The objective of this work was to estimate the coefficient of repeatability for production of fruits and solid albumen in coconut palm (cocos nucifera) using statistical methods of the variance analysis (ANOVA), main components (covariance and correlation) and structural analysis (covariance). Significant variability was detected between hybrids for fruits and solid albumen productions. Repeatability coeficients obtained by the variance analysis and structural analysis (correlation) showed the smallest values. If the level of 90% is considered sufficient to select hybrids with relative superiority for fruit and solid albumen productions based on the estimation of repeatability by the main components method (covariance), then it would be recommended to perform five and three evaluations for fruit and solid albumen productions, respectively.
Cocos nucifera; selection; breeding methods; quantitative genetics
