Distinguishing individual quality from habitat preference and quality in a territorial passerine.

Theory predicts that animals breeding in heterogeneous landscapes preferentially occupy habitats likely to maximize individual fitness, but identifying those habitats has proved problematic. Many studies develop metrics of habitat quality linked to site-specific reproductive output measured in successive years, but few separate the independent effects of individual "intrinsic quality" from those due solely to the attributes of the habitats themselves. In many populations, processes such as competitive territory defense, longevity, site-fidelity, and variation in breeding density and territory size over time have the potential to limit the degree to which individual and habitat quality will be positively related in nature. However, the effects of these processes on estimates of habitat or site-specific reproductive output have not been thoroughly investigated. We show that, in an insular population of Song Sparrows (Melospiza melodia), females nested preferentially in breeding sites with high mean reproductive output assessed over 35 years, and that variation in site-specific reproductive output was positively related to female intrinsic quality, measured here as the lifetime reproductive success of individual females relative to others hatched the same year (rLRS). In contrast, vegetation traits (shrub cover, edge, and soil depth) predicted female preference for breeding sites but did not predict site-specific variation in annual reproductive output. Female quality also did not predict which females occupied more- or less-preferred breeding sites over the study period. However, mean annual reproductive output of breeding sites estimated over 35 years was strongly positively related to the quality of the females that nested in them. Overall, these results indicate that site-specific estimates of habitat quality that do not consider the quality of the individuals occupying those sites may include substantial bias due to variation in occupant quality, and thus may not reliably predict the intrinsic effects of habitat quality on individual or population fitness.