Migratory movements of the American Redstart across North America.
This animation depicts weekly relative abundance estimates, with brighter color representing higher abundance. The data for this animation were generated using Spatio-Temporal-Exploratory Models to predict population relative abundance at specific times and locations by relating observations of American Redstarts from eBird to local environmental features derived from NASA remote sensing data. eBird (eBird.org) is a citizen science program run by the Cornell Lab of Ornithology.
American Redstart distribution in BirdVis
Use our distribution analysis tool to view predicted relative abundance in specific areas of the
American Redstart range.
Figure 3. Annual cycle of molt, breeding, and migration of the American Redstart.
Annual cycle of molt, breeding, and migration of American Redstart in ne. U.S. Thick lines show peak activity, thin lines off-peak.
Figure 4. Relative abundance of the American Redstart in U.S. and s. Canada, 2007-2013.
Relative abundance of American Redstart in U.S. and southern Canada, 2007–2013, based solely on data from Breeding Bird Survey. Numbers shown are average number of individuals detected per route per year. See Sauer et al. (2014) for details.
Figure 6. Migratory connectivity of American Redstart populations between summer and winter.
The distribution of the most likely breeding region (NW, Northwest; MW, Midwest; NE, Northeast; CE, Central-east; SE, Southeast) for individuals at each wintering region (M, Mexico; C, Central America; W, Western Greater Antilles; E, Eastern Greater Antilles; L, Lesser Antilles/South America). Black dots indicate sampling locations and bars indicate the proportion of individuals assigned to each breeding region (rounded to the nearest 5%). Figure from Martin et al. 2007, adapted from Norris et al. 2006a.
Representative breeding habitat of American Redstarts.
A northern hardwood forest at Hubbard Brook in New Hampshire. Dominant trees are sugar maple, American beech and yellow birch, with an understory of shrubs, ferns, and tree seedlings and saplings.
© Sara Kaiser, New Hampshire, United States, 25 May 2012
Shaded coffee habitat in Jamaica.
Shaded coffee habitat in Jamaica, where redstarts are abundant in winter. Shade trees are
Inga vera, a nitrogen-fixing legume that helps fertilize the soil and improve quality of the coffee by ripening the beans more slowly than in sun coffee.
© Thomas Sherry, Clarendon, Jamaica, 4 November 1997
Old-growth black mangrove swamp at Luana Point study site in Jamaica.
Black mangrove trees have distinctive pneumatophores, aerial roots emerging vertically above ground from the root zone and important in gas exchange. This habitat contains the highest density of redstarts documented to date in winter, and they defend territories and forage at all levels in this habitat from the pneumatophores to treetop. Photo documents old-growth black mangrove habitat at site prior to major damage from Hurricane Ivan in September, 2004.
© Thomas Sherry, Saint Elizabeth, Jamaica, 2 November 1986
Logwood coastal thorn scrub habitat at Luana Point study site in Jamaica, during the wet season.
Haematoxylum campechianum) coastal thorn scrub habitat at Luana Point study site in Jamaica, during the wet season (May through December?). These dense, thorny thickets contain abundant redstarts in winter, although this habitat can lose most of its leaves in typical dry seasons due to drought stress (see adjoining photo of same habitat), which can cause declines in food abundance with detrimental effects on redstarts.
© Thomas Sherry, Saint Elizabeth, Jamaica, 20 October 1991
Logwood coastal thorn scrub habitat at Luana Point study site in Jamaica, at peak of dry season.
Haematoxylum campechianum) coastal thorn scrub habitat at Luana Point study site in Jamaica, at peak of dry season (March?), showing almost complete absence of leaves resulting from drought stress. See adjoining photo of this habitat prior to peak of drought.
© Thomas Sherry, Saint Elizabeth, Jamaica, 26 March 1987
Tropical dry forest at Portland Ridge, Jamaica, in the rain shadow of the Blue Mountains.
Redstarts are abundant in this habitat following Fall migration and early parts of the winter, but they often abandon this habitat later in the winter season as it undergoes drought stress and declining insect abundance during the dry season. This photo documents the condition of this forest site prior to a fire in 2005 (following hurricane Ivan damage in September 2004 that left abundant dead trees and branches), which changed the vegetation substantially. Photo by R. T. Holmes.
© Thomas Sherry, Clarendon, Jamaica, 29 October 1992
Distribution in the Americas
Figure 1. Breeds across much of the eastern and northern U.S. and southern Canada. Breeds north to Newfoundland, southern Labrador, south-central Quebec, north-central Ontario, central Manitoba, north-central Saskatchewan, northeastern Alberta, west-central Mackenzie, and extreme southwestern Yukon. Breeds south locally along the Atlantic Coast to southeastern Virginia, central North Carolina, western South Carolina, northern and western Georgia, extreme northwestern Florida, southern Alabama, southern Mississippi, southern Louisiana; easternmost portions of Texas, Oklahoma, and Kansas; southeastern and northern Nebraska; southeastern and northwestern Wyoming; northern Idaho; northwestern Oregon; northern Washington; west-central British Columbia (not including coastal lowlands), and extreme southeastern Alaska ( ). Breeding distribution is patchy. Very small numbers also breed occasionally along eastern Cascade Mountains in southern Oregon ( American Ornithologists' Union 1983 ), on coastal slope of Humboldt and Del Norte counties in northeastern California ( Gilligan et al. 1994 ), in southwestern Idaho ( Small 1994 ), northeastern Utah and southwestern Wyoming ( Stephens and Sturts 1991 ), southern Apache County in east-central Arizona ( Behle et al. 1985 ), north-central Colorado ( Monson and Phillips 1964 ), in coastal plain of central South Carolina ( Andrews and Righter 1992 ), and locally elsewhere. Most abundant breeding from central British Columbia, southern Manitoba, southern Ontario, southern Quebec, Nova Scotia, and northern portions of Minnesota, Wisconsin, and Michigan, east within the U.S. through northern New York, northern Vermont, New Hampshire, and northern Maine (see McNair and Post 1993b Figure 4).
