Toxicology and carcinogenesis studies of pentachlorophenol (penta), a biocide used primarily as a wood preservative, were conducted by feeding diets containing a technical-grade composite or Dowicide EC-7 (a commercial grade with lower levels of contaminants) to groups of B6C3F1 mice. Based primarily on liver lesions (hepatocellular necrosis, degeneration, and cytomegaly) observed in 6-month studies, diets containing 100 or 200 ppm technical-grade pentachlorophenol or 100, 200, or 600 ppm EC-7 were fed to groups of 50 male and 50 female mice for 2 years. Control groups consisted of 35 animals. For the most part, mean body weights of mice exposed to technical-grade penta were comparable to those of controls. During the second year, the 600-ppm EC-7 female mice averaged 85% of the control body weights. Feed consumption by exposed mice was similar to that by controls. The average daily doses of penta were approximately 0, 17-18, 35, or 114-118 (EC-7) mg/kg. Survival of mice did not appear to be significantly affected by exposure to either technical penta or EC-7 at the doses used in these studies; survival of the control male mice (technical-grade) was comparatively low. Compound-related neoplasms were observed in three organs/systems: liver, adrenal gland medulla, and vascular endothelium. Dose-related increases of hepatocellular adenomas and of carcinomas were observed in male and female mice exposed to both technical penta and EC-7, although the increase was less marked in females exposed to technical penta. Pheochromocytomas of the adrenal gland in exposed male mice were significantly greater than those in controls for both technical penta and EC-7. These neoplasms were also increased in female mice exposed to EC-7 but not to technical penta. Hemangiosarcomas in the spleen and/or liver were increased in female mice that received technical penta and EC-7. The results of these studies show that both technical penta and Dowicide EC-7 are carcinogenic for mice, causing neoplasms in multiple organs/systems. In addition, the results suggest that the carcinogenic responses were due almost exclusively to penta itself, with possibly a minimal potentiating influence by the contaminants in the induction of liver neoplasms in male mice.