Tide gauge data from Key West to Norfolk were used to identify a monthly signal in sea level that is uncorrelated with local self-trapped process. Time series (1955-1975) of local winds, sea level slope, and river runoff were used in a regression model of sea level to separate a local response and a residual signal. The monthly means of the residual contribution were investigated for their relationship to the seasonal fluctuations of the Gulf Stream. In the Florida Channel, lower sea level is found to correspond to increased flow of the Florida Current. During July and August a marked fall in residual sea level, unrelated to the coastal winds, is found from Key West to Charleston, suggesting that the transport of the Gulf Stream increases both in and north of the Florida Channel during this time. Measured long-term monthly surface currents at Diamond Shoals, Cape Hatteras, which markedly increase in summer to high velocities, tend to substantiate this claim. An additional wintertime low in residual sea level occurs north of the Florida Channel. The wintertime low does not result from steric heating within the upper 100--150 m of water, nor does it, in contrast to the summertime low, appear to coincide with increased northward surface flow. Monthly mean Sverdrup transport was computed across the North Atlantic between 15¿N and 35¿N. When compared at the surface latitude, residual sea level and Sverdrup transport, both interpreted as indices as Gulf Stream transport, generally disagree in phase during summer. However, north of the Florida Channel they are consistent during winter, assuming that lower sea level at this time reflects increased flow in the stream.