It was discovered by chance that mice raised under otherwise entirely conventional conditions of husbandry but fed upon autoclaved diet (diet A) had stronger cell-mediated immune reactions than those of mice raised under the same conditions but with an unmodified diet (diet B) : skin allografts were rejected more quickly, transplantation tolerance was more difficult to procure and fibrosarcomas induced by the injection of methylcholanthrene (MCA) arose more slowly and less often. Analysis showed that these findings could be explained at least in part by the discovery of Mertin & Hunt (1976, p. 928) that a partial deprivation of polyinsaturated fatty acids led to an intensification of cell-mediated immunity; on the other hand, experiments with dietary mixtures made it seem unlikely that this was the whole explanation and pointed towards some positive immunopotentiation by an ingredient of autoclaved diet. This, it was proposed, might be a compound of unknown composition resulting from the interaction of vitamin A with other dietary constituents. This interpretation was not supported by direct evidence but by confirming that retinol derivatives, especially retinyl acetate, could exercise an immunopotentiation of the kind and degree under investigation: retinyl acetate could counteract the immunosuppressive action of linoleic acid, though retinyl methyl ether was ineffective. Although retinyl derivatives may protect against MCA tumours by impeding its metabolic conversion to an oncogenic form, the effects of an autoclaved diet upon skin allograft survival, the induction of tolerance and the formation of tumours is probably mediated through an immunological mechanism.