ABSTRACT
The aims of this research were to determine the sample size needed to estimate the mean of traits and evaluate the linear relationships between traits of Crotalaria spectabilis plants. An experiment with C. spectabilis was conducted in the 2014/2015 season, and the following characters were evaluated in 100 plants: plant height (AP), number of productive branches (NRP), dry matter of the aerial part (MSPA), dry matter of the pods (MSVG), number of pods (NVG), weight of each pod (PMVG), number of seeds (NS), average number of seeds per pod (NSVG), mass of hundred grains (MCG), and total weight of seeds (PTS). Measures of central tendency, dispersion and distribution were calculated, the normality of the data was checked, and the sample size needed to estimate the mean of each trait was determined, assuming estimation errors equal to 1, 2, …, 20% of the mean. Then, the Pearson’s linear correlation coefficients were calculated, and the direct and indirect effects of the explanatory variables on PTS were estimated by path analysis. In C. spectabilis, 121 plants are required to the mean estimation of NRP, MSPA, MSVG, NVG, NS, and PTS traits, with a maximum estimation error of 10% of the mean. In this same level of accuracy, 6 plants are needed for the estimation of the mean of AP, PMVG, NSVG, and MCG. The variable NS showed high correlation and direct effect on PTS, indicating a relation of cause and effect and the possibility of indirect selection. The other explanatory variables had indirect effects on PTS via NS.
Key words
experimental precision; experimental planning; path analysis; indirect selection
