Bedrock channels are a key component of the landscape. Pluvial bedrock incision sets the local base level, the boundary condition for hill slope processes, transmits tectonic and climatic signals throughout the landscape and transports sediments to sedimentary basins. Bedrock river networks are also responsible for the texture and relief of an unglaciated landscape, and in flexural landscapes bedrock channels significantly effect rates and patterns of erosional unloading and consequently long-term sediment fluxes to basins. This thesis investigates the mechanism and morphological expression of one of the key mechanisms of erosion in bedrock river incision: abrasion by bedload. Both long term landscape evolution and physical process based models of bedrock incision would benefit from a better understanding of the controls and behaviour of the abrasion process in natural substrates and settings. A field .campaign documenting evidence of bedrock channel erosion identified upstream facing convex surfaces as a common morphology found in bedrock rivers in Taiwan. An asymmetry of morphology and microscopic surface texture (microsurface texture) related to stream flow direction were identified as core characteristics of this bedrock bedform. A conceptual model is suggested explaining the formation of these features by impact abrasion from bedload particle streams. The three dimensional change of ten upstream facing convex surfaces were captured using three dimensional laser scanning in combination with a novel coreferencing system. This passive, non-invasive study established the kinematics of the eroding surface. Spatial variation of erosion is strongly linked to the aspect of the eroding surface in relation to the mean channel flow direction. The convex, upstream facing surface is found " to be the location of significantly elevated erosion relative to the companion downstream facing and horizontal lee surfaces. The process of fluvial bedrock abrasion by bedload was isolated in a flume study. In a parallel approach, the microsurface texture and morphological evolution of rock blocks abraded by cobble and pebble grade bedload was investigated. The control by grain calibre, grain lithology, bedrock lithology and initial form of the obstacle were investigated. A causal link between process and form was robustly established and some insights were gained into the importance of impact angle, impact frequencies and velocities of impact of a bedload particle stream. Deformation caused by single impacts of polished quartz spheres onto flat, polished marble plates was determined using optical white light interferometry in a controlled laboratory study. The influence of sphere diameter, impact angle and impact velocity were investigated and empirical relationships between these variables determined. In all 782 impact craters measured in this study, no significant erosion occurred. A final study examined the deformation of a surface that was repeatedly impacted until erosion occurred. Surface changes were mapped on an impact by impac~ basis and are described by statistical parameters. The marble plate underwent an incubation period during which no erosion was observed. At a critical number of impacts, catastrophic failure occurred in the deformed surface of the marble plate resulting in the generation of a wear particle (erosion). Abrasion by bedload was determined to be a key process in fluvial bedrock incision and the morphological signature of its action constrained. The outcomes of these projects require that channel roughness, bedload particle streams and properties of particle and substrate be factored into future physically based models of bedrock channel erosion for an accurate representation of channel evolution. A simple conceptual model of the nucleation and propagation of upstream facing convex surfaces originating at a lithological boundary is presented combining insights from all projects.