During the final months of 1978 and early 1979 large numbers o f small drifting buoys were launched all over the southern oceans in preparation for the First Global Atmospheric Research Program (GARP) Global Experiment. At the time, those of us who had been involved in the preceding years o f planning and development were mainly occupied with immediate questions—Would the Argos satellite tracking system work? Would the buoys survive long enough to form a useful array?—but our planning had included concern over the long-term use of the data for oceanographic research. The buoy system worked better than we expected and returned a major data set on the surface current sand temperatures of the southern ocean. Now, 5 years later, it seems timely to review our experience. I will examine what we learned about the potential of drifting buoys and how our experience may help us to plan future large-scale observing systems of all types.