Samples of tuff from boreholes drilled into fault zones in the Exploratory Studies Facility (ESF) and relatively unfractured rock of the Cross Drift tunnels, at Yucca Mountain, Nevada, have been analysed by U-series methods. This work is part of a project to verify the finding of fast flow-paths through the tuff to ESF level, indicated by the presence of 'bomb' 36Cl in pore fluids. Secular radioactive equilibrium in the U decay series, (i.e. when the radioactivity ratios 234U/238U, 230Th/234U and 226Ra/230Th all equal 1.00) might be expected if the tuff samples have not experienced radionuclide loss due to rock-water interaction occurring within the last million years. However, most fractured and unfractured samples were found to have a small deficiency of 234U (weighted mean 234U/238U=0.95??0.01) and a small excess of 230Th (weighted mean 230Th/234U 1.10??0.02). The 226Ra/230Th ratios are close to secular equilibrium (weighted mean = 0.94??0.07). These data indicate that 234U has been removed from the rock samples in the last ???350 ka, probably by pore fluids. Within the precision of the measurement, it would appear that 226Ra has not been mobilized and removed from the tuff, although there may be some localised 226Ra redistribution as suggested by a few ratio values that are significantly different from 1.0. Because both fractured and unfractured tuffs show approximately the same deficiency of 234U, this indicates that pore fluids are moving equally through fractured and unfractured rock, More importantly, fractured rock appears not to be a dominant pathway for groundwater flow (otherwise the ratio would be more strongly affected and the Th and Ra isotopic ratios would likely also show disequilibrium). Application of a simple mass-balance model suggests that surface infiltration rate is over an order of magnitude greater than the rate indicated by other infiltration models and that residence time of pore fluids at ESF level is about 400 a. Processes of U sorption, precipitation and re-solution are believed to be occurring and would account for these anomalous results but have not been included in the model. Despite the difficulties, the U-series data suggest that fractured rock, specifically the Sundance and Drill Hole Wash faults, are not preferred flow paths for groundwater flowing through the Topopah Spring tuff and, by implication, rapid-flow, within 50 a, from the surface to the level of the ESF is improbable. ?? 2002 Elsevier Science Ltd. All rights reserved.