1914] Sumner, et al.: Physical Conditions in San Francisco Bay 43 
It is quite probable that these figures do not represent at all ex- 
actly the average annual temperature cycle. In order to construct 
eurves which would be really typical of the local temperature con- 
ditions, it would be necessary to have data covering a considerable 
number of years. In particular, the temperatures recorded during 
January, 1913, are believed to be abnormally low, since the obser- 
vations were made shortly after an exceptionally severe spell of cold 
weather. 
It must likewise be borne in mind that we are dealing with only 
six periods of observation, separated by intervals of approximately 
two months. The six points which are connected by any one of these 
curves represent the mean temperatures for the respective periods 
of observation, referred to the mean dates of these periods. It is 
hardly probable that the time of highest water temperature for the 
year fell exactly on July 26, nor that of lowest water temperature 
on January 20. Indeed, the maximum and minimum points for this 
year may have been several weeks removed from the dates shown on 
our charts. 
The relations of surface and bottom temperatures during the 
different seasons are of interest. For the first of our periods (Feb- 
ruary 13-27, 1912) the two are practically identical (12°23 and 
12°21 C). During the next three periods of observation the mean 
surface temperature is higher, the difference being greatest in July, 
when the surface water was warmer, on the average, by 0°7 C. At 
the fifth period of our observations (November 25—December 5), an 
approximate equality is again reached, while at the sixth period 
(January 13-28) the relation is reversed, the bottom water being 
warmer by 0°22 C. 
The higher surface temperatures during the summer months are 
doubtless due to the absorption of solar heat by the superficial strata 
of the water, the heat rays failing, in large part, to reach the bottom. 
The lower specific gravity of this warm surface water retards its 
mixture with the layers beneath. 
In winter, when the radiation of heat from the water exceeds the 
absorption of solar rays, it is natural that the superficial strata should 
be cooled more rapidly than the underlying ones. But in this case 
the surface cooling would give rise to convection currents, so that the 
difference between surface and bottom temperatures would not be as 
creat as during the summer months. 
These differences between surface and bottom temperatures are 
