( <^5 ) 



Li(liii<l siil[)Iiiii' aiid li((iii(l ('lilorinc are iiiiscihlc in all pi'oportioiis. 

 If in these mixtures iio coinpoiiiul molecules were formed, two 

 regular boiling-point curves might be expected which would diverge 

 very much in the centre because the boiling points of the two 

 components lie far apart. 



In these mixtures, however, a fairly stable compound S^Clj is 

 formed. If this comi)ound were absolutely stable, that is if a liquid 

 and a vapour of the composition S.^C\^ consisted of nothing but 

 molecules of this formula, then the liquid and vapour would at this 

 point, have exactly the same composition. The system S-|-C1 would 

 then in reality be compounded of the two systems S -}- S^Cl.^ and 

 S3CI3 4" CI2 which could no doubt be represented in one figure, but 

 then the liquid- and the \'apour-pressure curves would not i)ass con- 

 tinuously into each other at the composition S.^Ol.^. 



As it is known that the dissociation of the vapour of ^.^jCl^ is 

 small it maj^ be anticipated that, in the system S -f- CI, the comiec- 

 tion at the composition S CI might become continuous, but in such 

 a way that the vapour and liquid curves nearly' coincide at this |)oint. 



This state of affairs was now confirmed and is indicated in the 

 figure by the liquid curve 1,3 and by the vapour curve 2,4. It will 

 be seen that the curves 1 and 2 and 3 and 4 nearly meet in a point 

 situated near the composition S CI, but in reality we ha\e here 

 continuity, from which it appears that S.^Cl2 is not absolutely stable 

 either in the form of liquid or vapour. The difference however, is 

 so small that this type really exhibits one of the smallest forms of 

 deviation. 



In the case of binary mixtures where the compound formed is 

 more strongly dissociated the divergence of the t\AO curves at the point 

 representing the compound will be much greater. The liquid curve 

 and the vapour cur>'e of the entire system will then more and more 

 assume the form which in the figure belongs to both halves. 



The investigation however, showed a further peculiarity in the 

 lo^ver half. The boiling-point curves 1 and 2 for the mixtures whose 

 composition lies between CI and SCI only relate to mixtures, which 

 are freshly prepared from li(iuid S^ C\.^ and li(piid chlorine. 



These mixtures at temperatures below 0' retain for a v(M-y long 

 time their yellow colour and then exhibit the boiling [)oint lines 

 indicated at 1 and 2. At higher tenq)erafures, and wry (piickly 

 aboxe 30°, the colour becomes darker and linally blood i-cd, chiefly 

 in the case of mixtures a[)pr<)aching the conqK)sition S(M^. 



The boiling })oints then rise, sometimes very considei'ably, lo a 

 maximum amount of about 70° so that the line 5 is found for the 



5* 



