New records, distribution and status of six seabird species in Brazil

Distribution records of poorly-known species are currently the most explored theme in the Brazilian seabird literature. If properly evaluated, this kind of information can improve our knowledge on distribution, migration and status of occurrence of these species. In this note we present new records for six species of poorly-known seabirds in the Brazilian coast, reviewing distribution records and defining their status of occurrence in the country. We consider Chionis albus (Gmelin, 1789) a pseudo-vagrant in Brazil and define its status as a scarce seasonal visitor from southern South America. We present the first records of Leucophaeus atricilla (Linnaeus, 1758) for Trindade Island, and of Leucophaeus pipixcan (Wagler, 1831) for the state of Rio Grande do Sul, and determined that the former is a vagrant in eastern Brazil and the latter a vagrant across the country. Anous stolidus (Linnaeus, 1758) is a vagrant in southernmost Brazil. We were unable to determine if records of Chlidonias niger (Linnaeus, 1758) for Brazil and southern South America refer to vagrancy or pseudo-vagrancy. Additionally, we verified the occurrence of breeding individuals of Anous minutus Boie, 1844 on Martin Vaz Island and confirmed that there is no evidence of breeding on neighboring Trindade Island.

Knowledge of seabirds inhabiting Brazil's extensive coastline and numerous coastal and oceanic islands has increased from the pioneering studies of IHERING (1888), NICOLL (1906) and MURPHY (1915) to recent reports focusing mainly on distribution, behavior and conservation (see VOOREN & BRUSQUE, 1999;BRANCO, 2004;DE LUCA et al., 2006;MORAES-ORNELLAS, 2009 and references therein). Despite these advances, the latitudinal gradient of 7,300 km spanning along the Brazilian seaboard (DE LUCA et al., 2006) and the isolation of some islands still pose major obstacles for the development of marine ornithology in Brazil. Not surprisingly, distribution records of poorly-known species are the most explored theme in the recent Brazilian seabird literature (MORAES-ORNELLAS, 2009). For some of these species, sound considerations on distribution and status are lacking simply because available records scattered in the literature are not pooled together for proper analysis.
In this study we report new records for six species of seabirds from the littoral of Rio Grande do Sul, Brazil's southernmost state, and Trindade and Martin Vaz Islands, off the coast of the state of Espírito Santo. We used background data and thoroughly reviewed previous distributional records following guidelines in FREY (2009) and DIAS et al. (2010) to elucidate if an extra-limital record refers to range expansion, range extension or vagrancy (see DIAS et al., 2010). We define background data as "records of species other than the taxon of interest that are likely to be documented using the same sampling methods", range expansion as "an enlargement of a species range through dispersal and then the establishment of new populations", and range extension as "the discovery of previously undocumented populations" (sensu FREY, 2009). In accordance to DIAS et al. (2010), we consider range extension analogous to pseudo-vagrancy (sensu GILROY & LEES, 2003), i.e., "nominally" extra-limital records of individuals that might actually be using previously unknown migration routes to unknown wintering grounds albeit at very low densities. In this sense, we deliberately apply the term range extension for individuals instead of populations. We see no incongruity in this approach, because pseudovagrant individuals are expected to meet and breed on their nesting grounds after migrating, which conceptually constitutes a population.
Any pseudo-vagrant species in a given region is in fact a regular migrant. Therefore, it can be formally attributed to a migratory system and have its regional status of occurrence determined. We classified pseudovagrant species in categories of regional status of occurrence following the scheme proposed by the COMITÊ BRASILEIRO DE REGISTROS ORNITOLÓGICOS (2011).

Species accounts
Snowy Sheathbill Chionis albus (Gmelin, 1789). Breeds on the Antarctic Peninsula, South Georgia, South Orkney, South Shetlands and Elephant Islands, and probably also on the South Sandwich Islands (HARRISON, 1983;BURGER, 1996a). Most birds leave their breeding grounds from late April to early July, while some may over-winter if food is available (BURGER, 1996a). Moves mainly to the Falkland Islands, Tierra del Fuego and southern Patagonia, some reaching Uruguay (BURGER, 1996a). Juveniles appear more migratory than adults (HARRISON, 1983). Vagrant to Brazil, South Africa, Santa Helena and Europe, the most extreme cases being shipassisted (BURGER, 1996a). Brazilian records are from the states of Rio Grande do Sul, Santa Catarina, Paraná, Bahia and Pernambuco, the latter being ship-assisted (Tab. I).
