PHYSICAL OCEANOGBAPHY OF THE GULF OF MAINE 843 
of the Gulf of St. Lawrence are similarly maintained by an inflowing bottom 
current of the same slope origin. 
The motive power that brings water of this character to the Gulf of Maine as 
a bottom current through the Eastern Channel (intermittently, it is true, but regu- 
larly enough to maintain the comparatively constant salinity and temperature 
actually recorded) is to be sought in the distribution of density along the edge of 
the continent. A considerable body of evidence has now been accumulated to the 
effect that the zone of contact along which coast and ocean waters mix, and where 
the slope water is manufactured, averages somewhat more dense (heavier) than the 
water in on the edge of the continent, except right at the surface. All the profiles 
that have been run out across the continental edge off Nova Scotia in summer, 
both those by the Canadian Fisheries Expedition (Sandstrom, 1919, pi. 9, sections 13, 
14, 15, 16, and 17) and by the United States Bureau of Fisheries, have shown 
something of this sort. Thus, on July 25 to 28, 1914, on the first Grampus profile 
out from Shelburne (stations 10231, 10232, and 10233), the stratum between the 
20-meter and 150-meter levels was more dense just outside the edge of the shelf 
than in over the latter, though the surface was less so. 
The Grampus again found the water heavier over the continental slope (station 
10295) than in over the shelf (fig. 168) along this same profile on June 23 and 24, 
1915, with a decidedly steep density gradient at the 50 to 100 meter level. Con- 
sequently, the whole mass of water on the shelf above 100 meters must have had 
a hydrostatic tendency to drift seaward, except immediately at the surface. 
A March profile along this same general line (stations 20073 to 20077) again 
shows higher densities at the outermost station, at 100 to 220 meters, than along 
the edge of the continent (fig. 169)— evidence of this same dynamic tendency for the 
water of low salinity and temperature to move out across the slope, though at the 
inshore end of the profile the dynamic tendency in the superficial stratum was the 
reverse. 
The water at 20 to 120 meters’ depth was likewise somewhat more dense over 
the southeastern slope of Georges Bank (station 10220) than in on the neighboring 
edge of the latter (stations 10221 and 10225) in July, 1914; again in April, 1920 
(stations 20109 to 20111), though our corresponding profile for March, 1920, 
crossed a more complex alternative of heavier and lighter bands there (stations 
20065 to 20069). 
The cross section of the western end of Georges Bank for July 20 and 21, 1914 
(fig. 170), is especially instructive in this connection, being the only one of our 
profiles that has reached water of oceanic salinity (36 per mille). Here, again, the 
upper 50 meters of water proved slightly more dense at the outer end (station 
10218) than over the neighboring edge of the bank (station 10216), resulting in a 
comparatively steep south-north gradient of density, though the relationship was 
just the reverse at a depth of 70 to 140 meters. A slight differential of this 
same order (density higher at the outermost stations than in on the bank) also 
prevailed in this same general region in February and again in May of 1920 
(stations 20045 and 20046 for February; 20128 and 20129 for May); but 
in the cold July of 1916 this seems to have applied only at depths greater than 
