Predicting pattern and process using the fewest possible explanatory variables is becoming increasingly important for lotic-system scientists attempting to maintain water quality, understand ecosystem functions such as nutrient retention, and protect species in rivers and streams. Successful prediction depends on our ability to delineate historical conditions prior to large-scale human impacts on water quality, community composition, and hydrology. Generalization (prediction) of stream network ecological properties is rooted in hydrogeomorphology, a search for scale-independent properties, and the concept of dynamic equilibrium. Overlaid upon this foundation is the River Continuum Concept, followed by the flood pulse concept and the idea of functional processing zones. Several regularities occur in downstream sequence including less intense or abrupt floods, smaller substrata, lower probability of drying, greater width to depth ratio, longer pools and riffles, a bigger discontinuity between size of entering side channels, less groundwater influence, and now, a higher probability of human pressure. Some of these patterns are tight (discharge versus channel area) and others while significant, have more variance (discharge versus velocity). I discuss these patterns as part of the adaptive template of stream organisms, and by necessity how the patterns scale with stream and river size.