The effects of sodium salicylate, sodium benzoate and 5-bromo salicylate on nerve fibres of the rabbit cervical vagus and of the frog sciatic nerve 'in vitro' were studied and using the single sucrose gap technique. The application of 5-bromo salicylate in concentrations between 2 X 10(-3) to 5 X 10(-3) M induced a slow fall in spike height accompanied by a small hyperpolarization (0.5--2.0 mV). Higher concentrations (1 X 10(-2)--2 X 10(-2) M) produced depolarization and a rapid block of nervous conduction. Concentrations of 1 X 10(-2)--2 X 10(-2) M of sodium salicylate were needed to induce the same hyperpolarizing and slow blocking effects in non-myelinated nerve fibres. Myelinated fibres of the frog sciatic nerve were also blocked by concentrations greater than 1 X 10(-2) M of 5-bromo salicylate, but showed an increase in the amplitude of the compound action potential and a prolonged falling phase when superfused with concentrations around 5 X 10(-3) M. A frequency-dependent decrease in the action potential amplitude was observed with sodium salicylate and 5-bromo salicylate and was shown to be more easily induced in non-myelinated than in myelinated fibres. 5-Bromo salicylate (2 X 10(-2) M) and 2,4-dinitrophenol (1 X 10(-4) M) reduced the amplitude and the rate constant of decay of the hyperpolarization produced by tetanic stimulation of non-myelinated nerve fibres. Sodium salicylate (2 X 10(-2) M) increased the time constant of decay of the hyperpolarization of the potassium-activated response and slightly decreased its amplitude. Sodium salicylate was 3--5 times less potent than 5-bromo salicylate; while sodium benzoate was practically ineffective in producing the phenomena under study.