Local and often absent as breeder along immediate coast of eastern North America from Rhode Island south (but abundant breeder on Nashawena Island, Buzzards Bay, Massachusetts; TWS). Absent as breeder from southeastern Alberta and much of southern Saskatchewan except Cypress Hills (
, Semenchuk 1992 ); southwestern Minnesota and portions of central South Dakota and North Dakota ( Smith 1996b , Stewart 1975b , Janssen 1987 ); and much of central and western Kentucky, and Tennessee, and southeastern Indiana ( Peterson 1995 , Indiana Breeding Bird Atlas unpublished), and locally elsewhere (see Historical changes, below). Absent as breeder throughout most of central and southern plains states, the arid Southwest, and southern and eastern California. However, nonbreeders occur in small numbers during summer in New Mexico, southern Arizona (especially in southeast), and northwestern California ( Palmer-Ball 1996 , Hubbard 1978c , Monson and Phillips 1964 ). Casual records in summer include northern Quebec, northern Mackenzie, and Banks Island and Greenland ( Small 1994 , American Ornithologists' Union 1983 ). Some individuals apparently remain on overwintering grounds ( Cramp and Perrins 1994b ), in some cases to breed, on basis of 3 records of American Redstarts nesting (April–July) in Cuba ( Stewart 1989a ). Kirkconnell and Garrido 1996 Overwintering Range
Figure 1 . Overwinters generally from western coast of northern Baja and from throughout southern Baja California, mainland Mexico (central Sinaloa on Pacific slope, central Veracruz on Atlantic slope, and Chiapas in interior), extreme southern Florida, Bermuda, and Bahama Islands south throughout Middle America (including offshore Caribbean islands), and throughout Caribbean (including all Lesser Antilles islands) to northern South America, including Venezuela, Colombia, eastern Ecuador, and northwestern Brazil east of Andes, and to northwestern Ecuador west of Andes, and all islands from Netherlands Antilles east to Trinidad and Tobago ( , American Ornithologists' Union 1983 , Evans 1990a , Amos 1991 , Arendt 1992 , Stevenson and Anderson 1994b ). Greatest overwinter abundance occurs in Caribbean, especially Greater Antilles and western Mexico ( Howell and Webb 1995 , Pashley and Martin 1988 ), but individuals can be locally abundant in Yucatán Peninsula ( Arendt 1992 ), Belize ( Lynch 1992 , Pashley and Martin 1988 ), and northern Venezuela in mangroves and other moist habitats ( Robbins et al. 1992a , Schwartz 1964 ). Lefebvre et al. 1992
Rare but regular during winter in the U.S. in Gulf states from Texas and Louisiana east to Florida (becomes increasingly numerous in Florida southward and near coasts;
); in Imperial Valley (Salton Sea) and along central and southern coasts of California ( Stevenson and Anderson 1994b ); in lower Colorado River valley of southeastern California and southwestern Arizona ( Small 1994 ); and in southern Arizona ( Small 1994 ). Monson and Phillips 1964
Casual overwinter records include Cocos Island (Costa Rica;
, TWS), and Clipperton Island and Revilla Gigedo Island, Mexico ( Slud 1967 ), coastal Oregon ( American Ornithologists' Union 1983 ), South Carolina ( Gilligan et al. 1994 ), and occasionally elsewhere north of regular overwintering range. Post and Gauthreaux 1989
Distribution Outside the Americas
Extralimital sightings (October–December) in Iceland, Britain, Ireland, and France, and over water near Azores (
, American Ornithologists' Union 1983 , Lambert et al. 1985 , Alström and Colston 1991 ). Cramp and Perrins 1994b
Neotropical–Nearctic migrant; breeding and overwintering ranges do not overlap (see
Figure 1), with only a few exceptions (See Breeding Range). A handful of long-distance band recoveries suggest a tendency to overwinter directly south of breeding areas; i.e., individuals breeding in eastern North America travel to Greater Antilles and northern South America, whereas those breeding in central and western North America migrate to Mexico and Central America ( , Stewart 1989a ). This pattern was essentially confirmed, and extended, by more recent studies using stable-hydrogen isotopes in feathers, sampled at 26 sites in 11 countries on the overwintering grounds, and linked to underlying hydrogen isotope maps for North America ( Holmes and Sherry 1992 ; see also Norris et al. 2006b Demography and Populations: Range). Additionally, demonstrated a “chain migration” pattern for American Restarts that breed in eastern North America; northern breeders tend to overwinter farther north, and southern breeders overwinter farther south; e.g., birds overwintering in Trinidad and Tobago breed in the southeastern U.S. (Figure 6). Norris et al. 2006b
Timing and Routes of Migration
Fall. Departs breeding areas beginning in July (e.g., Hubbard Brook Experimental Forest, White Mountains, New Hampshire; TWS) and arrives on northern Gulf of Mexico coast as early as late July ( , Lowery 1974 , Oberholser 1974c , Stevenson and Anderson 1994b ; see Figure 3). Most individuals, however, move south later than this, arriving south of U.S. by November. Peak movements occur from late August to mid-September in Minnesota ( Woodrey 1995 ), Wisconsin ( Janssen 1987 ), Lower Peninsula of Michigan ( Robbins 1991 ), and Massachusetts ( Granlund et al. 1994 ); throughout September and early October in Ohio ( Veit and Petersen 1993 ); and in first week of September in Cape May, New Jersey ( Peterjohn 2001 ). Most birds pass through the southern U.S. by October, peaking in numbers in the last week of September and first half of October (e.g., inland Louisiana; J. V. Remsen, unpublished data). Last stragglers depart the U.S. for overwintering areas by November, except for a few individuals that overwinter, e.g., in Florida peninsula ( Sibley 1997 ) and southern Louisiana ( Stevenson and Anderson 1994b , J. V. Remsen, unpublished data). Median 1990–1991 passage (capture) dates of migrant American Redstarts at Long Point (southern Ontario), Powdermill (southwestern Pennsylvania), and Fort Morgan (coastal Alabama) banding stations, respectively, were 1 September, 10 September, and 24 September (interpolated from data in Lowery 1974 ). Median migration capture dates were 30 August (range 16 August–24 September, Woodrey 1995 n = 527) in a fragmented forest site in Minnesota ( , and 5–10 September in northern West Virginia ( Winker et al. 1992d n = 192; ). Farther west, American Redstarts migrate through Alberta primarily from late August to early September ( Hall 1981 ), Kansas in late August to third week of September ( Semenchuk 1992 ), Oregon in first 3 week of September ( Thompson and Ely 1992 ), and California early September–mid October ( Gilligan et al. 1994 ). Migration through Bermuda lasts from mid-September (occasionally as early as August) to end of October ( Small 1994 ) and in Costa Rica from mid-August to late October ( Amos 1991 ). Overwintering residents (as opposed to transients) arrive on territories in Jamaica beginning mid-September (P. P. Marra, unpublished data). Stiles and Skutch 1989
Adult redstarts arrive synchronously with hatch-year (juvenile) birds at banding stations in Ontario, Pennsylvania, and West Virginia (
); but yearlings arrived earlier than older individuals in 1 of 3 years studied in coastal Alabama ( Hall 1981 , Woodrey and Chandler 1997 ). Yearlings arrive on overwintering sites in Jamaica before older birds (P. P. Marra, unpublished data), suggesting that yearling individuals progress to overwintering sites on average more rapidly than older birds, or at least leave Gulf Coast stopover sites earlier. Adults probably remain later on or near the breeding area to complete Definitive Prebasic Molt (see Woodrey and Moore 1997 Appearance: Molts).
