Atlantic salmon S1/2 pre-smolts from the VESO Vikan hatchery were assigned to study groups, i.p. immunized with commercially available, multivalent oil-adjuvanted vaccines with (Norvax Compact 6 - NC-6) or without (Norvax Compact 4 - NC-4) recombinant infectious pancreatic necrosis virus (IPNV) antigen. A control group received saline solution. When ready for sea, the fish were transported to the VESO Vikan experimental laboratory, where two identical tanks were stocked with 75 fish per group before being transferred to 10 degrees C sea water and exposed by bath to first passage IPNV grown in CHSE-214 cells. The third tank containing 40 fish from each group was challenged by the introduction of 116 fish that had received an i.p injection of IPNV-challenge material. The remaining vaccinated fish were transported to the VESO Vikan marine field trial site and placed in two identical pens, each containing approximately 53 000 fish from the NC-6 group and 9000 fish from the NC-4 group. In the experimental bath challenge trial, the cumulative mortality was 75% and 78% in the control groups, and the relative percentage survival (RPS) of the NC-6-immunized fish vs. the reference vaccine groups was 60% and 82%, respectively. In the cohabitation challenge, the control mortality reached 74% and the IPNV-specific vaccine RPS was 72%. In both models, the reference vaccine lacking IPNV antigen gave a moderate but statistically significant non-specific protection. In the field, a natural outbreak of infectious pancreatic necrosis (IPN) occurred after 7 weeks lasting for approximately 3.5 months before problems due to winter ulcers became dominating. During this outbreak, mortality in the NC-4 groups were 33.5% and 31.6%, respectively, whereas mortality in the NC-6 groups were 6.9% and 5.3%, respectively, amounting to 81% IPNV-specific protection. In conclusion, the IPN protection estimates obtained by experimental challenges were consistent between tanks, and were confirmed by the field results.