Examination of the correlations between estimated breeding values and the fitting of alternative models suggested, however, that such correlations were a consequence of other effects in the data, rather than evidence of a true negative genetic relationship