PHYSICAL OCEANOGRAPHY OF THE GULF OF MAINE 693 



(p. 636), the surface chills. At first this chilling chiefly reflects the convectional mix- 

 ing of the upper stratum, by which the substratum is warmed, in proportion as the 

 surface is cooled, combined with the effects of evaporation from the surface. Mean- 

 time the mean temperature of the whole column of water continues to rise slowly at 

 first, then remains stationary for a time as the sun continues to lose strength. At 

 the mouth of Massachusetts Bay, for example, the mean temperature of the upper 

 40 meters was slightly higher on August 31, 1912 (station 10045, about 12°), than it 

 had been on July 10 (station 10002, about 11°), although the surface had cooled from 

 18.3° to 16.1° in the interval. In 1915, too, the mean temperature of the upper 100 

 meters remained virtually unaltered at the mouth of the bay from August 31 to 

 October 1 (about 8° at stations 10306 and 10324), although the surface temperature 

 fell from 16.1° on the first date to 10.3° on the second, and the mean temperature of 

 the upper 40 meters from 11° to 9°. In fact, it is doubtful whether the column of 

 water, as a whole, actually commenced to lose heat at the mouth of the bay before 

 the end of that October (p. 638). In 1916, again, the mean for 80 meters was about 

 1° higher near Cape Cod on October 31 (station 10399, about 7°) than it had been 

 at the mouth of the bay near by on July 19 (station 10341, about 6°), the 80-meter 

 temperature having risen in the meantime from about 3.7° to about 5.8°, though the 

 surface reading had fallen from 16.4° to 10°. 



Thus, the heat received from the sun is sufficient to balance the loss of heat by 

 evaporation and by radiation at night, when the temperature of the air is cooler than 

 the water, imtU the date when the mean temperature of the air falls permanently 

 below that of the water, so to continue through the autumn and winter. Thereafter 

 the upper 100-meter stratum of water constantly loses heat, no longer merely simulat- 

 ing this loss by convectional equaHzation. As this loss of heat is chiefly the result 

 of radiation, out from the water into the air, the efficacy of this process deserves 

 a word. 



Although warm winds, as we have seen, heat the water below them to only a 

 small degree, and slowly, because of the very much higher capacity of the latter for 

 heat, cold winds, on the contrary, chUl the sm'face of any body of water, fresh or 

 salt, very rapidly because dry air is extremely transparent to radiation, especially 

 to the long wave lengths (Abbott, 1911; Harm, 1915). Because of this "diathermacy," 

 and because water is a good radiator," the surface radiates out very large amounts 

 of heat from September on, whenever the air is cooler than the water, dry, and the 

 sky clear of clouds, fog, or mist, very Httle of it being absorbed by the lower stratum 

 of the air. 



The greater the difl'erence in temperature between the air and the water, and 

 the drier the air, the more rapidly does the water lose heat in this way. When the 

 air is damp, or the sky clouded, the radiation from the surface of the sea is inter- 

 cepted by this water vapor, so that the water loses heat slowly under such circum- 

 stances even if the temperature of the air be considerably the lower. It happens, 

 however, that the humidity rules low and the sky usually is clear during the coldest 

 winter weather of New England and of the Maritime Provinces, especially at night. 

 Consequently, other conditions most favor radiation just when the differential 



« Schmidt (1915) found about 83 per cent as much radiation from a water surface as from a black surface. 



