560 
BULLETIN OF THE BUREAU OF FISHERIES 
its flow must then have slackened (or its temperature have risen), because the surface 
temperature of the critical locality rose to 4.6° by April 28 and to 7.8° on May 29, 
though the whole column of water on German Bank was still only 2.7° and 4.2°, 
respectively, on these dates (ice patrol stations 3, 21, 22, 37, and 38, p. 997). The 
seasonal time-table seems to have been about the same in 1915, when the cold Nova 
Scotian water was responsible for a temperature of about 3° from German Bank 
out across the eastern side of the basin on May 6 to 7 (fig. 27), suggesting that the 
Fig. 29.— Mean air temperature (solid curve) and water temperature (broken curve) for 10-day intervals at Ten Pound 
Island, Gloucester Harbor, Mass., from July 1, 1919, to June 30, 1920 
inrush into the gulf had reached its head some time in late March or April of that 
year. In 1920, however, it is certain that the cold current did not begin to flood 
past Cape Sable into the gulf in any considerable volume until after the middle 
of April. 
Water as cold as 0.27° to 0.56° had, it is true, spread westward past La Have 
Bank to within a few miles of the longitude of Cape Sable as early as the 19th of March, 
