of Edinburgh, Session 1881-82, 
811 
sure, if tlie temperature of the water experimented on he 15° cent, 
to begin with, there is on release a cooling effect of about 0 o, 68 ; 
while if the initial temperature be 0° cent., there is still a cooling 
effect of about 0 o, 15, thus showing that the maximum density point 
of water under 4 tons pressure (600 atmospheres) does not lie be- 
tween 0° and 15° cent., but is driven below the freezing point. 
Line I., however, crosses the axis at about 1 0, 5. That is to say, we 
found that water which, when released from 1 ton pressure (150 
atmospheres), cooled about 0°*13 when its temperature was 15°, did 
not change when its temperature lay between 1° and 2°, and at 
lower temperatures heated slightly on release. 
Some of the discrepancies observed by us on different days may 
probably be explained by convection currents in the vessel of water, 
due to the change of the maximum density point under varying 
pressure, and by changes in the delicacy of the galvanometer. We 
mentioned in our previous notice how we were disturbed by the 
constant change in the magnetic field by the passing of carts, &c., 
on the street. We found that even the motion of the handle of the 
pump affected the galvanometer, and this was indeed the cause (till 
it was detected) of our taking double readings in many of our 
observations. 
Some experiments we performed on salt water of the density of 
sea water, although quite in accordance with the received opinion 
