Atlantic salmon post-smolts were given an intraperitoneal (ip) injection of tissue homogenate of Atlantic salmon fry from an outbreak of infectious pancreatic necrosis (IPN), and cohabitants were given an ip injection of Earle's balanced salt solution (EBSS). Parallel treatment groups were exposed to recurrent episodes of environmental stress by water drainage twice a week. Fish injected with EBSS and non-injected fish were exposed to water drainage. The control fish were left untreated. Mortality due to IPN started 3 weeks after challenge in non-injected and EBSS-injected fish that had been exposed to water drainage. This showed that the fish used in the experiment were covertly infected with IPN virus (IPNV) prior to challenge, although no virus was detected in the fish sampled before the experiment. In fish that received an injection of IPNV, mortality started 5-6 days after challenge, regardless of the presence or absence of stress exposure. The EBSS-injected cohabitants started to die after an additional 5-6 days, also regardless of the presence or absence of stress exposure. The final cumulative mortality in the IPNV-injected fish was significantly lower than in the EBSS-injected cohabitants, thus suggesting that the secondary immune response after injection of IPNV provided more protection than the response after a water-borne infection. No disease outbreak was observed in the control fish.