Seed removal experiments by invertebrates of four Marantaceae species

The study was carried out in Reserva Ducke, a 10,000 ha (10 km x 10 km) protected area 26 km north of Manaus - Brazil (02o 55’-03o 01’ S, 59o53’-59o59’ W) in central Amazonia. The site is subdivided by an 8 x 8 km grid system for long-term ecological studies, with trails spaced by 1 km. The present study was carried out in 9 10 x 100 m plots, within a 2 x 2 km portion of the grid. The plots were evenly placed in the three main topographical microhabitats, three plots in plateaus, three in slopes and three in valleys. We choose four Marantaceae species with contrasting primary dispersal syndromes and seed mass: Goeppertia altissima (bird dispersed) and Ischnosiphon arouma (unknown vertebrate dispersed) produce large seeds (>0.3 g) while Monotagma densiflorum (ant dispersed) and Monotagma spicatum (ant dispersed) produce small seeds (< 0.07 g). We collected mature seeds directly from the infrutescence of several individuals (10-20) aiming to use them in removal experiments. In each plot, seeds were placed directly on the ground, in three observation points next to adult individuals of the focal species, separated by at least 10 m from each other. In the absence of adult individuals in the plot, seeds were placed at random locals respecting the minimum distance of 10 m. The number of seeds per observation point was standardized for each species and ranged between two and five, based on variable abundance of seeds between different species. Observations were performed for 2 hours, in two periods: diurnal, between 8:00 and 16:00 and nocturnal, between 18:30 and 1:00 in all plots. To facilitate visualization during night trials, all seeds were painted with a solution of fluorescent pigment powder diluted in acetone the day before the experiment. Observation points were constantly monitored by an observer that took notes of all interactions between invertebrates and seeds. In a seed removal event, the invertebrate was followed until the deposition of the seed in the ground, or the animal reached the nest (in the case of ants), or where the seeds were lost from sight. The measurement of removal distance was between the observation point and the last known location of the seed at the end of 2 hours of the experiment. One specimen of each ant and cricket species involved in the seed removal was collected whenever possible. Also, we used images of the seed removal events to help in the identification of species in the cases when it was not possible to collect a specimen.