Tamanho de amostra para estimação da média e do coeficiente de variação em caracteres de quatro espécies de crotalária

Abstract

The use of cover crops such as Crotalaria, is expanding, being an alternative to meet the demand for organic fertilizers, with high nitrogen fixation to the ground. For research on cover crops present valid, it must first establish the correct experimental design. The objective of this study was to determine the sample size needed to estimate the mean and coefficient of variation in four species of Crotalaria (C. juncea, C. spectabilis, C. breviflora and C. ochroleuca) at different precision levels. Four blank experiments (without treatment) were conducted - an experiment with each species -, at the Federal University of Pampa campus Itaqui, during the season 2014/15. At harvest, they were collected randomly 1,000 pods of each species and in each pod were measured the following characters: pod weight with seeds (PVGCS, in g); seedless pod weight (PVGSS, in g); pod length (CVG, in mm); pod width (LVG, in mm); pod height (AVG, in mm); number of seeds per pod (NS, in units); seeds weight per pod (PS, in g), and; mass of hundred seeds (MCS, in g). For each character, in each species, central tendency, variability, and distribution measures were calculated, the data normality checked and, then, determined the sample size necessary for the mean estimation with error of 1%, 2% ..., 10% of the mean, by Student's t distribution. Subsequently it was determined the sample size required to estimate the mean and coefficient of variation with amplitudes of the confidence interval of 95% (AIC₉₅%) of 2%, 4%, ... 20%, by resampling with replacement. The sample size variability between species and characters. For the same species, character and precision level, we need a larger sample size to estimate the average in relation to the need to estimate the coefficient of variation. To estimate the mean and coefficient of variation with AIC₉₅% of 10%, is required 195, 65, 95 and 80 pods, respectively, for C. juncea, C. spectabilis, C. breviflora and C. ochroleuca. At this level of accuracy is necessary to evaluate 20 pods for determining the mean and coefficient of variation of CVG, LVG and AVG, 85 pods for determining PVGSS and MCS and 195 pods for determining the remaining characters, regardless of the species considered.