Border effect on the agronomic evaluation of wheat cultivars
Alan Diego do Rosário Chaves, Camila Helena Teixeira, Ricardo Antonio Ruiz Cardozo, José Maurílio Moreira de Figueiredo Júnior, Dercio Antonio Banze, Maria Laura Gonçalves Cardoso de Mesquita, Kátia Regina de Andrade Campos, Giovane Bruno Savioli, Brena Kelly da Silva Almeida, Adilson Júnior Soares Alves, José Maria Vilella Padua, Aurinelza Batista Teixeira Condé, Fábio Aurélio Dias Martins and Jose Airton Rodrigues Nunes
Abstract: The aim of this study was to assess the border effect in wheat agronomic trials in Minas Gerais, Brazil. Experiments were conducted in three environments using 16 cultivars in plots of five rows (two outer rows as the border area). The traits evaluated were plant height (PH), grain yield (GY), hectoliter weight (HW), and thousand-kernel weight (TKW). Data from the useful, border, and total plot areas were analyzed separately and, in a split-plot-in space design using a mixed-model approach. The experimental precision and reliability in selection were similar among the areas. Differences were found among cultivars for all traits, except GY. A border effect was observed only for PH and GY traits. The use of a border area is recommended in wheat trials. However, the total plot area may also be used for evaluation, as the experimental accuracy and reliability of selection are comparable to evaluation using the useful plot area.