The microbial diversity in two deep, confined aquifers, the Grande Ronde (1270 m) and the Priest Rapids (316 m), Hanford Reservation, Washington, USA, was investigated by sampling from artesian wells. These basaltic aquifers were alkaline (pH 8.5 to 10.5) and anaerobic (Eh −200 to −450 mV). The wells were allowed to free-flow until pH and Eh stabilized, then the microflora was sampled with water filtration and flow-through sandtrap methods. Direct microscopic counts showed 7.6 × 105 and 3.6 × 103 bacteria ml−1 in water from the Grande Ronde and Priest Rapids aquifers, respectively. The sand filter method yielded 5.7 × 108 and 1.1 × 105 cells g−1 wet weight of sand. The numbers of bacteria did not decrease as increasing volumes of water were flushed out. The heterotrophic diversity of these bacterial populations was assessed using enrichments for 20 functional groups. These groups were defined by their ability to grow in a matrix of five different electron acceptors (O2, Fe(III), NO3−, SO42−, HCO3−) and four groups of electron donors (fermentation products, monomers, polymers, aromatics) in a mineral salts medium at pH 9.5. Growth was assessed by protein production. Culture media were subsequently analyzed to determine substrate utilization patterns. Substrate utilization patterns proved to be more reliable indicators of the presence of a particular physiological group than was protein production. The sand-trap method obtained a greater diversity of bacteria than did water filtration, presumably by enriching the proportion of normally sessile bacteria relative to planktonic bacteria. Substrate utilization patterns were different for microflora from the two aquifers and corresponded to their different geochemistries. Activities in the filtered water enrichments more closely matched those predicted by aquifer geochemistry than did the sand-trap enrichments. The greatest activities were found in Fe(III)-reducing enrichments from both wells, SO4-reducing enrichments from the Grande Ronde aquifer, and methanogenic enrichments from the Priest Rapids aquifer. Organisms from these aquifers may be useful for high-pH bioremediation applications as well as production of biotechnological products. These organisms may also be useful for modeling potential reactions near buried concrete, as might be found in subsurface waste depositories.