Development of linear models for estimation of leaflet area in common bean (Phaseolus vulgaris L.)

Abstract

Plant leaf area is an important physiological trait, and direct, non-destructive methods for estimating leaf area have been shown to be effective while allowing for repeated plant sampling.The objective of this study was to evaluate direct, non-destructive leaflet measurements as predictors of actual leaflet area (LA), to test previously developed models, and to develop genotype-specific linear models for leaflet area estimation in common bean (Phaseolus vulgaris L.). For development of appropriate regression models for leaflet area estimation, four common bean genotypes were evaluated under greenhouse conditions: BAT 477, 'Morales', SER 16, and SER 21. The greenhouse-derived models were evaluated under field conditions. Previously developed models were tested and found to overestimate or underestimate leaflet area. Leaflet measurements included maximum leaflet width (W) and maximum leaflet length (L) and L X W. The measurements with the highest values for the coefficient of determination (R²) were W or L X W for BAT 477, SER 16, and Morales (0.97, 0.95, and 0.95, respectively), and L X W for SER 21 (R² = 0.96). The linear models developed were shown to be effective and robust for predicting leaflet area under both greenhouse and field conditions during both vegetative and reproductive stages of plant development.