We studied the adsorption of human albumin onto synthetic hydroxyapatite, using a radiotracer technique and a special flow cell. Adsorption was studied under various conditions corresponding to different thermodynamic paths. It appears that (i) as is' the usual case, the isotherms obtained within a short time range (a few hours) do not correspond to a true equilibrium situation; (ii) when the adsorption process is followed for longer times, which is necessary at low bulk concentrations, one always reaches the plateau surface adsorption; (ii,) this plateau value is independent of the "history" of the adsorption process and corresponds well to the jamming limit predicted by the random sequential adsorption model; and (iv) surface denaturation, leading to enhanced surface binding and thus decreasing desorption constants, is the important phenomenon that can partly and qualitatively explain our observations. Its time dependence, however, remains to be clarified.The adsorption of proteins from solution onto solid surfaces is a fascinating and complex process that is known to have great biological impact (1). It exhibits both reversible and irreversible aspects, which often seem to be in mutual contradiction. For example, adsorbed proteins can be exchanged with similar or different proteins from the bulk solution (2-5). However, on the time scale of the experiments, only a limited population of adsorbed molecules seems to be "exchangeable." In addition, when a protein solution in contact with a solid phase is suddenly replaced by buffer, generally only a small fraction of the adsorbed molecules are desorbed (3, 5). Moreover, it has progressively been recognized that the irreversible aspects of these processes increase significantly with the mean contact time between the proteins and the adsorbing surface (6, 7). This is explained by interfacial molecular denaturation, leading to stronger interactions between the adsorbate and the surface. The observations of Jennissen and Botzet (8, 9), who found desorption hysteresis in the phosphorylase bc/butylSepharose system, have been interpreted in this way. Another typical consequence of the apparent and partial "irreversibility" of the adsorption/desorption process is that the amount adsorbed is not a unique function of the concentration of the solution in "equilibrium" with the surface; the "history" of the process plays a significant role. When adsorption is performed in one step, with the surface brought directly into contact with a solution of bulk concentration cb, the amount adsorbed is different (often higher) than in the case of a step-by-step experiment (the bulk concentration is then increased stepwise up to cb).The purpose of the present paper is to go one step further in the analysis of this puzzling phenomenon of protein adsorption. Two questions will be addressed in the discussion of our experimental results. (i) Is it still possible to speak about "isotherms" and surface/solution "equilibrium" when dealing with some protein/adsorbent systems? (ii) Does ...