The interactions between predators and their prey are key drivers of structure and functioning in many ecosystems. However, the ability of predators to effectively regulate prey abundance can be strongly modified by the context in which trophic interactions occur. My dissertation explores the effects of five factors which have the potential to mediate trophic interactions on nearshore reefs: prey density, organismal body size, habitat complexity, animal behavior, and fishery harvest. Working on both temperate rocky reefs and tropical coral reefs, I use field- and lab-based experiments as well as a numerical model to better understand the interactions among sea urchins, their finfish and invertebrate predators, and the nearshore reef-associated communities of which they are a part. Chapters 1 and 2 focus on the dynamics between sea urchins, spiny lobsters, and fish predators on the rocky reefs of southern California. Following the extirpation of the archetypal urchin predator, the sea otter (Enhydra lutris), top-down control of urchins in this system by spiny lobsters (Panulirus interruptus) and the labrid fish California sheephead (Semicossyphus pulcher), has been hypothesized, but rarely tested experimentally. Chapter 1 tests for density- dependent mortality of purple (Strongylocentrotus purpuratus) and red (Mesocentrotus franciscanus) urchins due to predation by finfish and lobsters. In laboratory feeding assays, spiny lobsters demonstrate a saturating functional response to urchin prey, whereby urchin proportional mortality is inversely density-dependent. In field experiments on rocky reefs near San Diego, CA, when purple urchins are offered alone, I find evidence of positive density-dependent urchin mortality at low densities, similar to those found within kelp forests. At higher prey densities, analogous to those found within urchin barrens, prey mortality is density-independent. When red and purple urchins are deployed to reefs simultaneously, urchin mortality is density-independent and fish do not aggregate to higher density patches. This shift in predation mortality is likely due to the increased biomass of the alternative red urchin prey rather than the increased structural complexity offered by their large spine canopy. Overall, results from Chapter 1 suggest that top-down control of urchins can occur only under limited circumstances, when predatory fish are abundant and large red urchins are absent. In Chapter 2, I develop a tri-trophic, size structured numerical model of a southern California rocky reef. The model includes multiple ecological processes that can drive feedbacks across trophic levels leading to alternative stable states, including recruitment facilitation and size-structured predation. I find that fishery harvest for the predator (spiny lobster) and prey (red urchins) interacts to determine the level of ecological resilience exhibited by the system, i.e. the likelihood of shifting between alternative stable states. Specifically, I show that predator harvest can drive the system from a kelp forest to an urchin barren, but that prey harvest determines the likelihood of this shift. Size structured predation on urchins is the feedback maintaining a given ecosystem state. This model suggests that ecosystem resilience depends on both predator and prey harvest in multi-trophic level harvest scenarios, which are common in marine ecosystems but are rarely accounted for by traditional single-species management. Collectively, my first two chapters demonstrate that predator regulation of urchins can occur only under limited circumstances which strongly depend on both predator and prey body size and species composition. These findings also have significant implications for the dynamics of alternative community states observed on rocky reefs, as harvesting predators and harvesting prey can interact to determine the ecological resilience of these important coastal habitats. In addition to density and organismal body size, habitat complexity can also play a vital role in shaping ecological communities. However, many coral reef ecosystems are shifting to alternative states with reduced structural complexity and altered community assemblages. Small- bodied herbivores, such as sea urchins, are common inhabitants of reefs, and their importance for controlling the distribution and abundance of algae in marine ecosystems is well understood. Less understood is the role of habitat complexity and species identity of foundational species in dictating the abundance of this increasingly-important group of herbivores. In Chapters 3 and 4, I explore the feedbacks between habitat complexity, herbivorous urchins, and their predators on fringing coral reefs of Bocas del Toro, located on the Caribbean coast of Panama. In Chapter 3, I use benthic surveys, tethering, and laboratory experiments to show that the structural complexity and species identity of three corals commonly observed on Caribbean reefs mediate the abundance, behavior, and demographic characteristics of an increasingly important herbivore, the reef urchin Echinometra viridis. Tethered urchins survive better on the more structurally complex coral Agaricia tenuifolia and hydrocoral Millepora alcicornis than on less complex branching Porites species. However, natural densities of urchins on these corals do not follow the same pattern, suggesting that coral identity, independent of complexity, also contributes to habitat associations. In habitat choice experiments, urchins prefer the structurally complex coral A. tenuifolia only when waterborne cues of predators are introduced. Despite minimal differences in the standing stock of algae associated with the different corals, urchins inhabiting Porites colonies have a marginally higher reproductive condition than those collected from the other corals, suggesting a fitness trade off to inhabiting the riskier coral. Understanding the drivers of herbivore habitat associations is vital for predicting the persistence of coral-dominated reefs due to feedbacks between changing coral reef communities (both species identity and habitat complexity) and shifts to algal dominance. In Chapter 4, I explore the potential for non-consumptive effects (NCEs) of predatory spiny lobsters on the grazing and movement behaviors of two urchins (E. viridis and Diadema antillarum) which contribute to Caribbean coral reef resilience. Non-consumptive effects of predators on their prey can be an important influence on ecosystems because predators can suppress the ecological function of far more prey than they can consume. However, herbivore responses to predatory risk cues can differ among species which otherwise could be functionally similar. Cues from a generalist predator, the Caribbean spiny lobster (Panulirus argus), strongly suppress grazing by Diadema but not Echinometra. Conversely, cues produced by simulated predation on conspecific urchins cause reduced grazing by Echinometra but not Diadema. In field tests for NCEs on movement behavior, Echinometra consistently move away from lobsters on coral colonies of a variety of structural complexity levels, but movement rates are reduced in response to lobster cues only when on highly rugose corals. Diadema movement is not affected by the presence of lobsters. The contrasting responses exhibited by these two urchins suggest that herbivore populations and their functional roles may respond in unexpected ways to anthropogenic changes to predator communities and reef structural complexity. Together, these chapters provide evidence of the importance of small-bodied herbivores to Caribbean coral reef resilience through feedbacks between herbivory and habitat complexity and improve our understanding of trophic interactions on degraded contemporary coral reefs.