Local anaesthetics of the amide type were studied in a modified rat infraorbital nerve block model, with which it was possible to determine varying degrees of sensory block. Of the agents investigated, 0.5% bupivacaine tended to give a longer duration of block than 2% prilocaine or 2% lidocaine, while 0.5% etidocaine had the shortest duration. The duration of prilocaine was prolonged by addition of adrenaline, 5 micrograms/ml, more than that of the other agents. Addition of dextrans of Mw 40-110 X 10(3) did not cause any prolongation of block induced by bupivacaine. When mixed with dextrans over a wide range of Mw (40-4900 X 10(3), prilocaine exhibited significant prolongations of its action by up to 200%. The extent of prolongation was dependent on the degree of block, the concentration of dextrans in the local anaesthetic solution, and the Mw of the dextran although in a less uniform way. An increase in the relative viscosity of the solutions might be a factor of importance for the prolonging effect of addition of dextran to local anaesthetics. Since a formulation providing analgesia of a long duration would be of clinical value, further studies on combinations of the comparatively low-toxicity agent prilocaine and macromolecular substances are of interest.