Purple Martin Progne subis Scientific name definitions

Charles R. Brown, Daniel A. Airola, and Scott Tarof
Version: 2.0 — Published September 10, 2021

Movements and Migration

Welcome to Birds of the World!

You are currently viewing one of the free accounts available in our complimentary tour of Birds of the World. In this courtesy review, you can access all the life history articles and the multimedia galleries associated with this account.

For complete access to all accounts, a subscription is required.

Subscribe Now

Already a subscriber? Sign in

Dispersal and Site Fidelity

Natal Philopatry and Dispersal

Yearlings wander widely (20), and relatively few return to the site of their birth. In Ohio, 5% of banded nestlings (n = 20) returned to the natal colony site to breed (124); in northern Texas, 1% of nestlings (n = 203) returned to the natal colony (CRB); and in western Texas, 0.6% of nestlings (n = ~10,000) returned (125). Among 24 individuals in northern Texas known to have returned to colonies within 8 km of their natal colony, 79% were males, suggesting that males show slightly greater natal philopatry than females (CRB). Among banded birds encountered in their first breeding season, 61% were found 0–1.6 km from their natal colony, 15.3% were 1.6–40 km away, 13.0% were 40–80 km away, 4.3% were 80–160 km away, 4.3% were 160–320 km away, and 2.1% > 320 km away (20). Other studies have shown similar results (126, 125 ), but such data show finite-study-area reporting bias: birds nearer the banding site are simply more likely to be detected because there is more area to search as one moves farther away from a central site (127).

Radio-telemetry and aerial flight monitoring of 15 radio-tagged fledglings in 2007 in northwestern Pennsylvania showed that young birds traveled on average 939 m on the day they fledged (128); a majority of the fledglings remained within 2 km of the natal colony until day 13; by 13–18 d after fledging, birds were detected on average 6 km from their natal site, after which they presumably moved to pre-migratory roosts (128).

Adult Fidelity to Breeding Site and Dispersal

Popular literature frequently suggests that whole colonies of Purple Martins return to a previous year's nest site, but there are few properly reported data with which to estimate breeding philopatry. Some adults return to the previous year's colony site, sometimes nesting in the same compartment they used the previous year (124, 129, 130 , 116, CRB). In Maryland, 55% of 527 breeders had nested at the same colony site in the previous year (123). In central California, 89% of individuals returned to the same colony site the next year (n = 25 colonies) and 17% to the same bridge nesting chamber (n = 19; DAA). Dispersal to different colonies tends to occur when reproduction is unsuccessful due to a major disruption by predators or human disturbance or when the number of breeding pairs decreases to < 4 (131, 132, 94). Across the range, new colony sites and previously abandoned colony sites tend to be first colonized or re-colonized by yearling pairs.

Data on apparent annual survival probability of known-age adults over consecutive years support the consensus that this species exhibits high breeding site fidelity (116). Inferring extent of site fidelity is impossible, however, because of the inability to separate mortality from long-distance movement. Among adults recaptured or resighted in the subsequent year in northern Texas, 86% were in the same colony used the previous year, and numbers declined at greater distances from that site (CRB), but these data probably reflect mostly reporting bias associated with a finite study area.

Fidelity to Overwintering Home Range

No information.

Migration Overview

Historically thought to migrate from the breeding range to the overwintering range via Mexico, the Central American isthmus, and northern South America. Huge late-summer roosts along the middle Gulf Coast from eastern Texas to northwestern Florida (133, 134) suggest that much of the eastern population migrates across the Gulf of Mexico (see also 135, 136, 137). However, while some birds do go directly across the Gulf of Mexico in both spring and fall (116, 138), others take a more easterly route through Florida and Cuba and then over water to the Yucatan or Central America, and birds from more westerly Progne subis subis populations take a westerly route along the eastern coast of Mexico (139). Reported from Bermuda, the Bahamas, Grand Cayman Island, Hispaniola, Aruba, Puerto Rico, and Netherlands Antilles (43, 27), although similarity to Cuban Martin (Progne cryptoleuca) and Caribbean Martin (Progne dominicensis) makes sight records of migrant Purple Martin in the West Indies questionable.

