We sampled macroinvertebrates in 51 rapids in regulated and unregulated north Swedish rivers. Our objective was to quantify the effects on the invertebrate fauna of altered flow patterns, habitat area and isolation. The remaining rapids in regulated north Swedish rivers have two types of flow regime: (a) Those with altered flow pattern, but unreduced flow, are characterized by low long-term variation and high day to day variation. Typically, the spring flood is strongly reduced, whereas winter flow is greatly increased in relation to the situation in unregulated streams. (b) Rapids from which a major part of the flow has been diverted usually show high seasonal and daily variation, although there are often long periods with constant flow within seasons. The effects of altered flow patterns were examined using the following method. Predictive models were built for biotic responses at unregulated sites. These models did not include variables affected by regulation, such as flow magnitude and variability. The models were then used to make predictions for affected sites. The differences between observed and predicted numbers were used as a measure of the effect of regulation. Sites with reduced flow had 0–30 species (mean=6) less than predicted, corresponding to a loss of 0–38% of predicted richness. The total abundance was 0–54% (mean=12%) less than predicted. Subdivision of the fauna into functional feeding groups showed that the abundances of collectors, grazers and predators, but not filter feeders and shredders,were negatively affected. For regulated sites,with no diversion of flow, significant losses were only recorded for the abundances of collectors and predators. A subsequent analysis of the relationship between the effects and the regulation-related variables, which had been left out in the first models, indicated that the occurrence of large and rapid changes of discharge was the most important factor. No effects on overall species richness of habitat size and isolation were found, suggesting that extinction and re-establishment of subpopulations are not prominent processes on the scale considered in this investigation. Our data suggest that avoiding large and rapid flow changes can increase both the abundance and the diversity of vertebrates. Increasing the flow will decrease flow variability but will also expand the habitable area and thus the production of invertebrates. The best effect is expected at sites where a considerable proportion of the flow has been diverted.