Non-destructive measurement of leaf area (LA) is preferred in growth analysis and plant physiological studies. Many regression-based models have been developed for estimating LA using leaf length (L), leaf width (W), or imaginary rectangle of L x W (LW) as predictor or independent variable. Objective of this study was to develop and validate appropriate regression models for estimating snap bean trifoliate LA using easily measured L, W, or calculated LW. Snap bean used in this research was PV072 cultivar. Trifoliate-leaf samples were purposively collected from different individual plants, to represent wide range of leaf sizes, from the smallest leaf with fully open blade to the largest available leaf. Snap bean trifoliate leaf consists of three leaflets. The sampled leaves were alternately divided into two subgroups, based on length of terminal leaflet, for developing and validating LA estimation models. Linear, quadratic, and power regressions were evaluated for their appropriateness to be used for estimating LA. Intercept (a) was forced to zero to make the models more geometrically realistic. Results of this research indicated that: (1) zero-intercept quadratic and power regression models performed well for length of leaflet (Lt) or width of leaflet (Wt) was used as predictor, whereas zero-intercept linear model was appropriate and geometrically-sound if imaginary rectangular Lt x Wt (LtWt) was used for estimating surface area of both terminal and side leaflets (LtA); (2) for a practical, fast, and accurate estimation of LA, LtWt of terminal leaflet was the recommended option among other single or combination of predictors; and (3) recommended empirical model for LA estimation of snap bean trifoliate leaf is LA = 1.5198 LtWt.
Artikel secara utuh bisa diunduh Artikel Lengkap