Existing models of mouse mammary carcinogenesis induced by the model polycyclic aromatic hydrocarbon 7,12-dimethylbenz[a]anthracene (DMBA) typically use a small number of bolus doses applied intragastrically. In contrast to this, typical human exposures to carcinogens are thought to be at lower doses and to occur with chronic or sporadic timing. When the classical dosage (1 mg DMBA given once a week for 6 weeks) was split into five daily doses of 200 microg given intragastrically to female SENCAR mice each week for 6 weeks, toxicity was high and the major tumor type seen was lymphoma. Lowering the dose to 60 microg/day gave less toxicity, a 75% incidence of lymphoma and a 30% incidence of mammary carcinoma. However, 20 microg DMBA given five times per week for 6 weeks resulted in a 65-70% incidence of mammary carcinoma within approximately 50 weeks. This represents a 50-fold lower daily dosage of DMBA than that used in the classical model. DNA was prepared from 10 mammary adenocarcinomas and 10 lymphomas and exons 1 and 2 of the H-ras1, K-ras and N-ras genes were sequenced using PCR techniques. Mutations altering codons 12 or 61 of one of the ras family genes were found in 4/10 mammary carcinomas and 5/10 lymphomas. Three mammary tumors exhibited codon 61 mutations, one in each of the genes studied, and a fourth tumor contained a codon 12 mutation in the K-ras gene. Among the lymphomas, two mutations in codon 12 of K-ras, one mutation in codon 61 of K-ras and two mutations in codon 61 of N-ras were also found. Each of the mutations could be interpreted as a G-->T or A-->T transversion. It is suggested that the high incidence of lymphoma at the higher, repetitive doses may be related to immunotoxicity. These low dose models of lymphomagenesis and mammary carcinogenesis should prove useful for tests of chemopreventive agents that target the initiation phase of carcinogenesis.