of Edinburgh, Session 1878-79. 83 
mum, were sealed up in glass tubes, and so completely protected. 
Those whose stems were bent had to be enclosed in metal cases 
closed with a screw. This form of protection never answered well, 
as it was impossible to screw on the cover so tight that water under 
the great pressures met with at considerable depths would not find 
its way in. In order not to have to abandon the use of thermo- 
meters of the convenient form of Six’s, the device of protecting 
the bulb only was hit upon, and it appears that the first ther- 
mometer of this kind was used by Captain Pullen on board 
H.M.S. Cyclops. The effect of pressure on the stem is quite 
insignificant, and under ordinary circumstances insensible. For, 
in nearly all seas where the surface temperature is over 40° F., 
the temperature of the water diminishes as the depth increases, 
and therefore it is the minimum leg which is used, and the 
effective part of it is that filled with spirit, which may have a 
length of at most three inches. The effect of pressure in diminish- 
ing the volume of a short piece of thermometer tubing must 
certainly be very small, but its actual value can only be determined 
by removing the bulb and taking the piece of the stem between 
the mercury and the neck of the bulb as the bulb of a new ther- 
mometer, and determining experimentally the effect of pressure on 
it. An approximation to the effect may be made by exposing the 
thermometer to various high pressures at known temperatures and 
observing the rise of the maximum index, then removing the bulb 
and calibrating the stem. Knowing, then, the ratio of the volume 
of this part of the minimum leg filled with spirit to the whole 
volume, from the bulb to the maximum index, it may be assumed 
that the compression will be in the same ratio. And this value 
will probably be greater than the real one, for the compression of 
the water produces of itself a certain rise of temperature, and con- 
sequently raises the maximum index. This can, however, be 
estimated either by comparison with a completely protected ther- 
mometer, or by bringing the minimum index also home on the 
mercury before raising the pressure. If, then, there has been a rise 
of temperature caused by compression, there will be a corresponding 
lowering of temperature on relieving the pressure. If the com- 
pression apparatus be allowed to stand, after the pressure is up, 
until it has dissipated the heat evolved by the compression, the 