Intraseasonal movements on the overwintering grounds in central South America can be extensive, with geolocator tracking showing that 44% of birds shifted roost sites by at least 500 km (mean 776 km ± 43 SE), with the mean duration spent at the first site being 66 d and at the second site 78 d (68). First roost sites tended to be in northern Amazonia and second sites in the eastern Amazon, and intra-seasonal movement often resulted in birds occupying sites with less forest cover than at the first roost site (68).

Timing and Routes of Migration


Most migrants, whether coming via Mexico or across the Gulf of Mexico, presumably follow the Central American isthmus between North America and South America, although apparent migrants are also seen on Caribbean islands and sometimes commonly in Bermuda (140, 116). The birds move rapidly during migration; for example, 2 females migrated south 2,500 km from northwestern Pennsylvania to Honduras in 5 d (500 km/d). The next spring one female flew 7,500 km in 13 d (577 km/d) to the breeding grounds in northwestern Pennsylvania; 4 of the 13 d were spent at stopover sites (116). Further work has shown that fall migrants move rapidly upon initial departure from breeding sites, with longer stopovers in the Yucatan peninsula, Cuba and the Caribbean, and Central America (especially Honduras and Nicaragua) before continuing on to South America (139, 141). The entire arrival period on the overwintering range in Amazonia spans 4 months (68).

Spring Migration

First migrants must leave the overwintering range in late December or early January — accounting for the birds that appear in the southern United States in mid-January. Unknown how late birds remain in southern Brazil (reported into May by L. D. Vizotto in 76); large numbers were roosting at Manaus, northern Brazil, as late as 25 March (72), and they were last recorded there 12 April (116). Species has been reported in late April in the eastern Andes of Colombia and 26 March in Venezuela (121, 76) but also as early as 8 February in Colombia (142). Purple Martin moves through Panama from mid-February to mid-March (120). One study reported Purple Martins migrating north through Panama in mid-late April and departing Honduras to cross the Gulf of Mexico as late as 23 April (116). They are recorded as early as 28 January in Costa Rica, 29 January in Honduras, 1 February in Nicaragua, 6 February in Belize, and 13 February in Yucatan, Mexico (27). The bulk of migration in Belize occurred during the first 2 wk of March (143), but they are seen there from early February to mid-April, primarily on the cayes (144). In eastern Mexico, migrants are frequently seen 3–30 April (117), with the latest on 30 May (27). Has been recorded as late as 8 May in Honduras, with large flocks moving north (145). In Bermuda, there are records from 15 February to mid-June (140).

The Purple Martin is one of the earliest migrants that overwinters in South America to return to breeding grounds in North America. Average first-arrival dates for the eastern subspecies subis (146) are southern Florida, 15 January; southern Texas, extreme southeastern Louisiana, and central Florida, 1 February; southern Georgia west to central Texas, 15 February; northern South Carolina west to southern Oklahoma, 1 March; extreme southern Virginia west to southern Kansas, 15 March; northern Delaware and Maryland west to southern Nebraska, 1 April; northwestern Pennsylvania, 25 April (116); central New Hampshire, central Vermont, and northern New York west to central South Dakota, 15 April; and southern Canada, 1 May (147). Birds continue to arrive for up to 8 wk after these dates in the northern United States (Pennsylvania; 148) and up to 14 wk after these dates in more southerly locations (northern Texas; CRB).

Spring arrival dates for western populations are less well documented. Individuals, probably subspecies subis or arboricola, have been seen as early as 19 March in Phoenix, Arizona, but most Arizona martins arrive in April (31). Desert-dwelling subspecies hesperia arrive at Tucson, Arizona, by 26 April, typically later than nearby montane populations and later than birds farther north. Reported first arrival dates are 22 April for Huachuca Mountains, southeastern Arizona; 21 April for northern Colorado; 10 May for north-central Montana (149); as early as 25 February in the lower Colorado River valley, Arizona (150); 8 March in central California (94); 15 March in eastern Oregon (48); 9 April for west-central Washington; and 20 April for British Columbia. Arrivals in British Columbia can extend over a 3-month period (52).

