Systematic sampling of 21 sites covering Hamilton Harbour (Lake Ontario, Canada) was carried out during the summer in 1990 and 1991 in order to study how well environmental factors, such as O-2, NO3-, and organic carbon, and the spatial structure can explain observed variation of potential denitrification, CH4 and CO2 production, as well as N-2 fixation in sediment slurries. Using canonical redundancy analysis and an extension of this method to partial out the variance into spatial and environmental components, we found that most of the explained fraction of potential microbial activities (70-90%) was accounted for by the significant environmental variables (NH4+, particulate carbon, dissolved organic carbon, dissolved O-2, depth, and temperature) and not much by the spatial polynomial trend surface. We found significant path coefficients (0.53 and 0.57 in 1990 and 1991) between CO, production and potential denitrification, which suggests that denitrifiers are dependent upon a heterotrophic bacterial population for directly assimilable carbon sources. We also found significant path coefficients between particulate carbon and both CH4 production (0.67 and 0.33) and CO2 production (0. 50 and 0.38), while significant path coefficients were also found between dissolved organic carbon and CO2 production (0.34 and 0.47). We conclude that beside well-known abiotic factors such as O-2, NO3- and organic carbon, a biotic factor involved in carbon metabolism may bk important in explaining the spatial variation of denitrification capacity in the sediment of Hamilton Harbour.