A single individual feeding on fish remains was observed and photographed ( Fig. 1)  This was the single record in a two-year monthly study.
Although common to fairly common year-round in some of the Falkland Islands, the species does not seem to be abundant on the Atlantic coastline of South America.
On Tierra del Fuego, where it occurs year-round, it is considered common only on the northeastern coast of Isla Grande in March (HUMPHREY et al., 1970;CLARK, 1986). Scattered individuals occur year-round associated to sea lion and fur seal colonies in Chubut and Santa Cruz, Patagonia (Fábio Olmos, pers. comm.). It is even less abundant to the north, as denoted by the status of scarce winter visitor in Buenos Aires Province, Argentina (NAROSKY & DI GIÁCOMO, 1993), and uncommon winter visitor in Uruguay (ROCHA, 2006). In the latter it is more numerous in islands with sea lion and fur seal rookeries and haul-outs, being regularly observed in small groups from May to September and also on November (ESCALANTE, 1970).
A close look on Brazilian records, especially those obtained from long-term bird monitoring campaigns (VOOREN & CHIARADIA, 1990;SOARES & SCHIEFLER, 1995;FONSECA et al., 2000;our records), suggests that C. albus occurs on a regular, albeit supra-annual, basis in the country's southern states. Taking into account that temporal and spatial fluctuations are expected at the margin of an organisms' range (LOMOLINO et al., 2006), we consider C. albus a pseudo-vagrant in Brazil, and not a vagrant or "exceptional" visitor as previously determined (BELTON, 1994;BURGER, 1996a;SICK, 1997). We propose that this species should be considered a scarce migratory visitor in Brazil, from the southern cone of South America.
Laughing Gull Leucophaeus atricilla (Linnaeus, 1758). Breeds locally on the Atlantic and Gulf coasts of North America, on the Pacific coast from southeastern California to western Mexico, and in the Caribbean from the Bahamas and the West Indies to Venezuela and French Guiana (BURGER, 1996b;HOWELL & DUNN, 2007). During the non-breeding period occurs in coasts, bays and estuaries along the Pacific from Mexico to southern Peru and on the Atlantic from North Carolina through the Caribbean to northeastern Brazil (BURGER, 1996b;HOWELL & DUNN, 2007;LIMA et al., 2010). There are inland records from central Mexico to Guatemala and, to a lesser scale, in the Andes and Amazonian lowlands of Ecuador and Peru (BURGER, 1996b;HOWELL & DUNN, 2007). Infrequently recorded from the northern Pacific coast of the United States of America, inland locations in North America, Hawaii, Clipperton Atoll, Cocos and Galapagos Islands, northern Chile, Greenland and Europe (HARRISON, 1983;BURGER, 1996b;BURGER & GOCHFELD, 1996;JARAMILLO, 2005;HOWELL & DUNN, 2007). Vagrants are also mentioned from Africa, Australia, Japan and Malaysia (see BURGER, 1996b;TEBB et al., 2003 and references therein).
In Brazil it is considered a northern visitor regularly recorded in coastal Amapá, Pará and Maranhão states during the non-breeding season (SICK, 1997;RODRIGUES, 2007;SOUZA et al., 2008;LIMA et al., 2010). There are also a few inland records from the state of Amazonas (including a banded individual), scattered observations from Ceará, Bahia, Rio de Janeiro and São Paulo, and a questionable undocumented record from the littoral of Rio Grande do Sul (LARA-RESENDE & LEAL, 1982;SICK, 1997;BENCKE, 2001;OLMOS, 2002;LIMA et al., 2010).
On the afternoon of 1st January 2007, L.B. observed and photographed (Figs 2, 3) an individual at Praia do Andradas (20º30'51''S, 29º18'23''W), Trindade Island, Brazil. The bird was feeding on the fat of triggerfish that lay on the beach after a mass mortality event (see PINHEIRO et al., 2010), and apparently on flies that gathered around. This individual was observed again on 4 and 6 January. This part of the island was visited at least twice a week by L.B. between December 2006 and April 2007 and despite intensive searches, individuals of this species were not recorded again. The bird was in late formative plumage, as denoted by the fresh medium upperwing coverts, juvenal feathers mixed with new plumage on the saddle and neck, and a complete distal tail band (HOWELL & DUNN, 2007). This is the first record for Trindade Is.
