1. We report here observations on the effects of muscle contraction history on thresholds for the detection of movements at the elbow joint of human subjects. Detection thresholds were measured in the direction of flexion or extension to movements of the relaxed forearm at a speed of 0.2 deg s-1 with the elbow at 90 deg. 2. As reported previously, thresholds for movements in the direction of extension were lower than in the direction of flexion after a conditioning co-contraction of elbow muscles with the arm flexed by 30 deg from its mid-position (hold-short). After a co-contraction with the arm held extended by 30 deg (hold-long), thresholds were lower in the direction of flexion. 3. Here we have made two additional observations. Thresholds for movements of the passive forearm after a co-contraction at the 90 deg test position (hold-test) were low, both in the direction of flexion and extension. Secondly, when thresholds were measured while subjects were carrying out a co-contraction of forearm muscles (15-20 % maximum voluntary contraction), thresholds were much higher. 4. It is concluded that muscle contraction history is an important factor to consider when making measurements of movement thresholds at the relaxed elbow joint. It is speculated that during an active contraction increases in muscle spindle discharges evoked by fusimotor activity lead to the rise in movement detection threshold.