A minimum flow of 40 cubic feet per second is required in the lower Bradley River, near Homer, Alaska, from November 2 to April 30 to ensure adequate habitat for salmon incubation. The study that determined this minimum flow did not account for the effects of ice formation on habitat. The limiting factor for determining the minimal acceptable flow limit appears to be stream-water velocity. The minimum short-term flow needed to ensure adequate salmon incubation habitat when ice is present is about 30 cubic feet per second. For long-term flows, 40 cubic feet per second is adequate when ice is present. Long-term minimum discharge needed to ensure adequate incubation habitat--which is based on mean velocity alone--is as follows: 40 cubic feet per second when ice is forming; 35 cubic feet per second for stable and eroding ice conditions; and 30 cubic feet per second for ice-free conditions. The effects of long-term streamflow less than 40 cubic feet per second on fine-sediment deposition and dissolved-oxygen interchange could not be extrapolated from the data. Hydrologic properties and water-quality data were measured in winter only from March 1993 to April 1998 at six transects in the lower Bradley River under three phases of icing: forming, stable, and eroding. Discharge in the lower Bradley River ranged from 33.3 to 73.0 cubic feet per second during all phases of ice formation and ice conditions, which ranged from ice free to 100 percent ice cover. Hydrostatic head was adequate for habitat protection for all ice phases and discharges. Mean stream velocity was adequate for all but one ice-forming episode. Velocity distribution within each transect varied significantly from one sampling period to the next. No relation was found between ice phase, discharge, and wetted perimeter. Intragravel-water temperature was slightly warmer than surface-water temperature. Surface- and intragravel-water dissolved-oxygen levels were adequate for all ice phases and discharges. No apparent relation was found between dissolved-oxygen levels and streamflow or ice conditions. Fine-sediment deposition was greatest at the downstream end of the study reach because of low shear velocities and tide-induced deposition. Dissolved-oxygen interchange was adequate for all discharges and ice conditions. Stranding potential of salmon fry was found to be low throughout the study reach. Minimum flows from the fish-water bypass needed to maintain 40 cubic feet per second in the lower Bradley River are estimated.