One hundred and twelve patients with proven or suspected serious intra-abdominal infections were randomized for treatment with either ticarcillin/clavulanate (3.1 g every 4 h) or gentamicin (dosage schedule adjusted according to serum concentrations) plus clindamycin (in most cases 900 mg every 8 h). Positive intraperitoneal and/or blood cultures were available in 47 cases. Considering only fully evaluable cases, 15/20 (75%) were clinically cured or improved in the ticarcillin/clavulanate group and 16/25 (64%) in the gentamicin/clindamycin group, the difference not being significant (P greater than 0.05). Bacteraemia was documented in nine of 45 fully evaluable cases. During the course of the study, the serum creatinine concentration increased by more than 5 mg/l in only one patient (ticarcillin/clavulanate group). Two patients (one in each group) had apparent allergic reactions. No other drug-related adverse effects were noted. In-vitro sensitivity testing of anaerobic isolates revealed that, of those tested, 5/38, 1/38 and 2/37 were resistant to ticarcillin, ticarcillin/clavulanate and clindamycin, respectively. Among aerobic Gram-negative isolates, 34/78, 6/78 and 4/78 were resistant to ticarcillin, ticarcillin/clavulanate, and gentamicin, respectively. Among aerobic Gram-positive isolates, 2/32, 2/32, and 5/14 were resistant to ticarcillin, ticarcillin/clavulanate, and clindamycin, respectively. We conclude that ticarcillin/clavulanate is a safe and efficacious preparation for treating serious intra-abdominal infections.