The study area lies within a 15 × 30 km quadrat (46^o50’N, 71^o50’W) in a forested landscape near Quebec City, Canada. This area is characterized by a hilly topography, with a mean annual temperature of 2.5°C and a mean annual rainfall of 1000 to 1600 mm (Robitaille and Saucier 1998). It is typical of the balsam fir-yellow birch bioclimatic domain, which is a heterogeneous mosaic of three forest types. Half of the study area is covered by mixedwood stands consisting mostly of yellow birch and balsam fir with a few stems of sugar maple ( Acer saccharum Marsh.) and red spruce ( Picea rubens Sarg.). The coniferous forest type covers approximately 25% of the area and is composed mostly of balsam fir, red spruce, and a few white birches ( Betula papyrifera Marsh.). Similarly, the deciduous forest type extends over the remaining 25% of the study area and is dominated by sugar maple and yellow birch, with red maple ( Acer rubrum L.) and American Beech ( Fagus grandifolia Ehrh.) as companion species. Most of the study area has been submitted to partial cuts or clearcuts over the last century. Nevertheless, a good proportion of well regenerated and mature stands persists through the area.

We used the typology reported on provincial forest maps to select 64 observation plots within mature coniferous, deciduous and mixedwood stands. Because forest maps are not sufficiently accurate at the local stand scale (Potvin et al. 1999, Dussault et al. 2001), we validated this first classification in the field using the provincial guidelines for ecological classification (Gosselin 2002). As each observation plot (100 m radius, about 3 ha) fit within stands (minimum area of 3 ha), only one forest type was associated with each plot.

In June 2000 and 2001, we observed eight species of birds: Black-capped Chickadee ( Poecile atricapillus ), Bay-breasted Warbler ( Dendroica castanea ), Blackburnian Warbler ( D. fusca ), Black-throated blue Warbler ( D. caerulescens ), Black-throated green Warbler, Golden-crowned Kinglet ( Regulus satrapa ), Magnolia Warbler ( D. magnolia ), and Red-eyed Vireo ( Vireo olivaceus ). Observations were almost exclusively performed on males because they are easy to locate when they sing and because females were mostly incubating during the study period. We made an exception for Chickadees because of the lack of sexual dimorphism in that species. Observations were conducted in the morning, between 4:00 and 10:00 EDT, only when the weather was appropriate (no rain; wind speed < 25 km/h). We first scanned the observation plot (100 m radius) to locate target species. Located birds were focally tracked, the observer focusing only on one bird at a time (Lehner 1979). Only one individual per species per point and per year was studied. This restriction was applied to ensure the independence of the data since the studied birds were not banded. We tracked birds for a minimum of three trees and a maximum of twelve trees, recording the time that the bird spent on each. We considered as negligible observer disturbance on bird behaviors since we waited few minutes after our arrival in observations plots to start tracking birds and because we were moving quietly and following birds within a distance of at least 5 m. To avoid detection bias, we did not consider the tree in which birds were first seen. We recorded bird activity on each tree (singing, foraging, singing and foraging, moving, resting, alerting) and the time of the observation. Data and commentaries were taped in real time and then transcribed in a data file.

In 2000, we inventoried trees within each observation plot to evaluate tree availability. We identified the species of each tree (alive or dead) within a 10 × 20 m plot and measured the diameter at breast height (with a minimum of 1 cm). We calculated the tree basal area of each species, then the total basal area of all coniferous and deciduous trees.

Birds from same species tracked in a same observation plot but during different years were considered as independent observations. We assumed that site fidelity was not high enough to expect, in the repeated points, observation of the same individuals each year. We used SAS 8.2 (SAS Institute Inc 1999-2001) to perform all the analyses.

Almost all species used both coniferous and deciduous trees for foraging and singing within the mixedwood forest but several species preferences were observed. The use of both deciduous and coniferous trees has also been mentioned in studies mostly performed in habitats dominated by deciduous or coniferous trees for Blackburnian Warbler, Black-throated Green Warbler, Black-capped Chickadee, Magnolia Warbler and Black-throated Blue Warbler (MacArthur 1958, Morse 1976, Holmes and Robinson 1981, Holmes 1986, 1994, Whelan 2001). However, the use of deciduous trees by Bay-breasted Warbler and Golden-crowned Kinglet was unexpected, since these species are known to be strongly associated with coniferous stands (MacArthur 1958, Morse 1976, Williams 1996a, Inglod and Galati 1997). Nevertheless, Sabo (1980) observed that the foraging "niche" of these two bird species included a certain volume of broadleaf foliage, which is consistent with our observations.

Among the species studied, only the Red-eyed vireo restricted its activities to deciduous trees. This observation is in accordance with previous observations made on this species in forests dominated by deciduous trees (Holmes and Robinson 1981). Moreover, Holmes (1986) observed that the Red-eyed Vireo foraging niche at the Hubbard Brook Experimental Forest in New Hampshire included less than 2% of coniferous foliage.

