Bee-Plant Interactions in Coffee Agroecosystems: Management and Matrix Effects on Mutualistic and Antagonistic Relationships

Abstract

Given the rapid pace of anthropogenic environmental change, understanding how such change influences biotic interactions and ecosystem functions is a key challenge for ecologists. My dissertation addresses this challenge by examining how land management practices in coffee agroecosystems affect multiple interactions between stingless bees and the plants upon which they feed, and explores the effects of the resulting restructuring of plant-bee interactions on the interacting species. To do this, I focus on two overarching questions: 1) How do farm management practices and landscape context affect stingless bee foraging patterns in coffee agroecosystems, and how do these foraging patterns in turn influence coffee pollination? 2) How is nectar robbing by stingless bees influenced by agricultural land use, what are the underlying drivers of land use-mediated changes to nectar robbing behavior, and to what extent does this behavior lead to adaptive pollen limitation in the plant firespike (Odontonema cuspidatum) by constraining floral display size? In Chapter 1, I introduce a framework for understanding the multiple pathways by which anthropogenic environmental change can influence the frequency, outcome, or consequences of interspecific interactions without changing species composition. In Chapter 2, I evaluate how the management of weedy herbaceous vegetation and canopy trees shaped the way coffee (Coffea arabica) interacts with neighboring plants for both pollination and abiotic resources. Co-flowering plants that share pollinators can interact with one another simultaneously for both pollination and abiotic resources, yet few studies have considered the joint effects of both interaction types on plant reproduction or yield. In this study, I tackle this problem by examining coffee and non-coffee pollen deposition, pollen tube formation, initial fruit set, and final fruit set in coffee plants with different background floral environments, using structural equation models. I find that coffee competes with neighbors for pollination, but that this has little effect on yield because the crop is not pollen-limited. Interactions with neighbors for abiotic resources have a stronger effect, and I find evidence for both competition (with weeds and co-flowering canopy trees) and facilitation (with non-co-flowering canopy trees). In Chapters 3-5, I examine nectar robbery (extracting nectar from a flower via an opening other than the corolla mouth) by stingless bees of firespike. In Chapter 3, I show that habitat-based heterogeneity in the intensity of nectar robbery is due to changes in floral traits and associated bee preferences, with plants growing in coffee fields producing more and more nectar-rich flowers and therefore experiencing more nectar robbery than plants growing in forest fragments. In Chapter 4, I use a reciprocal-translocation experiment to show that light availability drives differences in both floral traits and nectar-robbing behavior, and that light environment exerts clonal transgenerational effects on floral traits. In Chapter 5, I develop a novel conceptual framework to explain strong pollen limitation in firespike: conflicting selection on floral traits by pollinators and floral antagonists (nectar robbers). I develop this framework using data from field surveys and a field experiment, and explore its potential generality as a mechanism of pollen limitation. Finally, in Chapter 6, I situate the studies presenting in Chapters 2-5 in the context of the framework developed in Chapter 1, using them to highlight directions for future work.