672 BULLETIN OF THE BUREAU OP FISHERIES 



air averaged 10° the colder for 10 days in January, 9° the colder at Boothbay, and 

 it may be more than 20° colder than the water in the western side of the gulf on the 

 coldest days. Thus, on December 21, 1924, when the mean surface temperature of 

 the southern side of Massachussetts Bay was about 4.3° (p. 650), the air temperature 

 was — 18° C. at Boston (p. 650) . As another example I may cite December 17, 1919, 

 when the air temperature was about —21.5° C at Lubec (7° below zero F.), the 

 temperature of the surface water being 0°. 



In the winter of 1919-20 (a cold year) the air temperature averaged about 3.1° 

 colder than the surface at Gloucester from December 2 to March 1 and about 5° 

 colder than the water at Lubec. At Eastport the United States Army Signal Service 

 found the mean water temperature to average about 6.6° warmer than that of the 

 air for the period December to February during the 10 years 1878 to 1887. 



The temperatures collected by Rathbun at lighthouses and lightships do not 

 cover the months of January or February, and his statement (Rathbun, 1887, p. 

 166) that the reason for this omission is "the manifest errors of observation some- 

 times made during extremely cold weather" makes it doubtful how close an approx- 

 imation to the truth is given by his averages for the last half of December. Conse- 

 quently, it is necessary to turn to the observations taken on the Halcyon during 

 December to January, 1920-21, for the relationship between the air and surface tem- 

 peratures for the open gulf in midwinter; nor do these fairly represent its outer 

 waters, all having been taken within 30 to 40 miles of land. 



These Halcyon stations show the air 4.4° colder than the water off Boston Har- 

 bor (station 10488), but averaging about 2.5° colder than the w^ater in the northeast- 

 ern corner of the gulf and precisely the same as the water in the Fundy Deep 

 (station 10499). 



The records for this cruise would have been more fairly representative had it 

 included any severely cold daj's, which it did not, for the obvious reason that when 

 icy northwest gales sweep the gulf oceanographic research from a small ship becomes 

 impossible. Nevertheless, the regional difference just sketched does illustrate the 

 very important fact that the cold winds of winter are most effective as cooling 

 agents close in to the land. 



While no exact data are at hand for Georges Bank in early winter, general 

 report has it that the temperature of the au- is close to that of the water there in 

 December and January, except when cold northwest gales blow out from the land 

 or warm " southerlies " blow from the tropic water outside the edge of the continent. 



From the oceanographic standpoint, the most instructive conclusion to be drawn 

 from the relationship between the temperature of the air and that of the water is 

 that the surface of the gulf follows the aii'in its seasonal changes (p. 699; Bigelow, 

 1915 and 1917). This, of course, is a corollary of its situation to leeward of the 

 continent, with winds blowing from the land out over the sea for a much greater 

 percentage of the time than vice versa, especially in winter. It follows from this, as 

 I have emphasized in earlier publications, that the relation of sea climate to air 

 climate is, on the whole, the reverse here of what applies to northwestern Europe, 

 the surface of the sea responding rapidly in winter to the rigorous air climate. 



How closely the winter temperature of the water of the harbors and bays tributary 

 to the gulf depends on the influence of the land is illustrated by the fact that Gloucester 



