Some 114 patients with idiopathic Parkinson's disease (not demented, no tremor dominance; mean age: 67.0 years; mean duration of illness: 6.1 years) were subjected to the manual dexterity test "MLS" ("Motorische Leistungs Serie"). A retest was carried out 24 h later with no change in medication. Thus, adequate conditions were present for testing the stability of dimensions of manual dexterity. In addition, data from the initial measurement were normalized (T norms). The correlations of the values from both measurements (retest reliability) underscore the constancy of manual dexterity functions. Subtest show adequate stability of measurement so that they can be used for diagnosing dexterity disturbances in individual patients. Charts listing normal values and critical differences allow intra- and interindividual comparisons. It is thus possible to make statistically sound predictions for individual patients concerning the effectiveness of therapy or progression of illness with respect to manual dexterity. In addition, differences on the left and right sides can be tested for statistic significance.