Taking into account the fact that wintering areas of L. atricilla in South America are located in the northern portion of the continent; that the species has not been recorded from well inventoried localities in Brazil's eastern littoral where other similar-sized gulls are known (e.g., Chroicocephalus maculipennis (Lichtenstein, 1823), Chroicocephalus cirrocephalus (Vieillot, 1818); VOOREN & CHIARADIA, 1990;OLMOS & SILVA E SILVA, 2001;BRANCO, 2004BRANCO, , 2007CABRAL et al., 2006;LIMA, 2006;BARBIERI & MENDONÇA, 2008;COSTA & SANDER, 2008), and that a recent compilation of Brazilian records (LIMA et al., 2010) revealed that this gull is genuinely rare throughout this region, we consider our Trindade Is. record to be more parsimoniously attributable to vagrancy than pseudo-vagrancy.
We believe that other records from eastern Brazil mentioned in LIMA et al. (2010) also relate to vagrancy and not to range expansion. Most of them refer to immature individuals -like the Bahia bird in second basic plumage and the São Paulo specimen in formative plumage -or birds  Acronyms for states, departments and provinces (S/D/P): CE -Ceará; RN -Rio Grande do Norte; PE -Pernambuco; SE -Sergipe; AL -Alagoas; BH -Bahia; RJ -Rio de Janeiro; SP -São Paulo; PR -Paraná; SC -Santa Catarina; RS -Rio Grande do Sul; RO -Rocha; MA -Maldonado; SA -Salta; CH -Chaco; CO -Corrientes; CD -Córdoba; ME -Mendoza; BA -Buenos Aires; SM -Santiago Metropolitan Area. When the age of the specimens of A. stolidus was not mentioned in the original source, we defined that birds less than three years old (the minimum age of first breeding), and/or that lacked extensive white on the head, and/or displayed light brown barring on wing coverts, were young (Y) (sensu CHARDINE & MORRIS, 1996 LIMA et al., 2010, p.204, 205). Immature birds are vagrancy prone, and species with delayed maturation, like L. atricilla, may remain on their wintering grounds until first breeding (BERTHOLD, 1993;BURGER, 1996b). Correct aging of extra-limital individuals based on plumage characteristics, especially those in alternate plumage, and a sound knowledge on the timing of moult and migration, are necessary to unravel this situation.
Birds observed in eastern Brazil are probably "overshooters" (sensu GILROY & LEES, 2003) -migrating individuals that extrapolated their non-breeding areas in northern South America and ended up well beyond their normal range. Furthermore, the Trindade Is. record suggests that at least some vagrants recorded in the Old World could have crossed the Atlantic from South America, traveling up the coasts of Africa to Europe, as COTTRIDGE &VINICOMBE (1997) andFRASER et al. (2007) proposed for Leucophaeus pipixcan (Wagler, 1831) in the British Isles (see below).
Franklin's Gull Leucophaeus pipixcan (Wagler, 1831). This North American inland species breeds in Canada from British Columbia to Alberta, and in the United States of America from Montana to Minnesota, with scattered populations in the northern Rocky Mountains and the Great Basin (BURGER & GOCHFELD, 1996;HOWELL & DUNN, 2007). After breeding, it migrates through the United States of America across Mexico to the Pacific coast, spending the boreal winter on the western coast of South America (HARRISON, 1983;BURGER & GOCHFELD, 1996;HOWELL & DUNN, 2007). On its non-breeding grounds this gull is very common on sandy beaches, ports, river mouths and cultivated fields along the coast from Ecuador to central Chile, rarely to Tierra del Fuego (BURGER & GOCHFELD, 1996). The species is common in lakes in the Andes and vicinities, especially in Argentina (BURGER & GOCHFELD, 1994. This highly vagrant gull has been recorded in Europe, Africa, Israel, Marion Island, Australia, Japan, China, Hawaii and other islands in the Pacific, the Caribbean, French Guyana, Brazil, Paraguay, Patagonia, Tierra del Fuego, and South Georgia, South Sandwich, Gough, and Tristan da Cunha Islands (HARRISON, 1983;TOSTAIN & DUJARDIN, 1988;BURGER & GOCHFELD, 1994SICK, 1997;REID, 1998;MARIN & COUVE, 2001;SMITH, 2002;ALMEIDA, 2003;IMBERTI, 2003;OLSEN & LARSSON, 2003;GUYRA PARAGUAY, 