Sistemas de polinização em fragmentos de Cerrado na região do Alto Taquari (GO, MS, MT).

Abstract

The Cerrado Domain occupied originally 23% of the Brazilian territory (ca. 2 million km2), especially in the Central Plateau, being the second largest phytogeographic province of Brazil. The cerrado vegetation is not uniform in physiognomy, ranging from grassland to tall woodland, but most of its physiognomies lie within the range defined as tropical savanna. Is estimated that 3,000 to 7,000 vascular plant species occur in this vegetation type, from which 1,000 to 2,000 belong to the woody component. Different authors have attempted to use reproductive features to explain the general patterns of diversity and community structure found in tropical woodlands with the underlying idea that plant diversity and spatial distribution is dependent on reproductive processes. Studies on the reproductive biology of cerrado plant species have shown a great diversity of pollination systems, similar to those found in Neotropical forests. The data emerging for the reproductive biology of plants have important consequences for conservation and understanding of the organization of cerrado communities. We sampled five cerrado fragments in the Brazilian Central Plateau, in which we sampled woody individuals. Using the floristic data of all our field trips, we sampled 2,280 individuals, representing 121 species and 38 families. The richest families were Fabaceae and Myrtaceae, and Davilla elliptica A. St-Hill and Myrcia bella Triana were the best represented species. Most species presented open flowers, with diurnal anthesis, pale colors and with pollen as floral reward. In the cerrado vegetation, species with flowers visited mainly by bees and small insects were the main groups ecologically related to the pollination. Of the 121 species, 65 were pollinated mainly by bees; 30, by small insects; 15, by moths; five, by bats; three, by beetles; two, by hummingbirds; and one, by wind. The ordination analysis of floral characteristics and plant species showed that there was a grouping of species with some pollination systems, for which inferences based on floral characteristics are recommended, such as the species pollinated by bats, moths, and birds. On the other hand, for the species pollinated mainly by bees and small insects, these inferences are not recommended due their great dispersion throughout ordination axes and large overlapping. These dispersion and overlapping occurred probably due the absence of specificity between plants and pollinators. For four of the five pollination systems with at least ten individuals, we found no significant variation in relation to distance from edge, except for plants pollinated by beetles, for which there was a decrease in the frequency to toward the fragment interior. Similarly, we only found significant variation in relation to the height for plants pollinated by bats, for which there was an increase of the frequency with the height of the trees. In general, we found no horizontal and vertical variations in the pollination systems, contrary to what was found in forests and, probably, as consequence of the more open physiognomy of the cerrado fragments.