An experimental model of intraabdominal sepsis closely approximating the disease in humans was used to evaluate the efficacy of clindamycin, tobramycin, cephalothin, and cefamandole, alone and in combination, in reducing both early mortality due to peritonitis and the subsequent development of abscesses. Sepsis was induced in rats by intraperitoneal placement of gelation capsules containing various amounts of human stool inoculum. Ninety percent of the untreated animals given a high inoculum died of peritonitis within 24 hr, and the survivors showed significant intraabdominal abscesses at the time of sacrifice. Animals treated with clindamycin-tobramycin, cephalothin-tobramycin, cefamandole-tobramycin, or cefamandole alone in the high-inoculum group showed a significant decrease in early mortality. Only animals treated with clindamycin-tobramycin, however, had a significantly increased cure rate (survival with no abscess present at time of sacrifice). In low-inoculum groups, several antibiotics alone or in combination significantly decreased mortality and increased cure rates. The selective use of antibiotics is critical in the treatment of severe forms of intraabdominal sepsis.