The incidence and titer distribution of infectious hematopoietic necrosis virus in cavity fluid from spent female sockeye salmon (Oncorhynchus nerka) varied little when fish from a naturally spawning population were sampled three times on alternate days. However, when prespawning female sockeye salmon from a second population were individually tagged, penned, and sampled daily, the incidence and proportion of fish with high virus titer rose over a 6-d period. In 10 instances, consecutive cavity fluid samples from the same fish reverted from virus-positive to virus-negative. We suggest that spent fish should be sampled when accurate and quantitative data on the incidence and level of the virus are required.