2005;HOLT, 2005;BIRDLIFE INTERNATIONAL, 2009). The first Brazilian record refers to an adult bird seen on the island of Fernando de Noronha in May 1988 (ANTAS et al., 1988), but the identification was later questioned due to the lack of documentation and possible confusion with the more common L. atricilla   On 11 January 2009 ca. 16:00 h an individual in non-breeding plumage was observed by R.A.D. and Vânia M. Teixeira at the seaside resort of Cassino, Rio Grande municipality, Rio Grande do Sul. The bird was resting on the beach with a small flock of C. maculipennis at 32º10'57''S, 52°08'48''W, near the mouth of a small creek that drains the adjacent urban area. The gulls were quiet and tame, preening occasionally and allowing observation at close range from a car. The dark-grey mantle, ash colored partial hood, wide white forehead, throat and nape, marked white eye-ring, red tipped bill and white crescents on the scapulars and tertiaries distinguished it from the other gulls and were promptly noticed. After recording these distinctive characteristics, the observers left for a camera and binoculars, returning 10-15 min later. The bird, however, was no longer found. Intensive searches were conducted by car along a 10 km stretch of oceanic beach until ca. 18:00 h, but the bird was gone. Five additional surveys during January and early February also yielded negative results. On 26 April 2009, R.A.D., A. Gianuca and D. Gianuca (D.G.) found a dark-mantled gull resembling the above described individual at Cassino ca. 11:45h. The bird was standing amidst a loose group of C. maculipennis at the swash zone immediately south of the resort at coordinates 32º13'34''S, 52°11'56''W. It was approached by car and photographed at close range (Fig. 5). In flight, the conspicuous white trailing edge of the wing and the black primaries with white tips were clearly noticed (Fig. 6). The small amount of white separating the dark primaries from the gray upperwing suggests a bird in second basic plumage (sensu HOWELL & DUNN, 2007).
It is difficult to judge if the same individual was observed on both occasions. In this species, the third cycle pre-alternate molt starts in December and proceeds until April (HOWELL & DUNN, 2007). Second cycle birds may retain white flecking in black hood even in alternate plumage (HOWELL & DUNN, 2007), so a complete and fast darkening of the head is not expected from early January to late April. Since the January bird displayed a somewhat whiter forehead and throat than the specimen observed in April, the molt sequence is consistent with the hypothesis of the same individual being observed on both occasions. The fact that the records were made in the same locality further corroborates this hypothesis, despite a four-month hiatus and intensive searches between both records.
An additional record for northern coastal Rio Grande do Sul was made by C. E. Agne and A. B.-Silveira on a sandy beach at 30º07'05''S, 50º10'45''W, Cidreira municipality. An adult in alternate plumage (sensu HOWELL & DUNN, 2007) was observed and photographed on 30 May 2010 resting on the beach alongside L. dominicanus, T. acuflavidus and Thalasseus maximus (Boddaert, 1783). This was the only record in six months of fieldwork from March 2010.
Mar Chiquita and other lakes in Córdoba and Mendoza, Argentina -the only sites east of the Andes were the species is relatively common (NORES & YZURIETA, 1980;MARTÍNEZ et al., 1985;BLANCO & CARBONELL, 2001;NAROSKY & YZURIETA, 2003;CHEBEZ, 2009) -are the probable source of birds found in southern Brazil. Vagrants in South America east of the Andes may cross the cordillera from the altiplano, enter the continent via the Caribbean during their southward migration, or move around the southern tip of South America. HOWELL & DUNN (2007) stated that Andean and Caribbean records tend to be concentrated from September through December, which seems to be in accordance with the first two hypotheses. The rarity of this gull in southern Chile and Argentina (CLARK, 1986;IMBERTI, 2003;JARAMILLO, 2005;CHEBEZ, 2009) makes the last scenario less probable.