Spring. After spending up to 7 mo on overwintering grounds ( ; P. P. Marra, unpublished data), starts departing in late March, traveling to North America on broad front. More southerly breeding individuals pass through coastal Louisiana during spring migration earlier than northern breeders, based on stable-hydrogen isotope ratios in feathers, a method which also confirmed that western breeders do not migrate through the state ( Schwartz 1964 ). Departs Central America (e.g., Costa Rica) by late March ( Langin et al. 2009 ). Reaches Florida and Gulf Coast, arriving over water from both Yucatán Peninsula and Greater Antilles, beginning in late March ( Bent 1953b , Bent 1953b , Lowery 1974 ). However, most individuals do not depart Jamaican study sites until mid-April, continuing to depart through early May ( Langin et al. 2009 ); correspondingly, peak migration passes through Florida in last week of April and first week of May ( Studds and Marra 2011 ), and through Louisiana late April–early May (J. V. Remsen, unpublished data). Median spring passage date through coastal Louisiana was 28 April (range 31 March–13 May, Stevenson and Anderson 1994b n = 21 yr; ). The front moves gradually north, reaching Kentucky by late April ( Langin et al. 2009 ); Kansas by late April to mid-May ( Palmer-Ball 1996 ); and New Jersey, Ohio, and southern Michigan by first week of May ( Thompson and Ely 1992 , Sibley 1997 , Peterjohn 2001 ). Peak movements through northern U.S. occur in mid-May, including Massachusetts ( Granlund et al. 1994 ), Hubbard Brook Experimental Forest, New Hampshire (TWS), Minnesota ( Veit and Petersen 1993 ), Wisconsin ( Janssen 1987 ), and Alberta ( Robbins 1991 ). Median passage date at forest fragment in Minnesota was 19–20 May (range 5–28 May, Semenchuk 1992 n = 94; ). Farther west, arrives in coastal California beginning late April ( Winker et al. 1992d ). Passes through Oregon in late May and early June ( Small 1994 ). In Michigan, New England and southern Canada, older males arrive 7–10 d on average before females and yearling males ( Gilligan et al. 1994 , Ficken and Ficken 1967 , Francis and Cooke 1986 , TWS). Hahn and Silverman 2006
Timing of spring departure from overwintering grounds extensively studied in Jamaica (
, Studds and Marra 2007 , Studds and Marra 2011 , McKellar et al. 2013a , Tonra et al. 2013 ). Although increasing photoperiod stimulates the transition to migratory preparation in many migratory birds ( Cooper et al. 2015 ), environmental factors can modify this stimulus significantly in American Redstarts ( Ramenofsky 2012 , Studds and Marra 2011 , Studds and Marra 2012 , Tonra et al. 2013 ). Individual birds depart earlier in wetter springs and arrive earlier on breeding grounds, a pattern more or less consistent in eastern and western breeding populations around 3,000 km apart, although some differences found between these extremes, e.g., in calendar dates and strength of relationship ( Cooper et al. 2015 ; see McKellar et al. 2013a Demography and Populations: Range: Population Connectivity). Correspondingly, birds that winter in wetter habitats, at both coastal and higher elevation Jamaican sites, also tend to depart from Jamaica earlier in spring migration and arrive earlier in breeding areas, supporting the inference that wetter habitats in winter have more food and are thus of higher quality ( , Marra et al. 1998 , Reudink et al. 2009a , Tonra et al. 2011 ). Natural variation in local rainfall and food abundance on winter grounds can delay spring departure by 3–5 d, which may lead to later arrival in breeding areas and subsequent reduced breeding success ( Peele 2015 , Studds and Marra 2011 ). This inference about food controlling departure schedules is supported in older (definitive basic plumage) males by experimental food-reduction in high quality (black mangrove) Jamaican habitat: Food-reduced birds (insect abundance reduced by nearly 80% on winter territory) compared to controls increased subcutaneous fat (thought to be insurance against unpredictable conditions), but lost muscle mass and probably as a consequence departed about a week later in spring migration ( Studds and Marra 2012 ; see Cooper et al. 2015 Control and Physiology of Migration). Individually-banded redstarts arriving to breed in Ontario showed similar response to winter rainfall between years, i.e., similar timing vs. rainfall slope relationships, suggesting that annual changes in departure and arrival times in relation to changing rainfall are likely phenotypically plastic ( ). Migration timing is also plastic in relation to ecological conditions during migration, as evidenced by individuals arriving to breed in Michigan later during a cold, low-resource arrival period ( McKellar et al. 2013a , TWS). Smith et al. 2009b Routes
In eastern North America, migrates in fall primarily along coastal plain between the Appalachian Mountains and western Atlantic Ocean, on basis of (1) relatively large abundances and percentage of adults caught in coastal plain (rather than mountain or coastal) banding stations, (2) presence of many migrant individuals passing through Florida, and (3) scarcity of older individuals in Bermuda captures (
). Large numbers of juvenile birds, possibly disoriented, are found along eastern Atlantic Coast late August–early October ( Ralph 1981 , Ralph 1981 , Stewart 1986 ) and at sea ( Morris et al. 1996 ), suggesting that this coastline forms eastern edge of migratory route. Almost 97% of redstarts captured in a coastal Virginia study area were juvenile birds ( Ralph 1978 n = 1,057), suggesting that the coastal zone is not just a barrier but also an important migratory pathway for these first-time migrators ( , Stewart 1983b ). Calculations based on fat load also suggest that this species is not capable of long-distance flights over the western Atlantic Ocean, unlike the Blackpoll Warbler ( Stewart 1986 Setophaga striata; ). On basis of relatively large proportions of older birds captured at a site in fall, coastal Alabama (24% adults; Morris et al. 1996 ) and mountains of northern West Virginia (29% adults; Woodrey 1995 ) are probably closer to main southward migratory path than are coastal eastern U.S. stations (< 10% adults). More birds move south in western Florida than in eastern Florida; conversely, more move north in eastern Florida for spring migration, indicating seasonal differences in migratory path within that state ( Hall 1981 ). Stevenson and Anderson 1994b
Abundant individuals passing through the western U.S. indicate broad migratory front in both fall and spring (
