ABSTRACT Sandstone reservoirs from the Norwegian continental shelf commonly show varying degrees of postdepositional fractionation of Sm/Nd ratio, apparently due to preferential incorporation of Sm in the crystal lattice of diagenetic apatite. Based on data from leaching experiments, an apatite mineral separate, and petrography, it is proposed that fractionation occurred mainly during closed-system burial diagenesis, as phosphate deposited on the sea floor recrystallized to apatite and incorporated rare earth elements (REE) from the surrounding rock matrix. As a result, the Sm-Nd model age parameter in general contains two components of information: A) the true crustal age of the sediment and B) a secondary component of Sm/Nd fractionation, which can be either negligible or dominant. It is difficult to separate these components and thus correct for component B. Even in data showing no obvious signs of postdepositional fractionation, model ages may be variably disturbed by subordinate amounts of diagenetic alteration. However, the isotope ratio 143Nd/144Nd is expected to be relatively resistant to alteration by secondary Sm/Nd fractionation and should therefore be used rather than model age as the principal correlation parameter in stratigraphic studies. In this way, Nd-isotope stratigraphy can still be used as a valuable and unique tool for studies of provenance and correlation, even in strata showing strong Sm/Nd fractionation effects. Strongly Sm/Nd-fractionated horizons (recognized by bulk-rock Sm/Nd ratios > 0.23) may also be useful for stratigraphic correlation, but it is clear that their occurrence reflects special depositional conditions rather than shifts in sediment provenance.