The objectives of this study were: (i) to differentiate corn (Zea mays L.) cultivars in terms of yield and morphological traits and (ii) to verify a potential gain in experimental precision when yield covariance of plant population and/or ear number were used for statistical analysis of cultivar competition experiments. Three experiments (for middle, early-middle and early season maturity groups) were conducted at the Fitotecnia Department Experiment Station of Federal University of Santa Maria, Santa Maria, Brazil, in a non-tilling area over black oat. The experimental design was a complete randomized block with three replications. The experimental unit was two rows with 0.8m within-row and 5.0m length. Analysis of variance and covariance was performed for all evaluated traits and Scott & Knott was used to cluster cultivars in each maturity group. In conclusion, there are cultivar means differences adjusted for the same ear or plant number per plot and the covariance analysis improves experimental precision and reduce least significant differences.
precision; covariance; Scott & Knott; management
