1. The output effects of 214 cells in the magnocellular red nuclei of two rhesus monkeys (Macaca mulatta) were tested with spike-triggered averaging of electromyogram (EMG) activity from six forearm extensor and six flexor muscles. The monkeys performed an alternating wrist movement task (auxotonic paradigm) or generated wrist torque trajectories alternating between flexion and extension (isometric paradigm). 2. Sixty-five cells (30%) were identified as rubromotoneuronal (RM) cells on the basis of their postpike effects on forearm flexor and extensor muscles. Three major types of RM cell output organization were identified: 1) pure facilitation (28 cells), 2) reciprocal (18 cells), and 3) cofacilitation (16 cells). 3. RM cell output showed a strong preference for facilitation of extensor forearm muscles. This preference was reflected in the fact that 69% (43 of 62) of RM cells facilitated extensors exclusively or most strongly; 27% facilitated flexors exclusively or most strongly; and 5% facilitated flexors and extensors equally. Postspike facilitation (PSpF) was observed in 45% of the extensor muscles and 20% of the flexors tested. In contrast, postpike suppression (PSpS) was observed in 3% of the extensors and 7% of the flexors. 4. The mean number of extensors facilitated per RM cell was 3.1 (53% of tested) compared with 2.8 (51% of tested) flexors facilitated per cell. The extensor and flexor PSpS muscle field sizes were both 2.0 (35% of extensors and 36% of flexors tested). The mean number of muscles facilitated by cofacilitation cells was 5.8 (48%) per cell. No clear preference was found for facilitation of particular combinations of synergist muscles. 5. PSpF magnitude was assessed by measuring both the percent change of facilitation or suppression from baseline and the signal-to-noise ratio of effects. The overall average magnitudes of RM PSpF and PSpS were 4.1 +/- 2.0 and 4.0 +/- 2.3% change from baseline, respectively. The average magnitude of PSpF in flexors was not significantly different from that of extensors; neither was there a difference in the average magnitude of PSpS in flexors and extensors. 6. The mean onset latency of RM cell PSpS was greater than PSpF (9.2 +/- 3.0 vs. 5.7 +/- 1.8 ms; P less than or equal to 0.05). This can be attributed to an underlying minimal disynaptic linkage to motoneurons for suppression effects, whereas most PSpFs are probably mediated by underlying monosynaptic connections. The mean onset latency of flexor PSpFs was greater than that of extensors (6.4 +/- 2.3 vs. 5.4 +/- 1.5 ms; P less than or equal to 0.05).(ABSTRACT TRUNCATED AT 400 WORDS)