Golden-wings are habitat specialist. Dunn and Garrett (1997) state they prefer the 'early stage of successional growth.' Curson (1994) lists 'pastures that are reverting to woodland, in brushy fields, openings in deciduous woods and shrubby borders of streams.' Both Potter et al (1980) and Dunn and Garrett (1997) list powerline right-of-ways as breeding areas, but Dunn and Garrett go on to record 'wet fields with clumps of shrubs, swamp edges, and young stands of oak-hickory' as potential breeding sites.

Nests are located on or close to the ground, concealed by dead leaves, weeds, ferns, briars or tall grasses. (Curson et al 1994, Dunn and Garrett 1997, Ehrlich et al 1988, Hammel 1992, Potter et al 1980) Golden-wings forage low to middle levels, often gleaning leaves, twigs, and dead leaf clusters especially of deciduous shrubs and saplings. (Curson et al 1994, Dunn and Garrett 1997, Hammel 1992) Potter et al (1980) however states Golden-wings 'frequently feed very high in tall trees, foraging at the tips of branches . . .'

NATURE SERVE GLOBAL HABITAT COMMENTS:

BREEDING: Deciduous woodland, usually in dry uplands or areas of thick undergrowth in swampy areas; woodland edge with low cover; hillside scrub; overgrown pastures; abandoned farmland; powerline right-of-ways; recently logged sites; bogs; forest openings; territories usually have patches of herbs and shrubs, sparse tree cover, and a wooded perimeter (Confer 1992). Habitat tracts of 10-15 ha can support several pairs and are preferred over both smaller and larger areas (Confer 1992). Habitat can be created through logging, burning, and intermittent farming (Confer 1992). Habitat is ephemeral and requires periodic disturbance to return it to favorable early successional conditions. Nests on or a little above ground, in grass tuft, fern or weed clump, or concealed in herbage at base of shrub, tree, ferns, briars, or goldenrod (Harrison 1978, Confer 1992). Often the clump includes a taller stem used for descent to the nest. Nests usually at the ecotone of a forest with a field or marsh, or in a small opening in a forest (Confer 1992).

Nested abundantly in the chestnut-sprout (CASTANEA DENTATA) forests of West Virginia following the spread of the chestnut blight (Hall 1983). Commonly nest in upland sites on abandoned farmland in early stages of succession (e.g., Confer and Knapp 1981), or occasionally in logged areas (e.g., Will 1986). In the Canadian shield in Ontario, they nest '...in alder [ALNUS spp.] bogs, especially when a few taller species [of trees] are present' (Mills 1987). Several observers have mentioned nesting in powerline right-of-ways. In southern Michigan they nested in and around the edges of thickly wooded portions of tamarack (LARIX LARICINA) swamps as well as in small, brushy clearings (Will 1986). In northern Michigan, Will described their habitat as including dry fields overgrown with shrubs, and woodland clearings, as well as very wet areas that were recently logged and covered with felled trees and a homogenous cover of new saplings. Will suggested that, overall, they '...appeared to require proximal access to mature or second-growth woodlands as well as open areas in which there has been considerable invasion by brush, shrubs, and sapling trees.'

Vegetative characteristics of territories have been quantified for southern, central, and northern New York (Confer and Knapp 1981, Frech and Confer 1987). In southern New York and contiguous New Jersey, nesting takes place in the Ramapo Mountains (Confer and Knapp 1981, Skully, in press). In this rugged topography, territories occurred in marshes between rock outcrops often with a perimeter of alder surrounded by forest. In central and northern New York, territories usually were located on dry, upland sites of abandoned farmland but occasionally in wet sites. All territories had areas with dense herb growth without shrubs or trees. Herb growth of at least moderate density covered 60% or more of the ground, including the growth under woody plants. All territories had patches of dense shrubs which covered about half of each territory. Tree canopy covered less than 15% of the northern and central territories but up to 40% of the southern territories. Central and northern territories usually extended no more than 20 m into a forest, while southern territories frequently extended considerably further. In wetter sites sedges (CAREX spp.) were the dominant herb and alders were the dominant shrub. In upland sites a wide variety of herbs occurred while VIBURNUM spp., narrow-leaved meadowsweet (SPIREA ALBA), and dogwood (CORNUS spp.) were the dominant shrubs.

All New York territories had a similar vegetative pattern with patches of herbs and shrubs, a few trees scattered throughout, and a tree row or forest edge forming most of the perimeter (Confer and Knapp 1981, Frech and Confer 1987). In New York, abandoned farmland undergoing secondary succession has this distinctive pattern of vegetation for only about 10-20 years. Thus, golden-winged warblers at upland sites are restricted to a specific and brief stage of succession. Because of this restriction, Confer and Knapp (1981) suggested that this warbler was in some sense a habitat specialist. However, a species that can nest in chestnut-sprout forests in Virginia, tamarack bogs in Michigan, and alder swamps in Ontario clearly tolerates a wide range of conditions. It would be valuable to determine if nesting warblers require a specific plant profile but tolerate a wide range of plant species, or tolerate a wide range of both plant profile and plant species.