STOPOVER HABITAT USE BY SPRING MIGRANT LANDBIRDS: THE ROLES OF HABITAT STRUCTURE, LEAF DEVELOPMENT, AND FOOD AVAILABILITY

We used principal factor analysis and multivariate analysis of variance (MANOVA) to compare vegetation characteristics (means calculated over four plots per survey transect) between the five habitat types. If significant differences between habitats were detected, we ran a posteriori univariate ANOVAs for each habitat variable. We used stepwise multiple regression (SAS/STAT) to examine relationships between abundance of the two groups of transient Neotropical migrants (shrub-sapling-breeding and mature-forest-breeding) and the nine habitat characteristics. At each step, an F-statistic was calculated for each variable in the model. We allowed variables having P 0.10 as other variables were entered into the model. We considered full models with P

For each transect survey, we calculated a mean index of leaf development separately for understory and canopy-subcanopy vegetation using leaf-development scores of the common plant species. We used ANOVA to separately examine differences in canopy-subcanopy and understory leaf development among the three mature forest habitats and years. Using a series of ANOVAs run separately by year and visit, we tested for differences in leaf development among habitats. We used a generalized linear model for Poisson data (see above) to test for a relationship between leaf development and abundance of mature-forest-breeding transients and ran analyses separately by year and visit for each of the two forest layers.

RESULTS

Individual species patterns. - Overall, transient species were widely distributed among the five study habitats. For example, six of nine species were recorded in all five habitats, Blackburnian Warbler was recorded in only four habitats, Blackpoll Warbler was recorded in three habitats, and Northern Parula was recorded in two habitats (Table 1). Abundance of seven of the nine transient species differed among habitats (Table 1), and six of these seven species were, on average, most abundant in forest-agricultural edge or suburban forest or both. With the exception of Blackpoll Warblers, we recorded fewer transients in mature-forest interior. Most species were less abundant in younger forest habitats (pole-stage and shrub-sapling-stage), though Least Flycatcher and Magnolia Warbler were somewhat common in shrub-sapling-stage forest. Three of the seven species that differed in abundance among habitats exhibited a significant interaction between habitat and year, including Nashville Warbler (χ^sup 2^ = 18.9, df = 8, P

Group abundance and species richness. - Abundance of transient Neotropical migrants that breed in mature forests differed among habitats (χ^sup 2^ = 43.8, df = 4, P

Species richness of transient Neotropical migrants that breed in mature forests differed among habitats (χ^sup 2^ = 29.7, df = 4, P

Leaf development and migrant abundance.-The extent of canopy and subcanopy leaf development at mature-forest stopover sites was largely unassociated with site-use patterns of mature-forest-breeding transients within a visit interval (~1 week). However, annual differences in leaf phenology resulting from spring weather patterns were consistent with year-to-year patterns in the habitat use of mature-forest-breeding transients.