Brief tetanic stimulation of the preganglionic nerve induced a persistent potentiation of nicotinic synaptic transmission in the rat superior cervical sympathetic ganglion. Quantitative measurements of the post-tetanic increase in synaptic efficacy revealed two distinct time courses. The early, rapidly decaying component, termed post-tetanic potentiation (p.t.p.), had a decay time constant of 2-3 min, as reported elsewhere. The duration of the more persistent component, called long-term potentiation (l.t.p.), was extremely temperature dependent, lasting much longer at 32 degrees C than at 22 degrees C. In half of the experiments performed at 32 degrees C, l.t.p. showed no detectable decay over the course of 1 h or more after a brief tetanic stimulation. Other experiments were conducted at 22 degrees C. The induction of l.t.p. was dependent on the extracellular [Ca2+]. Transient elevation of the extracellular [K+] also produced a long-term enhancement of synaptic efficacy, and this effect was also Ca2+ dependent. The tetani that were effective in inducing l.t.p. (5-20 Hz for 5-20 s) were well within the physiological range of preganglionic activity. The magnitude and time course were related to frequency and duration of stimulation. The occurrence of l.t.p. was restricted to those preganglionic fibres that were tetanically stimulated. This lack of heterosynaptic or generalized effects was demonstrated by splitting the preganglionic nerve into two branches that could be independently tested and conditioned. Physiological activation of muscarinic or nicotinic receptors apparently does not play an essential role in causing ganglionic l.t.p., which is expressed as an enhancement of nicotinic transmission. A muscarinic antagonist (2 microM-atropine) did not block l.t.p. Preganglionic stimulation induced l.t.p. even when a high concentration of a nicotinic antagonist (3 mM-hexamethonium) was present during the tetanic stimulation. Furthermore, bath application of a cholinergic agonist (100-1000 microM-carbachol) could not substitute for tetanic stimulation in provoking l.t.p. Activation of adrenergic receptors also appeared not to play an essential role. Neither a beta-adrenergic antagonist (10 microM-sotolol or 1 microM-propranolol) nor an alpha-adrenergic antagonist (1 microM-phentolamine) had any significant effect on the magnitude or duration of l.t.p. The results indicate that ganglionic l.t.p. is a Ca2+- and temperature-dependent process that can be created independently of the activation of nicotinic, muscarinic or adrenergic receptors.