Vagrant L. pipixcan in Europe possibly originate in the Southern Hemisphere, traveling up the Atlantic after wintering in Africa (COTTRIDGE & VINICOMBE, 1997;FRASER et al., 2007). Birds may cross the Atlantic from northern South America, where vagrants are frequent (TOSTAIN & DUJARDIN, 1988;Olivier Tostain, pers. comm.), or even from North America in the wake of hurricanes, as FRASER et al. (2007) demonstrated. On the other hand, records from southern and southeastern Brazil suggest that birds may reach Africa after crossing the southern Atlantic Ocean, or vice-versa. Wind patterns in the Atlantic Ocean favor the dispersal of Palearctic and Ethiopian birds into the New World at low latitudes, whereas vagrants of South American origin usually appear in southern Africa (SICK, 1983;BENCKE et al., 2005). This gull is a strong flyer and has been frequently recorded in oceanic waters far from the coast (ALMEIDA, 2003;Álvaro Jaramillo, pers. comm.). Therefore, the species seems capable to perform transoceanic crossings from southern South America to southern Africa aided by the westerlies, or even northern South America and the Caribbean to the Old World against the prevailing trade winds.
Brown Noddy Anous stolidus (Linnaeus, 1758). Breeds in tropical and subtropical islands in the Atlantic, Indian and Pacific Oceans (HARRISON, 1983;GOCHFELD & BURGER, 1996). Breeding colonies in the Atlantic range from the Caribbean south to islands off Brazil and east to Ascension, Santa Helena, Tristan da Cunha and Inaccessible Islands, and in islands off Africa from the Gulf of Guinea to Cameroon (GOCHFELD & BURGER, 1996). Brazilian breeding colonies are known from São Pedro and São Paulo Rocks, Atol das Rocas, Abrolhos, Fernando de Noronha, and Trindade Islands (ANTAS, 1991;BRANCO, 2004). Records along the Atlantic coast of South America are more frequent in northeastern Brazil, especially in the vicinity of breeding colonies (Tab. I).
There are fewer records to the south. A bird of the genus Anous recently recorded in Uruguay has been identified as A. stolidus (ABENTE et al., 2010;Adrián Azpiroz, pers. comm.).
On 27 January 2005 a moribund individual of this species that had landed on a ship a few days before at an untraced locality off the littoral of Rio Grande do Sul was taken by the staff of CRAM (Centro de Recuperação de Animais Marinhos), a sea life rehabilitation center maintained by the Universidade Federal do Rio Grande (FURG) at the city of Rio Grande, Rio Grande do Sul. The bird died and its skin was later incorporated to the bird collection at FURG (number 427). The unsexed specimen was examined by R.A.D., L.B. and D.G. Bill depth and tarsus were smaller than the mean and near the lower limit reported for birds from the Caribbean and tropical Atlantic Ocean, while the exposed culmen and middle toe with claw were close to the mean for females from these populations (measurements from CHARDINE & MORRIS, 1996). The wing was also shorter, but the three outer primaries were much worn. Primary #7 was growing and the rest of the inner primaries were new. White on the head was restricted to a thin line stretching from the bill to the eye, and the whitish-grey wash was confined only to the forehead (Fig. 7). Despite worn, pale margins were visible on the scapulars. Based on these features and the fact that the bird was collected during the austral summer, when South Atlantic populations are breeding, we conclude that the specimen was immature, finishing its first or second pre-basic molt (in comparison to descriptions and measurements in CHARDINE & MORRIS, 1996).
Additionally, L.B. observed and photographed an individual that perched and remained for several minutes on the R/V Atlântico Sul on May 2002, off the coast of Rio Grande do Sul, in front of Mostardas municipality. The shore was at visual range.
The first mention for the state was made by IHERING (1888), who reported a black-brown tern that he could not collect and that could only have been A. stolidus. NASCIMENTO (1995) later mentioned a dead bird found at the mouth of the Lagoa do Peixe in April 1992. Despite the discrepancy regarding the collecting date, a skeleton in the collection of the Museu de Ciências Naturais of the Fundação Zoobotânica do Rio Grande do Sul (MCN 267) collected at this exact locality in March 1992 probably refers to Nascimento's specimen and constitutes the first documented state record (BENCKE, 2001).
The lack of observations of A. stolidus in southern Brazilian waters  indicates that records for southern Brazil and Uruguay are more parsimoniously attributed to vagrants. The species disperses to the open ocean after breeding, rarely straying out of warm waters (CHARDINE & MORRIS, 1996;GOCHFELD & BURGER, 1996). It is present most of the year near tropical breeding colonies, but seasonally absent from subtropical ones, such as Tristan da Cunha (HARRISON, 1983;GOCHFELD & BURGER, 1996). The considerable number of records of young individuals in the Brazilian littoral suggests that vagrants may refer to inexperienced or exhausted non-breeding birds blown off the course.
