A temporary but marked postnatal decline in UDP-glucuronosyltransferase activity occurs in homogenates and microsomes from rat liver. The profile of this trough and its time of occurrence (maximal over 13-16 days) are almost identical with the two substrates 2-aminophenol and 1-naphthol, whose rates of glucuronidation differ 10-fold. The trough is greatest with digitonin-activated preparations, least with fresh latent ('native') enzyme and intermediate when the native enzyme is treated with its specific activator UDP-N-acetylglucosamine (UDP-GlcNAc). Less detailed evidence supports similar conclusions with 4-nitrophenol as substrate. The trough is not due to the presence of an inhibitor of the transferase in rat liver at 15 days of age. Over the whole perinatal period, including the time of the trough, the enzyme in homogenates can be activated by UDP-GlcNAc; the microsomal enzyme is activated to a rather lesser degree perinatally, and evidence suggests this may be due to artefacts introduced during tissue fractionation. When the overall process of glucuronidation is studied in snips of intact liver offered high concentrations of the two different phenols, the trough is again evident over the same period as observed with broken cells, and of equal depth for both substrates. The infant rat is therefore probably less able to glucuronidate hepatically these phenols over the suckling or early weaning period than are the adult, late foetus or newborn, and may be especially incompetent at 13-16 days of age.