Rats with neurotoxic lesions of the dorsal ascending noradrenergic bundle (DB) were compared with sham-operated (SH) controls on the acquisition, steady state and extinction of response suppression maintained by a classical (conditioned suppression) or an instrumental (discriminated punishment) contingency. DB lesions interfered neither with the acquisition of the reference response of sucrose-rewarded barpressing nor with unconditioned responding to the overhead flashing light subsequently used as a signal of shock. The acquisition of discriminated response suppression was also unaffected by the lesion under both types of contingency. However, once discriminated suppression had stabilized, both the conditioned and the discriminative stimulus used were significantly less effective in maintaining suppression in DB animals than in SH controls provided that low intensity footshock (0.2 mA) was used as the unconditioned stimulus (UCS). Upon increase of UCS intensity (to 0.5 mA) normal suppression was observed in the DB group under both contingencies. Extinction of the classical contingency reinstated the difference between DB and SH performance: DB lesion resulted in significantly faster extinction of fear. In contrast, extinction of the discriminated punishment contingency was unaffected by the lesion, although generalized response suppression dissipated faster in the DB than in the SH animals trained under this condition. Our results offer no support for the reinforcement hypothesis of DB function (normal acquisition of barpressing and of discriminated suppression of barpressing); mixed support (greater initial generalization of suppression in DB animals) and contradiction (more rapid extinction of conditioned suppression in DB animals) for the attentional hypothesis; and weak support (reduced suppression and more rapid extinction of suppression in DB animals, but only within limited experimental parameters) for the anxiety hypothesis of DB function. Hence none of the extant theories of DB function offer a ready explanation of the pattern of results presented here. A simple interpretation which conforms with the sparsity of positive behavioural findings in the literature on DB lesions is that forebrain noradrenaline contributes to the detection and utilization of conditioned stimuli; but that this contribution is critical only for the detection of stimuli with low associative strength.