Long-distance migrant passerines are well known to often display high levels of philopatry to breeding and wintering grounds. One could expect that similar selective pressures and similar navigation skills would result in their being faithful to stopover sites, a pattern that has been described for several populations of migratory waders and waterfowl. In this paper, we develop the argument that passerines should suffer from higher costs and receive lower benefits from stopover site faithfulness than waterfowl and waders. Based on Alerstam's (1979)“optimal drift strategy” theory and other considerations, we predict that passerines should have lower stopover site fidelity than geese and waders, and that site faithfulness should decrease with increasing distance from either end of the migratory journey. We present results from a long-term study on the stopover ecology of migrant passerines in southern Portugal that support these predictions and show that, for species and populations that neither nest nor winter in this country, few individuals are faithful to the stopover site. On the other hand, populations that included individuals at (or near to) the start or the end of the migratory journey, had much higher return rates. We could not find any evidence that species linked to scarce habitats, such as wetlands, were more site faithful. Our results are in agreement with several other reports, but in apparent contrast to previous conclusions resulting from some studies involving Old World warblers. These differences are discussed and it is argued that there is no solid evidence to suggest that site fidelity should be important for passerines at stopovers far away from breeding or wintering grounds, meaning that there is a large within-individual variability in the precise migratory routes and stopover sequence used each year.