This experiment examined the effect of destroying central noradrenergic neurones, using the selective neurotoxin DSP4 (N-(2-chloroethyl)-N-ethyl-2-bromobenzylamine) on the acquisition and performance of discrimination between two time intervals. Rats that had received systemic treatment with DSP4 and vehicle-treated control rats were trained in a series of discrete trials to press lever A following a 2-s presentation of a light stimulus and lever B following an 8-s presentation of the same stimulus. Both groups acquired the discrimination (> 90% correct choices) within 15 sessions; however, the DSP4-treated group showed significantly slower acquisition than the control group. When stable performance had been attained, 'probe' trials were introduced in which the light was presented for intermediate durations. Both groups showed sigmoid functions relating percent choice of lever B to log stimulus duration. Neither the bisection point (duration corresponding to 50% choice of lever B) nor the Weber fraction differed significantly between the DSP4-treated and control groups. The levels of noradrenaline were markedly reduced in the neocortex and hippocampus of the DSP4-treated group, but the levels of dopamine and 5-hydroxytryptamine were not altered. The results indicate that noradrenaline depletion induced by DSP4 retarded the acquisition of temporal discrimination, but did not impair steady-state discriminative precision.