628 
Proceedings of the Boyal Society 
pair. However, any change in the electric field was at once felt. 
The rooms we worked in were on the ground flat looking into 
Chambers Street, and every passing vehicle, on account of its iron 
fittings, disturbed us. The reader of the galvanometer scale, who 
was in a darkened room by himself, could generally tell the direction 
in which a vehicle was moving from the motion of the image on 
the scale, and on once crying out that a man walking past affected 
the galvanometer, the experimenter at the pump immediately 
rejoined that he was carrying some iron on his hack. To avoid this 
change of the magnetic field, we took one set of observations at 
midnight. 
As it was generally impossible to determine directly with a ther- 
mometer the temperature of the junction inside the gun, to enable 
us to interpret the galvanometer readings there was an extra thermo- 
electric circuit, as nearly as possible the same as that connected with 
the gun, whose junctions were in vessels of water, so that the 
temperatures could be directly read with a thermometer and altered 
at pleasure. Any change in the sensitiveness of the galvanometer 
was each day detected by this arrangement. During the principal 
experiments we found it necessary to have three observers, as it was 
necessary to read the gauge, thermometer, and galvanometer, and let 
off pressure simultaneously. We think it sufficient in the meantime 
to indicate briefly the results of a few of the principal experiments 
made. For example, when the internal junction was at a tempera- 
ture of 8 0, 3, for a sudden decrease of pressure from 300 atmospheres 
to 1, there resulted a decrease of temperature = 0°T 7. As the 
pressure increased, so did the rate of cooling with pressure ; and at 
a temperature of 9 0, 3 for a sudden change of pressure from 750 
atmospheres to 1, there was a cooling = 0*°7 4. We next operated 
with the internal junction at a temperature less than 4° C. At a 
temperature of about 1 0, 3 for pressures less than 180 atmospheres 
there was heating instead of cooling for sudden decrease of pressure, 
which meant that e was - ; at the above pressure and temperature e 
apparently vanished, which meant that at a pressure of 1 80 atmo- 
spheres the maximum density point was about 1 ° ‘3 ; at greater 
pressures e was + , which meant that the maximum density point 
was still further lowered. 
The experiments we made at midnight corroborate these results. 
