The nonsteroidal anti-inflammatory drug aspirin and the spice curcumin retard adenoma formation when administered long-term to Apc(Min/+) mice, a model of human familial adenomatous polyposis coli. Both agents interfere with cyclooxygenase activity. When aspirin is administered to Apc(Min/+) mice only postweaning, but not before, it is inefficacious, while curcumin given postweaning is active. Here the hypothesis was tested that dietary aspirin (0.05%) or curcumin (0.2%) prevent or delay adenoma formation in offsprings when administered to Apc(Min/+) mothers and up to the end of weaning, but not afterwards. Whereas curcumin was without effect when administered in this way, aspirin reduced numbers of intestinal adenomas by 21%. When aspirin given up to the end of weaning was combined with curcumin administered from the end of weaning for the rest of the animals' lifetime, intestinal adenoma numbers were reduced by 38%. The combination was not superior to intervention postweaning with curcumin alone. These results show that aspirin exerts chemopreventive activity in the Apc(Min/+) mouse during tumour initiation/early promotion, while curcumin is efficacious when given at a later stage of carcinogenic progression. Thus, the results suggest that in this mouse model aspirin and curcumin act during different 'windows' of neoplastic development.