A galvanic stimulus (30- to 120-s, 0.3-mA constant current pulse) was used to depolarize the spike-generating region of horizontal and anterior canal afferent neurons. The galvanically induced spike activity from these neurons served as a driving input to the efferent vestibular system in the bullfrog. Efferent-mediated effects were assessed by intracellular recordings of posterior canal afferent spike activity, either ipsilateral or contralateral to the driving stimulus. Ipsilateral to the driving stimulus, efferent-mediated spike rate changes occurred in 62 (39%) of 158 posterior canal afferent neurons. Ipsilateral efferent-mediated effects were overwhelmingly excitatory (92%). Of responding units, 3% were inhibited during stimulus application and 5% showed mixed responses involving 3-20 s of inhibition followed by facilitation. Contralateral to the driving stimulus, efferent-mediated spike rate changes occurred in 18 (23%) of 77 posterior canal afferent neurons. Contralateral efferent-mediated effects were overwhelmingly inhibitory (95%). Only one unit was facilitated during stimulation and no mixed responses to contralateral stimulation were observed. Analysis of the coefficient of variation in interspike intervals (CV) before and during stimulation showed no significant efferent-mediated effects on spike train noise. Comparisons of resting spike rates between units showing efferent-mediated effects and those that did not were in general agreement with previous studies. Responding units had a lower mean spike rate (6.8 +/- 0.70 spikes/s, mean +/- SE) than did nonresponding units (10.7 +/- 0.42 spikes/s, mean +/- SE; P < 0.001; 2-tailed t-test of log-normalized data). Comparison between groups in the regularity of their resting spike rates, as quantified by CV, showed considerable overlap. When responding and nonresponding units with similar resting spike rates were compared, responding units had more irregular resting spike rates than did nonresponding units (P < 0.004; 2-tailed, paired t-test). In most cases (77%) the temporal pattern and general shapes of efferent-mediated responses mirrored the driving input of the galvanically activated afferent neurons. The other 23% of efferent-mediated responses exhibited a marked adaptation of the response. Adapting and nonadapting units were not significantly different in their mean resting spike rates or in the regularity of their resting spike rates.