The site dependence of the absorption-promoting actions of laureth-9, Na salicylate, Na2EDTA, and aprotinin was studied in rats. Insulin absorption was estimated on the basis of the cumulative hypoglycemic response from 0 to 4 hr postdose, relative to that after intramuscular insulin. Insulin was administered with or without adjuvants to isolated rectal, nasal, and buccal absorption sites. Laureth-9, a nonionic surfactant which irreversibly removes membrane proteins or lipids, promoted insulin absorption from each site. The rectal, nasal, and buccal routes were 30% as effective as the i.m. route. The enhancing effects of Na salicylate and Na2EDTA, which have reversible mechanisms of permeability enhancement, were specific for rectal absorption. With these adjuvants, rectal insulin was 30-40% as effective as i.m. insulin, but nasal and buccal doses were less than 5% as effective as i.m. doses. This specificity can be at least partly explained by considering the site-to-site differences in membrane histology, although differences in pore size and membrane biochemistry might also contribute. The protease inhibitor aprotinin was ineffective in increasing insulin efficacy via each route, either alone or in combination with laureth-9.