The impact of logging operations on bird species abundance and diversity was studied over an eight-year period using a variable-circular plot bird survey procedure and a Before-After-Control-Impact (BACI) design in karri (Eucalyptus diversicolor) forest in south-western Australia. The magnitude and spatial extent of impact was assessed from sample points in five bands parallel to the edge between clearfelled areas and mature forest (unburnt-forest, burnt-forest, forest-edge, gap-edge and gap). A total of 24748 birds of 53 species were detected within 30m of the observer during the study. Bird species abundance, and diversity, were similar in the five bands prior to treatment, and observed differences could be attributed to spatial patterns of flowering in the overstorey. Following logging, 19 of the 40 species that were detected more than five times showed changes in relative abundance (including both, increases and decreases) in one or more bands. Gaps showed the greatest number of significant changes (15) and these were typically of the greatest magnitude over the four years studied post-treatment. Gap-edges showed fewer changes (six), while both, the burnt-forest (three) and forest-edge (nil) bands showed few changes in relative abundance. Both, the number of species detected (for a given sampling effort), and species diversity declined in gap and gap-edge bands following logging. Nevertheless, variation in bird counts between successive seasons and years necessitates considerable replication before changes can be reliably assigned to disturbance impacts. In addition, the long period before maturation of the overstorey suggests that the local impact of logging and burning is likely to extend well beyond the life of experimental studies. Our results support the implementation of a well-dispersed system of retained mature forest along road, river and stream zones within the regeneration mosaic.