Leaf area estimation is an important biometrical observation recorded for evaluating plant growth in field and pot experiments. In this study, conducted in 2009, a leaf area estimation model was developed for aromatic crop clary sage (Salvia sclarea L.), using linear measurements of leaf length (L) and maximum width (W). Leaves from four genotypes of clary sage, collected at different stages, were used to develop the model. The actual leaf area (LA) and leaf dimensions were measured with a Laser Area meter. Different combinations of prediction equations were obtained from L, W, product of LW and dry mass of leaves (DM) to create linear (y = a + bx), quadratic (y = a + bx + cx2), exponential (y = aebx), logarithmic (y = a + bLnx), and power models (y = axb) for each genotype. Data for all four genotypes were pooled and compared with earlier models by graphical procedures and statistical measures viz. Mean Square Error (MSE) and Prediction Sum of Squares (PRESS). A linear model having LW as the independent variables (y = -3.4444 + 0.729 LW) provided the most accurate estimate (R 2 = 0.99, MSE = 50.05, PRESS = 12.51) of clary sage leaf area. Validation of the regression model using the data from another experiment showed that the correlation between measured and predicted values was very high (R 2 = 0.98) with low MSE (107.74) and PRESS (26.96). and R. Kumar, S. Sharma.
Nondestructive approach of modeling leaf area could be useful for plant growth estimation especially when number of available plants is limited and/or experiment demands repeated estimation of leaf area over a time scale. A total of 1,280 leaves were selected randomly from eight different morphotypes of som (Persea bombycina) established at randomized complete block design under recommended cultural regimes in field. Maximum leaf laminar width (B), length (L) and their squares B2, L2; leaf area (LA), and lamina length × width (L×B) were determined over two successive seasons. Leaf parameters were significantly affected by morphotypes; but seasons had nonsignificant impacts on tested features. Therefore, pooled seasonal morphotype means of each parameter were used to establish relationship with LA. L and its square L2 did not provide accurate models for LA predictions. Considerably better models were obtained by using B (y = 2.984 + 7.9664 x, R2 = 0.615, P≥0.001, n = 119) and B2 (y = 12.784+ 0.9604 x, R2 = 0.605, P≥0.001, n = 119) as independent variables. However, maximum accuracy of prediction of LA could be achieved through a simple linear relationship of L×B (y = 8.2203 + 0.4224 x, R2 = 0.843, P≥0.0001, n = 119). The model (LA:L×B) was validated with randomly selected leaf samples (n = 360) of som morphotypes and highly significant (P≤0.001) linear function was found between actual and predicted LAs. Therefore, the last model may consider adequate to predict leaf area of all cultivars of som with sufficient fidelity. and S. Chattopadhyay ... [et al.].