Black Noddy Anous minutus Boie, 1844. Breeds in tropical and subtropical islands in the Caribbean, central Atlantic, western and central Pacific, and northeastern Indian Oceans (GAUGER, 1999). Brazilian breeding colonies are known from São Pedro and São Paulo Rocks, Atol das Rocas, Fernando de Noronha and Martin Vaz Islands (ANTAS, 1991;GAUGER, 1999). Reports for Trindade Island (ANTAS, 1991) are equivocal (see below). Other colonies in the Atlantic are found in Ascension, Santa Helena and the Gulf of Guinea (GAUGER, 1999). The only records for the Brazilian coast are from the states of Ceará and São Paulo (GIRÃO et al., 2008;BARBIERI et al., 2010).
From 2-4 April 2007 L.B. observed a tiny breeding colony, not over a dozen nests, on a cliff on the northern face of Ilha do Norte (20º28'06''S, 28º51'17''W), Martin Vaz archipelago. The colony was characteristic, nests consisting of small elevations on a protected rocky shelf smeared with guano. The species was positively recorded only at this locality, although small groups of unidentified Anous were seen fishing nearby at sea. The diagnostic smaller size, blackish-brown plumage and whiter crown in comparison to A. stolidus (HARRISON, 1983) were noticed. The species was not recorded by L.B. during almost six months of fieldwork at Trindade Is., 45 km west of Martin Vaz.
Our observations confirm that A. minutus persists on Martin Vaz. The first mention of this species for the region was made by NICOLL (1906), who sighted an individual at Martin Vaz islets on 5 January 1906 (identified as Micranous leucocapillus). An observation of what probably was this species nesting at Martin Vaz was later mentioned in ROCKWELL (1932). LUIGI & NACINOVIC (1997) observed a breeding colony with about 30 nests on 25 November 1993, identifying the species as Anous tenuirostris (considered synonym with A. minutus).
Based on their description, the breeding colony was located at the same place where we observed it. ANTAS (1991) and GIRÃO et al. (2008) reported that there are no recent breeding records of the species for Trindade. However, as OLSON (1982) stated, there is no evidence that A. minutus ever occurred at Trindade. FONSECA-NETO (2004) correctly emphasized that reports of this species for Trindade Is. are erroneous and based on a misinterpretation of Nicoll's record for Martin Vaz. The absence of records for Trindade (NICOLL, 1906;ROCKWELL, 1932;OLSON, 1982;FONSECA-NETO, 2004;our observations) suggests that breeding is restricted to Martin Vaz.
Poorly-understood, long-distance inter-island movements of individuals from sedentary populations of this species have been recorded (GAUGER, 1999). Since the 45 km that separate Trindade from Martin Vaz pose no obstacle to the dispersal of this bird, it is possible that some individuals visit Trindade and escape identification due to large numbers of A. stolidus occurring there. We recommend careful identification of noddies at Trindade to solve this matter.
It is intriguing how this tiny population managed to persist on an islet in the mid-Atlantic for at least onehundred years. Studies on its breeding biology, nesting site fidelity and movements are badly needed.
Black Tern Chlidonias niger (Linnaeus, 1758). This species has two disjunct populations in the Northern Hemisphere recognized as subspecies: C. n. niger breeding in Eurasia from southern Scandinavia to southern Spain and east to Lake Balkhash and Altai; and C. n. surinamensis across North America from the Northwest Territories and British Columbia in Canada south to California and New England in the United States of America (HARRISON, 1983;DUNN & AGRO, 1995;GOCHFELD & BURGER, 1996). Palearctic populations winter mainly in western Africa, from Mauritania to Namibia and occasionally South Africa, whereas North American birds migrate along the Pacific to Mexico and Panama south to Peru, and through the Gulf of Mexico and the Caribbean to the coasts of Panama and Colombia east to French Guiana (DUNN & AGRO, 1995;GOCHFELD & BURGER, 1996). In South America it is considered rare/accidental in Chile, Argentina, Uruguay and Brazil (DUNN & AGRO, 1995;GOCHFELD & BURGER, 1996) (see Table I for a summary of records).
