The free drug hypothesis, which states that only the unbound moiety of drug in blood is available for tissue diffusion, is discussed according to recent investigations. In some experimental conditions, it must be assumed that part of the protein-bound drug in plasma is extracted during a single passage through the organ studied. The mechanisms underlying these observations are not unequivocal and remain hypothetical. In the liver, high-affinity binding sites for serum albumin have been demonstrated, and they would explain the high extraction by liver of endogenous and exogenous compounds. However, these experiments measure the unidirectional transfer of a drug from the vascular to the extravascular space in non-steady-state conditions. Hence, in steady-state conditions, the free drug hypothesis cannot be ruled out because it is supported by numerous pharmacokinetic studies.