( 654 ) 



states. Also for a neighbouring mixture this deviation is onlj percep- 

 tible in third approximation, while for mixtures with a small com- 

 position, i. e. on the edge of the if>surface, it exists already in second 

 approximation. 



The cause of this smaller deviation in mixtures near the special 

 mixture must be looked for in the circumstance that those mixtures 

 in all their qualities deviate only in second approximation from a 

 single substance; thus we deduce from equation (11) that the critical 

 points : plaitpoint, critical point of contact, critical point of the homo- 

 geneous mixture and point of maximum coexistence pressure, differ 

 only in second approximation, so that tlie four curves (in the space 

 with p, V and T as coordinates), which connect these critical points 

 of all mixtures touch each other at the critical point of the special 

 mixture, which in general is not the case at the two critical })oints 

 of the pure components. 



Application to mixture.s of hydrochloric acid and ethane. 



The experiments of Kienen with mixtures of ethane and nitrous 

 oxide, the first where the existence was shown of a mixture tliat 

 in its critical phenomena agrees with a simple substance, does not 

 allow us to form a complete image of the conduct of neighbouring 

 mixtures. Besides, his investigations were only aimed at the discovery 

 of the second kind of retrograde condensation, and the existence of 

 that special mixture was a new discovery, and not the object of the 

 investigation. Suitable data for our purpose are given by the measure- 

 ments of Quint on mixtures of hydrochloric acid and ethane; accord- 

 ing to Quint the composition of the mixture which behaves as a 

 simple substance is x/c = 0.44, i.e. 0.44 gram molecules ethane and 

 0.56 gram molecules hydrogen chloride. Mixtures behaving as a 

 pure substance have also been observed bij Caubet ^) in his experi- 

 ments with CH3CI and SO^ ; as Caubet however investigated only 

 two mixtures of this binary system, his data are insufficient for 

 our purpose. 



In order to proceed with the mixtures investigated by Quint in 

 the way indicated by Kamerlingh Onnes, we must determine in the 

 first place the critical elements of the homogeneous mixture T^k, Pxk, v^k- 



pv 

 Instead, however of drawing, the log pv, log v or — , log p diagrams 



it was sufficient, as in the case of my former investigations''') of the 



1) Liquefaction des mélanges gazeux, Paris, 1900. 



2) Arch. NéerL, (2), 5, 644, 1900. 



