1548 
limit-curve [consequently from the part ce on curve (Z,)|; above 
7. it rises starting from its limit-curves first up to its turning-line 
evh and afterwards it falls. This is represented again in fig. 4 by 
some lines, which touch the turning-line. Below 7, the region 
consists, therefore in stable condition of one single leaf only, above 
7T., however, it consists of two leaves. The one falls starting from 
the turning-line and it finishes in the curves e and 7; the other 
falls also starting from the turning-line, but it extends moreover below 
the curves ei and 4%. 
When we consider the region Z,LG in its whole extension then 
we may again represent the turning-line by curve ayzu from fig. 5; 
we imagine the curves (5 and vc in fig. 4 within this turning-line 
and the point of contact anywhere on branch zy of the turning-line. 
Here we have a deformation of a region, more important than 
in fig. 3. The region covers here, viz. its limit-curves (Z,) and (Z,) 
already in the vicinity of the invariant point, which is not the case 
in fig. 3. Also we see that this region does not occupy in fig. 4 
the whole space between the curves (Z,) and (Z,), but a part only. 
Consequently this is different to that which we should mean to be 
allowed to deduce from fig. 2. Also several other properties seem 
to be no more valid now. When we take e.g. the rule: each region 
which covers a curve (15) contains the phase /’,; the region Z,LG 
covers here viz. the curves (Z,) and (Z,) and yet it contains neither 
the phase Z, nor Z,. Also the property: a regionangle is always 
smaller than 180° seems to be no more true now; the region 
Z,LG extends itself viz. in fig. 4 over the invariant point 7, so 
that the region-angle is 360°. 
All those contradictions disappear, however, when we take into 
consideration the conditions. 1 and 2. E 
When we consider viz. in accordance with the first condition, 
only pressures and temperatures, which differ a little only from 
those of the invariant point or in other words, when we take from 
the curves (Z,) and (Z,) only parts in the vicinity of the point 2, 
then the region Z,LG occupies indeed the space between the curves 
(Z,) and (Z,). 
The other contradictions disappear when we take into consider- 
ation the second condition; this is apparent from the following. 
When we take away from fig. 4 the leaf cevhhqsu, so that the 
leaf evhbi remains only then all contradictions have disappeared. 
The region-angle is then smaller than 180° and the region ZLC 
covers no more its limit-curves (Z,) and (Z,). 
The liquids of the remaining region evh i in fig. 4 are repre- 
