The Musoshi copper deposit is a typical ore shale-type deposit, located centrally in the Zaire-Zambian copper belt. Cu-Fe sulfides are disseminated in a 20-m-thick laminated black siltstone unit, the ore shale, which occurs above a thick ( approximately 1 km) sequence of oxidized arkosic sediments of the Roan Group, at the base of the Katangan sedimentary basin.Uraninite occurs in veins which cut the ore shale and extend from major hydrothermal alteration zones in the footwall. Rutile accompanying this veining was dated, by Richards et al. (1988), at 514 + or - 2 m.y. (2Sigma ). In the present study, nine analyses of five uraninite samples give a U-Pb age of 514 + or - 3 m.y. (2Sigma ), in exact agreement with the rutile age.Information concerning the age and source of the copper mineralization at Musoshi has been sought by study of Pb isotopes in the Cu-Fe sulfides and other minerals. Leach-residue analyses of most ore samples show poor correlations on 207 Pb/ 204 Pb vs. 206 Pb/ 204 Pb diagrams, and scatter in a band. These samples typically display mature (recrystallized) textures and may contain thin concordant sulfide veinlets, with coarse secondary biotite, quartz, and carbonate. However, analyses of three samples from near the center of the ore shale, which show less evidence for recrystallization and contain no veinlets, form a wellcorrelated line (MSWD = 2.54; 15 analyses) corresponding to a two-stage model age of 645 + or - 15 m.y. (2Sigma ). Data from the other samples scatter below this line on the Pb-Pb diagram, and toward the secondary isochron obtained from the hydrothermal rutiles. Similar phenomena are observed on a 208 Pb/ 204 Pb vs. 206 Pb/ 204 Pb diagram, where addition of U to the disturbed samples is clearly demonstrated. We conclude that the samples record disturbance of Pb isotopes in a simple two-stage system by the late uranium-depositing hydrothermal event at 514 m.y.The significance of the 645-m.y. fractionation event is not clear, however. At least two interpretations may be given: that the event reflects primary Cu-Fe sulfide ore deposition, or that Pb isotopes were homogenized at this time in a preexisting orebody by some process which also introduced U and Th into the deposit. We are not able at present to distinguish between these two alternatives, but whichever is correct, it is clear that copper ore deposition occurred at or before 645 m.y. and is thus unrelated to the later 514-m.y. hydrothermal event.A high integrated U/Pb source ratio (mu = 10.2) obtained from the 645-m.y. sulfide isochron suggests an evolved crustal source for Pb in the orebody and rules out a significant contribution from mafic igneous rocks. If Cu and Pb are assumed to share a common source, these results do not support derivation of the metals from mantle-derived materials at Musoshi. This U/Pb ratio contrasts strongly with a value of mu = 9.7 calculated from the rutile analyses for the late hydrothermal overprint and further confirms that this later event was not involved in the formation of the copper deposit, other than in coarsening and locally concentrating the ore. The high-temperature (397 degrees + or - 5 degrees C) fluids associated with this event were probably derived from deep levels in the crust and appear to have been expelled during a late pulse of activity at the close of the Lufilian orogeny.