Forest fragmentation and the interaction between ants and fleshy diaspores in the Atlantic forest

Author(s)

Ana Gabriela Delgado Bieber

• Document type: Doctoral Thesis
• Press: Campinas, SP.
• Institution: Universidade Estadual de Campinas (UNICAMP). Instituto de Biologia
• Defense date: 2012-10-02

Examining board members

Paulo Sergio Moreira Carvalho de Oliveira; Inara Roberta Leal; Jean Paul Metzger; Marco Aurélio Pizo; Alexandre Vicente Christianini

• Advisor: Paulo Sergio Moreira Carvalho de Oliveira

Abstract

In tropical rainforests, ground-dwelling ants opportunistically interact with fleshy diaspores and, in some cases, this interaction can be classified as mutualistic. While ants gain from eating the nutritious fleshy parts, they may benefit the seed in two ways: (1) greater germination success; and (2) directed dispersal to nutrient-rich microsites where seedlings grow better. Thus, for some plant species, the loss of ant-diaspore interactions may negatively affect plant recruitment. Here, we explore ant-diaspore interaction patterns in two divergent forest types, representing what has remained from the Brazilian Atlantic Forest of São Paulo State: continuous old-growth forests (CFs) and secondary forest fragments (FFs) (four sites each). During this study, four complementary approaches were adopted. First, we surveyed ant-diaspore interactions in all forest sites during one year. Main plant families and ant genera registered during our survey correspond to previous studies performed in this biome. In spite of the similar abundance of ant-attractive diaspores on the forest ground, there were striking differences on ant-diaspore interactions between continuous and fragmented forests. Differences are most certainly attributed to the species-poor plant community, composed by less attractive species, found in the fragments. Based on the same dataset, we investigated whether the topology of ant-diaspore interaction networks differed between our two studied forest types. Three out of seven network-level metrics calculated diverged between fragments and continuous forests. The decrease in the number of interacting plant species observed in the fragmented forests appears to be a key-factor for explaining the observed results. Third, we offered a lipid-rich synthetic diaspore in experimental removal stations to compare ant attendance and behavior between forest types. Again, the experiment evidenced differences between the opposing forest types. In general, a higher number of ant species was recorded in CFs. The frequency of large ponerines (mainly Pachycondyla striata) was higher in CFs, corresponding to the higher frequency of diaspore removal and the higher removal distances observed in these sites. Finally, we experimentally evaluated if previous handling by birds (a condition more frequent in continuous forests) would affect ant attendance to fallen fruits. For this approach, we used as a model the bird-dispersed species Psychotria suterella (Rubiaceae), a common treelet in the understory of our forest sites. As expected, "feces-embedded" and "mandibulated" P. suterella fruits had increased frequencies of ant attendance than "intact" fruits. This result suggests that plant species in fragmented forests, together with the pervasive loss of their primary seed dispersers, may also be affected by the decreased attractiveness of unhandled fruits to ants (secondary dispersers), in spite of the abundance of these insects on the forest floor. Altogether, our results indicate that Atlantic Forest fragmentation negatively affects ant-diaspore interactions with respect to most of the studied parameters. Therefore, we expect that fragments will present a decrease on ant-derived potential benefits to seeds and seedlings (AU)