With the purpose of evaluating the accuracy of three methods to estimate area of individual coffee leaves a field experiment was carried out at Agricultural Institute (IAC) at Campinas, São Paulo State, Brazil, where 50 coffee (cultivar Mundo Novo) leaves were randomized collected. At three methods of measurement (SIARCS 3.0, Barros and leaves discs methods) were compared with the standard one (LI-COR) using simple regression analysis (Yi = betaXi + epsiloni). The three methods were evaluated using the following criteria: a) slope of the linear regression model (beta); b) determination coefficient (R²); c) mean relative error (EMR), and d) dispersion pattern of residuals. The uncertainty associated to R² and EMR was assessed using the bootstrap procedure. The SIARCS 3.0 method was the most accurate one, being useful to estimate leaf area (absolute values), or to calibrate expeditious methods, instead the LI-COR method. The Barros and the leaf disc methods showed a tendency of underestimating and overestimating the leaf area values, respectively.
Coffea arabica L.; leaf area; methodology; accuracy
