In the experimental planning the variability among portions can alter the estimates of the plot sizes and of sample, besides the number of repetitions and definition of the experimental designs. To verify the variability interference of the estimates of values used to determinate the plot size with different numbers of plants and groupings of crops, a two-experiment work was carried out. The first trial was carried with zucchini in winter/spring of 2004 and the second in the summer/autumn of 2005. The plants were cultivated in plastic tunnel, with drip irrigation with three lines with 24 plants per row, spacing among plants 0.8m and among lines 1.2m. Each plant was considered as a basic unit (UB) of the plot size. The sequence of two, three and four plants in the crop row formed the plots of two, three and four UB of size (X). Analyses were accomplished considering only considered individual and combined harvests. In both experiments, the estimative of variance, variation coefficient and relative information index were obtained. Bartlett's test was applied among the variances of the individual and combined harvests in each size of simulated plot. There were some increases in the estimative in the variance of fruits biomass with the increase of the plot size and/or the number of grouped harvests. The variances, among the grouped harvests, were homogeneous, considering the grouping of three harvests. The relative information index and the relative coefficient present values close to 100% with the groupings of crops. The arrangement of four groups of combined harvests, joined with plots of four plants for zucchini reduce the variability among plots.
Cucurbita pepo; horticulture; experimental precision; plot size
