In this paper, we examine the potential of electromyographic (EMG) and acceleration measurements in discriminating patients with Parkinson's disease (PD) from healthy persons. Two types of muscle contractions are examined: static contractions of biceps brachii muscles and elbow extension movements. Twelve features are extracted from static and ten features from extension measurements. These features describe signal morphology and nonlinear characteristics, power spreading in EMG wavelet scalograms and spectral coherence. Principal component approach is applied separately for static and extension trial to reduce the number of features before discrimination. The discrimination between subjects is done in a two-dimensional space by applying cluster analysis to the best discriminating principal components. The discrimination power of the used method was estimated with EMG and acceleration data measured from 56 patients with PD and 59 healthy controls. In the cluster analysis, three clusters were formed: one cluster with most (85%) of the healthy persons and two clusters with 80% of patients. Patients were divided into two clusters based on their type of motor disability (problems during movement and/or static contraction). Discrimination results show that EMG and acceleration measurements are potential for discriminating patients with PD from healthy persons. Furthermore, they have potential in the objective clinical assessment of PD.