Breeding

Colony Occupation

Birds in Northeast (U.S. and Canadian Maritime Provinces) return from late Apr to mid-May (see Fig. 5); often convene on nearby beach or rocks for 2–3 d, then begin spending some time on colony before spreading out over colony site (Cooper et al. 1970; see Migration: timing and routes of migration, above). Occupy colonies about 3 wk before egg-laying. Caribbean terns return to Culebra, Puerto Rico, in early May; first lay eggs in mid-May. At Dry Tortugas, FL, individuals arrive late Apr, on colonies in mid-May, begin nesting late May or early Jun. Why Caribbean and northeastern birds breed at same time needs study.

Pair Formation

Pairs form just before arrival at colony or in conjunction with territory acquisition (J. Spendelow pers. comm., ICTN, JB). At Bird I., MA, High Flights (see Behavior: sexual behavior, above) continue throughout season, even including nonbreeding 2-yr-olds in Jul; this behavior may function as mate selection for next year (ICTN).

Nest-Building

Nest initially a scrape in ground, made few days or hours before egg-laying. Pairs explore nesting area and make several scrapes before laying. Nest material is accumulated during incubation (see Nest, below).

First/Only Brood Per Season

Availability of small fish, particularly sand eels, is probably main determinant of timing. At Falkner I., CT, in 1981, Spendelow (1982) showed sharp early laying peak (73% of 215 nests initiated between 17 and 29 May), with smaller wave later (22% initiated 29 May–21 Jun), and few later nests into mid-Jul. At Cedar Beach, NY, earlier laying (median 20 May versus 23 May) and sharper peak (SD = 2 versus 7 d) for 17 nests studied in a good food year (1984) compared with 34 in a poor year (1985; Safina et al. 1988). In both years, laying continued very sporadically into early Jul (see also Eggs, below). Important to compare egg-laying phenology among colonies in the same years, since weather and food constraints influence both first and peak dates of egg-laying.

From 1987 to 1990, peak of egg-laying at Bird I., MA, was 24–29 May each year, with most pairs laying from 20 May to 4 Jun; whereas at Cedar Beach, NY, peak of laying was usually 30 May–10 Jun (Burger et al. 1996). At Bird I., laying period advanced by 13 d during 1970s, then regressed by 7 d during 1980s (Nisbet 1981a, 1989). Similar pattern at Cedar Beach, where modal date was 19 May in 1984 and 23 May in 1985, becoming early Jun by end of decade. This overall trend was correlated with change in abundance of sand eels (Safina et al. 1988). Mean laying date at Bird I. inversely related to age of parents: 28 May ± 1.6 d SE for 6- to 16-yr-olds, 3 Jun ± 1.5 d SE for 4- to 5-yr-olds, and 22 Jun ± 2.6 d SE for 2- to 3-yr-olds (Burger et al. 1996).

Breeding schedule in Caribbean similar to that of northeastern birds: egg-laying in May (first eggs 14–20 May), hatching in mid-Jun (Shealer 1992, JB). However, breeding delayed to Jul in some years (e.g., 1991 at Culebra, Puerto Rico; J. Saliva pers. comm.).

Most eggs hatch during a 2- to 3-wk period; stragglers up to 1 mo later usually less successful. Hatching dates for most nests: at Bird I., MA, 14–28 Jun; at Cedar Beach, NY, 17 Jun–4 Jul; at Falkner I., CT, 15 Jun–1 Jul (Burger et al. 1996, J. Spendelow pers. comm.).

Dispersal from colony can be rapid or prolonged depending on breeding synchrony and presence of late pairs (younger birds) or re-laying. Once they leave breeding colonies, birds disperse throughout breeding area (Nisbet 1984; see Migration: timing and routes of migration, above). Roseate Terns from at least 8 breeding colonies in ne. U.S. recorded in s. Maine each Aug (peak mid-Aug), and remained on staging grounds 1–26 d (confirmed by color bands; Shealer and Kress 1994). Birds roost on rocky islands and feed offshore.

Second/Later Brood Per Season

No evidence of second broods by successful pairs.

Colony-Site Selection Process

Almost exclusively on islands (Nisbet 1981a). Broad range of substrate tolerance (sand, shell, rock, vegetation), and uses traditional sites even as habitats change; occupies new or restored sites as other sites become less suitable because of predators, floods, or human disturbance. Some traditional sites occupied for decades, but many sites occupied only a few years (see Appendix 1).