), generally in relatively lowland areas, including eastern Kansas ( Curson et al. 1994 ), eastern Colorado ( Thompson and Ely 1992 ), and South Dakota ( Andrews and Righter 1992 ). In Oregon and California, most individuals move along the coast, west of Cascades and Sierra Nevada ( Peterson 1995 , Gilligan et al. 1994 ), but enough move along eastern lowlands of the Sierra Nevadas to suggest a regular migratory path ( Small 1994 ). California records indicate that the American Redstart is one of the most common “eastern” wood-warblers found in coastal California in fall—about 185 individual birds recorded annually as of 1980 ( Small 1994 , Pyle et al. 1987). The well-marked movements through Pacific coastal states probably connect breeding sites as far west as British Columbia, Washington State, Oregon, and northern California (see Figure 1) with overwintering sites in much of Central America and southern Mexico. Vagrants show up on all California offshore islands ( Roberson 1980 ), Revilla Gigedo Island and Clipperton Island off Pacific coast of Mexico ( Small 1994 ), and Cocos Island, 480 km off Pacific coast of Costa Rica ( Howell and Webb 1995 , TWS). Slud 1967
Analyses of eBird data (observations of individual birds by citizen scientists throughout the Western Hemisphere, 2002–2014 [
]) averaged for each species indicated that American Redstarts, like many other songbird species that migrate over the Gulf of Mexico or Atlantic Ocean, make “clockwise looped trajectories, which result in faster but more circuitous journeys in the spring and more direct journeys in the autumn”. This loop migration pattern results from northward migration in spring taking place further west of the southward migration path used in fall, and appears to be driven in part by seasonal patterns of productivity of insect prey, and atmospheric weather patterns including frontal systems ( La Sorte et al. 2013 ). Peak daily migration rate by species with clockwise migration pattern was greater in spring than in autumn, probably because of selection pressures to return to breeding grounds quickly (i.e., optimize migration, according to time-minimization hypothesis) plus slower movement rates of inexperienced juvenile individuals in fall ( La Sorte et al. 2014a ). Peak migration rate in American Redstart estimated at 44.4 and 32.5 km/d, respectively, in spring and fall, based on eBird data from 2007–2011 ( La Sorte et al. 2016 ). Redstarts migrate as part of an “eastern” (North American) flyway in Fall, and probably take advantage of a low-level jet stream in spring that is circuitous and somewhat to west of fall migration route but probably helps speed northward migration to breeding regions ( La Sorte et al. 2013 ). Although referred to as a “flyway”, this eastern migration pattern involves broad fronts of birds migrating within the region, primarily east of a line running between central Texas and the Mississippi River valley. La Sorte et al. 2014b
Intratropical migrations may occur in some non-breeding locations within the overwintering period. For example, abundance appears to shift between Pacific and Caribbean mangroves in Panama, possibly related to tracking of specific food resources (
). Lefebvre and Poulin 1996
As with most if not all parulid warblers, migration is nocturnal, on basis of abundant tower kills (see
Conservation and Management: Effects of Human Activity), and on identification of nocturnal flight-call notes ( ). Banded individuals known to migrate at rates of 66 km/d and 160 km/d ( Getty 1993 ), and 38 km/d ( Stewart 1989a ), often over water ( Robbins 1991 ). Often joins mixed-species flocks of chickadees, nuthatches, woodpeckers, and other warblers, particularly in fall, but also during inclement weather in spring ( Stewart 1989a , TWS). Tends to select treefall gaps more than closed forest while migrating through Illinois woodlots ( Morse 1970b ). Foraging behavior is variable and flexible during migration ( Martin and Karr 1986 , Martin and Karr 1990 ), as in other seasons (see also Woodrey 1995 Diet and Foraging: Feeding, and Distribution, Habitat and Migration: Spring and Fall Migration).
Stops en route in appropriate habitat and increases in mass (
, Morris et al. 1994 , Morris et al. 1996 ). During fall on Appledore Island in coastal Maine, American Redstarts (particularly lighter-weight individuals) fit general parulid stopover pattern of arriving lean, staying average of 2.7 d (range 1–18) to increase mass ( Woodrey 1995 , Morris et al. 1994 ). These behaviors allow rebuilding of body condition before continued long-distance southward flights (estimated 650–750 km total, on basis of fat levels); Morris et al. 1996 ). Juveniles in fall tend to arrive on Appledore Island with less fat and slightly lighter than older individuals (8.1 versus 8.3 g), and stay slightly longer (2.8 versus 2.4 d), consistent with hypothesis that younger individuals are less prepared for long-distance migration (see also Morris et al. 1996 ). Females also tended to arrive with lighter body mass and less fat than males, and to persist longer than males ( Woodrey 1995 ). Birds were less likely to be recaptured on Appledore in spring than in fall, and they stayed fewer days (2.6 d in spring; 3.4 d in fall), and correspondingly did not gain mass during spring stopover ( Morris et al. 1996 ). This greater extent of fattening during fall could result from earlier migratory stage in fall, whereas spring birds potentially benefit reproductively from rapid arrival on nearby breeding areas. Individuals weigh more in fall at an Alabama site, just before crossing Gulf of Mexico barrier, than at Appledore Island (9.4 versus 8.3 g in adults, Morris et al. 1994 n = 107, 346, respectively; 8.3 versus 8.1 in juveniles, n = 118, 1,515, respectively; , Morris et al. 1996 ); and unlike juveniles, Alabama adults in fall contain enough fat stores to fly the estimated 1,200 km across Gulf of Mexico (assuming calm air conditions; Woodrey 1995 ). Woodrey and Moore 1997
In spring migration, concentrated food comprised of abundant adult midges (Insecta: Chironomidae) emerging from Lake Huron in Michigan’s Upper Peninsula and spiders feeding on these midges create an important resource for stopover migrants (
): Individuals tended to glean more (relatively least energetically expensive feeding behavior—see Smith et al. 2007e ) in these food-rich habitats, where midges were most abundant in shoreline coniferous trees (primarily northern white cedar, Lovette and Holmes 1995 Thuja occidentalis) compared to more inland deciduous trees ( ); and consequently the birds tended to gain body mass in lakeshore stopover habitats ( Smith et al. 2004c ). Smith et al. 2007e
Possibly as a consequence of either individual quality and/or good feeding conditions during migration, American Redstarts arrive earlier on the breeding grounds and in better body condition, supporting the idea that ecological conditions during migration carry over into the breeding season (
). The importance of ecological conditions during migration is reinforced by tendency of redstarts to arrive earlier in Manitoba stopover site in warmer spring, and for earlier arriving individuals to be in poorer body condition as measured by relative body mass index ( Moore et al. 2005 ). However, the importance of ecological conditions during migration itself versus conditions in winter prior to the start of migration is not known (see González-Prieto and Hobson 2013 Demography and Populations: Population Regulation).
