Extensive animal studies are carried out during the development of new medicines to assess toxicity and predict their safety for use in man. There are, however, differences of opinion concerning the nature of safety evaluation studies required prior to marketing. Any prospect of rationalizing the number and design of animal studies must stem from a reappraisal of conventional animal testing procedures and better use of available data on the toxicity of compounds previously investigated. The Centre for Medicines Research has therefore established a toxicology databank containing comprehensive data from repeated-dose animal safety evaluation studies provided by pharmaceutical companies. Thirteen companies within the United Kingdom have provided toxicological data for 32 pharmaceutical compounds studied in the rat, dog, or primate, resulting in 45 case studies for which both short-term (less than or equal to 6 months) and long-term (greater than 6 months) tests had been completed. A comprehensive analysis of these studies has been carried out to determine what new findings, if any, become apparent in studies after 6 months. The results do not support the need for animal toxicity studies of longer than 6 months duration, apart from those designed to investigate carcinogenicity.