This study is concerned with explicit memory in both auditory and visual modalities in patients with non-demented(on DSM-III-R) Parkinson's disease. On some explicit memory studies, Parkinsonian patients were compared with normal controls matched for age and education. For assessment of recollection, recall and recognition were assessed using two clinical test batteries, Rey's Auditory-Verbal Learning Test and Rey's Visual-Design Learning Test. In addition to a comparison of recall and recognition, the present research inquired into the serial position data in free recall, analysis by applying a signal detection theory to the recognition data, and metamemory by using self-assessment of recognition. The results showed that the Parkinsonian group was significantly impaired on both tests of free recall compared to the normal controls. By contrast, when given tests of recognition memory for the same lists, their performance was almost identical only in assessment of correct scores(hits). There was a significant correlations between performances on achieved categories of the Wisconsin Card Sorting Test and on free recalls of the auditory-verbal learning test in the patient group. In recall, no qualitative differences of the serial position curves were observed between the two groups, as an increasing pattern of primacy and recency effect was preserved. In addition, the two groups performed equally well on both auditory (digit span) and visual(spatial span) short-term memory assessment. Moreover, on the trial-recall curves, from the first trial to the last two groups showed no significant differences in their learning effect and forgetting. Irrespective of modalities, however, the Parkinsonian group recalled less than the controls in the first trials. The poor performance of recall in the patients could be explained in terms of diminished attentional resources of the central executive system processing information beyond their short-term memory span within the framework of Baddeley's model of working memory which simultaneously controls two functions of encoding and retention. It was suggested that diminished central resources would be strongly associated with basal frontal dysfunction in Parkinson's disease. When recognition memory was further investigated, false alarms but not misses significantly increased despite normal hits in the patients on both tests. Analysis by using a signal detection theory produced significant differences, in that the Parkinsonian patients had lower values of beta and d' on auditory-verbal learning while they had lower tendency of beta on visual-design. The trend that they fell into errors by recognizing a distracter as the target presented previously was likely to be related a dysfunction of the supervisory system of Norman and Shallice's model which was assumed to be in the prefrontal cortex. Moreover, the present research disclosed that metamemory, i.e. the ability to assess their own memory, was disturbed in the Parkinsonian patients. In addition to the problem of recognition mentioned above, it was suggested that the impaired metamemory could be attributed to a deterioration of frontal lobe function.