Four decades after Belton : a review of records and evidences on the avifauna of Rio Grande do Sul , Brazil *

We present a new update of the list of birds of the state of Rio Grande do Sul, Brazil, based on a thorough review of new records and evidences accumulated from February 2011 to July 2017. This is the fifth update since the first compilation based on a reasonably complete geographic coverage of the state, published by William Belton in 1978, and the second produced by a regional (informal) committee. It is also the first to widely use citizen science contributions available on shared portals and digital databases on the internet. Forty-three taxa were added, resulting in a final list with 704 species, 6.5% more than in the previous assessment in 2010. Two species were replaced due to taxonomic changes. Documentation for inclusions based on unpublished records is indicated or published here. We also updated the documentation of another 20 species previously included in the list. Inclusions represent mainly migrants recorded in the state as vagrants or irregular visitors (22), but also cases of recent range expansion (especially from the north) and previously overlooked resident or migratory taxa. The average rate of additions (over six species per year) was 30% higher than in the previous period and is expected to accelerate. The percentage of accepted species without documented records in the state decreased from 1.8% in 2010 to 0.7% in the current list. We attribute these results to a better spatial and temporal coverage of the state in recent years, mainly due to the increasing contribution of amateurs, who accounted for 60% of the new occurrences. In contrast, the percentage of species documented by museum specimens has decreased steadily over time (currently at 84%). Investment in scientific research and collection of voucher specimens in the state should keep pace with the growing interest birds arouse in society, due to the importance and usefulness of museum specimens.

Unlike the scenario in most other Brazilian states, the avifauna of Rio Grande do Sul (hereafter RGS) can be considered reasonably well known (Bencke, 2001(Bencke, , 2010;;Bencke et al., 2010).The assembly of the existing knowledge on the birds of RGS goes back to the end of the nineteenth century, but this relatively short history of investigation is marked by a few decades of fruitfulness interspersed with long periods of minimal or no progress in the ornithological knowledge.Between 1880 and 1892, the state was residence to the German naturalist Hermann von Ihering, who published the first annotated checklist of the birds of RGS (Ihering, 1899).Although pioneering, his list covered only a small part of the state's territory.Three decades later, Rudolf Gliesch produced the second list of the birds of RGS based on information collected by himself and collaborators mostly on the outskirts of Porto Alegre and along the northern littoral (Gliesch, 1930), but once again the list did not provide a representative picture of the state's avifauna (for a summary of the early history of ornithological investigation in RGS, see Belton, 1984).The next few decades witnessed only modest progress, with the work of Camargo (1962) standing out alone as the only significant contribution.This phase lasted until the first publications of the American ornithologist William Belton in the mid-1970s.Belton's contributions were particularly central and in many ways pioneering because he used for the first time standardized survey methods and covered all ecosystems and physiographic regions of the state.His work resulted in the first compilation of the birds of RGS based on a reasonably complete geographical coverage of the state's territory (Belton, 1978).
Since then, four updates of the RGS checklist were published: (1) Belton (1984;1985) -the most important contribution to the ornithology of RGS, including distribution maps for all species listed, indication of museum specimens and original data on field marks, voice, seasonality, abundance, breeding, behavior, and ecology; (2) Belton (1994) -an updated Portuguese version of the previous work; (3) Bencke (2001) -the first review to adopt explicit criteria for the inclusion of species and to consider other types of documentation in addition to museum specimens; and (4) Bencke et al. (2010) -the first update produced by a regional, albeit informal, committee of ornithologists from some of the main research institutions in RGS.Therefore, in the last four decades the RGS checklist has been continuously revised and updated in a more or less standardized way, allowing comparisons between lists.
Since the first comprehensive assessment (i.e., Belton, 1978), almost 100 new species have been added to the state list (Bencke et al., 2010).Comparing their data with the results of previous reviews, Bencke et al. (2010) detected an accelerated increase in the number of species recorded in RGS (over four new species per year in [2001][2002][2003][2004][2005][2006][2007][2008][2009][2010].They also drew attention to the increased contribution of amateurs to the improvement of the list, either by contributing new occurrences or by providing better documentation.As in other regions, the proliferation of field-based observers in the state has catalyzed the emergence of important bird records and also enhanced the detection of rare species, including some previously unrecorded within the country or state boundaries.A large proportion of these records are accompanied by documentation in the form of sound recordings or, more frequently, digital photographs, which enable the confirmation of the identifications by experts.In addition to this greater involvement of amateurs, a significant increase in the number of professional ornithologists working in the state is also evident in the last three decades (I.Franz, unpubl. data).
The Brazilian ornithology has benefited from the contribution of bird watchers and photographers in recent years as well.Both birding and (especially) wildlife photography are fast-growing activities that have gained many enthusiasts in Brazil through the development of digital databases (e.g., WikiAves), specific events (e.g., Avistar and Festival Brasileiro de Aves Migratórias), and specialized magazines (e.g., Passarinhando and Uru).Besides being powerful tools for environmental education (Farias, 2007), these hobbies have great potential for generating large volumes of useful data (see Dickinson et al., 2010), for example, by nourishing national databases of bird records, which are key for spatial and temporal analysis, as well as for careful conservation planning (Giorgi et al., 2014;Lees & Martin, 2014;Pegas & Castley, 2014;Callaghan & Gawlik, 2015).Continuing Belton's legacy, and now making broad use of public-access databases nourished by both amateur and professional birders, we provide here the fifth update of the checklist of the birds of RGS.Recent records and new evidences are thoroughly revised and implemented.Furthermore, as recommended by Bencke et al. (2010), we tried to distinguish between range extension vs. expansion and between instances of vagrancy and pseudo-vagrancy when interpreting the new occurrences, in order to contribute to a better understanding of the composition and dynamics of the regional avifauna.

