50 
Fysik, Band II, No. 11, Uppsala, 1905) that the influence of a coast 
line upon currents set up by the wind is to produce a general move- 
ment of the water along in the direction of the coastal trend. The 
bulk of the current—i. e., the mid-water current—flows more or less 
parallel to the coast line, but as we approach the shore the mid-water 
current disappears and the surface and bottom systems merge into 
one with a consequent loss of character. This last-mentioned move- 
ment also tends to flow parallel to the shore line. A wind of given 
strength will produce a maximum effect, states Ekman, when di- 
rected 13° to the left of the coast line, which conditions, it is inter- 
esting to note, accord closely with relative directions of winter wind 
and coast line in the northwestern North Atlantic region. Figure 
24, page 49, illustrates this particular phenomenon. The long curved 
arrows represent the average direction of the wind during the Decem- 
ber-March period, and the short double arrows indicate the general 
movement of the surface layers. Offshore, where the depth to the 
bottom is relatively great, the wind current at the surface will be 
deflected nearly 45° to the right of the wind, and the mean transport 
of the water layer of the wind current will most nearly approach a 
direction 90° to the right of the wind. As we near the shore and 
shallow water the direction of the wind and movement of the surface 
water tend to become parallel. The gradient current, which is the 
most voluminous of the movements, arises whenever the sea surface, 
which has been deformed into a state of obliquity by the wind, 
tends to return to the level. A regular state of motion is soon estab- 
lished with a steady flow perpendicular to the pressure gradient, and 
we are permitted to draw a diagram showing the position of impelling 
and Ferrelian forces, AE and AC, respectively, with resultant direc- 
tion of the gradient current as line AB. The time required for such 
a development must be calculated in weeks or months depending 
upon the magnitude of the wind and the width of the current. 
Gradient currents are more or less independent of slight variations 
in the winds such as affect surface currents, but it is quite probable 
that variations in the circulation of the atmosphere are followed by 
corresponding variations in the gradient currents. As an example 
of such a phenomenon as described in the foregoing we point to the 
Labrador Current, which may be due to the melting of polar ice, 
but which, nevertheless, is controlled to a great degree by a system 
of seasonal winds, December—March, tending to aid the transport 
of cold water and ice out of the Arctic along the western side of the 
North Atlantic basin. 
O 
