( 850 ) 
ratures and the low three-phase pressures corresponding with them 
though perfectly sufficient qualitatively, could of course lay claim 
to but little accuracy, particularly because the pressures were read 
on a metal manometer, which indicated up to 250 atmospheres, and 
which can give only rough values below about 20 atmospheres, and 
because the slow setting in of equilibrium can cause a great relative 
error at these low pressures. 
I have now tried to find an improved method for the investigation 
at pressures below 20 atmospheres, which could yield more accurate 
results. I intend to give here a description of this method and to 
demonstrate its efficiency not by the aid of the system hydrogen 
sulphide + ammoniac, as this would only involve a repetition 
of former observations, which would not open new vistas *), but 
apply this method to another system with hydrogen sulphide as 
component, namely the system hydrogen sulphide + water. As will 
appear from the following description, mercury is not affected by 
moist hydrogen sulphide, at least below 30°, when presence of air 
is carefully excluded. 
2. Preparation of the mixtures. 
For the preparation of a hydrogen sulphide + water mixture the 
Fig. 1. 
1) These observations will shortly appear in the Zeitschrift für physikalische 
Chemie. It may only be remarked here that action of hydrogen sulphide on 
mercury was no longer observed, and that e.g. reduction of the volume under 
three-phase pressure to half its size, gave an increase of pressure, which generally 
amounted to iess, but in one case to slightly more than O,L atmosphere; these 
deviations are about as great as the errors of observation. 
