located near Porto Trombetas in Pará, Brazil. This company mines about 60 hectares per year. The topsoil (about 20 centimeters thick) is bulldozed into stockpiles during mining. After bauxite is removed, the mined trench is refilled with overburden and the stockpiled topsoil is respread on the site. The soil is then sliced at 1-meter intervals to a 90-centimeter depth to increase aeration and facilitate root penetration. Finally, the area is planted with nursery-raised seedlings of native forest species. Although growth is slow (most species attain about 20 centimeters in height per year), survival is good. The principal impediments to restoration are insect pests (e.g., leaf-cutter ants, grasshoppers, and caterpillars) and soil nutrient deficiencies.
Restoration of highly degraded Amazon pasturelands is also receiving attention through a study I am conducting in collaboration with Robert Buschbacher, Daniel Nepstad, and Adilson Serrao near Paragominas in northern Pará, Brazil. The vegetation of our study area is lowland evergreen rain forest. Rainfall is approximately 1,700 millimeters per year, and there is a distinct dry season from June through November. As a first step in determining how to rehabilitate degraded pastures, we are studying how natural forest and degraded pasture environments differ with respect to microclimate, soil water and nutrient availability, and physical properties of soil. This research sets the stage for our central research goal: to determine how biological and physical forces act to retard the establishment of rain forest tree species in highly degraded Amazon pastures. Our data suggest that forest trees have difficulty establishing in degraded pastures because of three factors:
few seeds of forest trees are being dispersed into pasture environments;
most seeds that do arrive are killed by seed predators; and
the few seeds that do manage to germinate eventually die because of harsh environmental conditions.
The fact that few seeds of forest trees are being dispersed into pastureland is a fundamental impediment to regeneration of these areas. Our floristic survey has shown that less than 15% of the forest species have fruits with adaptations for long-distance autodispersal (e.g., for dispersal by wind). The majority of tree fruits are fleshy and appear to be dispersed by birds, bats, and both arboreal- and ground-dwelling mammals. Hence, if seeds of forest tree species are to arrive at pasture environments, they will usually have to be carried there by animals. With this in mind, we are censusing and trapping (i.e., mist netting) birds and bats in forest edge, forest second-growth, and open-degraded pasture environments. In the case of birds, we have identified more than 150 species in our area, but only a subset of them are frugivores, and of the frugivores, fewer than 10 will move out into large openings. We are now studying the movement patterns of the few species that we have identified as potential seed vectors. Our goal is to critically evaluate the role that these species play in seed movement and to begin to elucidate the determinants of movement for these species.