914 
BULLETIN OF THE BUREAU OF FISHERIES 
coast at increasing depths in 1914, as was also the case in August, 1913 (fig. 146). 
The comparative uniformity of salinity recorded over a wide area in the western 
side of the gulf at the 40-meter level in August, 1914, contrasted with the definitely 
outlined tongue of high salinity in the eastern side, points to the north-flowing side of 
the eddy as much more definite than the south-flowing side. In August, 1913, how- 
ever, the distribution of salinity at 40 meters pointed to a closer approach to 
equality between the two sides of the eddy. The drift is not as clearly shown by 
the 40-meter salinities taken in August and September, 1915 (p. 990), except that 
the differential between higher salinities in the eastern side and lower ones in the 
western side of the gulf calls for some movement of the same anticlockwise sort, not 
being wholly explicable on the basis of upwelling, though assisted by that process 
(p. 768). 
In none of these years (1913, 1914, and 1915) did the 40-meter level show the 
expansion of water of low salinity off Cape Ann that involved the upper 40 meters 
in July, 1912 (Bigelow, 1914, pi. 2; isohaline for 32.6 per mille at 25 fathoms), in a 
definite easterly drift. Thus, the distribution of salinity reflects much more variation, 
from summer to summer, at the 40-meter level in the western side of the gulf than 
in the eastern side, as well as at the surface (p. 770). 
Unfortunately, the 40-meter chart for 1914 (fig. 145) does not so clearly show 
the dominant movement of water in the southwestern part of the area. However, 
isohalines closely crowded outside the 100-meter contour and the fact that they run 
parallel to the latter make it certain that no general drift was taking place trans- 
verse to the edge of the continent at the time, but that any dominant set that was 
then active roughly paralleled the latter. Consequently, the broad zone of 33 to 34 
per mille between it and Nantucket Shoals (much more saline than any part of the 
Gulf of Maine at this level, but less so than the tropic water outside the continental 
edge) did not reflect a direct encroachment of the latter at the time or even any such 
movement earlier in the season, but merely reflected (by its precise salinity) the pro- 
portionate amounts in which water of higher and lower values had mingled there. 
However this may be, the presence of water of this comparatively high salinity to 
the south and southwest of Nantucket Shoals, added to rather an abrupt transition 
to considerably lower values (about 32.8 per mille) on the neighboring parts of Georges 
Bank, is good evidence that the surface drift, which has carried so many bottles out 
of the gulf westward across or around the shoals (p. 881), was not then operative to 
as great a depth as 40 meters, but that it is deflected more to the eastward, as the 
depth increases, by the contour of the bottom. This suggestion is corroborated to 
some extent by the fact that the isohalines for 33 per mille or lower include the 
whole eastern end of Georges Bank on the 40-meter chart in question, with an abrupt 
transition to much higher salinities (34.5 to 35 per mille) off its southeastern slope. 
At first sight the presence of a tongue of water warmer than 10° running 
obliquely across Georges Bank from southwest to northeast at the 40-meter level, 
with lower temperatures within the gulf to the north as well as along the southeastern 
face of the bank (fig. 53), might seem to contradict this, but in this case salinity is 
the more reliable index to circulation, because the high 40-meter temperature at the 
station in question (10224), associated as it was with correspondingly low temper- 