Presence of Common Tern appears to be most important habitat feature in Northeast and Europe, where Roseate has never been recorded nesting without them (Bent 1921, Cramp 1985, Nisbet 1989). Where outnumbered by Common Terns, Roseates tend to cluster nests in groups of 2–10 (Northeast); average of 16 nests/subcolony in Europe (Cramp 1985). In Caribbean, nests in single-species clusters on islands with Sooty, Bridled, and Sandwich terns, and Brown Noddies (Gochfeld et al. 1994).

Colony Fidelity/Philopatry By Individuals

Strong site fidelity in Northeast, leading to estimate that only 10% of recruits breed initially in nonnatal colony (based on 1,636 chicks banded at Falkner I., CT, from 1978 to 1985; Spendelow 1991). Capture and recapture data from Metapopulation Study of adults banded at the 4 major northeastern colonies from 1988 to 1992 show that average of >97% (range 91–100) of surviving adults returned to same colony site each year (Spendelow et al. 1995). Variations in movement associated with destination colony rather than original colony site or distance between colonies (Spendelow et al. 1995). Roseate Terns shift colonies in response to heavy predation or reproductive failure in Northeast (Nisbet 1989) and to egging at Culebra, Puerto Rico (Burger and Gochfeld 1988b). Some colonies in Caribbean shift annually among individual cays, but retain some philopatry to groups of cays.

Nest-Site Selection Process

Nest sites range from bare sand and coral to dense rock or vegetation. In Northeast at Cedar Beach, NY: in 3 seasons, 50–80% of Roseate nests (n = 63–110) had <31% visibility from above, compared with 0–2% of Common Tern nests in same category. Only 5–25% of Roseate nests had >90% visibility, compared with at least 90–99% of Common Tern nests (Gochfeld and Burger 1987). Although cover within 5 m of nests was similar for the 2 species, Roseates had significantly more cover within 50 cm (Burger and Gochfeld 1988c). Almost all Roseate nests were within 20 cm of a vertical object, often seaside goldenrod (Solidago sempervirens). At Cedar Beach, defoliation of goldenrod by the host-specific chrysomelid beetle Trirhabda canadensis left Roseate nests exposed. Chicks hatching in such nests were called by parents to nearby cover within 3 d. Late-nesting Roseates used taller and thicker vegeta-tion, perhaps simply as a result of advanced vegetative growth during the season, and there was no evidence of competition for nest sites between Common and Roseate terns (Burger and Gochfeld 1988c). At Falkner I., CT, terns used rabbit burrows when available, and preferred nest boxes and tires when provided (Spendelow 1982).

In Caribbean, nests offshore on small rocky islands (Voous 1965, Burger and Gochfeld 1988b). Some birds nest on ledges of steep slopes or even cliffs. Substrates include barren or sparsely vegetated sites with bare limestone, shell, sand, marl, broken coral (Robertson 1978). In absence of Common Terns to provide early warning, Roseates nest in more open areas, even where vegetation is available nearby (Burger and Gochfeld 1988b). At Culebra, Puerto Rico, vegetation cover ranged from 5 to 90% and overall slope from 5 to 50%; visibility in direction of closest nest was 17%, compared to 25% toward other directions, suggesting preference for visual barrier between close neighbors (Burger and Gochfeld 1988b). For quantitative nest-site information from Europe, see Ramos and del Nevo 1995; for Australia, see Higgins and Davies 1996 .

Construction Process

Shallow scrape; ritual scrape-making is important during courtship; both mates explore nesting area, scraping alternately or side by side in prospective sites, giving soft, broody calls (see Sounds: vocalizations, above). They strut under vegetation, settle on sand, elevate tail, and wriggle sideways with breast, kicking backward with feet in an exaggerated sand-kicking maneuver, creating a depression in sand, loose gravel, coral rubble, or occasionally flattened grass. On harder substrates, no depression visible. Internal diameters range from 6 x 6 cm to 9 x 12 cm. Little or no lining at first, and often little is added, depending on sur-rounding habitat. At Bird I., MA, during incubation, nest material is added continually at the time parents exchange; bird leaving nest picks up loose vegetation and tucks it sideways behind body; brooding bird later picks up material and tucks it into nest (ICTN). At Cedar Beach, NY, nest material much less pronounced than in Common Tern nests, and often none present (MG). On pebbly substrates, nest has smaller stones than surrounding area.

