Field evaluation of type 2 modified augmented designs for non-replicated yield trials in the early stages of a wheat breeding program.

At the early stages (F 7 , F 8 ) of a wheat breeding program, a large number of test lines need to be evaluated for grain yield, but with limited seed supply per line. These lines are tested in non-replicated multi-location experiments. In this study the use of a modified augmented design for rectangular plots (type 2) was evaluated for grain yield in four multilocation experiments with each the same set of wheat lines, in total 10 experiments. Soil heterogeneity was investigated by ANOVA and existed for all experiments except one. Yield adjustment was based on either row-column effects (Method 1) or covariance (Method 3). Among 10 experiments, seven showed a good agreement between adjustment method as indicated by ANOVA and RE of adjustment. In all trials important changes in the ranking of the test lines did not occur. The MAD-2 provides a convenient and flexible means of measuring environmental heterogeneity, allowing yield adjustment and assessment of test lines for selection purposes. In case the RE does not support the choice for adjustment method based on ANOVA, further research of the experimental design, results and heterogeneity pattern is required.