of Edinburgh, Session 1871 - 72 . 795 
is, therefore, not determined by the mean annual temperature of 
air over this part of the earth’s surface. 
From Forbes’ “Climate of Edinburgh, ”it is seen that the tempera- 
ture there is under the annual mean from the 21st October to the 
26th April. Assuming that this holds good for Ballocli Castle, 
then the mean temperature for the cold half of the year will be, 
from — 
October 
21 to 31, . 
46°-0 
November 
1 to 30, . 
41°-7 
December 
1 to 31, . 
40°-9 
January 
1 to 31, . 
38°‘6 
February 
1 to 28, . 
39°-8 
March 
1 to 31, . 
40°-5 
April 
1 to 26, . 
45°-8 
The mean of these 188 days is therefore 41 0, 4. 
The close approximation of this temperature of 41 0, 4 to 42 o, 0, 
the deep-water temperature of the loch, is such as to suggest that 
it is the mean temperature of the cold half of the year which deter- 
mines the temperature of the lowest stratum of water at the bottom 
of deep lakes , so long as the deep-water temperature does not fall 
below that of the maximum density of the water. As this prin- 
ciple, if established, would be of great importance in many ques- 
tions of physical research, such as the deep-water temperature of 
the Mediterranean Sea, which Dr Carpenter has very accurately 
ascertained, in its connection with the larger question of general 
oceanic circulation, it well deserves further investigation. 