Nest Spacing And Density

Literature reports of density (nests/m2) and nearest neighbor distances are not directly comparable. Average densities reported (nests/m2): at Bird I., MA, 1.8 (n = 107; Nisbet and Drury 1972); in Virgin Is., 0.02–4.4 (n = 989; Norton 1988); in Britain, 0.4 (n = 20; Langham 1974), 0.3–1.0 (Aride; Warman 1979). At Bird I., average density now 0.6, ranging up to 4 in small patches (n = 1,200; ICTN). In Northeast, nests usually 50–250 cm apart; distribution dictated by available vegetation or rocks; sometimes as low as 30 cm when nesting under same bush or 10 cm for nests inside and outside a nesting box. At Cedar Beach, mean nearest neighbor distances (center to center) 143 cm ± 9 SE (n = 51; Burger and Gochfeld 1988c); at Culebra, Puerto Rico, 63 cm ± 4 SE to 97 cm ± 13 SE in 4 subcolonies (n = 177 nests; Burger and Gochfeld 1988b).

Microclimate

Not studied. Nests often sheltered by vegetation, rocks, or other cover that provides protection from inclement weather and sun. At Falkner I., CT, eggs lost 0.7–0.8% of fresh mass/day (J. Spendelow unpubl.). At Bird I., MA, eggs lost mass at rates of 0.685%/d (n = 97 A-eggs [first laid]) and 0.613%/d (n = 83 B-eggs [second laid]; Nisbet 1981a), indicating regulation of nest humidity similar to that effected by incubation behavior of other terns (Rahn et al. 1976).

Shape

Subelliptical; average longer, more conical, and less rounded than eggs of Common Terns, but often difficult to distinguish the two.

Size

In Northeast, mean egg length ranges from 41.9 to 43.9 mm, mean egg breadth from 28.8 to 29.9 mm (n = 676; see Table 2). Egg sizes quite similar throughout range of species. In most studies, B-eggs shorter and narrower than A-eggs (Nisbet 1981a). At Bird I., MA, mean breadth but not length declined significantly as season progressed; i.e., late pairs laid smaller eggs (Nisbet 1981a).

Mass

Little geographic variation. In Massachusetts, mean egg mass 18.9 for unsuccessful to 19.2 g for successful B-eggs (n = 54; Nisbet 1978), about 17–18% of female body mass; for fresh A-eggs, 20.6 g ± 0.09 SE (range 16.9–24.8, n = 244), and for B-eggs, 19.4 g ± 0.10 SE (range 16.3–22.8, n = 171); largest eggs laid in first few days of season (Nisbet 1981a, Nisbet et al. 1995). At Great Gull I., NY, mean egg mass 20.6 g (range 19.2–23.7, n = 20), about 19% of female body mass (Collins and LeCroy 1972). Older birds lay larger and heavier eggs (Burger et al. 1996).

Color

Variable, but ground color is various shades of brown, with speckles and streaks of blackish brown. Markings more uniform, with more fine spots and fewer blotches, and ground color darker and less likely to be greenish than Common Tern eggs. At Great Gull I., 1 nest held light blue eggs with no dark markings each year (1994–1996; G. D. Cormons pers. comm.). Eggs cryptic on rocky substrate, but most often partly or completely hidden from view. When 2 eggs of a clutch differ in shade, second egg is almost invariably the paler one (MG).

Surface Texture

Generally smooth; occasionally slightly rough, especially in late-season eggs.

Eggshell Thickness, Mass, And Porosity

Mean thickness of shells from unincubated eggs in 1981 0.213 mm (n = 44 eggs from 5 northeastern colonies) or 0.221 mm (n = 8 eggs from U.S. Virgin Is.; Custer et al. 1983). Hatched eggshells from Cedar Beach, NY, in 1990 were similar (mean thickness 0.20 mm, n = 23), but hatched eggshells from Cedar Beach in 1980 were significantly thinner (mean 0.13 mm, n = 7; Burger et al. 1995b). Mean shell mass 1.22 g (n = 44 A-eggs from northeastern colonies, 1981) and 1.22 g (n = 20 A-eggs from Muskeget, MA, 1875–1921; Custer et al. 1983). Eggs held in desiccator at 23°C lost mass at mean rate of 0.43%/d (n = 37; Custer et al. 1983), indicating shell porosity in same range as for other terns (Rahn et al. 1976); see also Nest, above.

