SAMPLING THE SPATIAL HETEROGENEITY OF THE HONEYCOMB MODEL IN MAIZE AND WHEAT BREEDING TRIAS: ANALYSIS OF SECONDARY DATA COMPARED TO POPULAR CLASSICAL DESIGNS

Abstract

SUMMARYThe systematic rather than random entry arrangement of honeycomb designs (HDs) has been deployed to sample the spatial heterogeneity. This hypothesis was studied in fairly homogeneous populations, assuming that their phenotypic variability stemmed absolutely from spatial rather than genetic heterogeneity. It was based on single plant performance in two separate trials of a maize hybrid and a wheat cultivar reflecting different level of spatial heterogeneity. In general, the HDs counteracted spatial heterogeneity well, particularly when the number of evaluated entries was limited. There was a suggestion that they do well even in a high number of entries with many replicates per entry. Distribution and layout of spatial heterogeneity across the experimental area did not affect the precision of the HDs. Standardized configuration, which ensures implementation of essential principles met in other experimental models such as blocking, replication and nearest-neighbour (NN) adjustment on the same baseline, renders the honeycomb experimental pattern advantageous over the classical experimental designs like the randomized complete block (RCB), the NN method and the lattice model.