S3-3 electric field observations for August 23, 1976, show a single convection cell engulfing the northern polar cap. The flow direction is that for a positive IMF By component. The particle data indicate that nearly half the duskside sunward flow occurs on closed field lines whereas the dawnside flow is entirely on open field lines. We interpret this in terms of an IMF By-induced deformation in the polar cap boundary, where the deformation moves with the convective flow. Thus convection streamlines cross the deformed polar cap boundary, but no flow crosses the boundary because it is carried by the flow. Since southern hemisphere convection is expected to occur with the opposite sense of rotation, we predict that closed field lines will be forced to tilt azimuthally. On the nightside the tilt produces a y component of the magnetic field in the same direction as the IMF for either sign of IMF By. Our interpretation is consistent with observations of a greater y component in the plasma sheet than the tail lobes, which are difficult to understand in the common explanation of IMF penetration. Alternatives to this interpretation are also discussed.