793 



Chemistry. — "On Critical Kndpoitits /» Ternary Si/stems". III. 

 By Prof. A. Smits. (Communicated by Prof. J. Ü. vanderWaals.) 



(Gommunieated in tlie meeting of October 30, 191.5). 



1. Projections on the concentration triangle. 



Already in a few earlier comnuinications I have written on the 

 occurrence of Critical endpoints in ternary systems '), specially because 

 this subject is of great importance for petj'oi/iriphi/ and particularly 

 for the chemistry of the magma. 



In my latest communication six cases have been successively dis- 

 cussed, the sixth of which referred to the occurrence of a binary 

 compound as solid phase. 



The consideration of a case that a ternary compound occurs as 

 solid phase was then postponed till later, because it seemed to me 

 that this case would not be a subject of study for the present. 



Soon after, however. Dr. Morey informed me that he was engaged 

 in the study of the ternary system H5O — SiO, — K^O in the Geo- 

 physical Laboratory of the Carnegie Institution of Washington, in 

 the course' of which research he met with a case that had not yet 

 been treated by me, so that it had become desirable to extend my 

 earlier considerations. 



I greatly desired, particularly because I knew Dr. Morey would 

 be interested, to undertake this work immediately after his commu- 

 nication, but want of time compelled me to put this off until now. 



7^^ Case. Continuing oui' earlier considerations we shall therefore 

 begin with the case that the volatile com|)onent A gives critical 

 endpoints neither with B nor with C, but that a ternary compound 

 Z), occurs, which more or less dissociates in the liquid and the 

 vapour phase, and which presents critical endpoints with A. This 

 case is schematically represented in fig. 1. 



In this figure in the first place the ternary eutectic liquid and 

 vapour lines are given, so the liquids and the vapours that coexist 

 with two solid substances. In the second place the critical endpoint 

 curve has been given. 



1) These Proc. XIII, p. 342 and XV, p. 184. 



51* 



