Modelos para la estimación no destructiva del área foliar de dos cultivares de mandioca (Manihot esculenta Crantz) en la Argentina

it is difficult to make the measurement of the cassava leaf area (la) due to the irregular shape of the lobes which need meticulous, time consuming and tedious methods and demand sophisticated and expensive instruments. the aim of this research was to adjust and to evaluate simple and precise statistical predictive models in order to estimate the la of two cassava cultivars grown in corrientes, argentina. in the different samples, leaves linear dimensions and leaf dry weight were determinated. lineal regression equations for two cassava cultivars were tested to estimate la though the stepwise variable selection method. the model selected due to its goodness of fit and precision to estimate la was: la = ？0 + ？1.pl(cm) +？2.cla(cm2 )+ ？3.σll(cm) , based on non-destructive and easy to measure dimensions. this model needs different estimations of the parameters for each cultivar. the results obtained in this research showed that under the evaluated conditions and for both cultivars, leaf area might be estimated using simple linear measurements, as petiole length, sum total of lobes length and central lobe area.