A test of the habitat amount hypothesis as an explanation for the species richness of forest bird assemblages

Aim For the past 20 years, researchers have been challenged to demonstrate that the spatial arrangement of habitat patches actually influences the distribution of organisms and the persistence of their populations, beyond the effects of its sheer amount. More recently, it has been argued that habitat amount in the ‘local landscape’ surrounding a site is sufficient to predict species richness (SR) in that site, irrespective of habitat configuration. Here, we tested four predictions derived from the habitat amount hypothesis (HAH). Location Eastern Ontario, Canada ( c . 44°55′–45°15′ N, 75°10′–75°45′ W). Methods Point counts ( n  = 157) were conducted in five subregions to estimate forest bird SR while accounting for detectability. Surveys were conducted in mature, deciduous‐dominated forest fragments, and landscape structure was quantified at three spatial scales (500, 1000 and 1500 m). Results Although we found a significant positive correlation between SR and either fragment area (FA) or habitat amount in the local landscape, predictions emphasizing the dominant influence of habitat amount and the lower influence of FA were either not supported or weakly so. Main conclusions Contrary to the HAH, we conclude that habitat amount in the local landscape is not a sufficient predictor of SR on its own. However, we agree with the contention that, in most landscape types, ‘local landscapes’ represent more natural spatial units than habitat fragments or ‘patches’.