Anticoagulant drugs are known to have an effect on tumour growth. However, the mechanisms by which they act are poorly understood, and have therefore been investigated in this study. Wistar rats were given eight weekly subcutaneous injections of azoxymethane, at a dose of 10 mg kg-1 week-1. Following this they were randomized into two groups: a control group, which received no further treatment, and a warfarin treated group, which received warfarin at 'non-therapeutic' doses in their drinking water, for a further 8 weeks. Pairs of rats from each group were killed at 5-weekly intervals from 10 to 35 weeks after the first azoxymethane injection. At 40 weeks all remaining rats were killed. Samples of colonic mucosa from the descending colon and rectum were taken for scanning electron microscopy. The number of microadenomas per low power field was determined in both groups at each time interval. Tumour incidence and distribution were noted in animals killed at 40 weeks. The median number of microadenomas was significantly lower in warfarin treated animals than in controls at all time intervals. Tumour number was also significantly decreased by warfarin treatment (27 in azoxymethane treated animals, 10 in animals receiving azoxymethane and warfarin, P less than 0.05). The distribution of tumours along the colon was similar to that seen previously, following 12 weeks of azoxymethane. These effects occurred despite the non-concurrent administration of azoxymethane and warfarin.