Control and Physiology of Migration
The Insurance Hypothesis (acquisition of body lipids in excess of that needed to complete migration) was tested using individual American Redstarts arriving in northern Michigan (
). Some support found, insofar as individuals of both sexes arrived with more body fat in a cooler spring, they tended to arrive with excess body fat, and those arriving early had more fat than those arriving later. However, evidence against this hypothesis consisted of birds arriving with excess fat even in a relatively benign spring period, and females arriving with more excess fat than males, suggesting that excess fat might play more of a reproductive than an insurance role. Smith and Moore 2005b
Some variation in body fat of migratory birds is determined, at least in part, prior to northward migration. For example, experimental food reduction on male territories in Jamaica prior to spring northward migration caused increased fattening but reduced pectoral muscle mass—and no net change of mass overall (
). These results indicate a tradeoff in fat deposition vs. muscle mass maintenance, and the experimental birds (with reduced food supply) were able to fatten in a resource-poor environment presumably in part by catabolizing muscle. Cooper et al. 2015
In multiple independent studies, older males arriving earlier on temperate breeding grounds have overwintered in wetter and thus higher quality habitats, as determined from stable-carbon isotope ratios in muscle (e.g.,
, but see Marra et al. 1998 ) and in toenail tissues ( González-Prieto and Hobson 2013 ). Individuals arriving earlier in breeding areas have higher circulating androgen levels in their blood, which is positively associated with more advanced physiological condition as evidenced by higher oxygen-carrying capacity (hematocrit levels) ( Tonra et al. 2011 ). Since the earlier arriving males do not have larger cloacal protuberances, their higher androgen levels appear to increase reproductive success via successful territory defense rather than accelerating reproductive condition. Thus, males with elevated androgen levels benefit directly from erythropoietic (increased oxygen capacity) effects during migration, and possibly for foraging and defending territories as they transition to reproductive activities upon arrival to breed. High androgen levels could also contribute to reproductive success in males via more opportunities for copulation ( Tonra et al. 2011 ), insofar as earlier arrival is associated with more extra-pair copulations and higher incidence of polygyny ( Tonra et al. 2011 ). Reudink et al. 2009a
Testosterone and other androgens appear to mediate these transitions from winter condition to migratory preparedness to breeding season, but only in older male American Redstarts (
). For example, only older males (and not first-winter males or females) in Jamaica increase circulating androgens from mid to late winter in preparation for spring migration and carry-over effects into the breeding season; and increased androgen levels in these older males are correlated with size-corrected body mass (i.e., body mass controlling for effects of wing, tarsus, and tail length). There was no relationship in this study between blood androgen level and distance to breeding areas (the latter determined using stable-hydrogen isotope ratio in feathers grown in breeding areas during the previous breeding season). Further, this study experimentally increased testosterone in older males in late winter just prior to migration period using silicone sub-cutaneous implants, which, compared to controls, increased body mass, lipid stores (needed for migration), breast muscle mass, and cloacal protuberance diameter; and sped up the departure timing of migration by about one week. Thus, endocrine physiology modulates transition simultaneously to migration and breeding condition, as indicated by several body condition indices, but this happened only in older males. Evidence also presented suggests that as long as older males exceed some threshold of body condition (depending on overwinter conditions such as rainfall and food availability), body condition and androgen levels feedback positively on each other in preparation for transition to migratory and breeding condition. The absence of similar mechanisms in first-year males (based on the observational components of the study; no experimental implants done yet with these first-year males to confirm this conclusion) is consistent with their later timing of migration (much as in females) and reduced reproductive success and probably reduced reproductive investment when reaching breeding grounds. Thus, first-year and older males appear to differ in the control of migration and reproductive behaviors. Tonra et al. 2013
Habitat in Breeding Range
Generally breeds in moist, deciduous, second-growth woodlands with abundant shrubs, often near water, especially in southern and western parts of range (
, Baker 1944a , Bent 1953b , Hamel et al. 1982 , DeGraaf 1985 , Peck and James 1987 , Sallabanks 1993e , Hunt 1996 , Manuwal 1986 ). Other frequently occupied habitats include alder ( Manuwal 2012 Alnus) and willow ( Salix) thickets, shrubby second-growth woodlands, thickets in treefall gaps within old-growth forest, shade trees and shrubby vegetation (e.g., fencerows), orchards, mixed deciduous-coniferous woodlands. Appears to be “area-sensitive” in parts of eastern North America; i.e., occurs disproportionately in interior woodland compared to habitat edges, and disproportionately in tracts of habitat > 4,000 ha in area ( , Robbins 1979 , Ambuel and Temple 1983 ). Most studies classify the redstart as a “mature forest species” ( Robbins et al. 1989a ) or as preferring older available vegetation ( Streby et al. 2011a ), but some classify it as “forest generalist” (e.g., in West Virginia; Norton and Hannon 1997 ) and as preferring early- to mid-successional woodlands ( Marshall et al. 2003b , Holmes and Sherry 2001 , Hunt 1996 ). Does not avoid edges in New Hampshire (TWS), or in Alberta (M.-A. Villard and S. Hannon, unpublished data). Types and arrays of habitats used locally may depend on landscape context. For example, forest disturbed by silviculture compared to agriculture tends to harbor more avian guilds that include redstarts, other long-distance migrants, and understory nesters ( Sodhi et al. 1999 ). In a two-year, before-after design logging experiment in Alberta, redstarts disappeared following clearcutting and declined in partial cuts of mixed deciduous–coniferous (aspen–poplar–spruce) forest ( Rodewald and Yahner 2001 ), although responses depended on habitat: redstarts prefer deciduous habitat in this boreal region, and actually increased in selectively-logged deciduous plots ( Norton and Hannon 1997 ). Norton et al. 2000
