The New South Wales sheep industry has developed a strategic plan to progressively eradicate ovine Johne's disease (JD) from the Australian sheep flock. This initiative relies on a suitable diagnostic strategy to identify both infected flocks and those flocks unlikely to be infected to provide a source of clean replacement sheep. Flock diagnosis of JD is limited by the unreliability of fecal culture, the poor sensitivity of serological tests, and the impracticality of sample collection for histology. We propose to base the flock test on ELISA testing a biased sample of the flock. We have determined the prevalence of JD in selected infected sheep flocks by histological examination of terminal ileum collected at slaughter. Prior to slaughter serum was collected from all sheep and body weight and condition score recorded. A significant but not unexpected finding was that sheep infected with JD are more likely to be found amongst sheep of low condition score within the flock. These results demonstrate that the sensitivty of serological diagnosis of JD in sheep flocks can be improved by deliberately biasing the sample of sheep tested to the tail of the flock. The reliability of this strategy is to be assessed in flocks of known JD status.