( 496 ) 



that if we come from low torn jx'i-atni'es the r-^onrve for the saturated 

 vapour ;iii<l thai for the saliiral(^<l iifpiid appi'dacli cacli otiici- more 

 and more, and finally pass willMiut a break ijiio each other. Al f/ 

 we f?et a repetition of what precedes in reversed ordei'. On the 

 left of /) and on the riuht of (/ we have the same thinjr as in 

 fig. 4; at p and rj h(n\e\'er. we see soiiiethinu' s|iecial, \i/. that 

 liere the critical phenomenon can he ol>sei'ved for a jnst s<itiir<ift'<l 

 solution. 



We ha\e further seen in fio:. 4 that the succession of cou(hlious 

 I^J^Si>-\- (r -^ S/i -\- (r—*(r could (tcciii- thcrc for C(»nditi(ins lying 

 above the curve hh^. h\ fig. 7 this takes |»lace. besides above the 

 curves hp and r/A, also between the concenti-ations .i\ and .I'j, corre- 

 s{)onding with the points p and y, for any '///y/V/v/y//// chosen volume, 

 because the region for SB-\-(r between p and y passes continuously 

 into the region for Sf--\-/.. Hence this phenomenon will occur much 

 more frequently for the tyjte of fig. 7 than foj- the type of fig. 4. 



The usual succession L -\- S/>-\- (r-^ L -{- (r^ (i is here only 

 possible for conditions lying within the /--/-cni-ve i'(-\'\_pf^.2f>i^'- 



The retrograde phenomenon discussed above will here be observed 

 for all conditions lying below the branch h^ h^ q of the second 

 iv^curve and above the tangent drawn to the branch >]<'zj\ from a 

 point ƒ (between t\ and /\). A very essential point of difference with 

 the case of fig. 4 is further to be found, first in the circumstance 

 that with a concentration ,i\ and a volume j\ v^ we pass here sud- 

 denly from the region L-^Sn-\-(T into the gas region and secondly 

 that for a composition sli(jhthj richer in B than .r, and with a volume 

 ,/', i\ just at the moment at which all the solid />' would evaporate, 

 a saturated solution is formed, which reaches its critical temperalurf^ 

 immediately after its formation. 



From fig. 7 follows that if we started from a concentration../', with 

 a volume j\ i\, with which therefore the point <i can l>c leached, 

 the transition L, -\- Sb-\- (r—^Su-^- G takes place at a temperature 

 lower than that corresponding to the point /;, so that the points p 

 and q can never be determined by one experiment, which it is 

 practically superfluous to mention. 



The curves piPp-t, r i\, tt^ and q^qq^, which denote the mol. 

 vols, and the concentrations of the "fluïde"" phases coexisting with solid 

 B, have still two vertical tangents, as tliey are in the immediate neigh- 

 bourhood of /y and <i, from wliicii the retrograde solidification follows. 

 For the curve *S'*S\, drawn halfway between p and q, the two 

 vertical tangents coincide in accoidance with the yu'-curve in Fig. 2, 

 which means, that there retrograde solidification is no longer possible. 



