The large area covered by the Amazonian domain and the limited financial support for biodiversity studies demand efficient research programs. Aiming to evaluate the consequences of reduced sampling effort on the retention of ecological information, we tested how differences in plot width affect the perceived relationship between environmental variation and the composition of the pteridophyte community in Central Amazonia. Measures of slope, canopy openness and soil clay content were taken in 37 terra-firme forest plots. All pteridophyte individuals were identified and mapped according to three sample strips. We tested the effects of environmental variables on pteridophyte composition of 250 x 2.5 meters-plots and on sub-samples of reduced width. The reduction of plot width from 2.5 to 1 m (60% reduction in sampling effort) corresponded to a reduction of 24% of the project's costs concerning field work expenses. This would reduce the number of sampled species from 52 to 44. The reduction of 20% plot width (from 2.5 to 2 m) would cause a reduction of two (4%) sampled species and a reduction of 8% of the costs. For all tested plot widths, community composition was associated with soil clay content and was not associated with the terrain slope. The effect of canopy openness on pteridophyte species composition was not consistent among sample sizes. These may be related to the relatively lower importance of light availability in determining community structure at the studied spatial scale. We concluded that the smallest plots were informative enough to detect the main gradients of composition and their association to environmental factors. This would allow a reduction in total costs, or the allocation of available budget to more plots, which could increase the power of the statistical analyses, reduce the confidence intervals and increase probability of detecting more species.
community; environmental gradients; ferns; sampling protocol; species richness; soil texture; tropical forest
