Area size is a matter of concern in field trials with perennial plants and fruits trees, as they usually require large spacings. Big plots and small number of replications are frequently used in these cases to justify the reduction in the experimental area, labor and consequently in the costs of the research. However, this procedure impairs the precision of parameter estimations and the efficiency in application of statistical tests. This work has been carried out with the aim of demonstrating that precision of trials and estimations of experimental error and effects of treatments cam be improved, a detection of significant differences among treatments are favored, and the experimental area is reduced by increasing the number of replicates and using small plots. A procedure has been developed that associates size of plot and number of replicates by the mean variance of each treatment, allowing the employment of a larger number of replicates to increase test precision, increase trial uniformity and improve quality of research. The results showed that the use of small plots favors the increase in number of replicates leads to the obtention of better estimations of experimental error effects of treatments and parameters besides allowing more efficiency to statistical tests used. A substantial reduction in the number of plants needed for the trial as well as in the size of experimental area was also observed.
Experimental area; treatment mean variance; intraclass correlation coefficient guard rows
