Precambrian cratonic pelites from the Kaapvaal Craton in southern Africa have similar REE patterns with relative LREE enrichment and absence of significant depletion in HREEs. They have a narrow range of Sm-147/Nd-144 ratios with a mean value of 0.118, which is identical to the mean value of approximate to 450 worldwide fine-grained samples of all ages obtained by isotopic dilution analyses. This value is probably the best estimate for the upper continental crust. The Kaapvaal pelites also have distinct Cr/Th ratios, but overlap in Eu/Eu* ratios, suggesting that variable provenance and sedimentary recycling were important both during and after the Archean. Because the light REE budget is controlled chiefly by granitoids, which mask contributions of mafic-ultramafic components, the relatively uniform Sm/Nd ratios in sediments do not indicate a near-constant composition for the upper continental crust. Most Kaapvaal pelites have negative epsilon Nd(T) values, indicating important contributions of older crustal sources. Overall, there is a slight decrease of epsilon Nd(T) values with decreasing age, but no clear distinction is apparent at the A/P boundary at 2.5 Ga. Almost all of Kaapvaal pelites have T-DM ages greater than their depositional ages but younger than 3.6 Ga, suggesting an absence of rocks older than 3.6 Ga in the Kaapvaal Craton. The debate on growth or no-growth of continents depends much on the choice of parameters in model calculations. The crucial parameters include sediment flux in subduction zones and delaminated lower crust, and the Sm/Nd ratio of continental crust. Unfortunately, the available data are ambiguous in modelling studies. Neodymium isotopic data and Sm/Nd ratios cannot be taken as a robust argument against the no-continental-growth model advocated by R. L. Armstrong (1991).