A study was conducted to examine the effect of using Henderson`s (1988) `average numerator relationship matrix` (ANRM) on the prediction error variance (PEV) of breeding values obtained using an animal model with maternal effects. Genetic trends with and without using ANRM were also estimated and compared. The results suggest that, in situations where multiple-sire joining is practised, the use of ANRM in mixed model equations will result in lower PEVs than when an ordinary numerator relationship matrix, treating sires as unknown, is used. Very little bias in genetic trend estimates was observed when ANRM was used.