Linking disturbance to stream structure and function has historically been approached by making bivariate inferences between suspected causes (stimuli) and effects of interest (responses).  Instead of focusing on bivariate relationships investigators are searching for more holistic approaches that integrate ecosystem components and address interactions among intermediate variables involved in the translation of disturbance.  I propose the cascade metaphor to organize research investigating complex interactions among stimuli, responses, and intermediate variables at the watershed scale.  The cascade approach is used to construct multiple hypotheses for initial testing.  Structural equations modeling offers one approach to constructing uni-directional hypotheses that link an ultimate stimulus (e.g., coal mining) to a response (e.g., metal contamination in stream sediments) through a set of intermediate variables that ‘translate’ ultimate causes to the effect of interest.  Components can be quantified to determine the relative strength of each variable and holistic models identified based on how well field data ‘fit’ hypothesized structural models.  Strongest models can be directly translated into management protocol by focusing resources (money, person power) toward the components that are most likely to affect translation of cause to effect.  Here I outline cascade terminology and provide examples for discussion.