In this study we present further evidence supporting the reciprocal nature of output effects on forearm flexor and extensor muscles from single corticomotoneuronal (CM) cells. Spike-triggered averaging of rectified EMG activity was used to test the output effects of 105 motor cortex cells in two rhesus monkeys (Macaca mulatta) trained to perform alternating wrist movements and power grip. The electromyographic (EMG) activity was recorded from six forearm flexor and six forearm extensor muscles through pairs of percutaneously inserted intramuscular stainless steel wires. CM cells were identified by their characteristic postspike facilitation (PSF) in spike-triggered averages of agonist muscle EMG activity. Agonist muscles are those which coactivate with the cortical cell during movement. Of 105 motor cortex cells tested, 56 (53%) had no effect on either agonist or antagonist muscles. Of 49 cells that produced PSF of the agonist muscles, 14 (29%) also produced clear postspike suppression (PSS) of the antagonist muscles. Reproducibility of postspike effects was demonstrated by comparing spike-triggered averages of full-wave rectified EMG with averages of the same EMG activity triggered from randomly generated pulses. Consecutive averages from random triggers never showed consistent postspike effects. As a further test that our postspike effects were real, we computed averages of simulated EMG activity from the spikes of CM cells with reciprocal output effects. None of these averages showed consistent postspike effects. The mean onset latency of PSF calculated from 14 reciprocal CM cells yielding 51 PSF effects was 6.3 ms compared with 10.1 ms for 28 PSS effects from the same cells. PSS effects from a particular CM cell were nearly always longer in latency than the cell's PSF effects; only 2 of 28 PSS onset latencies were shorter than the longest latency PSF onset from the same cell. Average peak latencies for PSF and PSS were 8.6 and 11.6 ms, respectively. The magnitude of postspike effects was expressed as the percent of peak facilitation above the base-line mean for PSF or peak suppression below the base-line mean for PSS. With this measure, the average magnitude of PSF was 7.0% compared with 4.1% for reciprocal PSS. There was no correlation between onset latency and magnitude of PSF or PSS, although strong PSFs tended to have shorter latencies. Concerning the distribution of postspike effects, the average reciprocal CM cell facilitated 3.8 agonist muscles and suppressed 2.1 antagonist muscles. EDC was facilitated by all extension-related reciprocal CM cells.(ABSTRACT TRUNCATED AT 400 WORDS)