Demography and Populations

Age Of First Breeding; Intervals Between Breeding

Reports of breeding at age 1 yr erroneous; few breed at 2 yr (to date, recorded only in males: n = 4; J. Hatch pers. comm.). Most breed first at age 3, but some delay until ≥4 yr old (Nisbet 1984, 1989, J. Hatch and J. Spendelow pers. comm.). Breeding presumed to occur annually, but trapping data at Cedar Beach, NY, suggest that some pairs may skip breeding in poor food years (JB). Additional details in Spendelow 1991 .

Clutch

Only 1 clutch (mode 2 eggs) in season if successful; replacement clutch (often only 1 egg) may be laid if eggs or chicks lost. See Breeding: eggs, above.

Intraseasonal variation: At Cedar Beach, clutch averaged >1.73 for most of season, but 1.47 late in Jun (Burger et al. 1996), down to 1.2 for Jul nests. At Falkner I., CT (1981): 1.94 (n = 78 nests) to 1.2 (n = 5 nests; Spendelow 1982). From 1987 to 1990, clutch size at Bird I., MA, averaged ≥1.75 for most of peak egg-laying period; decreased to 1.6–1.25 for pairs nesting very late (Burger et al. 1996). Similar fluctuations reported elsewhere (Nisbet 1989).

Tropical Roseates have lower average clutch size; range of 1.00–1.71 from 6 sites in Caribbean (n = 599 clutches); 1.49 ± 0.50 SD in e. Puerto Rico to Virgin Is. from 1980 to 1985 (n = 6,969; Norton 1988), but 1.63 in 1988 (n = 1,528; J. Pierce in litt.). At Culebra, Puerto Rico, 389 nests studied from 1983 to 1989: lowest average 1.09 (1989), highest 1.47 (1988; JB, MG). At Aruba, 1.08 (n = 25; Voous 1957).

Hatching Success

Hatching success generally very high, but can be as low as 58.6% (LeCroy and Collins 1972), or 0% in the case of heavy predation, as at Cedar Beach, NY, in 1995 (JB). In 1987–1990, averaged >80% for most of egg-laying period at Bird I., MA; 75–85% for most of egg-laying period at Cedar Beach; pairs nesting very late have hatching success of 20–40% (Burger et al. 1996). At Bird I., main cause of hatching failure is infertility. Hatching success about 98% in nests attended by males, 46% in nests without males (female pairs or female trios; Nisbet and Hatch in press).

Annual Reproductive Success

More information available for Roseates than for most species of birds (see Appendix 2), but estimating productivity is challenging for this species, for which nest sites are often inaccessible (Great Gull I., NY), chicks hide in dense vegetation (Bird I., MA) or under rocks (Falkner I., CT; Great Gull I.), or predation occurs (Cedar Beach, NY; Falkner I.); hence several methods used by different researchers (Nisbet et al. 1990). Estimates of productivity depend on end point chosen, methodology, habitat, and season (Nisbet 1989, Nisbet et al. 1990, Burger et al. 1996). Average overall reproductive success less variable than for Common Terns in same colony, but can vary from 0.0 to 1.6 young fledged/nest, depending on food supply, egg size, parental performance, year, colony, predation rates and other factors (LeCroy and Collins 1972, Nisbet 1978, 1981a, 1989, Spendelow 1982, Nisbet et al. 1990, Burger et al. 1996). Low productivity (<1 chick/pair) generally limited to small colonies and/or transitory incidents of predation; hence average productivity of northeastern population usually high (>1.1 chicks/pair; see Appendix 2; Nisbet 1981a, 1989).

Productivity varies within as well as between years (Nisbet 1989). Late-nesting pairs almost always less successful than earlier-nesting pairs (Burger et al. 1996). At Falkner I. (1981), early nests had mean of 1.26 chicks survive to 4 d of age (n = 186), compared with 0.7 for nests begun after 10 Jun (n = 29) and zero in 5 nests after 9 Jul (Spendelow 1982). From 1987 to 1990, hatching success declined seasonally from more than 90 to about 25% at both Bird I. and Cedar Beach; overall productivity declined from 1.50 to 0.40 young/nest (Bird I.) and 0.25 young/nest (Cedar Beach) late in season (Burger et al. 1996). Laying date, clutch size, hatching success, and productivity change with age (Bird I.; Burger et al. 1996): Productivity (chicks fledged/nest) 0.33 for 2- to 3-yr olds (n = 6), 1.14 for 4- to 5-yr olds (n = 14), 1.53 for 6- to 16-yr olds (n = 17). At Bird I. (1987–1990), 92–98% of A-chicks (n = 123–166) survived to 5 d, versus 31–57% for B-chicks (n = 74–90); at Cedar Beach (1987–1990), 4377% of A-chicks (n = 60–75) survived to 20 d, versus 38–46% of B-chicks (n = 30–48; Burger et al. 1996). Survival of 3 chicks is rare.

