Background/Question/Methods This study was carried out in a 5-ha permanent plot established in the Amacayacu National Park, Colombian Amazonia. We assessed the extent at which floristic patterns of tree species were determined by either the environmental variation or the spatial configuration in tierra firme forests. Floristic and environmental variation was assessed on quadrants of 20x20 m. To run the analysis, four different categories of abundance (all, high, medium and low) were considered. We used Detrended Correspondence Analysis and Canonical Correspondence Analysis, followed by a variation partitioning, to analyze the patterns of species distribution in order to quantify the role played by both environmental filtering and dispersal limitation. The spatial template, which represents dispersal limitation, was evaluated in two different ways: using a third degree polynomial and using the Principal Coordinates of Neighbour Matrix (PCNM) method. Results/Conclusions Beta diversity was relatively low within the plot. PCNM appears as the most appropriate method to analyze the spatial template at this scale. The floristic differences explained along the 5-ha plot were mainly associated to biological processes, such as dispersal limitation. The largest proportion of community variation in our dataset was unexplained by either environmental or spatial data. In conclusion, these results support random processes as the major drivers of the spatial variation of tree species at a local scale on tierra firme forests in the Amacayacu National Park.