976 
BULLETIN OF THE BUREAU OF FISHERIES 
slope the year round, veering through west to southwest across the basin toward 
the offing of Massachusetts Bay; and though variations in salinity and temperature 
prove this drift intermittent, its stream track seems comparatively constant from 
season to season during its periods of activity. 
The correspondence between the dominant circulation of the gulf, as established 
by direct evidence, and the dynamic gradient is close enough to show that the 
former is essentially dynamic, set in motion by the regional inequalities in density, 
but given its eddylike character by the confining effect of the bottom contour of 
Georges Bank to the south. 
Deflection of the horizontal tidal oscillations by the rotation of the earth simi- 
larly tends to produce an anticlockwise movement around the basin of the gulf, and 
with the effect of the wind consistent with this, the several motive forces are parallel 
in effect. 
The westerly drift of slope water along the slope of the continent is also dynamic 
in source, and available evidence suggests the same motive power for the “Gulf 
Stream ” drift abreast of the gulf. 
TABLES OF TEMPERATURE, SALINITY, AND DENSITY 
Temperature is in degrees Centigrade, salinity in parts per mille, and density is 
at the temperature in situ but without correction for compression. The tables on 
page 977, summarized from Ekman’s (1910) tables 2, 4, and 5, give a close enough 
approximation to the latter for general purposes in depths as small as those of the 
Gulf of Maine. For computations involving the specific volume, Smith’s (1926, p. 
19) simplification of Hesselberg and Sverdrup’s (1915) tables are to be preferred. 
STANDARDS OF ACCURACY 
The old type reversing thermometers used in 1912 and 1913 were accurate only 
to within about ±0.15° C., but with the instruments used subsequently for the 
subsurface readings the probable error in temperature determination is less than 
0.05° C. As the surface readings have often been taken under difficulties and by 
various persons, accuracy is not claimed for them beyond about ±0.3° C. 
All the determinations of salinity, except some for the winter of 1925 (noted 
below under the respective stations), have been by titration. So far as personal 
and instrumental errors are concerned, the results are reliable considerably within 
the requirements of the International Committee for the Exploration of the Sea — 
probably to ±0.03 per mille of salinity. However, as Giral (1926) has recently 
emphasized, regional or seasonal variations in the relative proportions of the various 
solutes in sea water, such as are known to occur, introduce another source of error, 
which makes it unsafe to claim accuracy closer than about 0.05 per mille even for 
waters as nearly uniform in their saline content as the Gulf of Maine probably is. 
The accuracy of the calculated densities depends, of course, on that of the deter- 
minations of temperature and salinity on which they are based; and while errors in 
these two may partially offset each other, they may, on the contrary, be cumulative. 
Allowing as the probable range of error 0.05° and ±0.3 per mille, the probable error 
