( 637 ) 



Foi' not \er\ (liliile solutions this is liowcnci' no loiiiici- the case 

 and as van Laar') tbnnd, for tlicni llic follow iiiu- relation holds: 



• p rt;' t 



or 



A/> 1 /A/A- 



(5) 



+ 0. - 



LT 



S 7', 



(^) 



m\^ r 



It' therefore we wish to compare the resnlts of Jion-dilnled solutions, 

 we ean e.g. ealonlate instead of i(^=zijj according to (J) and (2) the 



following Mxlnes : 



S T„ N 



B.T 



or 



/T) = 



A;, 



N 



'Ap 1 



which theoretically must have exactly the same value. 



T have ab-eady shown l)efore '), that the error committed by |)ntting 

 i.G= ID instead of I q =i I £> for not very dibde solntions, is not 

 sensible before Ave reach the concentration ± 1 gr. 



If we now compare the resnbs obtained by Raout.t, Jonks, GktjMan 

 and myself by means of the factor /, ^\Q get the following table: 



T A H L E lY. 

 Na 01. 



Concentration /(? — Raoult. 



/g-JoNES and 

 Getman. 



I J, — Smits. 



0.0.") 

 0.40 

 0.50 

 1.00 

 2.00 

 3.00 



1 ül 

 -i.SO 



l.s.-. 

 1.8() 



l.SO 



i.X7 



1.004 



1 .'J81 



2.064 



2.293 



i.si 



1.7-) 

 l.SO 

 1.8S 

 2.03 



It is obvious that the second colunui is now qnalitatively in 

 concordance with the two following; also the factor T derived from 

 the observations of Raoitt, gives a niinimnm, bnt it is very taint, 

 so faint, that it did not a|)pear in the calculation of /(-; (table III)''). 



1) Zeitschr. f. physik. Chem. 15, 457 (1894). 

 ■i) These Proceedings III, Febr. 23 1901 p. 512. 



■5) This lias already been shown by van La.vu in a sonicwliat diirercnl way. 

 (Archives TeijltM' 8 (1903) j. 



