Tamanho ótimo de amostra para análise da qualidade fisiológica de diásporos de espécies florestais nativas do cerrado

Abstract

What are the consequences of working with different sample sizes? What are the consequences of this on quality of results? What is the minimum sample size required for the results of germination measures to be representative of the population? Thinking about this, this study evaluated different sizes of samples to obtain the optimum sample size for germination or emergence testing with diaspores of Savannah‟s species, and also to study the behavior of germination/emergency measures as a function of the number of seeds composing the sample. For this, eight trials were conducted in a 3x3 factorial, completely randomized design (3 sample sizes x 3 different quality lots), conducted in a growth chamber (model MPG Seedburo Equipment Company - 2000) under continuous fluorescent white light and average temperature of 26.2 ± 2.5ºC. The results from these tests were studied in two manners. The behavior of germination/emergence measures as a function of sample size variation was analyzed by a conventional statistical study (ANOVA followed by Tukey or Kruskal-Wallis followed by Dunn and their assumptions). The optimum sample size was obtained, in principle, from the independent coefficients of variation (CVX), of germination and/or emergency percentage, which were submitted to the Meier and Lessman‟s exponential model. From the estimation of this model, optimum size of sample was determined with the mathematical expression Xc = (-d⁄b CV1 )[1/(-b-1)]. In general, it was found that germinability and/or emergency percentage is the most stable germination measure in relation to changes in sample size, while Maguire‟s speed and uncertainty were the most unstable. Most species have optimum sample size between 114.25 and 172.11 diaspores.