Five categories of evidence of carcinogenicity in rats and mice were used to group interpretative results on 86 chemicals studied in recent carcinogenicity tests carried out by the National Toxicology Program (NTP). Of these studies, 50% (43/86) were regarded as showing carcinogenic effects, 42% (36/86) gave no evidence of carcinogenicity, 6% (5/86) showed equivocal evidence of carcinogenicity, and 2% (2/86) were regarded as inadequate experiments. The liver was the most frequent site of cancer in male and female Fischer-344 rats and in male and female B6C3F1 mice. Male rats appeared more sensitive than female rats to the induction of neoplasia, while for mice the females seemed more responsive. The routes of administration yielding the highest percentage (80-83%) of positive studies were gavage and inhalation; approximately one-third of the feed, drinking water, and dermal studies showed carcinogenic effects. In feeding studies, overall survival in dosed and control groups were similar, while the majority of gavage studies showed significantly reduced survival in one or more dosed groups relative to the corresponding controls. The overall percentage of studies showing carcinogenic effects (50%) agrees closely with the rate reported by other investigators for nearly 200 earlier carcinogenicity experiments conducted by the National Cancer Institute.