Clutch Size

In temperate zone, clutch ranges from 1 to 4 eggs, very rarely 5 (mode 2). Older females usually lay 2 eggs; young females (3–4 yr old) usually lay 1 egg (J. Hatch unpubl.). Proportion of 1-egg clutches varies widely among colonies, among years, and seasonally, from <10% among early layers in good years at large temperate colonies to near 100% among late layers, in poor years, and/or at small or tropical colonies. Hence, wide range in mean clutch size from 1.00 to near 2.00, reflecting variations in phenology, parental quality, food supply, or other environmental characteristics (Nisbet 1981a, Safina et al. 1990).

Reports on clutch size require cautious interpretation, since a single or a late visit may result in spurious estimate. Data on mean clutch size compiled by Nisbet (1981a): 1.47 at Great Gull I., NY (n = 5,005); 1.80 at Monomoy I., MA (n = 1,139); 1.51 at Coquet Is., United Kingdom (n = 724); 1.79 at Bird I., MA (n = 13,943; Nisbet and Hatch in press). Nisbet (1978, 1981a) found increased mean clutch size in good food years, occasionally >1.9 (Nisbet 1989), but this increase was not detectable in a 2-yr study at Cedar Beach, NY (Safina et al. 1988).

“Supernormal” clutches of ≥3, often attended by ≥2 females, constitute about 7% of clutches at Bird I. (Nisbet and Hatch in press) and 1.4–4% of clutches at Great Gull I. (range of means for 1990–1996; G. D. Cormons pers. comm.).

Nineteenth-century reports mention greater proportion of large clutches. At Muskeget, MA, in 1890s, 60% of clutches contained ≥3 eggs (Mackay 1897). Two clutches of 7 eggs reported in Puerto Rico and Virgin Is. (Norton 1988, Shealer and Zurovchak 1995). See also Demography and populations: measures of breeding activity, below).

Egg-Laying

Mean laying interval 2.83 d (n = 73), increasing from about 2 to 4 d during laying season (Nisbet 1981a). Shorter and longer intervals (<2 or >5 d) usually reflect attendance at nest by ≥1 female, and/or intraspecific nest parasitism (Nisbet and Hatch in press). See Phenology: first/only brood per season, above.

Onset Of Broodiness And Incubation In Relation To Laying

Although incubation begins with laying of first egg, it may be sporadic until laying of second egg. However, hatching intervals average about the same as laying intervals (Nisbet and Cohen 1975).

Incubation Patch

Both sexes incubate, and both develop 2 brood patches; refeathering starts late in incubation.

Incubation Period

Mean (for several years) 23.1–23.4 d for both A-eggs and B-eggs. At Bird I., MA, decreased from about 24 to 21–22 d during the season in 1971–1972 (Nisbet and Cohen 1975), but no significant changes in 1980 (Nisbet 1981a). Incubation periods increased to 23–31 d (mean 27) when adults deserted colony at night to avoid nocturnal predation (Nisbet 1981a).

Parental Behavior

When not disturbed, adults cover eggs >98% of time during incubation, both day and night. Incubating adults retrieve their own eggs displaced up to distances of about 50 cm (MG). Both parents take turns incubating; dura-tion of shifts similar for male and female. During daylight hours in last 5 d before hatching, male incubates 47% of time, female 53% (n = 15 pairs, 105 pair-hours; J. Hatch pers. comm.). Median duration of incubation bouts 26 min; exchanges frequent when both parents present at nest. Occasionally both parents attempt to sit on eggs simultaneously, especially in female-female pairs (J. Hatch pers. comm.). Exchanges occur fairly regularly at dawn and dusk, irregularly during day, and seldom at night. In Azores, bouts shorten through incubation period, becoming very short (minutes) during hatching (A. del Nevo pers. comm.).

Hardiness Of Eggs Against Temperature Stress; Effect Of Egg Neglect

Resistant to chilling; hatching success >80% even when incubation prolonged by nocturnal desertion (ICTN).

Preliminary Events And Vocalizations

Starlike pattern of cracks appears toward large end of egg about 72 h before hatching. Eggs pipped about 24 h before hatching. Chicks give sharp cheep calls after pipping; more intense near hatching. Parents become more agitated, incubate for shorter periods, and stand more often as hatching approaches.

Shell-Breaking And Emergence

Chicks continue to enlarge pipping hole with egg tooth, resting between bouts of chipping. Emerge about 18–30 h after pipping. In Azores, B-eggs take longer to hatch (67–89 h) than A-eggs (58–82 h; A. del Nevo pers. comm.).

