Intracellular recordings were made from rabbit nodose neurones in vitro. Two temporally distinct spike after-hyperpolarizations (a.h.p.s) were identified in a subpopulation of C-type neurones. The fast a.h.p. after a single spike lasted no longer than 500 ms, while the slow a.h.p. persisted for seconds. Both a.h.p.s. were increased in amplitude in low K+ (0.56 mM) solutions and decreased in amplitude in high K+ (11.2 mM) solutions, and both were reversed at hyperpolarized membrane potentials. The slow a.h.p. was reduced in low Ca2+ (0.22 mM), in the presence of Ca2+ antagonists (Ni2+, 1 mM; Cd2+, 100 microM; or Co2+, 1 mM) and was enhanced in tetraethylammonium (5 mM). In approximately half of the cells tested, the fast a.h.p. was reduced in low Ca2+ and in the presence of the Ca2+ antagonists. In the remaining cells the fast a.h.p. was insensitive to these procedures. Prostaglandin (PGE1, 1-10 micrograms/ml) reduced the slow a.h.p. in all cells tested. Neither the Ca2+-sensitive nor the Ca2+-insensitive fast a.h.p. was affected by the prostaglandin. It is concluded that there is a subpopulation of C-type nodose neurones possessing a slow a.h.p. which is due to a Ca2+-dependent K+ current. This subpopulation of neurones can further be divided on the basis of the presence of a Ca2+-sensitive fast a.h.p. Furthermore, PGE1 pharmacologically separates the fast and slow a.h.p.s by selectively blocking the slow one. The blockage by the PGE1 is most probably not due to a reduction in Ca2+ influx.