First arrivals are often single adult males, perpetuating the myth in the popular literature of “scouts” who return south to guide other birds to nest sites. At least in the eastern United States, dark-bellied males usually outnumber females in the local population for several weeks after first arrival, suggesting that older males migrate earlier than females or yearling males. In northern Texas, first yearlings of both sexes arrive on average 2 months later than the first adult males (CRB). Among adults, arrival date tends to correlate directly with age (147). Average arrival dates for yearling males in central California were 26–59 d after the average arrival dates for the first adult birds in the same years (DAA).

Fall Migration

Begins migration after nestlings fledge, so departure may be staggered within a locale and may vary from year to year. Peak migration in the United States is late July, August, and September, although fall migration begins as early as late May along the Gulf Coast (151). Migration south from Pennsylvania was recorded on 26 and 29 August for 2 females, arriving in Honduras in late August to early September, and reaching Brazil on 10 and 13 October (116). Occasionally, some birds remain into October, especially in the southern half of the United States; from 1973 to 1986, latest date in north-central Texas was 11 October (152, CRB), and 1 November in northeastern Oklahoma (153). Some birds may remain along the Gulf Coast until November; last seen at a Louisiana roost on 17 November (154). Migratory departure of western birds in central California occurs from early August through mid-September (94). Stragglers reported on Christmas Bird Counts along the Gulf coast in December, if valid, are probably late-fall migrants that delay their departure in response to mild weather. Near Tucson, Arizona, subspecies hesperia departed 23 September–4 October (155).

Transients in Mexico have been recorded 10 July–12 November (117, 27), with large numbers in the Gulf lowlands of eastern Mexico in August. Birds reach Belize by late June (144), Honduras and Nicaragua by 19 August (145, 119), Panama by early August (120), central Colombia by 21 August (73), and northeastern Peru by 13 September (118). Recorded in Bermuda 1 July–24 October (140). Rare transient in Guayas Province, western Ecuador (30 individuals, 25 August) and Orellana Province, eastern Ecuador (5 individuals, 15 August) (156). Reports in coastal Peru of 45 individuals in August 1997 and 15 individuals in August 1998, may have been related to El Niño years (157).

Most eastern birds probably do not start arriving in Brazil until late September, although there is one record as early as 2 August for Manaus (66). Band recoveries suggest that most eastern birds are in Amazonia at least by mid-November and reach southern Brazil in large numbers at least by early December (73)—some apparently as early as October (L. D. Vizotto in 76). A bird banded in northern Texas was recovered 65 d later, on 21 August, in central Colombia, and another bird banded in northern Oregon was recovered 110 d later, on 4 November, in south-central Arizona. This suggests that some individuals move south relatively rapidly, while others migrate more leisurely. Variation in arrival at overwintering locations was best predicted by when a bird departed from the breeding grounds, with birds departing earlier arriving earlier, but whether birds went across or around the Gulf of Mexico had no effect on their arrival time at overwintering sites (139).

Migration dates divided by subspecies are given in Phillips (27).

Migratory Behavior

Flocking begins as soon as nestlings fledge; birds of all ages assemble in roosts before fall departure. This may represent nonbreeding activity rather than a specific response to upcoming migration, because the species is highly social and flocks in large roosts throughout the overwintering period (see Behavior: Social and Interspecific Behavior). Some late-summer roosts are enormous; one at St. Charles in east-central Missouri contained about 100,000 birds in mid-August (158), one at Lake Pontchartrain in southern Louisiana, contained more than 200,000 birds in July (154), and one on a 4.9-ha island at Lake Murray in central South Carolina, was estimated at 703,000 birds in late July and early August (159). Roosts in western populations are much smaller (160) likely reflecting a smaller population size in western North America, and roosts in some areas such as Seattle, Washington, have disappeared over the last 50–70 years (161, 50). About 50% of roosts in the eastern United States persist at the same site for at least 4 yr, with roosts in urban areas having a slighter greater persistence from year to year; roost persistence seems unrelated to regional population trends (114). Western martins occupy small roosts in the southwestern United States and northern Mexico for 3–4 wk during September before continuing migration to South America (B. Cousens, personal communication).

