This study was undertaken to compare the ability of a series of compounds from different chemical classes to induce lung tumors in strain A/J mice after either ip or po administration. 3-Methylcholanthrene, benzo(a)pyrene, urethan, diethylnitrosamine, ethylnitrosourea, and dimethylhydrazine induced a significant (p less than 0.05; t test) increase in the lung tumor response when given both ip and po. 2,4-Dinitrotoluene, 2,6-dinitrotoluene, and a 2:1 mixture of 2,4-dinitrotoluene and 2,6-dinitrotoluene were inactive by both routes of administration and at all dose levels. The lung tumor response to all doses of 3-methylcholanthrene and benzo(a)pyrene, the highest dose of diethylnitrosamine, and the middle doses of both ethylnitrosourea and dimethylhydrazine varied as a function of the route of administration. This finding was most evident for the polycyclic hydrocarbons, e.g., the average number of lung tumors per mouse in animals that received the middle dose of 3-methylcholanthrene or the highest dose of benzo(a)pyrene by the ip route exceeded that by the po route by factors of 12 and 13, respectively. Tissue distribution and elimination studies were conducted in an effort to determine the basis for the observed difference in lung tumor response to 3-methylcholanthrene after ip or po administration. The data indicated that 3-methylcholanthrene persists for longer periods in the animals when given ip, thus potentially providing an extended carcinogenic stimulus. Extrapulmonary lesions observed at a higher than normal frequency at necropsy included peritoneal sarcomas (in 3-methylcholanthrene-treated mice), and both squamous cell carcinomas of the forestomach and abnormal lesions of the liver (in diethylnitrosamine-treated mice).