|Ramirez, Victor - GRAD STUDENT, UNIV P.R.|
|Harmsen, Eric - UNIV. P.R., MAYAGUEZ|
Submitted to: Journal of Agriculture of the University of Puerto Rico
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: May 5, 2008
Publication Date: May 1, 2009
Citation: Ramirez, V.H., Porch, T.G., Harmsen, E.W. 2008. Development of linear models for estimation of leaflet area in common bean (Phaseolus vulgaris L.). J. Agric. Univ. P.R. 92(3-4):171-182. Interpretive Summary: The primary function of leaves in plants is the generation of nutrition through photosynthesis. The area that leaves occupy is directly correlated with the amount of sunlight they are able to capture. This study developed linear models for estimating leaf area in common bean varieties using a ruler to determine simple width and length measurements of leaves. Leaf area estimates from these simple measurements were found to show a high correlation with actual leaf area as determined using digital image analysis. This technique is advantageous because it is fast, simple, and accurate and because it does not damage the plant, thus multiple measurements can be taken at different developmental stages. Four common bean genotypes were evaluated, BAT 477, ‘Morales’, SER 16, and SER 21, under greenhouse conditions and two genotypes, ‘Morales’ and SER 16, were selected for verification of the model under field conditions. The study found that a single measurement, leaf width, was sufficient for predicting leaf area in common bean, but that different linear models were necessary for each variety.
Technical Abstract: Plant leaf area is an important physiological trait used in plant physiology, soil and crop science, and agroclimatology studies. Direct, non-destructive methods for predicting leaf area have been shown to be correlated with actual leaf area and also allow for repeated plant sampling. The objective of this work was to evaluate direct, non-destructive leaf measurements as predictors of actual leaf area and to develop genotype-specific linear models for leaf area estimation in common bean (Phaseolus vulgaris L.). Four common bean genotypes were evaluated, BAT 477, ‘Morales’, SER 16, and SER 21, under greenhouse conditions for model development. Two genotypes, ‘Morales’ and SER 16, were selected for model verification under field conditions. Leaf measurements included maximum leaf width (W) and maximum leaf length (L), which were used to calculate a third variable, LxW. The highest correlation determined using the coefficient of determination (R2) value, although not significantly different, was observed between leaf area and a single predictor, leaf width (R2 > 0.94 in all genotypes). Leaf width was most highly correlated with leaf area for BAT 477, SER 16, and ‘Morales’ (0.97, 0.95, and 0.95, respectively), however, LxW was found to be most highly correlated with leaf area for SER 21 (R2=0.96) and LxW was equal to W for ‘Morales’ and BAT 477. The linear models developed were shown to be effective and robust for predicting leaf area under both greenhouse and field conditions (‘Morales’ and SER 16) and during vegetative and reproductive stages of plant development.