( 832 ) 
excluded and the mixture is only in contact with a small quantity 
of pure mercury; so contact with not entirely air-free mereury. and 
rubber joints are entirely avoided in this way. 
3. Determination of pressure and temperature. A cylindric vessel 
conically narrowed at the lower end was fastened to the Cailletet- 
tube by means of a cork, and filled with water. The heating took 
place electrically; the temperature was regulated by means of an 
incandescent lamp resistance, and read on an Anschützthermometer, 
which had been compared with a normal thermometer. The stirring 
took place in the waterbath by means of lead plates, which moved 
vertically up and down, in the test tube by means of a KUENEN 
stirrer; the electromagnetic coil required for this was vertically moved 
round the heating-vessel. 
For the determination of the pressure I used two air-manometers, 
one of which indicated a minimum pressure of about 3, the other 
of about 8 atmospheres. The errors of the method remain in this way 
below O°.1 and 0.1 atmosphere (errors of the manometer and difference 
of position between the mercury in the test tube resp. manometer 
and in the pressure cylinder); the errors which can be made in the 
determination of the three phase curves, may be generally estimated 
at about 0.1 atmosphere, as appears from the concordance of the 
results. J shall, however, return to this, when discussing the results. 
4. Results. In fig. 2 the P-7-projection of the spacial figure is 
represented. In the first place we see drawn in it the vapour-pres- 
sure curve of hydrogen sulphide, which had been determined before 
by many observers. Just as in my previous determinations concern- 
ing the system hydrogen sulphide + ammoniac | arrived again at the 
result that the values given by RrexauLr, are too high; in the range 
of temperature investigated by me the deviations vary: between 0.8 
and more than one atmosphere. The only value which was deter- 
mined by Onszewskt in the range examined by me, presents a 
deviation of less than 0.1 atmosphere from mine, and has been denoted 
by OQ in the graphical representation ; RraNauLt’s deviating values 
I have omitted in the figure for the sake of clearness. 
The hydrogen sulphide showed a variation of pressure of less 
than 0.1 atmosphere on isothermal compression to a fourth of the 
total volume, while also the pressure at which the last quantity of 
vapour disappeared differed less than 0.1 atmosphere from that at 
the greatest possible volume. 
Immediately below this line two three phase curves appear in 
the P-7-projection, one of which indicates the coexistence of the 
hydrogen sulphide hydrate by the side of vapour and a liquid rich 