On 23 March 1999 L.B. observed through binoculars and a telescope an individual in basic plumage among a large group of C. maculipennis, L. dominicanus, Sternula superciliaris (Vieillot, 1819), S. hirundo, S. trudeaui and T. acuflavidus on a sandy beach (32º09'33''S, 52º04'57''W) of the mouth of the Lagoa dos Patos at the base of Molhe Leste, São José do Norte municipality, Rio Grande do Sul. Another individual in basic plumage was observed by L.B. amidst a group of S. superciliaris on 10 May 2002 at Praia do Cassino, Rio Grande municipality, Rio Grande do Sul, at 32º09'39''S, 52º05'5''W. The small size, round head, slender, delicate bill and the diagnostic dark smudges on sides of breast in front of wings (DUNN & AGRO, 1995;OLSEN & LARSSON, 1995) were noticed and allowed identification. These are the first records for the estuarine region of the Lagoa dos Patos.
Brazilian records come mainly from the northeastern and southern sections of the littoral. The species has not been recorded during inventories carried out in localities dispersed throughout the Brazilian littoral (e.g., VOOREN & CHIARADIA, 1990;OLMOS & SILVA E SILVA, 2001;BRANCO, 2004BRANCO, , 2007ARAÚJO et al., 2006;CABRAL et al., 2006;LIMA, 2006;BARBIERI & MENDONÇA, 2008;SOUZA et al., 2008), with the exception of northern Rio Grande do Sul (COSTA & SANDER, 2008). Small terns, especially S. superciliaris, were recorded in many of these localities, suggesting that C. niger is genuinely rare in Brazil.
There is an unusual concentration of records from Lagoa do Peixe, southern Brazil, as inferred from a personal communication of Scherezino Scherer to BENCKE (2001, p.63) stating that the species had been captured and banded at this locality on many occasions from 1986 onwards during banding campaigns carried out by CEMAVE (a governmental bird banding organization where S. Scherer worked). However, based on a recent analysis of all CEMAVE banding information from this site from 1984 to 2008 (SIMÃO & SERAFINI, 2010;Isaac Simão, pers. comm.), the only banding campaign record for Lagoa do Peixe refers to a bird captured in October 1991. Therefore, or additional captures occurred during banding campaigns carried out by other institutions (I. Simão, pers. comm.) -an unlikely hypothesis since BENCKE (2001) states that captures were made by CEMAVE staff, and because CEMAVE was the major institution banding birds at Lagoa do Peixe -or the information provided by S. Scherer is equivocal. In light of this new information, the only confirmed record for Lagoa do Peixe other than the abovementioned October 1991 capture is a sighting of a bird in alternate plumage made by Suzana M. Lara-Resende on 9 June 1986 and reported in RESENDE & LEEUWENBERG (1987) and BELTON (1994).
Neighboring Uruguay and Argentina also hold several records, which is quite surprising since the main area of occurrence of C. n. surinamensis during the nonbreeding season is in the Northern Hemisphere. The only South American region where this tern is abundant is the coast of Colombia, especially along the Pacific (HILTY & BROWN, 1986). It is uncommon along the Peruvian coast, and rarely recorded inland in the Amazon and Andean lakes (HILTY & BROWN, 1986;HILTY, 2003;SCHULENBERG et al., 2007). There are no records for Paraguay (GUYRA PARAGUAY, 2005) or Bolivia (MAYER, 2009).
The origin of birds recorded in Argentina, Uruguay and Brazil is intriguing. Nearly half of the records for these countries are for the austral spring, suggesting that vagrants originate during the southbound migration. They may be "over-shooters" (sensu GILROY & LEES, 2003): true vagrants from North America that extrapolated their wintering areas and ended up in the Southern Hemisphere. Furthermore, some of these birds, especially those recorded in northeastern Brazil, may be of Palearctic origin, as inferred by the encounter in Rio Grande do Norte of a specimen banded in Berlin (SICK, 1997). A series of Old World species have been recorded from oceanic islands off northeastern Brazil (SOTO & FILIPPINI, 2003;BRANCO, 2004;BENCKE et al., 2005;SILVA E SILVA & OLMOS, 2006), their dispersal to South America facilitated by tradewinds (SICK, 1983;BENCKE et al., 2005; see above). It is also possible that Palearctic individuals blown off-course to northeastern South America keep migrating south along the Brazilian coast as they would do in Africa, finally reaching coastal Rio Grande do Sul, Uruguay and Argentina. On the other hand, these southernmost records may refer to pseudo-vagrancy, being attributable to an unknown migratory population that spends the boreal winter in the wetlands of Argentina, Uruguay and Rio Grande do Sul. Careful observation and documentation of individuals recorded in southern South America is needed to enlighten the origin of these individuals.