Some remarks on the biometrical evaluation of unorthogonal series of field trials.
1984
Rundfeldt H.
Selection in plant breeding is firstly done by means of field trials. In these trials genotype-environment interactions are always present, therefore breeders have to lay out series of field trials. In practice the trials in these series may be different in several respects i.e. size of plots number of replications, choice of varieties, error mean squares, treatment mean squares and part of interaction between treatments, locations and years. These differences lead to the fact that simple unweighted general means can't give the best informations. Therefore it becomes necessary to calculate weighted means. In the paper the influence of the different respects of non-orthogonality on the formula for the computation ot most efficient general means is discussed. The conclusions are that the use of such means leads to a somewhat higher probability for the selection of good lines out of the breeding material. This is equivalent to lowering the costs of breeding.
显示更多 [+] 显示较少 [-]