164 TRANSACTIONS LIVERPOOL BIOLOGICAL SOCIETY. 
at ten metres is slightly colder than that at top or bottom. 
In May, at Station VII, the intermediate water is slightly 
cooler than that at top or bottom, while at Station VIII 
it is shightly warmer. The most striking case, however, 
is that of Station V in July, when the water at 30 metres 
is considerably colder than that at any other depth. Such 
variations are not at all surprising when one considers the 
way in which changes of temperature at the surface affect 
the temperature of the water below. 
A rise of temperature at the surface will slowly be 
transmitted by conduction to the water at the bottom, the 
time it takes to reach the bottom depending upon the 
depth. <A fall of temperature will, as a rule, be trans- 
mitted more quickly to the bottom, especially when, as 
in the Irish Sea, the water has almost the same salinity 
from top to bottom, for if the temperature of a layer of 
water falls its density increases and, if 1t becomes greater 
than that of the warmer water below it, it will smk. This 
process will continue with other layers, and so the cooling 
of the water proceeds from the top chiefly by means of 
convection currents, and far more rapidly than it would 
by mere conduction of heat. It is for this reason that 
one hardly ever finds that the temperature rises when 
evoing from the surface to the bottom, but that it nearly 
always falls. 
Now, the changes of temperature require time to 
reach the bottom, so it is easy to understand that before 
the effect of a spell of warm weather has had time to get 
to the bottom the effect of a succeeding spell of cold 
weather may have begun to travel downwards, or vice 
versa. It is in this way that the above-mentioned inter- 
mediate warmer or cooler layers are to be explained. On 
referring to the tables it will be seen that in some 
cases these intermediate layers are clearly in unstable 
