694 
BULLETIN OF THE BUREAU OF FISHERIES 
between sea and air temperature is widest, as it is from November on through 
the winter over the Gulf of Maine (p. 671). 
W ater itself is so opaque to radiation that only the thin surface film that is actu- 
ally in contact with the air loses heat rapidly when the air is the colder of the two, 
for it effectually insulates the deeper strata. Consequently, the rate of radiation 
from water to air depends on the activity of vertical circulation; the more actively 
the water is stirred by tides or waves, and the more constantly the surface layer is 
replaced by water from below, the more rapidly will the column give off its heat to 
the colder air and so cool off with the advance of autumn and winter. 73 For this 
reason it would be reasonable to expect the gulf to reflect the autumnal cooling of 
the air most closely where tidal stirring is most active, and temperatures taken by 
Vachon (1918) in the St. Andrews region in 1916 prove this to be the case. 
The coldest winter winds of the region blow from the land out over the gulf, and 
these cold westerly winds predominate in the western side of the gulf during the three 
winter months (p. 965) . Consequently, the water loses heat most rapidly in the coast- 
wise belt around the western and northern shore of the gulf, over which a fresh sup- 
ply of icy air from the land is constantly passing, as long as the cold winds blow from 
the quadrant between north and west. The wind, in turn,iswarmed by the absorption 
of radiant heat from the surface of the water in its passage over the latter; for although 
the lower stratum of air absorbs but a trifling percentage of this total radiation, its 
capacity for heat is so low that but little heat need be intercepted by it to raise its tem- 
perature considerably. This interception is favored, furthermore, by the increasing 
humidity given the air by the evaporation that is constantly taking place from the 
surface of the water. The result is that by the time the air has traveled a certain 
distance out from the land, its temperature rises so close to that of the water, and the 
air is made so humid, that the sea loses heat by radiation but little faster than it 
gains heat from the sun, even in midwinter. 
In any sea exposed to a rigorous air climate, winter chilling may be expected to 
proceed much more rapidly in inclosed harbors, among the islands, and close in to 
the land generally, than it does only a few miles out at sea. This general rule is 
exemplified in a typical way by the Gulf of Maine, where the stations closest to the 
land have proved considerably the coldest in late autumn, winter, and early spring. 
The thermal history of Massachusetts Bay during the winter of 1924-25 affords a good 
example of this. 
Storm winds also hasten the winter chilling of the water by the stirring action 
exercised by the waves, which may reach down to very considerable depths at this 
season, when the water has little vertical stability. In severe winter storms the 
whole upper stratum, 100 meters thick, may be mixed in this way and a constant 
supply of new water thus brought up to the surface, there to give off its heat to the 
icy air. 
Were vertical stirring not so active in autumn, the immediate surface would 
cool off even more rapidly than it actually does, and the whole coastwise belt of the 
gulf, if not the entire area, would freeze over in winter. At the same time, however, 
the surface film would interpose so effective a barrier to the radiation of heat upward 
73 See Nansen (1912) for an illuminating discussion of the loss of heat from the surface of the Northern Atlantic in winter, and 
on the extent to which this is governed by the freedom of vertical circulation. 