Prefers deciduous over coniferous vegetation in the northeastern U.S. (
, Ficken and Ficken 1967 , Morse 1973 , Sherry and Holmes 1985 , Sherry and Holmes 1989 ). In the southeastern U.S., favors bottomland hardwoods and swamps; e.g., within Piedmont and Coastal Plain physiographic regions ( Hunt 1996 , Hamel et al. 1982 ) it prefers levee habitat ( Sallabanks 1993e ). Most abundant breeding species overall in bottomland hardwoods forest in southeastern Louisiana (Pearl River floodplain), and increased following Hurricane Katrina’s impacts on vegetation, although not significantly ( Sallabanks et al. 2000 ; see also Brown et al. 2011a Breeding: Nest Site: Site Characteristics). In western North America, prefers willow and alder thickets, deciduous riparian woodlands (particularly alder), and coniferous woodland ( , Sallabanks 1993e ; M.-A. Villard and S. Hannon, unpublished data). In Alberta, redstarts defend territories, nest, and forage disproportionately in shrubby willow thickets, and less in aspen compared with availability, both within and among quaking aspen ( Manuwal 2012 Populus tremuloides) forest fragments ( ; see also Sodhi et al. 1999 ). Also breeds in coniferous woodlands in northern parts of range, where it occurs most frequently in willow/tall shrub habitat within aspen forest, including in New Brunswick (R. E. Lemon, personal communication) and Alberta (M.-A. Villard and S. Hannon, personal communication). Huffman 1997
Occurs in a range of successional stages, especially in mosaics of forested plots of different ages (
). In New England, the habitats with the highest densities were early- to mid-successional, relatively deciduous (vs. coniferous) forest stands ( Yahner 2003 ). Redstarts in these stands were predominately two-year or older males that occupied relatively small territories and were nearly all mated; yearling males present in these stands had a greater likelihood of obtaining a mate than those in other habitats. Moreover, the amount of early-successional vegetation in the landscape predicted higher abundance of redstarts, and older forests appear to act as habitat sinks for redstarts in part due to inferred poor nesting success ( Hunt 1996 ). Collectively, these observations suggest that early- to mid-successional habitats are for unknown reasons more suitable and thus preferred by this species. Further research on what constitutes high quality breeding habitat is needed. Hunt 1998c
The ecological basis for breeding habitat choice has not been well studied. Since nest predation overwhelmingly impacts nesting success and even local population dynamics (
), nest predation risk and nest predators, particularly scansorial mammals including red squirrel ( Sherry et al. 2015 Tamiasciurus hudsonicus), likely contribute to breeding habitat choices. The preference redstarts show for woodlands dominated by deciduous vs. coniferous trees in New Hampshire, for example, could arise because of greater threat to nesting redstarts from red squirrels in areas with more conifers, which provide an important seed resource for the species. Additional support for the influence of nesting success on habitat choice comes from nest tree preference for yellow birch ( Betula alleghaniensis) at Hubbard Brook and correspondingly successful nesting therein (TWS, RTH, S. Wilson, and C. S. Hunter, unpublished data).
Social cues are also important to habitat use. In an experimental study in mixed deciduous–coniferous forest in the Upper Peninsula of Michigan, newly arriving older (but not yearling) male redstarts, responded positively to conspecific songs (
, Hahn and Silverman 2006 ), and response was greatest in areas with lower resident male density ( Rushing et al. 2015b ). Yearling males were uninfluenced by experimental playbacks of conspecific song, and reached highest density on plots with more older males, prompting the hypothesis that because they return to breed 6–10 days later than older males, yearlings use older males’ song as a settlement cue ( Hahn and Silverman 2006 ). However, in an experimental study in bottomland hardwoods in Maryland, both yearling and older males responded about equally during the spring settlement period to conspecific vocalization playbacks, about 8 times more strongly than to control treatments, but not to public information cues played during fledgling period ( Hahn and Silverman 2006 ). This same study also found impacts of habitat on settlement patterns, but different habitat features were important to older vs. yearling males. Consistent with the lack of response to simulated public information during fledgling period, this study found little evidence of prospecting behavioral responses to potential public information about nesting success. Thus, site selection by breeding redstarts in Maryland was affected by both the presence of conspecifics and by habitat features such as plant species composition, tree density and shrub cover, with responses differing regionally. Rushing et al. 2015b
Breeding yearling males often occupy territories in lower quality habitat, and thus analyses of their distributions and patterns of habitat use can be helpful in identifying both physical habitat and social factors that are important. For example, yearlings are often found in less frequently occupied habitats, such as coniferous or mixed-coniferous stands in the northeastern U.S., because of preemption by dominant older males occupying nearby more frequently occupied deciduous sites (
, Ficken and Ficken 1967 , Morse 1973 , Howe 1974a , Sherry and Holmes 1988 ). Similarly, aggressive interactions with an interspecific competitor, the Least Flycatcher ( Sherry and Holmes 1989 Empidonax minimus), can indirectly influence yearling male settlement patterns and habitat use, subjecting yearling males to more interspecific attacks compared to older male redstarts, who better avoid the flycatchers ( , Sherry 1975 , Sherry 1979 , Sherry and Holmes 1988 ; see also Fletcher 2007 Behavior: Social and Interspecific Behavior). In New Brunswick, yearling males are found more often than older males in dry deciduous woodland (R. E. Lemon, personal communication), possibly also because these areas are less preferred by older males, thereby constraining yearlings to accept these less preferred sites.
