Eficiência de cinco modelos de armadilhas etanólicas na coleta de coleoptera: Scolytidae, em floresta nativa no município de Itaara, RS

Abstract

This work was developed with the objective of testing the efficiency of five models of ethanolic traps to collect specimens from the Scolytidae family in a native forest. The study was made in a native forest near of the CIPAM field seat in Itaara-RS. The models of trap tested were: PET Santa Maria, Roechling, Marques-Carrano, Escolitídeo-Curitiba and Marques-Pedrosa. 35 traps were installed in the forest, distributed in seven blocks. The collects occurred during 2006, since February up to December, each two weeks, totalizing 23 collects. To calculate the efficiency were used the mean test, the cost/man/hour, the relative variation, the liquid relative precision and diversity indexes for each model studied. The indexes calculated to analyze the data were frequency, dominance, constancy and abundance. The capture of the insects was correlated with the precipitation, the medium temperature and the air relative humidity. The models Marques-Pedrosa and Escolitídeo-Curitiba were most efficient in number of insects captured; the models PET Santa Maria and Roechling captured more species. The PET Santa Maria model had lower cost and was handier. The best values to the liquid relative precision were for the Marques-Carrano and Marques-Pedrosa models. The diversity indexes were high for all models of traps. 2.256 scolitids were caught in the traps, and included in 31 species. Among these species, six were considered abundant and 25 rare or disperse. The species H. eruditus was the most collected for all models; the species H. bolivianus, X. gracilis, C. letezeneri and X. retusus were classified as constant and abundant for the most of the models. The population fluctuation shows the same pattern for all the studied traps; the biggest number of collect insects occurred in the months of September, October and November, with significant correlation with the meteorological factors, but with low values for the most of the models.