We investigated transport of viruses through saturated and unsaturated sand columns. Unsaturated experiments were conducted under conditions of uniform saturation and steady state water flow. The water saturation ranged from 1 to 0.5. Bacteriophages MS2 and $phi$X174 were used as surrogates for pathogenic viruses in these studies. Phosphate-buffered solutions with different pH values (7.5, 6.2, 5.5, and 5) were utilized. Virus transport was modeled assuming first-order kinetic adsorption for interactions to the solid-water interface (SWI) and the air-water interface (AWI). Under saturated conditions, virus retention increased as pH decreased, and a one-site kinetic model produced a good fit to the breakthrough curves. Under unsaturated conditions a two-site kinetic model was needed to fit the breakthrough curves satisfactorily. The second site was attributed to the adsorption of phages to the AWI. According to our results, $phi$X174 exhibits a high affinity to the AWI at pH values below 6.6 (the isoelectric point of $phi$X174). Although it is believed that MS2 is more hydrophobic than $phi$X174, MS2 had a lower affinity to the AWI than $phi$X174, presumably because of the lower isoelectric point of MS2, which is equal to 3.9. Under unsaturated conditions, viruses captured within the column could be recovered in the column outflow by resaturating and immediately draining the column. Draining columns under saturated conditions, however, did not result in any recovery of viruses. Therefore the recovery can be attributed to the release of viruses adsorbed to the AWI. Our results suggest that electrostatic interactions of viruses with the AWI are much more important than hydrophobicity.