Reply-Enhanced heritabilities and best linear unbiased predictors through appropriate blocking of progeny trials

Abstract

was deliberately set aside for the purpose of improving breeding efficiency. Such plain, unbalanced, incomplete block designs and mixed models, as described by Ericsson (1997), do not fulfill the prerequisites needed when it comes to using the methods discussed by Williams and Fu (1999). The model used with fixed block effects is incompatible with their method of analysis, since fixed block effects are simply “adjusted for” in the solution of the mixed-model equations before REML (restricted maximum likelihood) iterations as well as the final BLUP (best linear unbiased predictors) computation.