Arrangement of check plots in augmented block designs when spatial analysis is used

With 5 figures and 6 tables Abstract Early generation plant breeding trials are often laid out in unreplicated designs with replicated checks. This study compares the systematic arrangement of check plots to an allocation, which has two or more different arrangements of check plots within blocks. Four uniformity trials were overlaid with augmented designs, which differed in block size, number and type of checks. Another aim was to compare different spatial models and a baseline model, which did not consider spatial trend. Model fit was assessed using the average empirical variance (EMP) for each arrangement. For three of the four uniformity trials the lowest EMP was found for the arrangement of check plots with different defined distributions of checks within blocks. The systematic arrangement was superior only in one uniformity trial. We propose therefore to allocate checks with different arrangements within blocks. The comparison of the spatial models showed that the linear variance model, the AR(1) model, and the spherical model, all assuming a nugget, were the most promising candidates. A further gain in precision could be detected by a two‐dimensional analysis.