MATHEMATICAL MODEL FOR PREDICTING LEAF AREA OF OCIMUM GRATISSIMUM (HAFENDAHL FW) USING LINEAR MEASUREMENTS

Abstract

An experiment was conducted to develop a mathematical model for predicting leaf area for Ocimum gratissimum using linear measurements. A total of 300 leaves, representing five various leaf sizes, were randomly selected from the field over a period of three months. The leaf sizes and number of leaves collected per size were as follow: very small (0.5cm width) and 75 leaves; small (1.2cm width) and 70 leaves; medium (2.6cm width) and 60 leaves; large (4.5cm width) and 50 leaves; very large leaves (6.5cm width) and 50 leaves. The maximum lamina length (L) and lamina width (W) of the leaf samples were measured with a well-graduated meter rule and the selected leaves were also traced on a standard graph paper. The square, sum and product of the L and W were calculated and recorded as the leaf area estimates while the number of squares within which the trace of the leaf fell on the graph paper were counted and also estimated as a leaf area. The best-fit model was selected based on F-test, mean square error (MSE) and coefficient of determination (R²). The results of statistical analyses showd that correlation coefficient of all the parameters were highly significant at 1% level of  significance. Linear regression indicated that L, L², W, W², L+W, L*W and graph paper were 91 %, 92 %, 89 %, 93 %, 95 %, 98 % and 98 % respectively to the actual leaf area. The regression model of  Y= 0.5466(L*W) + 0.7501, such that the actual measurements of L and W are simply inserted into the equation and leaf area computed.