Despite promising preclinical evidence, dozens of new anti-sepsis agents have failed to demonstrate clinical efficacy. One of the reasons may be that the preclinical trials were conducted using animal models that did not adequately reflect clinical realities. Various kinds of experiments utilized for the development of new agents were carried where bolus or short term continuous infusions of large doses of bacteria or endotoxin were administered intravenously. For the preclinical testing of agents, (1) i.v. bacteria and endotoxin models in which the total challenge dose of adequate bacteria or endotoxin is reduced and/ or the length of administration time is increased, and (2) peritonitis models, e.g., cecal ligation and puncture and peritoneal implantation of bacteria or endotoxin, will become reasonable choices.