The effect of trimebutine 100 mg i.v. and placebo on colonic myoelectrical activity was investigated in 10 patients with the irritable bowel syndrome (IBS) (5 constipated and 5 diarrhoeic), using an intraluminal probe supporting 8 groups of electrodes. At each site examined from transverse to sigmoid colon, the electromyograms exhibited two kinds of spike bursts: short spike bursts (SSB) localized at one electrode site and appearing rhythmically at 10.3/min, and long spike bursts (LSB), isolated or propagated orally or aborally. Computerized analysis of the duration of each kind of spike burst showed that, as compared to the control, trimebutine 100 mg, selectively inhibited by 43 to 73% the mean duration of LSB activity in the transverse, descending and sigmoid colon, from 0 to 30 min after administration. The inhibitory effect was similar in constipated and diarrhoeic patients. Placebo injection did not significantly affect (p greater than 0.05) the duration of LSB and SSB activity. Variance analysis indicated that the inhibitory effect of trimebutine was significantly greater (p less than 0.05) on LSB activity in the transverse than the descending colon, and that it was absent from the sigmoid colon. The results suggest that trimebutine selectively inhibits the propulsive activity of the proximal two thirds of the colon in IBS patients, and that this effect cannot entirely explain its therapeutic efficacy in IBS.