Twenty helminth-naive Scottish Blackface lambs were given three infections with 10,000 infective larvae of Haemonchus contortus at 8 week intervals. An additional six lambs served as uninfected controls and eight lambs were infectivity controls. The lambs were 7 months old at the start of the infection. Four of the 20 lambs developed severe haemonchosis and were put down during the experiment. The remaining 16 lambs plus uninfected controls were necropsied 8 weeks after the third infection. The mean faecal egg count peaked 6-8 weeks after the first infection, gave a second smaller peak 6-8 weeks after the second infection but produced no peak after the third infection. Mean red blood cell counts fell rapidly during the first infection, then rose gradually during the second and third infections. The mean values suggested that two infections were sufficient to produce effective immunity in the sheep population but they masked considerable individual variation. Eleven animals appeared relatively resistant following the first infection, while two animals were relatively susceptible to even the third infection. The repeatability of mean faecal egg counts or mean red blood cell counts for each animal during the replicate infections were very high, because the rankings of the individual sheep remained remarkably stable. Faecal egg counts were very strongly correlated with red blood cell counts. Multiple regression analysis showed that four factors--faecal egg counts, red blood cell counts, weight and sex--accounted for essentially all of the observed variation in worm burdens among the lambs. Therefore, under these controlled experimental conditions, additional markers appear unnecessary for the detection of resistance status.