Field and laboratory experiments were conducted to investigate inactivation of viruses attached to mineral surfaces. In a natural gradient transport field experiment, bacteriophage PRD1, radiolabeled with ³²P, was injected into a ferric oxyhydroxide-coated sand aquifer with bromide and linear alkylbenzene sulfonates. In a zone of the aquifer contaminated by secondary sewage infiltration, small fractions of infective and ³²P-labeled PRD1 broke through with the bromide tracer, followed by the slow release of 84% of the ³²P activity and only 0.011% of the infective PRD1. In the laboratory experiments, the inactivation of PRD1, labeled with ³⁵S (protein capsid), and MS2, dual radiolabeled with ³⁵S (protein capsid) and ³²P (nucleic acid), was monitored in the presence of groundwater and sediment from the contaminated zone of the field site. Release of infective viruses decreased at a much faster rate than release of the radiolabels, indicating that attached viruses were undergoing surface inactivation. Disparities between ³²P and³⁵S release suggest that the inactivated viruses were released in a disintegrated state. Comparison of estimated solution and surface inactivation rates indicates solution inactivation is ∼3 times as fast as surface inactivation. The actual rate of surface inactivation may be substantially underestimated owing to slow release of inactivated viruses.