No species showed preference indices that were significantly different between the studied forests, after accounting for multiple testing. Such results suggest that foraging preferences of no species counteracted changes in substrate availability, i.e., no species exhibited a rigid, habitat-independent, foraging profile. Instead, their choice of tree species in mixedwood forests was plastic, and not in accordance with an interpretation of mixedwoods as sub-optimal habitats from the foraging point of view, as often suggested in the literature (Morse 1976, Sabo 1980, Holmes 1986, Inglod and Galati 1997, Cimprich et al. 2000). The fact that Blackburnian Warbler used neither deciduous nor coniferous trees with constancy, but rather in proportion of their availability, provides evidence against foraging behaviour as a cause for their abundance in mixedwood forests in comparison with other forest types (Morse 1994, Gauthier and Aubry 1995, Girard et al. submitted). Blackburnian Warblers' foraging plasticity here also contrasts with accounts of its foraging behaviour described as "stereotyped" by Morse (1971), who interpreted the replacement of the Blackburnian Warblers by the Bay-breasted Warblers on small islands by the inflexibility of the former to markedly modify their foraging regime.

We offer two alternative explanations to the results obtained in this study. First, it is possible that the driving factor bringing some of these species into mixedwood forests is not their intrinsic need for both deciduous and coniferous trees but their behavioral plasticity (Lefebvre et al. 1997) that may confer them fitness advantages relative to strict conifer or deciduous specialists. Deciduous and coniferous trees have very different leaf and branch architecture. Such architectures force birds to use different chasing manoeuvres and approaches (Holmes and Robinson 1981, Parrish 1995a, Whelan 2001). In a mixedwood forest, adaptability skills could decrease the waste of energy and time spent on searching for the right type of tree since both tree types may be adequate. Whelan (1989) showed that temporary changes in bird foraging preferences are usually conditional to sufficient benefits. We believe that such benefits may also motive the development of plastic behaviours.

A second explanation to the apparent lack of preference and constancy by the species may be inter-specific competition, which may influence the behavior of subordinate birds as well as that of dominant birds (Alatalo et al. 1987). Birds sometimes change their foraging behaviour apparently to reduce their interactions with others (Inman et al. 1987). In the past decades, studies have focussed on the relationships among several of the species that we have studied here. Among others, it is known that Magnolia Warbler socially dominates Black-throated Green Warbler, which in turn dominates Blackburnian Warbler (Morse 1993). These two latter species were observed being chased by Golden-Crowned Kinglet, as was Black-capped Chickadee (Inglod and Galati 1997). The Black-throated Blue Warbler is not known to interact with the other seven bird species studied, but it does with American Redstart ( Setophaga ruticilla ) and Ovenbird ( Seiurus aurocapillus ) (Holmes 1994), both present in our study area. Similarly, Bay-breasted Warbler and Red-eyed Vireo were both known not to frequently interfere with the other studied species, the former interacting only during spruce budworm outbreaks (the studied year corresponds to an endemic period) (Williams 1996a) and the latter usually interfering with its family members, as Philadelphia Vireo ( Vireo philadelphicus ) (Robinson 1981). Given the subordinate status of Blackburnian Warbler and to a lesser extent of Black-throated Blue Warbler and of Black-throated Green Warbler, they may have been excluded from the optimal habitats occupied by dominant species, the deciduous or the coniferous forests, and forced to use the suboptimal habitat of their dominant species, the mixedwood forest. Moreover, this may have driven them to develop adaptation skills that give them today the best fitness in mixedwood forest. Morse (1971) and Pierce and Grubb (1981) mentioned that subordinate birds could be more plastic in their foraging behavior and occupy larger niches than dominant ones.

The majority of the species present in mixedwood forests exhibit preferences for either deciduous or coniferous trees, but they do accommodate to the non-preferred tree type for foraging and singing. We did not find any evidence for dependency on the proximity of deciduous and coniferous trees, i.e. a requirement for both deciduous and coniferous trees in similar proportions across the three forest types, nor did we find strong dependency on deciduous or coniferous trees through all forest types for a given species. Consequently, we suggest that other processes such as behavioral plasticity or inter-specific competition may explain the presence of Bay-breasted Warbler, Blackburnian Warbler, Black-capped Chickadee, Black-throated Blue Warbler, Golden-crowned Kinglet, Magnolia Warbler, and Red-eyed Vireo in the mixedwoods.

Our inferences are valid only for males since we did not track females. Moreover, they only considered a small sample of the numerous factors defining niche of species; we did not study the tree type use for nesting, nor did we evaluate insect availability, two factors that may also influence tree preference. Thus, our study is one of many steps that will be required for understanding whether and how microhabitats attract species to mixedwood forests.