Different methods have been proposed to study spatial and temporal variability of species diversity and community composition in order to understand the processes underlying spatial and temporal patterns. We studied the patterns of species assemblage of canopy and understory butterfly communities over four years in natural and secondary-growth forests in Costa Rica in order to compare the performance of different techniques, including statistical analyses as principal coordinates analysis (using abundance-based Jaccard similarity index), correspondence analysis, and cladistic analysis. Similar general patterns were identified using traditional ordination techniques and cladistic analysis. The main component of variability identified by both techniques was the vertical spatial axis indicating that the organization of species composition over the years differed between canopy and understory communities. Cladistic analysis showed an important advantage over traditional ordination techniques: species can be related to the general patterns observed. This helps to answer questions as which species follow the general patterns and which do not.