( 381 ) 



(3.48 atnis. at 9.5(!.)^), and (ho Iwo values louiid Uv us (§ 8). Extra- 

 polation to — -38". 5 ('. gives 0.53 atms. For carbon dioxide followB 

 from Kuknen's observations') at — 25° C, /> =r 1(^.5 atms., at 



' I max 



— 38°.5C., /) =10.4 atms. 



' wax 



From our model for — 25° C. (Comm. Suppl. N". 8) follows for 

 methyl chloride at — 25° C. p = 0.59 atms. and for carbon dioxid^y 



•' ' max 



p ^10 alms, (instead of 0.96 and 16.5). The begin condensation 



' max ^ 



pressure of the mixture x = 0.5084 is accoi*ding to the model 1.30 atms. 

 instead of 2.05 atms., as was found by us. 



The insufficient concordance on the side of methyl chloride shows 

 once more clearly that for methyl chloride and ether the agreemeut 

 required for the validity of the law of corresponding states leaves 

 much to be desired. 



In order to judge about the degree of deviation of the mixtures' 

 from this law, we must first free ourselves as much as possible from- 

 the deviations of the separate components. We have tried to do so 

 by raising the ju'-curve 0.4 atms. Then w^e get for j>^^^^ 0.99 atms. for 



methyl chloride, 16.4 atms. for carbon dioxide, and 7;^^= 1.7 atms. for 



the examined mixture. The i-emaiuing difference with the observed 

 p = 2.05 atms. is, no doubt, partly due to the fact that the surface 



had not yet been constructed with sufficient care, to determine 

 with sufficient certainty the place of the points of contact when 

 the plate of glass is rolled over the ridge and over the convex part. 

 More points should have been calculated if we wished to render the 

 model sufticieutly precise for an accurate determination. 



This nncertainty when rolling, makes it also doubtful whether on 

 account of the law of corresponding states, w ith an accurately executed 

 theoretical model, a straight line would really be found for the liquid 

 branch of the border curve in the /Af-diagram, as Avas derived from 

 our model. We got the impression, that a curve of the same character 

 as that derived by Hartman in his experiments and from hi« model, 

 ought to be found also by us, on our model (Comm. Suppl. N". 8), 

 but we had not enough data for ascertaining the deviations from the 

 straight line, and had therefore to adopt the straight line us the 

 simplest approximation. 



That the c.xperimeulal li(iuid branch of the border curve in the 



1) Comm. N". (U. 



■^} I'liil. Mag. Jan. 1002. 



