( 196 ) 



iilready be examined in main lines. Bv tiie aid of these determina- 

 tions I have already schematically f»iven the shape of the P-, /'-section 

 thi'ough the spacial tigure for about 20'', which is reproduced in 

 ligure 1. From the slight differences in pressure between the tension- 

 line of the liquid hydrogensulphide on one side {a in Fig. 1), and 

 the two three-phase lines SLJjr(J)) and L^L^G{c) on the other side 

 it was already vei-y i)robable that the concentrations of L^ for the 

 two mentioned three-phase pressures would lie near the hydrogen 



P 



^ 



]{,Z 



lO -h Li, 



3.G 



VG 



H,0 



Fi.2. 1. 



sulphide side, while the gas phases corresponding to them must 

 contain >till less water than />,. With regard to the situation of 

 G and L, on the thi-ee-phase line SL^G{(I) we may derive from 

 the relative volatility of the two components that G lies on the 

 side of the first component; the situation of L.,, however, cannot 

 be found from the determinations of vapour tension. 1 had already 

 drawn A^ iu the P-,i'-section on the side of water, as some {)reli- 

 iniiiary experiments had already proved that the liquid coexisting 

 with hydrate on SLJt{(1) contains only little hydrogensulphide. 



To obtain further certainty I have determined the situation of the 

 liquid bi-anches on the three-phase lines SLJjrih) and SL.,G [d] by 



