We examined relationships between the biomass of trout (species of Salvelinus and Salmo) and physical and biological habitat variables in streams to identify habitat factors that might limit trout biomass. Thirty sites were chosen to span a wide array of habitat types. At each site we measured a large number of habitat variables representing instream cover, substrate, stream morphology and velocity, stream temperature and food availability. Two habitat quality index models developed by Binns and Eiserman for Wyoming streams accounted for only 6.7 and 9.2% of the variation in trout biomass at Ontario stream sites. Different factors must limit trout biomass in Wyoming streams than in Ontario streams. Regression and discriminant function analyses indicated that trout biomass in southern Ontario is correlated with microcommunity biomass (measured as ATP of the suspended solids, and representing bacteria, fungi, and algae), percent pool area, mean maximum summer temperature, biomass of small benthic invertebrates, presence of piscivorous fish, and a variable representing pools and overhead cover. Microcommunity biomass was the most important habitat variable in these analyses and it was significantly correlated with basin yield. We hypothesize that microcommunity biomass is a surrogate measure of localized groundwater inflow, which, in turn, has a beneficial effect on the microhabitat of trout eggs and underyearlings.