Understanding the mechanism that governs the transport of long chain free fatty acids (FFA) across lipid bilayers is critical for understanding transport across cell membranes. Conflicting results have been reported for lipid vesicles; most investigators report that flip-flop occurs within the resolution time of the method (<5 ms) and that dissociation from the membrane is rate limiting, while other studies find that flip-flop is rate limiting and on the order of seconds. We have reinvestigated this problem and find that the methods used in studies reporting rapid flip-flop have not been interpreted correctly. We find that accurate information about transport of FFA across lipid vesicles requires that FFA be delivered to the vesicles as complexes with albumin (BSA). For example, we find that stopped-flow mixing of uncomplexed FFA with small unilamellar vesicles (SUV) containing pyranine yields the very fast influx rates reported previously (>100 s(-1)). However, these influx rates increase linearly with lipid vesicle concentration and can therefore not, as previously interpreted, represent flip-flop. In contrast, measurements of influx rates in SUV and giant unilamellar vesicles performed with oleate-BSA complexes reveal no dependence on vesicle concentration and yield influx rate constants of approximately 4 and approximately 0.5 s(-1), respectively. Rate constants for efflux and dissociation were determined from the transfer of oleate from vesicles to BSA and reveal similar influx and efflux but dissociation rate constants that are approximately 5-10-fold greater. We conclude that flip-flop is rate limiting for transport of FFA across lipid vesicles and slows with an increasing radius of curvature. These results, in contrast to those reporting that flip-flop is extremely fast, indicate that the lipid bilayer portion of biological membranes may present a significant barrier to transport of FFA across cell membranes.