Approximately 10 % of vascular plants can grow as epiphytes, but the impact of epiphytism is more remarkable because there is an unknown percentage of re-terrestrialized taxa that evolved from epiphytic ancestors. Two main hypotheses have been proposed to explain the evolution of epiphytes: one suggests that they evolved from humid, umbrophilous environments (i.e., forest understory; Schimper hypothesis), and the other that they came from an open-infertile environment (Tietze-Pittendrigh hypothesis). The core evidence supporting these hypotheses is the frequency of terrestrial plants that accidentally grow as epiphytes (i.e., accidental epiphytes), because these plants should be abundant in the habitats that promote the colonization of the canopy by the terrestrial plants (i.e., the early state of epiphyte evolution). In a landscape with both environments (humid and umbrophilous vs. open and infertile), we tested the affinity of the flora to the epiphytic habitat and the association of the accidental epiphytes with each environment. We found 71 plant species. Forty-one percent were terrestrial; among the rest, 39 %, 3 %, and 17 % were accidental, facultative (growing equally as terrestrial or arboreal), and true epiphytes (preferentially living as arboreal), respectively. Discounting plants that exclusively grow terrestrially in the sample, in this landscape, a large proportion of the species living in tree crowns are accidental epiphytes (28), and 16 species belong to Asteranae and Rosanae, superorders where epiphytism is infrequent. Nine accidental epiphytes were associated with the open environment and none with the forest, supporting only the Tietze-Pittendrigh hypothesis. Our results support the idea that species from open environments could be preadapted to grow as epiphytes, and landscapes combining forested areas with open environments (e.g., xerophytic scrub, savanna) could promote the colonization of the canopies from terrestrial plants.