The nigrostriatal degeneration underlying Parkinson's disease (PD) is commonly modeled in experimental animals by injection of the neurotoxin 6-hydroxydopamine (6-OHDA). Although a wide variety of simple behavioural screens exist to assess the impact of such dopamine lesions, more complex tasks that assess multiple parameters of an animal's performance may provide a more sensitive measure of the resulting functional impairment. This study assessed the performance of two unilateral lesion models of PD in a lateralised nose-poking task in the nine-hole box test apparatus. This task assesses the accuracy and speed of movements to either side of a rats' head, as well as a number of errors of performance. Rats with complete unilateral dopamine depletion (induced by injection of 6-OHDA into the medial forebrain bundle (MFB)) attempted fewer trials and committed more procedural errors than controls. They developed a marked ipsilateral responding bias, with a reduced accuracy for contralateral stimuli. They were also slower to react to contralateral stimuli and to complete movements bilaterally. Rats with unilateral nigrostriatal terminal lesions (induced by multiple injections of 6-OHDA in the striatum) developed a similar pattern of deficits, but they were significantly less impaired and spontaneously recovered to pre-operative levels by 4 months post-lesion. This experiment confirms that the lateralised nose-poking task may be a powerful tool for assessment of the nature of deficit and recovery in rats with complete but not partial unilateral dopamine lesions.