The most recent updates to the avifauna of continental and insular Colombia listed 1,909 species (Avendaño et al. 2017) and 1,941 species (Donegan et al. 2018), respectively. Of these, Avendaño et al. (2017) included 57 and 63 hypothetical and vagrant species, respectively. The former involve birds whose presence in the country is based solely on published sight records without supporting materials (i.e., museum specimens, published photographs, videos or audio recordings). These authors called upon the ornithological community ‘to publish their data and to improve the evidence supporting future new records for the country'. Documentation and formal publication is necessary to consolidate knowledge of Colombian avifaunal diversity and distribution patterns (Arbeláez-Cortés 2013, Avendaño et al. 2017). Here we report new documented records of two seabirds that are currently classified as ‘hypothetical’, ‘unconfirmed' or ‘vagrant’ in the Colombian bird lists (Avendaño et al. 2017, Donegan et al. 2018).
On 22 November 2017 several seabirds were observed and photographed by one of us (ABA) a few km north-east (10°27′02″N, 75°31′06″W) of Cartagena de Indias, dpto. Bolívar, on the Caribbean coast of Colombia. The birds were feeding on fish scraps discarded by fishermen, and subsequently many of them gathered in a flock nearby on the beach. The group was dominated by Sandwich Terns Thalasseus sandvicensis and Laughing Gulls Leucophaeus atricilla but also included several Royal Terns Thalasseus maximus, two Common Terns Sterna hirundo, a Roseate Tern S. dougallii and a Ring-billed Gull Larus delawarensis. The latter two represent noteworthy records for Colombia.
RING-BILLED GULL Larus delawarensis
Breeds in Canada and the northern USA, and winters mainly throughout the USA, West Indies and Mexico south to the Pacific coast of Costa Rica; it is casual in central Panama and exceptional in Colombia, Ecuador and Amazonian Brazil (Howell & Dunn 2007). The bird at Cartagena de Indias was apparently a first-winter based on the dark scales on the breast and belly, pink bill with a black tip and retained juvenile wing-coverts (Sibley 2000, Howell & Dunn 2007; Fig. 1). Its overall appearance and size made the bird easily separable from the nearby Laughing Gulls.
The species' first record in Colombia was a specimen collected on 8 December 1964 (Instituto de Ciencias Naturales, Bogotá; ICN 14953) at Lago de Tota, dpto. Boyacá, recently reported by Donegan et al. (2010). Subsequently, this gull has been observed occasionally on both coasts of Colombia. On the Pacific coast, it was first recorded at Buenaventura, on 19 January 1978, 9 February 1984 and 28 September 1990 (Hilty & Brown 1986, Naranjo & Franke 1995); and inland at Laguna de Sonso in the Andes of dpto. Valle del Cauca on 10 March 2001 (Downing 2005). There are at least seven additional sight records between early November and late March on the Pacific coast in recent decades (eBird); however, the only documented records (i.e. supported by photographs) are from Tumaco, where it appears to be a rare but annual visitor (V. E. Góngora Fuenmayor unpubl., eBird). These records are in line with the presence of vagrants on the northern Pacific coast of Ecuador (Nilsson et al. 2014). On the Caribbean coast, it was first recorded at Camarones, near Riohacha on 7 January 1998, then at Santa Marta on 9 January 2007 (Mazar Barnett et al. 1996, Donegan et al. 2010). Seven additional sight records are available in the last decade from between Sincelejo and Riohacha, all during late August to mid March (eBird), with one photographed at Camarones on 12 March 2015 (L. E. Ureña unpubl., eBird). Our documented record at Cartagena adds to the increasing evidence from both coasts of Colombia that suggest that this gull is probably a rare but regular winter visitor. However, additional field work is necessary to understand the precise status and distribution of this and other ‘rare’ gulls, especially on the Pacific coast (Ellery & Salgado 2018) and along the coast from the Gulf of Urabá to Barranquilla, a well-known knowledge gap for seabirds in Colombia (Estela et al. 2010).
ROSEATE TERN Sterna dougallii
New World populations of the species breed from Nova Scotia to New York and Florida, south through the Gulf of Honduras and West Indies to islands off northern Venezuela. They winter primarily on northern and eastern coasts of South America, with records in Brazil south to 21°04′S (Gochfeld et al. 2020). The US Fish & Wildlife Service lists the northeastern population as ‘Endangered’ and the Caribbean population as ‘Threatened'. Hays et al. (1997) reported recoveries of Roseate Terns in South America, noting the predominance of records (n = 146) between 10°N and the equator (along the north coast of South America), with most in November–December, suggesting that they moved elsewhere in December. More than half of these recoveries were of first-year birds. In another paper, Hays et al. (1999) described a major non-breeding area for the species on the east coast of Brazil. Current work in Brazil by P. Lima (pers. comm.) has confirmed that these remain important areas for non-breeding Roseate Terns in December–March.
