QUANTIFICATION OF THE OPTIMAL NUMBER OF REPLICATES FOR STUDIES OF SORGHUM AGRONOMIC TRAITS

Abstract

Agricultural experimentation is a crucial step in cultivar recommendation; however, it is necessary to determine how many repetitions are needed to achieve the ideal coefficient of determination for selecting genotypes with high precision and lower costs. This study aimed to estimate the repeatability coefficients and the number of evaluations needed to obtain a coefficient of determination equal to 85% for sorghum genotype variables. The trials were conducted in Balsas, Maranhão, during the second growing season of 2021 and 2022, arranged in a randomized complete block design with three replications and ten treatments: Sudan 4202, AGRI 001-E, AGRI 002-E, BRS 658, BRS 810, BRS Ponta Negra, BRS 373, BR 509, BR 506, and IPA 467-4-2. The following were evaluated: plant height, stem diameter, number of leaves, leaf area, flowering, total cycle, green matter production, dry matter production, relative chlorophyll index, and Brix. The repeatability coefficients (r) and minimum number of repetitions (η0) were estimated using analysis of variance; principal component analysis of correlations and variances and covariances; and structural analysis of correlations and covariances, using the GENES program. High values were observed for the repeatability coefficient (r) and coefficient of determination (R²), greater than 0.55 and 78.54%, respectively, with the total cycle variable standing out at 0.99% and 99.65%, respectively, demonstrating high precision in predicting the actual number of individuals from the three replications. Therefore, increasing the number of replications would not result in significant improvements in precision. Experimental trials of sorghum, for the studied cultivars, with three replications, can identify superior genotypes related to all variables studied