Historical data for Leydig cell tumors from untreated or vehicle treated rats from carcinogenicity studies collected in the RITA database are presented. Examples are given for analyses of these data for dependency on variables considered to be of possible influence on the spontaneous incidence of Leydig cell tumors.In the 7453 male rats available for analysis, only one case of a Leydig cell carcinoma was identified. The incidence of Leydig cell adenomas differed markedly between strains. High incidences of close to 100% have been found in F344 rats, while the mean incidence was 4.2% in Sprague Dawley rats and 13.7% in Wistar rats. Incidences in Wistar rats were highly variable, primarily caused by different sources of animals. Mean incidences per breeder varied from 2.8% to 39.9%.Analyses for the dependency on further parameters have been performed in Wistar rats. In breeders G and I, the Leydig cell tumor incidence decreased over the observation period and with increasing mean terminal body weight. The incidence of Leydig cell tumors increased with mean age at necropsy and was higher in studies with dietary admixture compared to gavage studies. These parameters had no effect on Leydig cell tumor incidence in breeders A and B. Animals from almost all breeders had a considerably higher mean age at necropsy when bearing a Leydig cell adenoma than animals without a Leydig cell adenoma. Studies with longitudinal trimming of the testes had a higher incidence than studies with transverse trimming.The observed dependencies and breeder differences are discussed and explanations are given.Consequences for the use of historical control data are outlined. With the retrospective Page 3 of 43 A c c e p t e d M a n u s c r i p t Nolte et al: RITA -The application of historical control data for Leydig cell tumors in rats -3 -analyses presented here we were able to confirm the published features of Leydig cell adenomas and carcinomas. This indicates that the RITA database is a valuable tool for analyses of tumors for their biological features. Furthermore, it demonstrates that the RITA database is highly beneficial for the definition of reliable historical control data for carcinogenicity studies on a scientifically solid basis.