Two models of colitis produced in rats that have received significant attention over the past few years are the acetic acid and trinitrobenzene sulfonic acid (TNBS) models. The objective of this study was to quantify and compare the temporal relationship among mucosal permeability, epithelial injury, and inflammation induced by acetic acid, ethanol (vehicle), ethanol plus TNBS (unbuffered, pH 1.0), and ethanol plus TNBS (pH 7.4). Data obtained show that the inflammation induced by these four irritants results from caustic injury to the colonic epithelium and interstitium as measured by the rapid and dramatic increases in mucosal permeability and tissue water content as well as by histological analysis. The injurious nature of TNBS was confirmed in a separate series of studies showing that buffered TNBS (pH 7.4), in the absence of ethanol, is toxic to cultured rat intestinal epithelial cell monolayers. Only after 1-2 days of the initial insult, were signs of classical inflammation observed, including increases in colonic myeloperoxidase activity (neutrophil infiltration) and colon weight as well as hyperemia and mucosal ulcerations. Although ethanol plus TNBS (pH 1.0 or 7.4) tended to produce higher mucosal permeabilities (epithelial cell injury) at 1-2 weeks after the enemas than acetic acid or ethanol groups, only the ethanol plus TNBS (pH 7.4) permeabilities were found to be significantly enhanced. In addition, all four groups showed significant elevations in colonic myeloperoxidase activity and colon weight at 1-2 weeks after enema. It is suggested that these models of colitis are useful to study events that occur at the time of inflammation and repair. However, these models may have significant limitations in understanding events that initiate inflammation of the intestine in human inflammatory bowel disease.