study area: wolf river, mackenzie river, spruce river and nipigon 10 burn zone watersheds. these are tributaries of lake superior and lie within the northwestern ontario boreal forest near thunder bay, ontario. sites were classified according to catchment size and disturbance adjacent to the stream... the effectiveness of macroinvertebrate indices for monitoring the effects of timber harvesting on stream habitat were determined as well as the influence of stream size and disturbance on community structure and physical habitat...macroinvertebrates were collected from fifty-six stream segments (study sites). in addition, collections were gathered from nine streams yearly and from five streams repeatedly throughout one season. all stream segments or study sites were classified according to catchment size and to disturbance adjacent to the stream (forested, harvested or burnt). the study had two components. the first component determined the effectiveness of macroinvertebrate based indices for monitoring the effects of timber harvesting on stream habitat in ontario’s boreal forest. the yearly and seasonal variability (coefficient of variation (cv)) of various biomonitoring metrics were determined from yearly and seasonal data sets. seasonal and year to year variability was high in all metrics except percent dominance, percent model affinity and percent diptera. to assess the sensitivity of various metrics to impairment, metric values from disturbed sites were compared with values from reference sites and established scoring criteria. when considered as a group, harvested sites did not differ from reference sites for individual metrics. however, when sites were considered individually, impairment was detected at some harvested locations. to assess the influence of stream size on the various biomonitoring metrics two-way analysis of variance was used. all metrics differed among catchment areas. similarly, classification of reference communities by catchment size reduced the standard error of metrics in many cases. the second component of this study used more detailed multivariate and taxonomic analyses to investigate harvesting impacts. several specific questions were addressed. firstly, do stream size and disturbance (timber harvesting or burning) influence macroinvertebrate community structure and physical habitat (discharge, temperature, closed cover and substrate profile)? secondly, do the aforementioned physical variables influence macro invertebrate community structure? finally, is the magnitude of a disturbance impact dependent upon the size of the stream and type o f disturbance? a relationship between macroinvertebrate community and disturbance adjacent to the stream (forested, harvested or burnt) was not detected. invertebrate community structure was correlated to the physical habitat variables measured. analyses detected variation in macroinvertebrate community and habitat structure in relation to catchment area. the variations observed in the macroinvertebrate community in relation to stream size were similar to macroinvertebrate communities influenced by size selective predation patterns in brook trout (bechara et al. 1992).