Validação de métodos para teste de germinação de sementes de espécies florestais com madeira exportada

Abstract

Brazilian sawn wood exports of tropical species between 2006 and 2011 was over 1.4 billion kilograms and generated a profit of more than US$1.08 billion. Researchers are concerned over regulation of forest tree seed trade. Therefore, were conducted a research to verify seed quality through secure and robust analysis methods aimed at evaluating normal and abnormal seedling sand dead seeds from the validation process for the germination analysis of six forest tree species. Pre - germination tests with species Anadenanthera macrocarpa, Cedrela fissilis, Cedrela odorata, Handroanthus impetiginosus, Peltophorum dubium and Schizolobium parahyba var. amazonicum were conducted in order to establish protocols based on samples of different quality (high, medium and low quality). Seeds were sent to at least six laboratories approved by the Ministry of Agriculture, Livestock and Supply (MAPA) to execute the analysis of germination repeating the protocol method. The standard ISTA procedure was used for statistical analysis of seed germination (normal seedlings): outliers on variances; effects of labs and lots (ANOVA); average test for batch and laboratory (Tukey), repeatability, reproducibility, accuracy and robustness (Mandel h and k statistics). From these tests, effects of labs and lots and repeatability and reproducibility were done to check for abnormalities in seedlings. The dead seeds were observed in graphic scheme. Abnormal seedlings and dead seeds had discrepant results between and within laboratories. The abnormality result is probably due to the difficulty of the analyst to distinguish this characteristic and the dead seed result is probably due to heterogeneity within the plot itself and genetic and environmental variability. However, the methods proposed for testing the seeds of the six species were validated because they presented themselves precise, accurate and robust.