The motor effect of erythromycin was characterized in conscious rabbits chronically fitted with electrodes and strain-guage force transducers implanted along the proximal and distal colon. Fecal pellet output was also evaluated as an index of propulsive activity. In order to get an insight into the pathways involved in mediating the effect of erythromycin, the macrolide was also administered after pretreatment with atropine, nifedipine or ondansetron. Furthermore, in vitro experiments with erythromycin alone and in the presence of atropine, nifedipine, tetrodotoxin or ondansetron were carried out with circular muscle strips taken from rabbit distal colon. In vivo, erythromycin (0.087-5.6 mg/kg i.v. bolus) dose-dependently stimulated spike and mechanical activities at both colonic levels, with a more marked effect on the distal colon. Erythromycin also dose-dependently increased the number of aborally migrating long spike bursts and fecal pellet output. The reproducibility of the response to erythromycin was confirmed by experiments with the dose of 2.8 mg/kg i.v. bolus, repeated in five consecutive experiments at 48-hour intervals. Nifedipine, but not atropine or ondansetron, significantly reduced the colonic motor response to erythromycin. In vitro experiments gave results in line with the in vivo data: the concentration-dependent contractile effect of erythromycin was almost suppressed by nifedipine, but resistant to atropine, tetrodotoxin or ondansetron. In conclusion, this study provides evidence that: (1) erythromycin is a prokinetic drug at the colonic level in rabbits, and (2) both in vivo and in vitro, the effects of erythromycin are exerted at the smooth muscle level by mechanisms depending on influx of extracellular calcium, while muscarinic and 5-HT3 receptors are not involved, at least in this model.