A number of nitrosamines that have been studied by administration to rats as solutions in drinking water have been examined by gavage administration of similar doses, for assessment of the role of pharmacokinetics in organ-specific carcinogenesis. Methylnitrosoethylamine was more effective as a liver carcinogen by gavage than in drinking water and gave rise to tumors of the lung and nasal mucosa by the former route, but not the latter. By gavage, methylnitroso-2-oxopropylamine and methylnitroso-2-hydroxypropylamine induced mainly tumors of the esophagus, as they did when given to rats in drinking water, but the potency was greater by gavage. At a higher dose rate (85 micromoles per week) methylnitrosohydroxypropylamine induced a high incidence of mesenchymal tumors of the kidney and lung tumors, in addition to esophageal tumors, but methylnitrosooxopropylamine did not. The tobacco-specific carcinogen NNK induced tumors of the liver, and to a lesser extent, of the lung and nasal mucosa when given by gavage to rats, as it did by other routes of administration. The similarly basic nitrosamine methylnitroso-N, N-dimethylaminoethylamine was equally potent, whether administered by gavage or in drinking water to rats, and gave rise only to tumors of the esophagus. The cyclic nitrosamine nitrosomorpholine was equally effective by gavage and in drinking water, but induced in rats more esophageal tumors by gavage in addition to a high incidence of liver tumors. Its 2-hydroxy derivative, a postulated metabolic intermediate of nitrosodiethanolamine, was a very much weaker carcinogen than either the latter or nitrosomorpholine, and induced low incidences of liver and lung tumors toward the end of the lifespan of the rats.