It has been suggested that local anesthetics may block sodium conductance through nervous membranes also by hydrophobic interaction, e.g., by expanding the membrane. Decreased anisotropy (fluidization) and depressed phase transition temperatures have been shown by relatively high local anesthetic concentrations. We studied the dose dependence of the effect of three clinically used local anesthetics, with different lipid solubility, on lipid fluidity parameters of four different model membranes. With stearic acid spin labels in dipalmitoyl lecithin vesicles etidocaine (1-5 mM) had the clearest fluidizing effect followed by bupivacaine (5 mM); 2-chloroprocaine was without effect on lipid fluidity. In synaptic plasma membranes a fluidizing effect near the hydrophilic part of the lipid bilayer was similar with etidocaine and bupivacaine (5-10 mM); 2-chloroprocaine had no effect. Bupivacaine at 125 and 250 muM had a small ordering effect, which was not seen at a more hydrophobic site of the membrane. Etidocaine had the strongest fluidizing effect at the latter site of the synaptic plasma membranes. In erythrocyte ghost membranes, probed by stearic acid spin labels near the hydrophilic end, none of local anesthetics affected fluidity at 24 degrees C, while at 37 degrees C etidocaine (1-5 mM) and bupivacaine (5 mM) had a fluidizing effect. Dimyristoyl lecithin vesicles were probed by cis- and trans-parinaric acid. Etidocaine and bupivacaine (5-10 mM) clearly depressed the phase transition temperature evaluated from fluorescence intensity scans. The effect was most marked with bupivacaine (1-10 mM) when cis-parinaric acid was used. While isolated mammalian nerves are blocked by local anesthetic concentrations below 100 muM, this study shows that the clinically used local anesthetics increase fluidity and depress phase transition temperature only at 10-100 times higher concentrations at physiological pH. This kind of hydrophobic membrane interaction may not be important for the nerve blocking effect.