Evidence from a number of laboratories suggests that membrane proteins may meditate the transport of physiologic fatty acids (FA) across cell membranes. However, studies using lipid membranes indicate that FA are capable of spontaneous flip-flip, raising the possibility that rapid transport through the lipid phase obviates the need for a transport protein. Determining the rate-limiting steps for transport of FA across lipid membranes, therefore, is central to understanding FA transport across cell membranes. The transport of long-chain FA across lipid membranes, from the aqueous compartment on one side of the lipid bilayer to the aqueous phase on the other side, has not been measured previously. In this study, we have used the fluorescent probe ADIFAB to monitor the time course of FA movement from the outer to the inner aqueous compartments and from the lipid membrane to the outer aqueous compartment of lipid vesicles. These two measurements, together with measurements of the lipid:aqueous partition coefficients, allowed the determination of the rate constants for binding (kon), flip-flop (kff), and dissociation (koff) for the transport of long-chain natural FA across lipid vesicles. These rates were determined using large unilamellar vesicles (LUV) of approximately 1000 A diameter, prepared by extrusion and giant unilamellar vesicles (GUV), prepared by detergent dialysis, that are >/=2000 A diameter. The results of these studies for vesicles composed of egg phosphatidylcholine (EPC) and cholesterol reveal kff values that range from 3 to 15 s-1 for LUV and from 0.1 to 1.0 s-1 for GUV, depending upon temperature and FA type. For these same vesicles, dissociation rate constants range from 4 to 40 s-1 for LUV and from 0.3 to 2.5 s-1 for GUV. In all instances, the rate constant for flip-flop is smaller than koff, and because the rate of binding is greater than the rate of transport, we conclude that flip-flop is the rate-limiting step for transport. These results demonstrate that (1) kff and koff are smaller for GUV than for LUV, (2) the rate constants increase with FA type according to oleate (18:1) < palmitate (16:0) < linoleate (18:2), and (3) the barrier for flip-flop has a significant enthalpic component. Comparison of the flip-flop rates determined for GUV with values estimated from previously reported metabolic rates for cardiac myocytes, raises the possibility that flip-flop across the lipid phase alone may not be able to support metabolic requirements.