The aim of this study was to determine the sample size needed to evaluated the mass, length and diameter of pineapple fruits in large experimental units under different fertilizations. In the experiment with pineapple (cultivar Perola) it was evaluated two sources of nitrogen (urea and ammonium sulfate) and five doses of potassium (zero, 350, 700, 1,050 and 1,400 kg ha-1). Each plot had five double strips of four meters length, being the three central strips taken as useful area. The plant spacing was 1,2 x 0,4 x 0,4 m, making 60 plants per useful area of each plot. It was used a randomized blocks design with four replications. The sample size ( η ) was calculated for the semiamplitude of the confidence interval (SA) equal to: 2, 4, 6, 8 and 10% of the average estimate, with degree of confidence (1-α) of 95%, using the Student's t distribution. Afterwards, η was set as the total amount of fruits per plot for calculating SA as percentage of the average estimate for each character. With an average estimate error of 4%, should be sample 83, 35 and 10 fruits in each experimental plot for evaluation of mass, length and diameter, respectively, for pineapple fruits, cultivar Perola, in experiments with fertilization.
Ananas comosus; sampling; experiments with large experimental units