After young fledge, individuals of all ages, including independent fledglings, usually move to early-successional forest habitats (e.g., 2–10 yr-old regenerating clearcuts) in Minnesota (
). Both invertebrate food abundance and canopy height are important in models of post-fledging season abundance, but food is most important, probably because of its importance to survival and, in the case of adults, to molt and preparation for fall migration ( Streby et al. 2011b ). Forest birds with habitat requirements similar to American Redstarts also show post-fledging (mid-July through August) shifts in habitat use to regenerating clearcuts with dense woody vegetation and berries in New Hampshire ( Streby et al. 2011a ). Post-breeding individuals in heavy molt become vocally inconspicuous and sedentary, occupying thickly vegetated, food-rich habitats, which in Adirondack Mountains, New York, and probably elsewhere include shrubby lakeshore habitats with abundant emergent aquatic insects. Because young and molting individuals are likely more vulnerable to predators, dense habitats where individuals can remain inconspicuous without having to fly often is probably advantageous (TWS, unpublished data). Chandler et al. 2012
American Redstarts use variety of shrubby and wooded habitats during migration (
, Parnell 1969 , Martin and Karr 1986 ). During fall migration, use of foraging habitats differed by age at a Gulf Coast stopover site in Alabama; older individuals selected sand live oaks ( Woodrey 1995 Quercus geminata), whereas first-year individuals used both oaks and slash pines ( Pinus elliottii; ). Use of slash pines by hatch-year individuals may entail greater risk of predation, but appears to be constrained by social dominance from older birds. Social dominance was implicated by the fact that older birds won 13 of 15 intraspecific aggressive interactions with known outcome, and that the number of such interactions increased significantly with the density of migrant redstarts at the site ( Woodrey 1995 ). Woodrey 1995
Habitat in the Overwintering Range
Habitats occupied by American Redstarts during the overwintering period are diverse, usually native forest and woodland as opposed to agricultural and non-forested habitats, and generally at low and middle elevations (
, Sliwa 1991 , Greenberg 1992b , Lynch 1992 , Sliwa and Sherry 1992 ). Found in montane forest up to 3,000 m in elevation in South America ( Confer and Holmes 1995 ), but is rare above 1,500 m in the Blue Mountains in Jamaica (TWS). On Caribbean islands, relatively catholic in use of habitats compared to most other overwintering warblers, but habitat use differs among islands; redstarts are particularly abundant in mangroves in the Greater Antilles, and generally abundant in shade coffee plantations, lowland second growth, and dry scrub on New Providence in the Bahamas ( De Schauensee and Phelps 1978 ). In Puerto Rico, shade coffee provides habitat of similar quality to natural forest ( Wunderle and Wade 1993 ). On San Salvador Island, Bahamas, body mass was significantly greater in mangroves compared to three other habitats studied, and females were significantly more frequent than males in all habitats except the most wooded, second-growth habitat, where the sexes were equally frequent ( Wunderle and Latta 2000 ). In Jamaica, based on local abundance of individuals, prefers black mangroves, native wet and dry limestone forest, mesic forest, and some secondary growth and scrub thickets (most to least frequent occurrence per point count; but individuals occur in almost any woody habitat, including fencerows and isolated trees and gardens in downtown Kingston and Montego Bay; Murphy et al. 2001a , Sliwa 1991 , RTH). Uses shade coffee in the Venezuelan Andes ( Confer and Holmes 1995 ). In southwestern Costa Rica, a number of individuals used a 6 yr-old regenerating pasture (1,100 m elevation) dominated by Bakermans et al. 2009 Cecropia spp. and Heliocarpus appendiculatus, although never > 100 m from primary rainforest ( ). Reid et al. 2008
Sexes tend to segregate among overwintering habitats—males in relatively moist and more forested sites, and females in shorter-stature woodlands and scrubby thickets, as evidenced at a variety of overwintering locations (
, Lopez-Ornat and Greenberg 1990 , Marra et al. 1993 , Parrish and Sherry 1994 , Sherry and Holmes 1996a , Murphy et al. 2001a ). Males dominate shade coffee in Puerto Rico by up to 79% ( Peele 2015 ). In a study of carry-over effects of winter and spring migration in redstarts and other migrant warblers arriving at Manitoba stopover sites, the effects of the wet–dry overwintering habitat gradient on migration timing and spring body condition appears to be weaker than what’s been found for Caribbean birds, possibly because sexual habitat segregation may not be as strong in overwintering areas in Mexico and Central America ( Wunderle and Latta 2000 ). González-Prieto and Hobson 2013
In Jamaica, where much of the research on overwinter habitat segregation in redstarts has been conducted, males are most prevalent in wet limestone forest (> 80% of individuals are males, almost all older males) and in black mangroves (65–70% males). Other habitats there tend to be predominantly female-inhabited: only 40–45% of individuals were males in a coastal thorn scrub habitat adjacent to the black mangroves, and 45% of individuals were males in dry limestone forest (
). The most intense studies of the mechanisms of habitat segregation in Jamaica are based on comparisons of black mangroves with adjacent thorn scrub habitat at Luana Point, on the south coast. Upon arrival to this overwintering site, hatch-year individuals of each sex settle equally in the two habitats; but then gradually segregate as older individuals arrive and evict some of the younger individuals ( Sherry and Holmes 1996a ). The mangrove habitat appears to be higher-quality than thorn scrub based on the presence of a higher percent of older, socially dominant individuals after the arrival period in Fall and on evidence that vacancies in mangroves fill more readily than in thorn scrub and mangroves have higher displacement rates (usually older birds displacing first-year individuals ( Marra 2000 , Marra 2000 ). Individuals occupying mangroves also tend to be physically and behaviorally dominant in part due to larger body size of both males and females that prevail most of the overwintering period in this habitat ( Studds and Marra 2005 ; see Marra 2000 Behavior). These patterns involving sexual habitat segregation appear to be related to higher insect abundance in mangrove habitat ( , Parrish and Sherry 1994 ), and probably in part to the seasonal stability of the habitat and its food abundance arising from relatively persistent moisture throughout the dry season in the mangroves ( Studds and Marra 2005 ). Redstarts also segregate by age, with proportionately more older individuals in mangrove compared to thorn scrub ( Angelier et al. 2011 , Marra and Holmes 2001 ) implicating site fidelity of older individuals as a factor in overwintering habitat selection. This sexual habitat segregation in Jamaica has fitness consequences ( Peele et al. 2015 , Marra and Holberton 1998 , Marra and Holmes 2001 , see Studds and Marra 2005 Demography).
