Whole-rock shale and < 0.5-mum-sized clay separates from a well spanning 3500 m of burial in south-central Texas have been analyzed for there REE, Sm-Nd and Rb-Sr contents. Clay Rb-Sr and Sm-Nd HCl leachate-residue pairs yield ages which decrease with depth, implying that these samples have not experienced a ''punctuated'' diagenetic event. Rb-Sr ages from the most deeply buried clay samples are younger than their depositional ages, but Sm-Nd leachate-residue ages are always equal to or greater than depositional age. Leachate Sr appears to be in isotopic equilibrium with present-day sandstone formation waters, implying that pore fluids readily pass between the two lithologies during, and after, diagenesis. The REE are redistributed and fractionated on a local, mineral-grain scale during burial diagenesis. The HCl-soluble REE reservoirs decrease from 20% to 8% in whole rock and from 35% to 7% in <0.5-mum clay, and become progressively enriched in the middle REE relative to the light and heavy REE. Conversely, the HCl-insoluble (silicate) reservoirs become progressively enriched in the light and heavy REE relative to the middle REE. This fractionation, which is more pronounced in the clay than in the whole rock, appears to be a result of authigenic illite and phosphate formation during burial diagenesis. Although the REE are redistributed on a local scale, there is little evidence that the REE are gained or lost by the murdocks during diagenetic redistribution. Combined Nd and Sr isotopic data show that the whole-rock shale system may be modelled as a three-component system composed of young volcanic-derived and old craton-derived detritus, and a seawater-derived component. These three components are partially homogenized during diagenesis, and as a result the Sr-Nd provenance information originally held in the silicate fraction of whole-rock shale becomes obscured.