-a(omic isopentane molecule. This 

 circumstance is clearly shown at (he smaller volumes. Along (he 

 whole of (he liquid branch of the boundary curve the argon volume 



Pk''k 



1) The quantity ; = --- (sec Proc. June 1001 Comm. No. 71) which was used 

 ^k 



/? 7' 



in out- former paper is connected with the f|uantily /v|. = "^ by the simple rc- 



lalion A = -77- . The property that is from the point ot view of Ijie law of 



corrcsponding states an invariant was lirst utilised in Proc. fSept. 1000 Comm. 

 N". 50a, while in Arch, necrl. (2) 5. 644. 1900 Gomni. N". G5 ;t was first applied 

 to mixtures. In Arch, necrl. (2). 5. 665. 1000 (Comm. N'. 66 and Proc. Jan. 1004 

 ii: appeared as C4 alona; with the quantities (7i C^ ('< which had been introduced 

 before, and which, from the point of view of the law of corresponding slates, are 

 also invariants, in Proc. March 1011, Comm. 120(/, and Proc. Febr. 1911, Comm. 

 N'^. 117 Ki is written for Cj. to prevent confusion with the coellicient (' of the 

 general equation of stale, and in future this change will be continued. 



Comp. also Ph. A. Guye, Ann. d. Chim. e. d. phys. (6), 21, 211, 1890 and Arch, 

 d. Sc. phys. e. nat. (3), 23, 204, 1800. 



~) As Plate II of this paper we print an improved block of the diameter graph 

 (sec Proc. Dec. 1910, Comm. N'. ll.Sa, and C A. Crummelin, Diss. Leiden 1910) 

 as certain inaccuracies appeared unnoticed in the original block. 



