This paper demonstrates how interactions between host genotype for resistance to an infectious disease and the epidemiology of that disease can have large influences on animal productivity and hence on breeding goals for domestic livestock. This is illustrated for the case of gastro-intestinal parasitism in lambs. A model of the parasite infection was developed to include between-animal variation (genetic, permanent and temporary environmental) for live-weight gain, food intake, larval establishment rate in the host, worm fecundity and worm mortality rate. Achieved live-weight gain was defined as the sum of potential live-weight gain under conditions of no parasite infection, a trait correlated with food intake and growth-rate reduction due to the infection. The reduction in growth-rate was calculated from cumulative larval challenge and cumulative worm mass in the lamb. Genetic parameters were then estimated for the output traits of observed live weight at 6 months of age, growth rate reduction and faecal egg count. Model parameters were chosen so that the output means and heritabilities for faecal egg count and live-weight gain mimicked field data for Scottish Blackface lambs and growth-rate reductions were proportionately 0·25 , on average. The model predicted a weak phenotypic correlation (mean = -0·10) between observed live weight and faecal egg count, the indicator of resistance but a stronger favourable (negative) genetic correlation between these traits (mean = -0·27). The severity, or epidemiology, of the disease greatly influenced the results - the genetic correlation between observed live weight and faecal egg count strengthened from -0·02 to -0·46 as the disease severity changed from mild to severe. Selection for reduced faecal egg count resulted in large correlated increases in live-weight gain, more than twice that predicted by quantitative genetic theory, due to the reductions in growth rate losses as the disease challenge to the animals decreased. Conversely, selection for increased live-weight gain resulted in reductions in faecal egg count close to expectations. This asymmetry of selection response emphasizes the epidemiological benefits obtainable from selection for resistance to infectious chronic diseases - such selection will result in improvements in both animal health and productivity not seen when selection is for improved productivity, alone. Breeding goals should be designed to take account of such effects.