Sample size for experiments with bean pods in different environments

Aiming to estimate the sample size for the fresh weight of pods of bean pod, five experiments were performed with blank Macarrao at autumn / winter 2009 and spring / summer 2010, in greenhouse, tunnel and field. The basic units (BU) were pairs of plants, resulting in 36 per line of UB cultivation in the greenhouse, and 42BU in tunnel and field. For each row of plants in each harvest and group of crops it was estimated the mean, variance and coefficient of variation (CV%). When the variances were not heterogeneous lines, it was used for calculation the of the sample size, the CV% of the experiment, when they were heterogeneous, it was used the CV% of the line in which this was higher It was concluded that the variability of production of fresh pods increases in adverse weather conditions. The analysis of total production enables the use of smaller sample sizes. In this condition, for a semi-amplitude of the mean confidence interval (D%) of 10 in fall / winter, it is necessary to sample in the row of plants, 28 plants in the greenhouse, 21 in the tunnel and 26 on the field. In the spring / summer for a D=10%, it was necessary to sample in the row of plants, 22 plants in tunnel cultivation and 24 on the field.