Demographic correlates and ecological bases of variation in habitat quality have been extended by multiple Jamaican studies, using multiple habitat-quality indicators (summarized mostly qualitatively in
Table 2). The most consistent, and thus probably best indicators of habitat quality are annual survival (predicted well by body mass maintenance over the overwintering period; ), proportion of older males, and relatively early departure from sites at start of Spring migration northward. Density of individuals is sometimes an unreliable indicator of habitat quality ( Johnson et al. 2006b ). Based on these criteria, heavily forested wet limestone forest characterized by higher rainfall than the coastal sites and with older males comprising up to 80% of all individuals appears to be even higher in quality for overwintering redstarts than black mangroves ( Peele 2015 ). Similarly, coffee habitat is better quality agricultural habitat than citrus based on the above indicators ( Peele 2015 ). The poorest habitat of those studied is dry limestone forest occurring in south coastal areas in the rain shadow of the Blue Mountains. These differences in the quality of overwintering habitat affect reproductive success in the subsequent breeding season due largely to differences in the timing of northward departure ( Johnson et al. 2006b , Marra et al. 1998 , Norris et al. 2004a , Reudink et al. 2009a ; see also Cooper et al. 2015 Demography and Populations: Range: Carry-over Effects); and although results differ slightly by sex, redstarts overwintering in wetter habitat compared to thorn scrub were predicted to fledge 1–2 more offspring ( ). Norris et al. 2004a
The ecological determinants of overwinter habitat quality for female American Redstarts have received little attention, with exception of a study of females overwintering in the black mangrove-thorn scrub gradient in Jamaica (
). These researchers documented an interaction between rainfall and habitat quality in these habitats, affecting females’ body condition, which was shown to change within habitats over short (i.e., monthly) time periods ( Angelier et al. 2011 ). Angelier et al. 2011 implicate survival differences between the black mangrove and more stressful adjacent thorn scrub habitat in Jamaica, based on their study of telomeres (see Angelier et al. 2013 Demography and Populations: Causes of Mortality: Ecological Conditions in Winter).
Since females in some migratory bird species tend to overwinter farther south of males (e.g.,
), it is interesting that near the southern limit of the overwintering range (e.g., Venezuela, Panama) redstarts studied in coastal habitats (essentially only mangroves) are almost exclusively females and yearling males (B. Poulin and G. Lefebvre, personal communication), unlike that observed in Jamaican mangroves. However, in moister habitats such as Venezuelan shade coffee, the sex ratio is approximately 50:50, and 68% of individuals in adjacent primary forest are male ( Komar et al. 2005b ), similar to proportions of males found in Jamaica wet limestone forest. Thus, it is not clear that redstart sexes segregate geographically as opposed to occupying different habitats—a topic for future study. Bakermans et al. 2009
Based on available studies, overwintering habitat selection in redstarts involves multiple stages and behavioral processes, including social dominance hierarchies, site tenacity, and shifts of individuals among habitats varying in quality, especially changing food abundance, during progressive droughts (
, Marra et al. 1993 , Sherry and Holmes 1996a ). To the extent that habitat quality deteriorates during the dry season in some habitats, individuals abandoning one habitat move during the overwintering period, possibly over large distances, and may not be able to find or defend territories in arrival habitats. These non-territorial, presumably transient or floater individuals comprise variable and sometimes non-negligible proportions of overwintering populations in some habitats and regions ( Johnson and Sherry 2001 , Lefebvre et al. 1994a ). For instance 21% of redstarts in Puerto Rican shade coffee (mostly females) were considered transients ( Toms 2011 ). Similarly, transients in Jamaica—also disproportionately females and yearling individuals of both sexes—comprised up to 50% of individuals even in good (black mangrove) habitat where their abundance varied annually; few were detected, however, in somewhat higher elevation, wet limestone forest ( Wunderle and Latta 2000 , Peele 2015 ). Relatively little is known about habitat use and selection processes of such transient individuals, in part because they are more difficult to detect and study than territorial individuals (see also Peele et al. 2015 Priorities for Future Research).
Historical Changes to the Distribution
Appears to have declined patchily and at different times over the last 50 yr across breeding range, largely because of loss of preferred habitat—but not so much yet as to amount to a significant range shift (See
Demography and Populations: Population Status: Trends). Declines of spring migrant redstarts noted in eastern Massachusetts since the early 1950s ( ). Declines also noted since the 1930s in central Tennessee ( Hill and Hagan III 1991 ). Urbanization accompanied by habitat loss or fragmentation, particularly in latter part of twentieth century, accounts for localized declines in the Washington, D.C., area ( Robinson 1990a ), in central and western Kentucky river bottoms ( Johnston and Winings 1987 ), in Allegheny and Beaver counties of western Pennsylvania ( Palmer-Ball 1996 ), in various parts of Ohio, including western Ohio, where agriculture is most intensive ( Brauning 1992a ), in panhandle region of Florida because of both loss of bottomland timber and possibly brood parasitism by Brown-headed Cowbirds ( Peterjohn 2001 Molothrus ater) ( ), in the most urbanized coastal regions of Connecticut ( Stevenson and Anderson 1994b ), and in southern Louisiana around New Orleans (J. V. Remsen, unpublished data). Declines in Ontario and northern New England were attributed to habitat loss due to forest maturation ( Zeranski and Baptist 1990 , Cadman et al. 1987 ). Hunt 1996
A westward expansion between 1957 and 1983 was suggested by comparison of the fifth and sixth editions of the American Ornithologists' Union checklists (
), and probably reflects improved coverage of western parts of range, since no local range increases are reported in western states. Possible expansion onto Piedmont and Coastal Plain ecoregions in South Carolina ( Morse 1993 ) should be better substantiated with more breeding records. For other local decreases and increases, see McNair and Post 1993b Demography and Populations: Population Status: Trends; . Sauer et al. 2014b
No information available.