( 589 ) 



Then we measured olF for the isotliennals of carbon dioxide which 

 lay between those of the mixtures, the situation of the former with 

 regard to the two neiglibonring isothermals of tiie latter ; this was 

 done for a number of points distributed at regular distances. Let 

 2\ and 1\ be the temperatures of the two neighbouring isothermals 

 of the mixture, and T' the temperature which for the mixture cor- 

 responds to that of the isothermal of carbon dioxide considered, then 



r—2\ 



we obtain a great number of values for the relation tp = — — -. 



We assumed that the isothermals coincided best when the mean 

 square of the differences of </) and the mean tp were as small as 

 possible. 



Table XXIV, which refers to the comparison of the /c^ ^;-diagram 

 of the mixture 0.1 with that of carbon dioxide shows in the 3^' colunm 

 the afore-mentioned mean square of the differences, viz. for the 

 isothermals of 41.95^ C. and 37.09° C of carbon dioxide (after reduction 

 to the same temperature interval 1\ — 1\) coml)ined, for the different 

 superimposings of the coinciding systems indicated in the 1^' colunm, 

 while 1.85 of the mixture coincided with the values of log p for 

 carbon dioxide given in the 2"^ column. 



TABLE XXIY. 



In table XXV we find the data for the critical points of the mixtures 

 thus found, combined witli tlie data about the plaitpoints and the 

 points of contact, and also the critical points of carbon dioxide and 

 oxygen (the latter according to Oi.szewski). ^) 



1) It here appears that in the diagram given in § 4 (p. 587 ) in order to 

 obtain the best agreement for the area of the larger volumes, the isothermals of 

 the mixture 0.1 must be moved 5 mm. to the left, those of the second mixture 

 G mm. to the right. The conclusions about the non-correspondence at the smaller 

 volumes, however, still hold, tlie deviations even increase. 