Birds in late-summer roosts generally disperse from the roost site before dawn, presumably spreading over large surrounding areas while foraging, and begin reassembling 1–2 h before sunset. Morning departures from the roosts are generally more synchronized than evening arrivals, with departures concentrated enough to be visible on weather radars (113,114 ). It is unknown how long individual birds remain in a roost before migrating or the rate of daily turnover among roost residents, but assemblages gradually decline through August and September as more birds depart. Birds apparently travel long distances to join roosts; in northern Texas and southern Oklahoma, juveniles joined a roost about 80 km north of their fledging site, and martins from this roost apparently traveled at least 48 km from the roost during daily foraging (162). In some areas, birds roost throughout the nesting season (see Behavior: Social and Interspecific Behavior).

Migratory behavior south of the United States is poorly known. Purple Martins apparently flock with smaller swallows less frequently than do other species, but they often assemble with other Progne martins, both during migration (117) and at overwintering roosts (122). At night, migrants join roosts of Gray-breasted Martin (Progne chalybea) in Mexico (117). Purple Martins migrate closer to beaches than other swallows and apparently avoid the highlands, at least in Mexico and Central America (27, 144). Probably exclusively a diurnal migrant, foraging as it migrates. Individuals often remain at stopover sites in the Yucatan, Cuba and the Caribbean, or Central America during fall migration, with geolocator studies showing an average duration of 16 d at stopover sites (139). Stopovers may allow birds to refuel for the flight to Amazonia or to undertake some molt.

Control and Physiology of Migration

Spring migrants departing the Yucatan for trans-Gulf flights were sensitive to wind speed and direction, with most departures occurring on days with winds blowing from the east to southeast (138). Smaller-bodied birds, as inferred from tarsus length, made fewer stopovers during fall migration than did larger-bodied birds, whereas in spring larger-bodied birds made fewer stopovers and spent less time at stopovers than smaller-bodied birds (163). Wing length had little effect on migratory performance such as flight speed or the number of stopovers (163). Breeding latitude appears to predict length of stopover on fall migration, with more northerly breeding populations tending to have later and longer stopovers; habitat quality of the stopover site (as inferred from extent of vegetation and insect abundance based on rainfall) did not affect how long birds remained (141). Time of departure from overwintering sites, speed of migration during spring, and arrival date at breeding sites did not seem to be affected by unusually warm weather on the breeding grounds (164), suggesting possible constraints on the species' ability to respond to climate change. Individuals exposed to more artificial light at night at more urban overwintering sites tended to depart on spring migration earlier, and arrive on the breeding grounds earlier, than birds not exposed to artificial light at night (165). Repeated tracking of the same individuals across years showed high within-individual variability in the timing of departure from both overwintering and breeding grounds, time to reach the Gulf of Mexico in spring and fall, and arrival at both breeding and overwintering grounds, with slightly greater individual consistency in these traits in spring than in fall (166). Birds arriving first on the overwintering grounds tended to settle in the core overwintering range in northwestern Amazonia, whereas later arriving individuals were more likely to overwinter in areas outside the core range (68).

Recommended Citation

Brown, C. R., D. A. Airola, and S. Tarof (2021). Purple Martin (Progne subis), version 2.0. In Birds of the World (P. G. Rodewald, Editor). Cornell Lab of Ornithology, Ithaca, NY, USA. https://doi.org/10.2173/bow.purmar.02