AbstractOver the past two decades of refinement and application of instream flow evaluations, we have examined the hydraulic habitat of aquatic macroinvertebrates in a variety of conditions, along with the role of these macroinvertebrates in sustaining ecosystem integrity. Instream flow analyses assume that predictable changes in channel flow characteristics can, in turn, be used to predict the change in the density or distribution of lotic species or, more appropriately, the availability of useable habitat for those species. Five major hydraulic conditions most affect the distribution and ecological success of lotic biota: suspended load, bedload movement, and water column effects, such as turbulence, velocity profile, and substratum interactions (near‐bed hydraulics). The interactions of these hydraulic conditions upon the morphology and behavior of the individual organisms govern the distribution of aquatic biota. Historically, management decisions employing the Physical Habitat Simulation (PHABSIM) have focused upon prediction of available habitat for life stages of target fish species. Regulatory agencies have rarely included evaluation of benthos for flow reservations. Although ‘taxonomic discomfort’ may be cited for the reluctant use or creation of benthic criteria, we suggest that a basic misunderstanding of the links between benthic macroinvertebrate and the fish communities is still a problem. This is derived from the lack of a perceived ‘value’ that can be assigned to macroinvertebrate species. With the exception of endangered mussel species (for which PHABSIM analysis is probably inappropriate), this is understandable. However, it appears that there is a greater ability to predict macroinvertebrate distribution (that is, a response to the change in habitat quality or location) and diversity without complex population models. Also, habitat suitability criteria for water quality indicator taxa (Ephemeroptera, Plecoptera, and Trichoptera; the so‐called ‘EPTs’) may also provide additional management options to stream regulators. The greatest application for macroinvertebrate criteria will be in low‐order streams where a more immediate link to fish communities can be established. We present an example from Queens Creek, in North Carolina, USA, in which monthly allocations required to preserve the integrity of the benthic macroinvertebrate community were significantly higher than for the target benthic fish species, Cottus bairdi. In the months when both Cottus and community diversity of macroinvertebrates were the ‘bottleneck’ life stages, preservation of only fish species could result in an additional 5–25% loss in macroinvertebrate habitat. We suggest that, as there becomes an increased emphasis on maintaining macroinvertebrates as monitors of stream health, there will be a concurrent emphasis on incorporating hydraulic habitat conditions as a part of bioassessment. Copyright © 2001 John Wiley & Sons, Ltd.