( 237 ) 
pressure does not bring about mixing. Whether such “open’ plaits 
occur is in no way to be decided at the present state of our know¬ 
ledge. Theoretically not because the supposition that h depends linearly 
on x leads to the existence of an open plait, but the supposition 
that b is a quadratic function of x 1 ), on the other hand, excludes 
the possibility of such plaits. Now it is certainly true that both 
these suppositions are wrong, because they do not take account of 
the variability of b, and as long as we do not know the form of 
this variability it is certainly not possible to decide which of the 
simple suppositions mentioned agrees most closely with the truth. 
Experimentally not because then not only excessively high pressures 
would be required, but moreover in most cases the appearance of 
the solid state renders any investigation impossible. Hence whether 
open plaits are possible or not will be left entirely an open question 
in what follows; in our figures, where necessarily either the one 
or the other supposition had to be made we shall assume that open 
plains do not occur, because in this way we get the simplest and 
clearest figures, which are then easily modified in such a way that 
they hold for the opposite supposition. 
4. Purpose of the investigation. 
In this first investigation we propose to examine a great part of 
the binary mixtures, of which it is known that they possess a critical 
point of mixing within easily attainable limits of temperature, for so 
far as is necessary to determine to What category they belong in a 
simple classification, which gives a survey of the possible cases. Our 
attention was chiefly directed to those systems that consist of two 
normal substances and for which also the occurrence of chemical 
compounds may be considered as excluded; but we have also 
examined other systems of very different composition in the hope of 
finding instances of all those types that are predicted by the theory. 
This investigation, which must necessarily be rather cursory on 
account of the great many systems which are to be examined, will 
be followed by accurate measurements on some of the systems which 
seem to us the most interesting. 
The substances used were partly taken from the store of the 
Laboratory, partly bought at Kahlbaum’s; none of them have been 
purposely purified, but the comparison of the critical end-points 
which we found with those given by others, proves that the purity 
l ) With the well-known value of 