At Bird I., chick survival weakly related to egg size: Successful B-eggs averaged 19.1 g ± 0.26 SE (n = 14) versus 18.9 g ± 0.29 SE (n = 17) for unsuccessful B-eggs (Nisbet 1978).

Fencing was recommended for productivity studies by Nisbet and Drury (1972), but not by Langham (1974). Fencing not practical in dense vegetation at Bird I. (Nisbet et al. 1990), but used at Cedar Beach. Slight mortality due to trauma or entanglement in fences (0.6%; Nisbet and Drury 1972). In Florida, poor success attributed to flooding and predation (Smith 1995). In Caribbean, complete failures in some years due to egging; up to 0.4 fledged/pair in good years (Gochfeld et al. 1994). Virgin Is. colonies had poor reproductive success because birds nested on very steep, loose cliffs (R. Philibosian in litt.). In sw. Puerto Rico, however, a 400-pair Roseate Tern colony fledged 1.56 young/pair in an unusually good season (1991; Shealer and Burger 1992). See also Causes of mortality, below.

Recruitment rate (survival from fledging to first breeding) has been estimated at about 20% on basis of data from Falkner I., CT (Nichols et al. 1990, Spendelow 1991). Metapopulation Study at Bird I., Great Gull I., Cedar Beach, and Falkner I. using capture and recapture models indicated annual adult survival estimates of 0.74–0.84 from all 4 sites (lower than most comparable species); 91–100% of surviving adults from each colony returned to same colony the following year (Spendelow et al. 1995). See also Range, below.

Band loss is a problem for longevity studies; early data derived from aluminum bands were biased. More durable bands have been used since 1975 (Spendelow et al. 1994). Oldest recorded banded birds: 25.6 yr (banded as a chick in Massachusetts chick and netted 25.6 yr later in Brazil; H. Hays in litt.), 23 yr (Ireland; O. Merne pers. comm.); and several >20 yr in Northeast (ICTN). Oldest in Caribbean 17 yr (D. Shealer pers. comm).

Little information from North America. Chewing lice (Mallophaga) recovered from Roseates at Bird I., MA. Roseate Terns and 13 other marine bird species in Seychelles colony afflicted with ticks (Amblyomma loculosum) that carry Aride arbovirus; effect unknown (Converse et al. 1976). In Azores, ectoparasite load in chicks negatively correlated with growth rate; colony sites with high parasite rates were more likely to be abandoned (A. del Nevo pers. comm.).

Egg Mortality

Minor in most colonies; includes embryonic death, hatching death, disappearance (mainly due to predation), flooding, and abandonment. Eggs predated occasionally at Cedar Beach, NY, but not at Bird I., MA. Mortality of eggs at Cedar Beach greatest during egg-laying period, when nests sometimes unguarded. Desertion of eggs may increase with heavy predator or human activity. Ants sometimes enter pipping eggs and kill chicks before they hatch (Jones 1906, Safina et al. 1994).

Chick Mortality

Mortality of chicks greatest during first 5 d after hatching (Nisbet et al. 1990, Spendelow 1982). Mortality due to ants occurs in chicks up to 2 d old. Mortality due to starvation and predation varies by year and colony (Nisbet 1981a, Spendelow 1982, Nisbet et al. 1990, Safina et al. 1994, Burger et al. 1996). Flooding infrequently significant, except at low-lying or salt marsh colonies.

No evidence of cannibalism or siblicide.

Adult Mortality

Few data on adult mortality. In northeastern colonies, mortality of adults appears to be very low during breeding season (<0.3%/yr), except in sporadic incidents of predation by rats, owls, harriers, or falcons (see Behavior: predation, above). Mortality by fishing line, kite strings, and plastic six-pack holders not documented for Roseates. Human predation outside breeding season (see Conservation and management: effects of human activity, below) may be major source of adult mortality (Urban et al. 1986) and may have contributed to population declines in 1970s (Nisbet 1981a).

