Tamanho amostral para ensaios em esteira de distribuição longitudinal de sementes de milho e soja

Abstract

To evaluate the performance of precision sowing, it is necessary to collect a sample of seed spacing to be classified in accordance with current standards, thus demonstrating the performance of the machine as its quality that is expressed in regularity distribution . Generally, how much higher percentage of acceptable spacing between seeds, will be better the performance of the feeder. Due to the great heterogeneity of sample sizes used in these evaluations, the objective of this study was to estimate the sample size for laboratory experiments with horizontal plate and pneumatic seed meters, under the conditions described in rules, still evaluating the regularity of distribution of seeds of corn and soybeans. Experiments were conducted under laboratory conditions, according to the international standards ISO, UNE and design of ABNT, with four different seed feeders, two pneumatic and two horizontal plate. In test 1, the treatments comprised the combination of the four metering mechanisms with three seeding rates of 2.5, 5.0 and 7.5 km h-1. In test 2, were combined the four metering mechanisms and four peripheral speeds of the plate, 0.09, 0.18, 0.28 and 0.38 m s-1, both in a factorial design with three replications. We used three methods for determining the sample size: sampling intensity, bootstrap adapted and practical method. Regarding the distribution of maize seeds, the meter mechanism studied showed no differences in the regularity of seed distribution and response to elevated peripheral speed was negatively linear for the percentage of acceptable spacing. For soybean, the meter mechanism valued tires perform better mechanisms for horizontal plate, no direct correlation with the regularity of the peripheral speed of the plate. About the sample size, it is possible to reduce both cultures, but in a more pronounced for maize, the reduction of which can reach 73% of the value above the standard. For soybean, it is possible to reduce by 27% the sample size for testing in the laboratory with a high level of accuracy.