A neutralizing monoclonal antibody against infectious hematopoietic necrosis virus (IHNV) was used to select neutralization-resistant mutants from isolates of virus obtained from adult steelhead Oncorhynchus mykiss returning to the Round Butte Hatchery (RB mutants) on the Deschutes River in Oregon, USA, and from rainbow trout (nonanadromous O. mykiss) at a commercial hatchery in the Hagerman Valley of Idaho, USA (193-110 mutants). Two of the mutants, RB-1 and 193-110-4, were significantly (P < 0.001) attenuated compared with parental strains. Vaccination of rainbow trout by waterborne exposure to the mutants conferred solid protection against challenge with wild-type virus. In some trials, fish vaccinated with the RB-1 mutant at 50% tissue culture infectious doses (TCID50) of 1 X 104-1 x 105 TCID50/mL or with the 193-110-4 mutant at 1 x 102-1 x 103 TCID50/mL, held for 14 d, then challenged with the homologous wild-type strain at 1 x 105 TCID50/mL showed relative percent survival of 95-100% (P < 0.005). There was no significant difference (P > 0.05) in protection among fish exposed to the RB-1 vaccine strain at a dose of 1 x 105 TCID50/mL for periods of either 1, 12, or 24 h, held for 14 d, and then challenged with the wild-type RB isolate, although the 1-h exposure seemed to be somewhat less effective. Fish were vaccinated with the RB-1 strain at 1 x 103-1 x 105 TCID50/mL for 24 h then challenged after 1, 7, 14, or 21 d with the wild-type RB isolate. No significant (P > 0.1) protection was observed at 1 d postvaccination, but the relative percent survival increased progressively at each subsequent challenge period, becoming statistically significant by day 7 (P < 0.001) and beyond. These results suggested that resistance to challenge with wild-type virus resulted from development of IHNV-specific immunity and not from viral interference or interferon induction, and they reinforce the potential of an attenuated vaccine to control this important disease.