Influence of gene dispersal and environmental heterogeneity on spatial and genetic patterns of the understory herb Heliconia acuminata across a fragmented landscape in central Amazon, Brazil

Understanding how plants are spatially and genetically distributed in the environment can be a challenging task given the difficulty to characterize ecological processes, such as gene flow, and to disentangle the relative importance of multiple factors underlying the generation of distinct patterns. In this dissertation, I study different populations of the understory plant Heliconia acuminata L.C. Richard (Heliconiaceae) distributed across 1-ha fragments and continuous forest sites in the Biological Dynamics of Forest Fragments Project (BDFFP), an experimentally fragmented landscape in central Amazonia. I characterize a set of ten microsatellite markers developed for Heliconia acuminata to first evaluate gene flow, which is one of the main processes influencing genetic structure and spatial patterns of plants, and second to assess the potential influence of endogenous (e.g., seed dispersal) and environmental factors on spatial patterns of plants and genetic relatedness distribution. I combine genetic and ecological data in a novel and comprehensive Bayesian model to estimate parentage to more fully characterize the contribution of pollen and seed dispersal to H. acuminata gene flow. I then compare metrics of gene flow between fragments and continuous forest, while taking in consideration the variation in abundance of reproductive plants in each population. I tested the conservation genetics prediction that gene flow is interrupted in fragmented landscapes. Contrary to this hypothesis, I found that that both fragmentation and low population densities were associated with greater immigration rates and longer pollination and seed dispersal distances. My results are one example of how fragmentation does not limit gene dispersal. I suggest that conservation genetics predictions are reformulated by taking in consideration the variation in the behavior of pollinators and seed dispersers across heterogeneous landscapes in response to habitat configuration and to the spatial and temporal availability of food resources. To investigate the influence of endogenous factors (plant-plant interactions) and environmental covariates (light, slope and soil characteristics) on spatial patterns of seedlings and adults, I use a new statistical methodology to model marked point patterns. Using this flexible approach, I also evaluate whether local spatial genetic structure is associated to spatial distribution of plants. The results show that H. acuminata seed dispersal is contagious, but not distance-restricted or genetically structured (presence of highly related plants). The absence of an association between spatial pattern and local genetic structure for adults also suggest the absence of genetic structuring in seedlings over time. Light and zinc availability are positively associated with spatial patterns of seedlings and adults, which may indicate carryover effects of seedlings on recruits over time. Carbon is negatively associated with adults, which may be evidence of competition with large dominant trees. I finally propose a new mechanistic framework to the studies of frugivore-mediated seed dispersal. I conduct a qualitative analysis of existent studies explicitly linking frugivores, fruiting plants and seed shadows and propose a frugivore-centered, process-based view of seed dispersal that integrates animal movement and seed dispersal ecology across multiple spatio-temporal scales. This critical analysis provides the empirical foundation over which we can build a more comprehensive, multi-scaled, research approach to the study of seed dispersal, process which is known to play a crucial role in the dynamics and evolution of plant populations.