The rare earth elements (REE) are frequently used as indicators of sedimentary provenance (e. g. McLennan, 1989), and they are strongly associated with the clay mineral component of mudrocks. However, there has never been a comprehensive study of REE in a large number of sedimentary rocks from a range of tectonic settings, critically assessing their use. The present study reviews data for several mudrock suites from various settings; including the Boso Peninsula, Japan; the Nankai Trough, Japan (ODP Leg 131); the Mississippi Fan, Gulf of Mexico (ODP Leg 96); Alai Range, southern Tien Shan, Kyrghyzstan; Point Leamington, Newfoundland; and the Rybachi Peninsula, Northern Russia; to determine how accurately REE content and fractionation patterns correspond with provenance. The provenance was initially determined using major element data and the discriminant function diagrams of Roser and Korsch (1988), supported by previously published interpretations of tectonic setting where available. As expected, sediments derived from mafic igneous rocks are generally less fractionated and have lower total REE concentrations than those derived from rocks with a more intermediate provenance. Some samples do not fit this trend, often due to them having a mixed sediment source. In this case, the provenance signal from major elements is indicative of the larger mineral grains, whereas the REE signal is representative of the clay mineral source. This therefore allows the separation of the provenance of the fine fraction of a sediment with a mixed source, which is not possible using major element techniques. It is necessary to examine these issues in detail before proposing a definitive scheme for determining provenance using REE, but it seems that such a scheme would be useful as a means of interpreting mudrock sequences where a small, infrequent input of course-grained material from a second source may considerably alter the provenance signal recorded by major elements.