Relationships between forest structure and avian communities have been documented in many northern temperate and tropical studies. This study tests the applicability of conclusions from those studies to New Zealand forest bird communities. For 29 months from September 1982 to February 1985 we monitored the bird community in part of the Ruamahanga Ecological Area, Tararua State Forest Park. We used a modified 5 min bird count to determine the relative abundances of each bird species. The modification consisted of separating the counts into birds near to (<20 m) and far from (>20 m) the observer. The near/far ratio proved useful in assessing the distribution of common bird species but was inadequate for rarer species. Twentynine bird species were seen, 21 of which probably bred within the study area. Eleven forest bird species were seen with sufficient frequency to apply the near/far ratio.Principal component analysis was used as a graphical tool to describe the inter-relationships of bird species distribution with plant species and plant structure. Plant species' composition was a better predictor of bird species' composition than was forest structure. Although the distributions of 10 common bird species were positively correlated with high canopy forest, five of these species were associated with red beech dominated forest and the other five species with podocarp/broadleaf dominated forest. Introduced birds were largely confined to forest margins, suggesting either that the *To whom correspondence should be addressed. Received 20 May 1991; accepted 12 November 1993remaining native birds are successfully excluding introduced birds, or that the forest habitat is unsuitable for the introduced species.