Author(s): PEDROZO, C. A.; ALCOFORADO, A. T. W.; MAYER, M. M.; SANTOS, R. F. dos; PESSONI, L. A.

Summary: The objective of this study was to estimate the spatial repeatability coefficients of Bertholletia excelsa fruits and nuts using different statistical methods, as well as to define the minimum number of fruits and nuts to be evaluated for the selection of superior genotypes. Twenty-one genotypes in the 2017 crop and 28 genotypes in the 2018 crop were evaluated. From each genotype, 10 fruits and 10 nuts per fruit were sampled. Seven fruit traits and three nut traits were measured, which were used to estimate repeatability coefficients by four statistical methods: 1) analysis of variance; 2) main components based on the correlation matrix; 3) principal components based on the covariance matrix and 4) structural analysis based on the correlation matrix. Estimates of repeatability coefficients varied among crops, traits and estimation methods. In general, the coefficients presented moderate magnitudes, and the methods based on principal components provided the highest estimates. For the most economically important traits for the Brazil nut market, minimum sampling of 8 fruits and 7 nuts per genotype is required to select genotypes with determination coefficients above 85%.

Publication year: 2021

Types of publication: Journal article

Unit: Embrapa Roraima

Keywords: Bertholletia Excelsa, Genetic improvement, Variability