Massive, tabular sulfide bodies commonly occur within sequences of subaqueous volcanic rocks. They are believed to form as chemical sediments when hydrothermal solutions which derived metals from interaction with the volcanic rocks flow out into sea water. Such a water-rock interaction, hydrothermal exhalation model implies that changes in the composition of the hydrothermal solutions result from changing" conditions within the hydrothermal system and that the different metal contents of the resulting chemical sediments are related to different alteration reactions. It is suggested that the rare earth elements (REE), because of their coherent and somewhat predictable geochemical properties, can serve as tracers of the alteration reactions which take place within these hydrothermal systems. Changes in the reactions should be reflected by the REE pattern of the solution which will in turn influence the REE pattern of the resulting deposit.In order to test this hypothesis the REE patterns of massive sulfide deposits and associated rocks from the Bathurst-Newcastle District, New Brunswick, Canada, were determined by instrumental neutron activation analysis (INAA). Interpretation of the REE data was aided by a simple computer model which predicted solution REE patterns resulting from various solution-rock interactions.Chlorite-rich samples have REE distributions and abundances that are very similar to those of shales which suggests a detrital source for their REE. Other chemical sediment samples, except Cu-rich sulfides, have positive europium anomalies. The magnitude of the Eu anomaly and the amount of heavy REE enrichment increase with increasing content of oxide and sulfide minerals. Samples with high contents of Pb sulfides have the highest Eu anomalies and the strongest enrichments of heavy REE and these characteristics do not correlate with the abundance of any other mineral.The observed patterns are best explained by an input of hydrothermal solutions to the site of chemical sedimentation. Because of the characteristic REE patterns and consistent stratigraphic positions of the chemical sediment zones (Cu-rich sulfides, Pb + Zn-rich sulfides, and iron-formation) the following sequence of events apparently took place within the hydrothermal system: 1. Alteration of ferromagnesian minerals and/or volcanic glass released Cu and Fe to the solution (Cu-sulfide zone). 2. Preferential alteration of feldspar by a very concentrated brine enriched the solution in Pb, Zn, Ba, and heavy REE (Pb + Zn-sulfide zone). 3. Continued removal of Fe from the volcanic rocks without alteration of a large amount of the solid phases (iron-formation). Results of the REE study are consistent with the formation of massive sulfide deposits by precipitation from solutions which derived metals from volcanic rocks during water-rock interactions. In fact, the data require that water-rock interactions take place and are incompatible with direct release of magmatic water onto the sea floor. Generation of solutions capable of precipitating Pb-rich massive sulfide deposits requires that phenocrysts of feldspar are present in the volcanic rocks, that the rocks are felsic and that conditions within the hydrothermal system favor alteration of feldspar. A relation should then exist between the metal contents of sulfide deposits and the composition of the host volcanic pile. As expected, deposits in mafic lavas are Cu rich while those in mixed, mafic-felsic piles have Pb + Zn/Cu ratios which increase with the felsic rock content of the volcanic pile.