15 



and for water to be pumped toward the right side of the current and away from 

 the left. In this manner a topographic trough may develop on the left and a 

 ridge on the right of the current which would locally steepen the pressure gradi- 

 ent across the Gulf Stream and give rise to temporary countercurrents. There 

 is evidence of such effects but it is too early to be certain of this interpretation. 



While these effects may help to maintain a constant pattern of circulation 

 under a wind pattern of variable intensity, there are anomalies in both the wind 

 pattern and the flow pattern. In the equilibrium state it is considered that the 

 climatological mean wind torque supplies the energy needed to overcome the 

 frictional losses accompanying geostrophic motion across the pressure gradi- 

 ents. Frictional losses are thought to be greatest by far in the region of west- 

 ward intensification of the current systems. Since the wind torque is accumu- 

 lated over the whole ocean surface, the braking action in the zone of westward 

 intensification must be very intense. Braking seems to occur not only as a 

 result of water shearing over itself or against the continental boundaries, but 

 by meandering and the shedding of eddies. The nneander patterns in the west- 

 ward intensified flow are well recognized (Iselin and Fuglister, 1948) (Fuglister 

 and Worthington, 1951). The centrifugal accelerations accompanying meander 

 curvature may sometimes amount to more than 15 per cent of the local pressure 

 gradient acceleration (von Arx, 1952). 



There may be other accelerations associated with the displacement 

 meanders or their form drag as they move along the length of the current. Since 

 the water masses on the right and left of the current are of different character, 

 a meander represents a pair of waves moving in phase along both the right and 

 left-hand density interfaces. The left interface separates the water masses 

 having the greatest difference in density, so that waves on this interface, de- 

 scribed theoretically by Haurwitz and Panofsky (1950), may dominate the wave 

 pattern and motion. In order for a meander to grow or to progress it is neces- 

 sary for the wave forms on both interfaces to remain approximately in phase 

 and for the water masses adjoining the current to move about harmoniously as 

 the waves pass. Through a combination of vertical and horizontal motions the 

 water mass on the convex side of the current must withdraw, and that on the 

 concave side must advance as a meander deepens. At the same time the re- 

 verse must occur as the meander progresses. This may give rise to charac- 

 teristic circulation patterns in the adjacent water masses. It seems unlikely 

 that with an inequality in the density differences on either side of the current the 

 rate of progress of the waves along the borders of the adjacent water masses 

 can be so harmonious that there will be no change in either the pressure differ- 

 ence across the current or of the width and depth of the current. Localized 

 convergence and divergence of flow and accelerations of other kinds may influ- 

 ence the details of motion even if the systenn is broadly in geostrophic equilibri- 

 um. That is to say, the equilibrium state is not necessarily a steady one. 



Finally, currents like the Gulf Stream are not driven primarily by local 

 winds. The passage of storm centers across these currents and neighboring 

 water masses may influence the results of direct current measurements near 

 the surface and cause temporary departures from geostrophic motion, probably 

 involving cyclic motion in inertia circles. These add many difficulties to the 



problem of relating field observations to the broad ideas that have been dis- 

 cussed, and suggest both the length of time observations must be carried on to 

 obtain a fair sample and the detail that would be desirable in describing the 



transient state. 



As a step to broaden the scope of observations, the tendency recently 

 has been to measure the properties of the surface layer from ships moving at 