Parental Assistance And Disposal Of Eggshells

Parents remove eggshells very short distance; bite around perimeter of shells and deposit them just outside nest cup (ICTN). Because Roseate adults often do not remove eggshells immediately, some B-eggs become encased within the slightly larger shell of a hatched A-egg, occasionally interfering with hatching (ICTN).

Hatching Interval

At Bird I., MA, mean 2.95 d (n = 73; Nisbet 1981a); at Falkner I., CT, 2.74 d (n = 552; Nisbet et al. 1998).

Condition At Hatching

Semiprecocial, covered with down; eyes open at a few hours, and able to waddle at 3–6 h, and can walk erect by end of second day. Mean mass at hatching: at Great Gull I., NY: 14.8 g (LeCroy and Collins 1972); at Bird I., MA, and Falkner I., CT: A-chicks 15.4 g (n = 740), B-chicks 14.2 g (n = 222; Nisbet et al. 1995). Egg tooth remains at least 5–6 d; trace visible up to 13 d (LeCroy and Collins 1972).

Ability To Get Around And Care For Self

From Nisbet 1981a . Newly hatched chicks remain quietly in nest without moving while they dry and until eyes open. Usually are fed within 2 h of hatching. Within first day, body movements are coordinated; chicks are capable of walking, but usually remain crouched in nests. Utter soft cheep calls, which become louder with each day (see Sounds: vocalizations, above). Undisturbed chicks remain in sheltered nest or within 10 cm until nearly fledged; usually simply move to back of nest against vegetation, and remain hidden there or under other nearby debris, but may move up to 60 m to denser cover within 10 d after hatching of B-chick. B-chicks occasionally left behind. Parents stand and call chicks to them; then move farther and repeat process. Probability of movement depends on availability of hiding places, density of neighbors, and/or disturbance by humans or predators.

Growth And Development

Fledging period (until flight): 22–30 d in Northeast (Nisbet and Drury 1972) and in Caribbean (D. Shealer pers. comm.). At Cedar Beach, NY, mean 23 d ± 1.6 SE in 1984 (n = 19) and 25 d ± 2.5 in 1985 (n = 32; Safina et al. 1988). Take short hopping flights by day 20; some capable of sustained flight by day 24. In Azores, some A-chicks flew at 18 d (A. del Nevo pers. comm.). Tarsus is 13.0–14.8 mm long at hatching and 20.5–21.5 mm long at fledging (n = 10; compared with 19–21 mm in adults, n = 146). Culmen is 9.0–9.5 mm long at hatching and 23.0–24.0 mm long at fledging (compared with 36.5–40.0 in adults, n = 246). Longest outer primary is 6–18 mm on day 10; 52–71 mm on day 20; 95 mm on day 27 (sample size declined from 9 at day 10 to only a single surviving bird on day 27; Great Gull I., NY; LeCroy and Collins 1972). For asymptotic mass, see Table 3 . In egg-switching experiment (n = 22 clutches) at Bird I., MA, large-egg chicks that were raised by parents that laid smaller eggs grew faster (7.0 g/d ± 0.4 SD between days 4 and 14, n = 5) than did small-egg chicks raised by large-egg parents (5.3 g/d ± 0.6, n = 4; Nisbet 1978); greater chick survival from larger eggs. European data fitted to logistic models by Langham (1983) and Ratcliffe (1996): asymptote A, 105–106 g; growth constant k, 0.260–0.263 (determines the shape and slope of the curve).

For Bird and Falkner Is. (1987–1990), A-chicks from broods of 1 and 2 had similar linear growth rate and asymptotic mass, while B-chicks had more variable parameters, slower growth (mean 5.4 versus 6.0 g/d, n = 794[A] and 193[B]) and lower asymptotic mass (mean 95 versus 100 g). Growth rates of A- and B-chicks within a brood were highly correlated, and survival of B-chicks was correlated with growth rate (Nisbet et al. 1995). At Falkner I. (1987–1996), growth during first 4 days fitted a quadratic model; early growth of B-chicks predicted later growth and survival. Growth parameters similar in chicks raised by the same pairs in successive years, suggesting that chick growth and survival are determined mainly by parental quality (Nisbet et al. 1998).

Molt Into Juvenal Plumage

Table 4 . Peak of fledging is at 27–30 d of age (70%; Nisbet and Drury 1972), by which time Juvenal plumage is complete, although flight-feathers not fully grown. Captive Roseate had wing 160 mm long at 5 wk and 212 mm long at 12 wk (LeCroy and Collins 1972).

