( 920 ) 
we shall easily be able to reach the value 0.001 and smaller. But 
the maximum quantity of the dense liquid, which was even difficult 
to observe for our last mixture, gets so small at still smaller a that 
it is lost to sight. Already for quantities as they are found in our 
mixture VI we were in danger of being misled about the appearance 
of the homogeneous state by the exceedingly slight quantity of liquid 
being spead over the wall of the tube by the stirring, which is 
necessary for the equilibrium. Only when after violent stirring we 
waited a long time, we saw a trace of liquid slowly collect again 
on the bottom of the tube. Perhaps we may sueceed by special 
experimental contrivances (e.g. by increasing the quantity of sub- 
stance without enlarging the wall of the tube in the same proportion 
nor the diameter of the tube at the place where the liquid 
collects) to surmount this difficulty. For the present we prefer to 
try and ascertain the double retrogade condensation in other systems 
where it may not take place with such a small rz as in the 
investigated one. As such, systems of ethane and nitrous oxyde with 
little volatile substances suggest themselves in the first place. 
If the phenomena deseribed up to now are in perfect concordance 
with the theoretical anticipations — and not the least in this that 
double retrograde condensation is a phenomenon exceedingly difficult to 
observe — in one respect the obtained experimental result was 
different from what we had expected. On p. 825 loc. cit. VAN DER 
Waars says: “We might begin to try and show that after the 
termination of the. first retrograde condensation of the equilibrium 
3.1 another condensation makes its appearance with further raised 
pressure.” This sentence and the preceding one: “The fact that 
it only undoubtedly exists for such small values of «7 which do 
not only lie below wv, but must moreover be smaller than the 
value of « of the plaitpoint of the equilibrium 2.1 ete”, suppose 
— a supposition which naturally suggests itself — that it will 
be comparatively easy to realise the retrograde condensation of 
the second liquid, but very difficult to realise the circumstances in 
which after this the condensation of the lighter liquid takes place in 
a retrograde way. To our surprise it appeared, however, that for 
the condensation 2.1 the retrograde phenomenon appears with a 
distinctness as is only met with for very concentrated mixtures as a 
rule. Not only did it appear that for the mixture “= 0.00875 the 
point of contact and the plaitpoint were still about 1°.65 apart (or 
more; the piaitpoint lay at 34°.5, and it is the question whether we 
have really reached the temperature of the point of contact, this may 
remain in the metastable region); but also the increase of the liquid, 
