Food Habits

Main Foods Taken

Almost exclusively small marine fish; very rarely small crustaceans. More specialized on schooling fish species and tends to catch smaller fish than Common Tern does; diet much less diverse than that of Common Terns in same colony. Strong preference for sand eels (sand lance [Ammodytes spp.]). In Puerto Rico, primarily dwarf herring (Jenkinsia lamprotaenia) and anchovies (Anchoa spp.; Shealer 1992).

Microhabitat For Foraging

Forages over shallow sandbars, shoals, inlets, or schools of predatory fish, often in mixed flocks with other terns and noddies (Safina 1990a, 1990b, Heinemann 1992, Shealer and Burger 1993, 1995); also feeds pelagically over schools of predatory fish. Compared with Common Terns, Roseates on Long I., ,NY. foraged over shallower water (mean 7 m), closer to shore and mainly in inlets, and less often over schools of predatory fish (Safina 1990a, 1990b). At Bird I., MA, and Great Gull I., NY, Roseates fed farther from colony than Common Terns (Nisbet 1981a, Duffy 1986); at Bird I., feeds mainly over 3 small sandbars, but also over tide rips up to 30 km from colony (Heinemann 1992). At Falkner I., CT, travels at least 25 km to sandbars off northern shore of Long I. (J. Spendelow pers. comm.). In Caribbean, feeds primarily over schools of predatory fish, or along reef margins (Shealer and Burger 1993, 1995).

Food Capture And Consumption

Forages mainly by plunge-diving, but also by contact-dipping or surface-dipping over shallow sandbars, reefs, or schools of predatory fish. Tends to fly into wind, hover, and dive, usually from height of 1–6 m, but up to 12 m (mean 4.2 [n = 232 dives, England] and 4.4 m [n = 31, New York]; Duffy 1986). Immersion often complete down to at least 0.5 m and for >1 s (Duffy 1986), but depends on depth of prey. Capture success 48% (Duffy 1986). Captures prey in bill and carries it to surface. Then springs into air and either swallows prey with few tosses of bill or carries it crosswise back to its mate or chick(s). Usually carries single fish, uncommonly 2–4, and rarely up to 9 at one time (Hays et al. 1973). Factors influencing transport of multiple fish need study. Fishing inhibited by winds that disturb surface and interfere with diving (Dunn 1973a). Wind >19 km/h depressed chick growth rates by interfering with adult foraging (Europe; Dunn 1975), but did not affect foraging success in Puerto Rico (Shealer and Burger 1995).

In Caribbean, Roseates often forage in flocks with Sandwich Terns or Brown Noddies (Anous stolidus) over shallow lagoons or just off rocky islands (Shealer and Burger 1993, 1995). In Puerto Rico, up to 5% of feeding is by contact-dipping, mainly in mixed-species flocks (Shealer and Burger 1993); capture success 18–75% (Shealer 1995). Foraging success of Roseates decreased with increasing flock size, because of passive interference competition; negative correlation (–0.71) between feeding attempts and flock size. Size of single-species flocks did not relate to foraging success of Roseates. Terns captured fish on about 75% of attempts, with correlation of +.95 between capture rates and attempt rates. Mean flock size for Roseates at Culebra, Puerto Rico, was 6.9 birds, compared with optimum of 8 (Shealer and Burger 1993). One-year-old Roseates had lower overall foraging success than adults (Shealer and Burger 1995).

Frequently reported feeding over predatory fish schools in many parts of range (Bent 1921, Britton and Brown 1974, Safina 1990b, Shealer and Burger 1993). Also feed over foraging Double-crested Cormorants (Phalacrocorax auritus; Heinemann 1992). On Long Island, bluefish (Pomatomus saltatrix) competed with terns for small fish, but also drove fish to surface, increasing availability for terns; bluefish activity right at surface depressed tern capture success (Safina 1990a). In Caribbean, capture rate 4 times higher over predatory fish schools than in other situations (Shealer 1996).

Duffy (1986) and Safina (1990a, 1990b) compared foraging of Roseate and Common terns (from Great Gull I. and Cedar Beach, NY). Roseates plunged from greater heights and remained submerged longer and fed in smaller, less dense flocks, where they were more successful (Safina 1990a, 1990b). Common Terns may outcompete Roseates and exclude them from dense flocks (Duffy 1986, Safina 1990b). Nisbet (1981a) suggested that Roseates preferentially feed in dispersed groups; hence differential feeding sites related to specialization rather than to competition. In mixed flocks, Roseates dived more frequently (both more successful and more unsuccessful dives) than Common Terns (Safina 1990b). Roseates are less able to hover for long periods over patches of prey (Kirkham and Nisbet 1987). Unlike Common Terns, Roseates often accelerate downward with rapid wing-beats (Kirkham and Nisbet 1987).

Feeding habits render this species vulnerable to capture by humans on wintering grounds in Africa and South America. Captured by netting or snaring after being attracted by small fish scattered on beach. Food and feeding of European Roseates summarized in Cramp 1985 and del Nevo et al. 1994 . For data on Australian Roseates, see Hulsman 1987 and Higgins and Davies 1996 .

Foraging Range

Depends on local fish availability, which in turn is influenced by bottom topography and occurrence of predatory fish. Northeastern breeders may fly 30 km to feed (J. Spendelow pers. comm., Duffy 1986, Heinemann 1992). No data on mean foraging distance from colony. At Ram I., MA, many fish caught within 300 m of colony (J. Hatch pers. comm.). Cedar Beach birds feed mainly in inlet 5–8 km from colony. At Culebra and other Puerto Rican islands, Roseates fed mainly within 2 km of colony, often within 200 m (Shealer and Burger 1993, 1995).