Control Of Body Temperature

By day 3, all chicks maintained stable body temperature of 38–39°C, though lower than adults (average 42.6°C; LeCroy and Collins 1972). Tern chicks ≥1 d old that are exposed to sun seek shade or gular-flutter. Newly hatched B-chicks may succumb to chilling when adults are away catching food for larger sibling. Begging vigor declines when chicks become chilled (LeCroy and Collins 1972).

Behavior

Chicks capable of locomotion within hours of hatching, and are agile by day 2. If nest is in cover, chicks remain partly hidden even when being fed. If removed from nest cover, they retreat rapidly and unhesitatingly back to nest or nearest cover, a distinctive behavioral characteristic recognized by Jones (1906). When disturbed, may be called and herded by parents to new hiding place (see above).

Brooding

Parents take turns brooding. At hatching, chicks are brooded 90–95% of time; brooding declines gradually as chicks age; ceases at age 6–8 d, except on cold, rainy days. During first 3 d after A-chick hatches, female broods 58% of time, while male brings 62% of food; sex roles in feeding approximately equal by day 4 (J. Hatch pers. comm.). Adults aggressive toward strange young that enter territory, and occasionally attack chicks on neighboring territories (J. Zingo and J. Spendelow unpubl.).

Feeding

Parents take turns feeding chicks until fledging. Chick initially pecks at parent’s bill; begging starts at 1 d old. Parent holds fish in bill and offers it to chick, moving fish until chick takes it; picks up and reoffers fish if chick drops it. After 1–2 d, chick quickly seizes fish from parent’s bill. At Bird I., MA, in first 3 d after A-chicks hatched, male brought 0.51 ± 0.06 SE fish/h, female 0.31 ± 0.05 (n = 21 nests, 183 nest-hours; J. Hatch pers. comm.). At Falkner I., CT, A-chicks were fed 1.2 fish/h (n = 18 pairs), surviving B-chicks 0.7 fish/h (n = 14; D. Shealer pers. comm.).

Food Of Young

At Cedar Beach, NY, prey items (n = 1,026) delivered to chicks included 72.6% sand eel, 13.4% bluefish, 6.8% herring (Clupea and Alosa spp.; Safina et al. 1990; see also Food habits: diet, above). Proportion of sand eels remained constant during day; proportion of other species varied. During postfledging period in Aug in Maine (Shealer and Kress 1994) and Nantucket, MA (J. Hatch pers. comm.), adults fed young exclusively sand eels. Young depend on parents entirely for food, even after fledging; first start to catch fish about 3 wk after fledging, but still dependent on parental feeding at least until migration in Sep (ICTN). Azorean chicks are fed on wintering ground to age 7 mo (A. del Nevo pers. comm.). Comparable information needed for North American birds.

Nest Sanitation

Incubating birds fly or walk from nest to defecate; chicks back out of nest to defecate.

At Bird I., MA, unmated birds (probably mainly females) frequent nesting area, visit nests, and sometimes attach themselves to pairs and attempt to incubate eggs or feed chicks. These birds often are driven off by parents during incubation, but where tolerated or accepted, may form established trios or quartets. No information on relatedness of these birds (J. Hatch pers. comm.). Both trios and female pairs occur, and some raise young (see Behavior: sexual behavior, above).

At Bird I., MA, third or fourth eggs rarely appear in nests midway or late in incubation (Nisbet and Hatch in press). No clear distinction in these cases between brood parasitism (egg-dumping) and cooperative breeding. Large clutches in Caribbean attributed to egg-dumping (Norton 1988, Shealer and Zurovchak 1995), but may also represent attendance by ≥ 2 females.

At Bird I., MA, fledglings do not return to densely vegetated nesting area and leave island within 5–15 d to gather in large mixed-species flocks near favorable feeding locations (J. Hatch pers comm., ICTN). At more sparsely vegetated Cedar Beach, NY, adults may feed fledglings on territory or adjacent open areas for 1–2 wk (JB).

When 2 chicks raised, male usually leaves with A-chick while female remains for 3–7 d until B-chick fledges; not known if family reunites later (M. Watson pers. comm.). Parent closely attends and defends fledgling, which accompanies parent to fishing grounds within 4–10 d.

Chicks leave colony with parents and join other Roseates at coastal staging areas, particularly in s. Maine and Massachusetts, developing foraging skills while still being fed by parents (Shealer and Kress 1994; see also Migration, above). Family groups remain together at least until migration. No behavioral information from wintering area.