1. The behaviour of motor units in the m. biceps brachii (long head), in the m. brachialis and in the m. supinator during slow isometric contraction and relaxation was studied when subjects were performing different motor tasks. These tasks were: flexion of the elbow joint, supination of the forearm and exorotation of the humerus. Motor unit activity was recorded by means of bipolar fine wire electrodes. In the long head of the biceps, motor unit activity was recorded at medial, central and lateral sites.2. When the subject relaxed from flexion, the firing rate of motor units located in the biceps and the brachialis was always found to be lower than that at the corresponding level of flexion force during contraction. The firing rate during relaxation decreased slowly and almost linearly with force. However, during relaxation from supination or exorotation, the firing rate of motor units at medial and central locations in the biceps was more or less constant until decruitment. The firing rate of motor units of the supinator during relaxation from supination decreased slowly and was lower than during contraction.3. Motor units located medially and centrally in the biceps had decruitment thresholds for flexion that were lower than their recruitment thresholds. Motor units on the lateral side of the biceps did not show such a difference. In the brachialis decruitment thresholds for flexion were usually higher than the recruitment thresholds. Differences between decruitment and recruitment thresholds for motor units in the biceps were much more pronounced for supination and exorotation than for flexion. For motor units in the supinator the decruitment threshold during relaxation from supination was higher than the recruitment threshold. 4. The time that had passed after the onset of firing of a motor unit did not influence its decruitment threshold. If, after complete relaxation, the exerted force was increased again, it appeared that the recruitment threshold was changed. It took about 4 s to reach the original recruitment threshold.5. It is concluded that the relation between the firing rate of a motor unit and totalThe authors' names are in alphabetical order. * To whom correspondence should be addressed. J. J. DENIER VAN DER GON AND OTHERS exerted force depends on the phase of contraction. This relation varies within a muscle and between muscles. Furthermore, the results indicate an interchange of activity within the motoneurone pools of the synergists involved in isometrical motor tasks.