Piracy

In Great Britain, robbed other terns by diving from above, attacking from rear, or swooping down at terns on the ground (Dunn 1973b). Piratical terns patrolled solitarily 10–20 m above colony and tried to seize fish from beak of victim. Piracy success averaged 7.5% (range 2–18, n = 2,358) of attempts in England (Dunn 1973b) and 2% in Australia (Hulsman 1976). Attack success inversely related to fish size, ranging from >20% for 2.5-cm fish to only 7% for 10-cm fish, but payoff greater from stealing larger fish. Piracy on Common Terns also noted regularly in North American colonies (J. Spendelow pers. comm., ICTN). In Caribbean, steals from Brown Pelicans (Pelecanus occidentalis, Shealer 1996).

Major Food Items

Diet consistently less diverse than that of Common Terns in same colony or area. Sand eels (Ammodytes americanus) predominate in diet. Adults trapped on nests in northeastern colonies typically regurgitate 5–20 or more sand eels <7.5 cm in length (Nisbet 1981a). Chicks are fed on sand eels up to 13 cm long and on several other fish species, depending on availability. Also takes various small fish including bay anchovy (Anchoa mitchelli and A. hispidus), Atlantic herring (Clupea harengus), blueback herring (Alosa aestivalis), Atlantic menhaden (Brevoortia tyrannus), Atlantic mackerel (Scomber scombrus), Atlantic silversides (Menidia menidia), bluefish, and white hake (Urophycis tenuis). Only rarely takes insects, squid (Loligo sp.), or small crustaceans (shrimps, Alephidae), even when these are the main food item carried by Common Terns in same colony.

Quantitative Analysis

Data available for main colonies studied by observation of adults with food items in their bills:

Maine. In 1991 and 1992, postbreeding adults at Stratton I. fed 100% sand eels to fledglings (n = 34 and 62 feedings) and 92% sand eels to other adults (presumably mates, n = 12; Shealer and Kress 1994). In e. Maine, fed mainly on Atlantic herring and white hake (S. Kress unpubl.).

Massachusetts. Courtship-feeding in May and early Jun consists of >95% sand eel; diet diversifies in late Jun to include herrings, bluefish, and mackerel (Nisbet 1989), although sand eel remains the most important prey fish throughout the season (71% in 1990 and 69% in 1991; Heinemann 1992). In some months, herring predominate, but overall made up 8% of diet in 1990 and 11% in 1991, while silversides constituted 10 and 11% and anchovies 6 and 4%, respectively (Heinemann 1992).

Connecticut. In 1984 at Falkner I. (n = 327), sand eel (59%), bay anchovy (29%), bluefish (6%), and silversides (5%; Richards and Schew 1989); also herrings and four-bearded rockling (Enchelyopus cimbrius). One adult on Falkner I. specialized on northern pipefish (Syngnathus fuscus; D. Shealer pers. comm.).

New York. Main food at Cedar Beach (1984–1987) was sand eels (mean 73%, n = 1,026; Safina et al. 1990), but sand eels often constituted 100% of diet for days at a time. Juvenile bluefish 13% of adult prey; clupeids <10%. Several other species made up a small percentage of diet, including herrings (7%), bay anchovy (4%), and Atlantic mackerel (1.6%; Safina 1990b, Safina et al. 1990). Regurgitation from trapped adults consisted of >85% sand eels in all years from 1985 to 1992 (MG). Little diurnal and seasonal variation in diet: Bluefish declined from 15 to 4% after early morning, and herrings increased in late morning from 2 to 12% of feedings, but decrease after Jul (Safina et al. 1990).

Caribbean. In Puerto Rico, main prey is small silversides, apparently dwarf herring; adults regurgitated anchovies (Anchoa spp.; 39%), dwarf herring (35%), sardines (Harengula sp.; 10%), and reef silverside (Hypoatherina harringtonensis; 9%; Shealer 1995).

Food Limitation

At Cedar Beach, NY, Roseates laid eggs earlier and delivered more fish per hour to chicks, and young grew faster and had higher survival in a year when prey populations were high, as measured from sonar transects at sea near the colony (Safina et al. 1988). At Culebra, Puerto Rico, Roseates may be limited by food; sizes of fish brought to chicks did not increase during breeding season; harassment by Laughing Gulls (Larus atricilla) kept adults in high level of aggressive responsiveness, detracting from parental care and feeding (JB, MG). Interannual variability in breeding performance seemed related to prey abundance, suggesting limited food in some years (Shealer 1995). At Falkner I., CT, food limitation in 1995 resulted in slower growth and longer foraging range of adults than in previous years (J. Spendelow and D. Shealer unpubl.).

Opportunistic; searches for schools of prey fish that are either forced to surface by predatory fish or are already near surface over shallow sandbars. See Shealer 1995 for experiment on food selection. No food storage.

No quantitative data.

Some Roseate chicks able to maintain body temperature 10–12°C above ambient shortly after hatching; by day 3, all chicks tested maintained nearly stable body temperature independent of ambient temperature but still below the 40.9–43.6°C temperature of adults (LeCroy and Collins 1972). Both chicks and adults rest in shade of rocks or vegetation during heat of day. Adults and chicks (1–2 d old) gular-flutter when air temperatures are high. Adult mean body temperature is 42.3°C during breeding season (n = 73; Cedar Beach, NY; MG, JB). See also Breeding: young birds, below.

Drinks salt water by dipping bill in water while skimming surface. Physiology of supranasal salt gland not yet studied. Adults normally walk or fly from nest to defecate; chicks usually defecate when picked up. Chicks at many colonies do not have access to water and do not drink before fledging. Pellets rarely found at nest site.