The performance of sires across flocks and years are likely to vary due to factors such as interactions with the environment, differences in ewe genotypes, errors in recording, non-random mating, and preferential treatment of progeny. Research in other breeds and species has shown that these sire by flock-year interactions typically account for less than 5% of the phenotypic variance. This paper examines the significance of sire by flock-year effects for weight traits in Poll Dorset sheep. The results demonstrate that while the sire by flock-year effects explained between 2 and 4% of the phenotypic variance, they significantly improved the fit of the model and resulted in a direct-maternal genetic correlation closer to zero. However, heritabilities were reduced significantly by up to 50%, indicating that sire by flock-year effects may be removing too much genetic variation for traits with maternal effects. On balance however, it seems advisable to include sire by flock-year interaction effects in the across flock evaluation.