Exposure

Eggs and young chicks vulnerable to extreme cold and rain, but usually protected by vegetation or rock cover. Parents may desert at night, causing death of small chicks from exposure.

Predation

In some colonies, predation is main source of egg and chick mortality (see Behavior: predation, above).

Initial Dispersal From Natal Site

Young from at least 8 northeastern breeding colonies disperse with adults to Massachusetts (Nisbet 1981a) and s. Maine (Shealer and Kress 1994) in Aug–Sep before migrating to winter quarters. Roost on islands and feed in offshore waters and tidal inlets (ICTN). See also Breeding: phenology, above.

As Breeders, Fidelity To Natal Area

See Breeding: nest and colony site; colony fidelity, above.

Fidelity To Breeding Site And Winter Home Range

Intercolony movement documented in Metapopulation Study (Spendelow et al. 1995). Adults have moved among Bird I. (MA), Falkner I. (CT), and Great Gull I. (NY; and to lesser extent Cedar Beach) between breeding seasons, and even within a breeding season after initial nesting failure. This movement amounts to about 3% emigration, taking into account the possibility of dispersal to yet unknown sites (Spendelow et al. 1995).

Philopatry to nest site less studied. At Cedar Beach, 6 pairs returned for at least 3 consecutive years to same artificial nest sites; 18 pairs returned to same goldenrod area in 2 consecutive years, and high tenacity may occur for most terns nesting under goldenrod bushes that are not destroyed by beetles (MG); at Falkner I., many pairs reuse nest sites or territories (J. Spendelow unpubl.).

No information on fidelity to winter range.

Home Range

Inadequate information. May forage up to 30 km from breeding colonies. Winter home range unknown.

Numbers

Total northeastern population recently fluctuating around 3,500 pairs; low of 3,125 in 1992, high of 3,775 in 1996 (J. Spendelow in litt.); includes about 125 pairs in Canada (Kirkham and Nettleship 1987). Current numbers shown in Appendix 1 .

Various estimates place Caribbean population at about 4,000 pairs in >40 known colonies (Gochfeld 1983, van Halewyn and Norton 1984); total probably closer to 5,000 pairs and almost certainly less than 8,500 pairs (USFWS 1993). Florida breeding population about 350 pairs (Smith 1995); Dry Tortugas population fluctuates between 15 and 225 pairs (Robertson 1978, Smith 1995). E. Atlantic population about 1,600 pairs (del Nevo et al. 1994).

Trends

Northeastern population declined to about 2,000 pairs in 1890 after persecution for millinery trade; recovered rapidly to about 8,500 pairs in 1930; declined to 4,800 pairs in 1952 and 2,500 pairs in 1979; ranged from 2,500 to 3,300 in 1978–1988 (Nisbet 1980; USFWS 1989); but in 1997 approximately 3,500 pairs, with most birds clustered in a few large colonies; 90% in 4 colonies (see Appendix 1). Number of colonies declined from 21 in 1987 to 16 in 1995. Sharp decrease by about 15% between 1991 and 1992, perhaps because of Hurricane Bob in Aug 1991; some recovery since. Formerly bred sparingly in New Jersey, Maryland, and Virginia (Nisbet 1980).

Total world population estimated between 20,000 and 30,000 pairs (Gochfeld 1983), but now probably closer to 40,000 (A. del Nevo pers. comm.).

Principal limiting factors are loss of suitable nesting habitat (to humans, gulls, erosion), food supply (particularly of sand eels), and human predation on wintering grounds. Secondary factors include predation at some colonies, and hurricanes in some years, causing heavy mortality. In most years at large colonies, chick production usually >1.1 chicks/pair. In some years, particularly at small colonies, predation and/or human disturbance significantly reduce productivity (see Measures of breeding activity, above). In poor food years, food availability may limit productivity (Safina et al. 1988). Productivity also limited by unbalanced sex ratio in some colonies (see Behavior: sexual behavior, above). Adult mortality of Roseate Terns greater than for other seabirds (Spendelow and Nichols 1989); if estimate of annual adult survival of 0.82 (Spendelow et al. 1995) is representative, then excess adult mortality may be the most important factor regulating population size, and this mortality occurs mainly away from breeding grounds, where terns are taken for food. More details on winter mortality needed.