Abstract:
Plant breeding experiments require the use of appropriate experimental designs that will efficiently block variation
due to wide heterogeneity nature of tropical soils. The primary objective of this study was to assess the
effectiveness of eight different alpha-lattice designs relative to randomized complete block design for evaluating
108 genotypes of maize under rainforest agro-ecology. The maize genotypes were field-tested using three replications
at two locations. Data were collected on grain yield and other agronomic traits. Data collected were
subjected to analysis of variance (ANOVA) assuming randomized complete block design (RCBD) and eight alphalattice
designs. Pearson's correlation and stepwise multiple regression analyses were used to analyze relationship
among different designs and efficiency of the lattice designs over RCBD was computed. Result showed that all the
alpha lattice designs except 27 4 were effective in evaluating the genotypes for plant height. There was significant
difference (p < 0.001) among genotypes for grain yield only when data were analyzed based on 9 12
alpha lattice design. In addition, results showed that the proportion of variation due to blocking and R-square
values of the model increased with increase in the number of blocks for grain yield. In contrast, coefficient of
variation decreased with increase in the number of blocks. The result showed an increase in efficiency of the alpha
lattice designs as the number of blocks increased. It could then be concluded that the more the number of blocks
within replicate, the proportion of total variation due to blocking increased, the coefficient of variation (CV)
reduced, coefficient of determination (R-square) increased and thus, effectiveness increased. Appropriateness of
designs was trait dependent. The 9 12 alpha lattice design was identified to be the best in the evaluation of
grain yield for the maize genotypes.
