| Sumario: | In order to determine optimum plot size, the method of max R(2) was used, for which 6 regression models were tested on the basis of the following selection criteria: the coefficient of determination, error mean square and the coefficient of variation. Optimum plot size was 7 m(2). To determine the no. of replications, the methodology proposed by Romero, Carney & Rojas (1966) was followed. Three replications were found to be adequate for this trial. As for plot shape, it was found that a rectangular design minimized exptl error. From a statistical standpoint, the estimation of production on the basis of 5 plants does not fulfill the desirable characteristics due to the great variability, bias and low degree of association with the total plot production. (AS- CIAT)
|
|---|