Although biomonitoring approaches are being increasingly used in the measurement of stream and river health, critical assumptions about the nature of biological populations and communities that underpin them are often ignored. Many approaches based on pattern detection in plant and animal communities assume high temporal persistence in the absence of anthropogenic disturbances. However, this has been rarely tested with long-term data sets and there is evidence that this assumption is not true in some river systems. Biological processes, such as predation and recruitment, can account for considerable spatial and temporal variation in the structure of some stream communities. These processes may prevent the development of robust predictive models or indices based on pattern detection. Measurements of population or community attributes also are often used to infer ecosystem processes, yet the link between pattern and process has rarely been demonstrated. Many goals of river management relate to the maintenance of natural ecological processes and ecosystem function; direct measurement of these processes is, however, often neglected in assessment programs. Such measures are often sensitive to causal factors that are known to affect river health and it is possible to develop simple but powerful predictive models. Perhaps more importantly, should an impact to be detected, strategies for remediation are more obvious as the causal processes are generally better known. The ultimate success of biomonitoring approaches depends on how well we understand the biophysical processes that influence the structure and dynamics of stream and river systems, and the way they function.