Utilization of alpha lattice design and trend analysis for controlling the experimental error in Soybean variety trials [electronic resource]

Includes bibliographic reference

Spatial plot to plot variability is a real problem perhaps faces the agronomists and plant breeders in variety trials especially those contain a large number of genotypes. Although, using the replication system by complete block design may partly account for a proportion of this local heterogeneity, a considerable amount of intra-block variability still unaccounted for which may mask the significance of small differences among genotypes means. To hold this undesirable part of variability, the seed yield data of 24 soybean genotypes were analyzed using randomized complete block design RCBD), alpha lattice design and trend analysis. The field experiments were conducted using alpha lattice design with three replications at Sakha Agricultural Research Station, Kafr El-Sheikh Governorate, during the two successive seasons of 2014 and 2015. Four statistical criteria being Coefficient of Variation (CV %), Relative Efficiency (RE%), Type I and Type II errors were used to investigate the validity and usefulness of alpha lattice design and trend analysis over RCBD in accounting for the spatial variability. Also, to identify the effect of the adjustments by the two proposed models on the rank orders of the estimated genotype means, Pearson and Spearman rank correlation coefficients were computed among these means. Results showed that alpha lattice design and trend analysis were more precise and effective in reducing the experimental error mean squares compared to RCBD indicating their ability to detect the significance of small differences among genotypes means. The superiority of alpha lattice design and trend analysis over RCBD was clear in both seasons due to the lower values of each of CV%, Type I and Type II errors beside the high values of RE%. There was inconsistency in the rank orders of the genotype means resulted from alpha lattice design and trend analysis compared to
RCBD. This result might be expected due to the different mathematical background of the three used models in removing plot to plot heterogeneity. Methods of analyses, it was observed that the two genotypes; Giza111 and H6L48 produced the highest seed yield that ranged from 2.09 to 2.36 and from 2.07 to 2.34 (ton/fed), in the two growing seasons, respectively.
Finally, it could be concluded that alpha lattice design and trend analysis appeared to be effective diagnostic and remedial tools to account for intra-site heterogeneity especially when the pattern of this variation.

Summary in Arabic