The surface circulation is investigated during the autumn of 1993 in the vicinity of the Azores Current 18–28°W 31–36°N during the SEMAPHORE experiment. A total of 125 drifters were tracked by satellite, some of which were drogued, some not, some with long tethers and some were mixed layer drifters drogued at 15 m depth. The different types of drifters respond differently to the wind, the wave action and the mixed layer currents, resulting in differences in drifts sometimes exceeding 10 cm s−1. These effects can be partially taken into account statistically. After doing that, the different drift data were combined to map the near-surface quasi-geostrophic flow. The major currents in the area have a width of 100 km or larger, and changes in the shear happen over a shorter scale. The data density is sufficient to resolve scales of 50–100 km for 11 days averaged circulation and sea surface temperature (SST). The maps produced every 5 days portray the surface evolution of the Azores Current and of the eddy field for 2 months. The changes are large in the area where the Azores current interacts with a large anticylonic eddy A1 associated with a Meddy. There are some closed eddy circulation present throughout the period, whereas other structures are relatively short-lived (a month or less). The general features of this analysis are reproduced in analyses of the circulation from altimetric data. The resolution of this analysis is much larger than what was earlier obtained from the hydrological surveys, enabling us to estimate the quasi-geostrophic mesoscale divergence field from the horizontal vorticity budget. The quasi-geostrophic horizontal vorticity advection provides the largest contribution to divergence and contributes to a series of convergence and divergence patches in the order of 1×10−7 s−1 associated with the various ridges and troughs of the Azores Current. Although uncertainty is high in these estimates, this suggests that vertical temperature advection does not contribute much to the surface temperature evolution at the mesoscales of these surveys. Horizontal temperature advection is estimated from the current and surface temperature analysis. Away from the frontal area near 25°W, horizontal temperature advection is significantly non-zero, but does contribute only a small portion of the observed temperature change at the mesoscales. The budgets suggest that the area south of A1 was the site of subduction in October–November.