Birds are useful for testing ecological theory, monitoring environmental change, and establishing the basis of a national conservation programme. However, there are difficulties in estimating the numbers of terrestrial birds. Observer error and differences in spatial and temporal detectability are problems common to the census of birds throughout the world. Asynchronous reproduction and a high level of communal or co-operative nesting are special problems encountered in censuses of Australia’s terrestrial avifauna. Despite these unique features and the relatively recent use of bird censusing procedures in Australia, it appears that the census methods developed in the Northern Hemisphere are equally valid in Australia, suffering only from the same sources of error encountered in their use in America and Europe. This paper reviews the history of terrestrial bird censuses in Australia, describes the special problems encountered by Australian workers, and shows how counts of the terrestrial avifauna have been used to test ecological theory and develop conservation and management programmes for the continent’s biota. It concludes with some simple guidelines which, if followed, will facilitate the use and comparison of census results between projects and localities.