High-K granitoids in the Archaean Congo craton are products of crustal differentiation by partial remelting of Archaean TTG, and in addition the remelting of charnockites at its northern metacratonic margin, the Ntem complex, in southern Cameroon. At the Sangmelima region of this metacratonized complex, high-K granites mostly occur as small intrusions and as pods, dykes and veins in TTG and charnockites. Zircon U Pb analyses of these high-K granites show an important radiogenic Pb-loss, indicated by highly discordant U-rich fractions. Pb-loss in these smaller granitic intrusions appears to be due to continuous than discrete Pan-African diffusion and leaching, that characterizes larger charnockitic and tonalitic intrusions in the region, or a combination of both. The contributive effect of leaching due to later reactivations during the Phanerozoic cannot be ruled out. These multiple Pb-loss causes, and apparently disproportionate amounts of 206Pb and 207Pb loses, probably explain the irregular clustering of the highly discordant U Pb data. However, an upper intercept age of 2751±32 Ma and a lower intercept age of 300±29 Ma are obtained. Most of the zircons, especially those from high-K granite veins, lost Pb from the entire grain with no domains remaining that reflect the original Pb isotopic composition. It was therefore impossible to obtain meaningful 207Pb/206Pb evaporation ages for such grains. Thus caution should be taken against use of the evaporation method for zircons that experienced severe Pb-loss. However, some whole zircons and zircon domains from larger granite bodies apparently retained their primary Pb isotopic composition and were dated successfully by the evaporation technique. They yielded reproducible 207Pb/206Pb ages between 2717±9 Ma and 2724±3 Ma, with a weighted mean of 2722±2 Ma, interpreted as the minimum estimate of the crystallization age of the Sangmelima high-K granites. Similar Sr and Nd isotope signatures and Nd-model ages (3.04 3.25 Ga) as for the TTG and charnockite protoliths indicate that the high-K event did not add new juvenile material to the Archaean crust of the Congo craton but was the result of crustal differentiation.