Butyrylcholinesterase (BChE) displays hysteretic behavior with certain neutral and charged substrates in the approach to steady state. Previous studies led us to interpret this phenomenon in terms of slow transitions between two enzyme conformers E and E'. This kinetic peculiarity is observed in human, horse and rat BChE. Oscillations that superimpose on the hysteretic lag are observed when benzoylcholine and N-alkyl derivatives of benzoylcholine are used as substrate. Hysteresis of BChE can be modulated by medium parameters (pH, salts, temperature, and pressure). Though mutant enzymes show different hysteretic behavior, so far attempts to provide a molecular mechanism of BChE hysteresis from mutagenesis studies have been unproductive. However, the substrate dependence of the hysteretic induction times, using wild-type BChE and several mutants, allowed us to build a general, mechanistic model for the hysteresis. In this model, substrate can bind to E, E', or both conformers, and ES and/or E'S can be catalytically active. The exact pathway followed depends on both the nature of the substrate and the structure of the BChE mutant under study. We propose that oscillations develop when substrate exists in different, slowly interconvertible, conformational and/or aggregation forms, of which only the minor form is capable of reacting with BChE. In support of this proposal, NMR studies have provided direct evidence for slow equilibria between monomeric and micellar forms of long-chain, alkyl derivatives of benzoyl-(N-substituted) choline. There is no direct evidence that hysteresis plays a role in BChE function(s). However, the "new view" of protein dynamics proposes that proteins are normally in equilibrium between pre-existing, functional and non-functional conformers; and that binding a ligand to the functional form shifts that equilibrium towards the functional conformation. Therefore, a physiological or toxicological relevance for the hysteresis in BChE cannot be ruled out.