( 294 ) 



In tlie preceding communication I have concluded to the same 



/O/A , dp, . , . . , . 

 value from tlie equalitv of r^ and -— m the critical circumstances, 



\óxjü dx 



p. T, — T 



and by means of the empirical foi-muhi — 1^ —f -~^- For from this 



p,. T 



formula follows 



dpc dpy. . d Ty. 



Pcdx p,.dx Tdx 



or 



1 dpc _ dTy 1 db , dT, 



p,, dx Tydx h dx ' Tydx 

 or 



pydxj,2^ ^•' ^ {Tydx'^ f-lhdxy 



But we could arrive at this equation in a much simpler way still. 

 From: 



follows, when clT is put equal to cITy (taking for dTy the variation 

 of the critical temperature of the unsplit mixture) 

 1 dp 1 /ö/>\ f 2' dp \ dTy, 



py dx p yè^vJ^-T \p dT) Tydx 



and so 



1 /ö/A __ 1 dpy , dT, 



p \dxJvT py. (^'^ ' ^'//^'^ 



And this equation is not only preferable because it is shorter, but 

 also because it is independent of the circumstance whether the law 

 of corresponding states is applicable or not. The value of ƒ in this 

 derivation is that of the component. 



Besides ( — 1 we have to determine the value of t— r- . 

 \dxJ,T \dxdvjT 



For tiiis quantity we find : 



d'b da 



ÖV^ ^^^^ f èh\ fdb\ dxdv dx 



ö>\_2ajl da j^_j^/ d^^ya^-N MRT v' Ó'M 



dxd'vjfv^l'^dx '^ {v-by\ dvXdxJ, 2a {v—bydxdv\' 



or 



(: 



In this expression only the quantity ^-y is unknown. We deter- 

 mine it from : 