The bird seen at Cartagena de Indias was in non-breeding plumage (Fig. 2). Overall appearance in flight was reminiscent of a Sternula tern. At rest, the most obvious characteristic was its size: the bird appeared significantly smaller than the nearby Common Terns. Based on the bird's blue plastic field-readable band (CS2) and a partial reading of the USGS Bird Banding Lab band number (1402-18260), it was identified later (J. Spendelow in litt. 2017) as a first-year Roseate Tern. The bird was banded as a chick by GC under H. Hays' permit, on Great Gull Island, New York, USA, on 22 June 2017, making it five months old when photographed in Colombia.
Roseate Tern can be difficult to distinguish from Common Tern in non-breeding plumage: adults are similar in size and both have a dark bill and legs, and a white forehead. Roseate Tern has pale upperparts and a white underwing, which separate it from Common Tern (Harrison 1983), but these differences are subtle. At this season, all Roseate Terns lack the long tail-streamers typical of breeding plumage. A dark carpal bar, as seen on CS2, is characteristic of a first-year. The Colombian individual probably appeared smaller than the nearby Common Terns, not only because Roseate Tern is slimmer than Common, but also because this young bird's tail feathers may not have been fully grown.
The bird reported here is the second verified record for Colombia to be published. The first was a bird that was also banded as a chick at Great Gull Island, New York, on 8 August 1969, and was recovered on Gorgona Island, off the Pacific coast, on 27 October 1969 (Hays 1971). In an analysis of Roseate Tern recoveries in the Western Hemisphere, Nisbet (1984) reported seven additional records from Colombia (six from the Caribbean and one from the Pacific coast during the period 1927–79). When not individually verified, however, banding data may include errors in reporting or processing that can result in significant uncertainties regarding actual geographic locations (Nisbet 1984). For example, at least one of Nisbet's Colombian records referred to a bird on a ship at an unspecified distance off the coast, while some other observations might have been of birds found dead or dying onshore (Nisbet 1984: 11). Other occasional sight records of Roseate Terns have been made along the Caribbean and Pacific coasts of Colombia; records are from Isla Salamanca on 14 January 1975, Cartagena harbour on 21 January 1978 (Gochfeld et al. 1980), Punta de la Cruz, Casablanca and Buritaca on the Guajira Peninsula in February 1979 (Naranjo 1979). Recently, Arzuza et al. (2008) reported the species at Ciénaga Grande de Santa Marta, Isla Salamanca, and Gorgona Island, but without dates or documentation.
Earlier references to ringed birds (Hays 1971, Nisbet 1984) were apparently overlooked by Donegan et al. (2010) and Avendaño et al. (2017), who considered the species hypothetical or unconfirmed in Colombia, which may reflect differences of interpretation regarding supporting materials, with photographs and specimens being perceived as better verified than banding or radio-tracking data (Donegan et al. 2018). It is noteworthy that, to date, the only two Colombian records verified by colour-banding are those reported by Hays (1971) and the individual described here.
Although Gochfeld et al. (2020) speculated that the few Roseate Tern sightings in Caribbean Colombia might involve birds from the Venezuelan colonies, both verified records have involved birds from the Great Gull Island (USA) colony. That described here may have been blown off course by either of the two major hurricanes (José and María) in the Caribbean during September 2017. Satellite-tracking data from four Common Terns reveal that two flew through the hurricane, to the coast of Venezuela; one was only c.100 km from the Colombian border (P. Loring pers. comm.). Perhaps Roseate Tern CS2 encountered one of these hurricanes and was blown towards Colombia. Like the other birds in the flock at Cartagena, it was presumably attracted by the fish scraps. H. Hays (pers. comm.) and GC have observed Roseate and Common Terns in Brazil feeding on bycatch and fish scraps thrown overboard.
In sum, the records summarised here suggest that this tern is a vagrant to Colombia. However, given the identification challenges involved with mid-sized terns (especially non-adults), it is possible that the species has been overlooked. That Roseate Terns seem to feed some distance offshore and visit land primarily at night (Nisbet 1984, Hays et al. 1999) may also explain the scarcity of records. Additional observations of banded birds might enable a better understanding of the occurrence of Roseate Tern in Colombian waters.
Juan Freile and David Shoch assisted with contacts in Colombia and the USA. David also contributed helpful comments regarding the identification of the birds reported here. Jeffrey Spendelow identified the Roseate Tern, and he and Helen Hays provided guidance on reporting procedures. The Bird Banding Laboratory, US Geological Survey, provided details of the tern reported here. Helen Hays, Pam Loring and Pedro Lima contributed other data and observations included herein, while eBird records by Luis Eduardo Ureña and Vinicio E. Góngora Fuenmayor helped contextualise our Ring-billed Gull observation. Ian C. T. Nisbet provided insights regarding his 1984 article and possible errors related to recovery procedures. Finally, Trevor Ellery and Guy Kirwan provided useful criticism that helped improve the submitted draft.