Migration strategies of sylviid warblers: chance patterns or community dynamics?

The effects of community dynamics in birds on the optimisation of their mig ratory strategies is a neglected area. For three years, we captured migrati ng warblers on autumn passage at a coastal site in western Britain. We used canonical correspondence analysis (CCA) to assess spatio-temporal patterns of occurrence, and principal components analysis (PCA) to assess morpholog ical variation. We calculated Euclidean distance in ordination and morpholo gical space to assess separation between species pairs. and used Monte-Carl o simulations to assess the probability of pattern occurring by chance. Ordination revealed five species-groups separated by habitat type and time of passage. Reed Warbler Acrocephalus seirpaceus and Sedge Warbler A. schoe nobaenus (Group I) occurred in wet habitats and peaked simultaneously. In d rier habitats with scrub, a first wave of Blackcap Sylvia atricapilla (Grou p 2) significantly preceded Grasshopper Warbler Locustella naevia. Willow W arbler Phylloscopus trochilus, Whitethroat Sylvia communis and Lesser White throat Sylvia curruca (Group 3). which in all but one case (Lessor Whitethr oat) significantly preceded Garden Warbler Sylvia borin (Group 4); peak num bers of Chiffchaffs Phylloscopus collybita and a second wave of Blackcaps ( Group 5) occurred later still. Age effects were found only in Acrocephalus with adults peaking before juveniles. For seven out of eight pairings within genera, separation in time of passag e increased significantly in species that were morphologically similar. The only exception was Blackcap and Lesser Whitethroat which differed substant ially in both passage time and morphology. Monte-Carlo simulations showed t hat chance was unlikely to be responsible for ordination patterns, nor for inter-specific variation in passage time and its relationship with species morphology. These data provide annually consistent evidence that migrating sylviid warb lers are separated ecologically by habitat use, time of passage and morphol ogy: we cannot refute the hypothesis that community dynamics have influence d niche use and autumn migratory strategy. We call for further tests of the 'migrant interaction' hypothesis in other geographical locations and tars, particularly where migrants are allopatric and interact ecologically only an migration.