PHYSICAL OCEANOGBAPHY OF THE GULF OF MAINE 
911 
The summer salinities of 1912 (p. 770) pointed very clearly to a longshore move- 
ment of this sort around the northern and western margins of the gulf, setting west- 
ward along the coast of Maine, southward to Cape Ann, and spreading eastward off 
the cape in a rather definite tongue, outlined (at the surface) by the isohaline for 
31.8 per mille (Bigelow, 1914, pi. 2). It was the presence of this tongue which 
established the direction of flow beyond dispute, because considerably higher salinities 
in Massachusetts Bay to the south of it, as well as offshore, left the coastal belt to 
the northward as its only possible source. 
On the other hand, the salinity of the surface then afforded little evidence of 
river water in the northeastern corner of the gulf, in spite of the proximity of the St. 
John River. This, however, can be explained by the active mixing that takes place 
there, for while the mean salinity of the upper 50 to 60 meters was slightly higher 
(about 32.5 per mille) in the Grand Manan Channel and at its western end that 
August than it had been at the mouth of Massachusetts Bay a month earlier (about 
32.2 per mille), the difference is no greater than can be explained as due to the reg- 
ular seasonal succession (p. 799). A detailed discussion of the salinities for that 
summer, given in an earlier report (Bigelow, 1914, p. 90), leads to the conclusion that 
water of high salinity was being drawn into the eastern side of the gulf while the 
coastwise belt was dominated by a nontidal set alongshore from north and east to 
south and west, with expansions of water of low salinity off Penobscot Bay and off 
Cape Ann suggesting two separate anticlockwise eddies. 
The subsequent summer cruises have expanded this preliminary concept of a 
general circling movement around the northern and western shores of the gulf to the 
domination of the surface over the entire basin by a great anticlockwise eddy, par- 
alleling the land northward along Nova Scotia and swinging westward and then 
southward toward Cape Cod (Bigelow, 1917, p. 340), this being the only assumption 
on which the distribution of surface salinity can be rationalized. 
This, it will be noted, has since been corroborated by the bottle drifts just 
described. A comparison between the recurving tongues of low salinity off Cape Ann 
and off Penobscot Bay, when such phenomena develop there, with the drifts from 
the Mount Desert, Cape Elizabeth, and Cape Ann lines, is especially instructive, 
for we find in such tongues a rational explanation for the tendency of the bottles to 
veer out from the land on successive radii. If, for example, bottles had been put 
out off Mount Desert in the summers of 1912 or of 1913, salinity suggests that the 
majority would have turned southward, abreast of Penobscot Bay, and that few, 
if any, would have stranded along the coast farther west. This actually happened 
in 1923 (p. 902, fig. 183). The tendency for bottles put out near land on the Cape 
Elizabeth and Cape Ann lines of that year to veer offshore from the beginning of 
their drifts would similarly find a reasonable cause in expansions of low salinity out 
toward the basin from the offing of Cape Ann, such as were actually recorded in July, 
1912, and in August, 1914 (p. 763, fig. 136). But the distribution of surface salinity 
in August and September, 1915, when scattered observations outlined a band of low 
salinity of comparatively uniform breadth as paralleling the coast line from Nova 
Scotia to Cape Ann (fig. 137), would be compatible with drifts hugging the shore 
