Artrópodes em plantas de sub-bosque: descrição de um novo método de coleta e as relações entre predadores e guildas funcionais em dois gêneros de plantas

Abstract

Plants are a habitat model for ecological communities study, mainly to investigate the invertebrate occurrence. In this paper, we described a new method for arthropods sampling in plants, the Socket-trap, we evaluated their efficiency by comparing with the most commonly used collection methods for this sampling type and investigated the relationships between individuals abundance in arthropod guilds and predators abundance (spiders and ants) in two genera of plants (Protium and Inga). Samples were taken in 10 permanent plots, in Adolpho Ducke forest reserve, Manaus. In each plot we sampled 18 plants of Protium (Burseraceae), 6 tree beating, 6 manual collection and 6 using Socket-trap. In each plot we also sampled 6 specimens of Inga (Fabaceae) using Socket-trap. All insects collected in Protium were identified at the family level and those belonging to the order Hymenoptera were identified at the species and morphospecies level. All invertebrates collected using Socket-trap in Protium and Inga were identified and grouped into one of the following guilds: Detritivores, Ants, Herbivores (chewers and suckers), Omnivores, Parasitoids and Predators (arachnids and predatory insects). New method sampled more insect families and Hymenoptera species than tree beating and manual collection. A similar pattern was found for abundance: Socket-trap sampledmore individuals (617), followed by tree beating (295) and manual collection (127). The other methods performance was similar for insects richness and abundance. The highestrichness of Hymenoptera and insect families sampled with Socket-trap is due to plant bagging, which prevents invertebrates scape, especially those very agile or winged. Probably during the plant agitation on tree beating some specimens escaped, in addition, very small insects can also be sub-sampled with this method. The smallest number of individuals, species and families sampled manually is related to the difficulty of specimens capturing,directly in the plant, especially those very agile. For all evaluated parameters, Socket-trap was the best collecting invertebrates method in plants. In both plant genera, the predators guild was the highest abundant, followed by detritus, ants, herbivores, omnivores and parasitoids. Despite the phenotypic difference between plant genera, arthropods proportion by guilds was similar, but the relationship between guild abundance and predator abundance was plant genus and predator type dependent. In general, the predators abundance was negatively correlated to arthropods abundance per guild. This pattern was major in relationship between ant abundance and detritivore abundance in Inga, probably due to presence of extrafloral nectaries that maintains higher ants abundance in the plant. Similarly, spider abundance was negatively correlated with the sucking herbivores abundance. However, some positive relationships between predators abundance and arthropods abundance by guilds were also found. Positive relationship between sucking herbivores abundance and ants is an indication of hemipterans and ants interaction due to production of honeydew.Positive relationship between spiders and detritivores suggests that spiders choose plants with more food. Our results demonstrated that spiders and ants act differently on arthropod guilds in plants, and this pattern depends on plant characteristics.