MATERIALS AND METHODS
As defined by Bencke et al. (2010), the geographic coverage of the checklist includes the continental territorial area of RGS (281,748.54km 2 ), as well as the corresponding territorial sea and Exclusive Economic Zone (EEZ) up to the limit of 200 nautical miles, or 370 km, from the coastline.
The list revised and updated by Bencke et al. (2010) served as the basis for the present review, which covers the period from February 2011 through July 2017.We incorporated all new information and updated the taxonomy and classification according to the latest checklist of Brazilian birds published by the Brazilian Ornithological Records Committee -CBRO (Piacentini et al., 2015).The only exceptions were the separation of the willets (Scolopacidae) into two distinct species, Tringa semipalmata (Gmelin, 1789) and T. inornata, in which case we followed Oswald et al. (2016), and the treatment of Thalassarche steadi Falla, 1933 as a full species, following Abbott & Double (2003a,b), Chambers et al. (2009), ACAP (2011) anddel Hoyo et al. (2017a).
Criteria for inclusion of species in the list were the same as in Bencke et al. (2010).Accepted species must have (i) occurrence in the state documented by verifiable physical evidence in the form of a study skin or complete specimen, skeleton, photograph, video or audio recording, published in bibliographic sources or available for examination in scientific collections or archives of public access, or (ii) at least one state record supported by non-material evidence allowing an unequivocal diagnosis of the taxon involved, such as a circumstantiated written report containing a detailed description of, or reference to, the diagnostic features observed.
We compiled new information from published and unpublished sources.For new occurrences and documentation updates compiled from the scientific literature (including 'in press' manuscripts), we provide only basic metadata for each record in summary tables.Additional information on these published records can be obtained by consulting the original sources.For unpublished records, the information is presented or discussed in more detail and, as a rule, the available evidence is fully referenced or published here.In a few cases, however, we intentionally chose not to publish the evidence in order to preserve its novelty, as the authors of some of the new state records expressed the intent to publish their findings in the near future.In any case, we made sure that supporting evidence for these records is available for examination in archives of public access.
In general, only the evidence of higher support is indicated for each species, according to the following order of importance: specimen (complete or partial, in the form of dry study skin, skeleton or specimen preserved in liquid ["spirit specimen"]) > photograph or video > audio recording > circumstantiated sight record (sensu Bencke et al., 2010).More than one evidence is mentioned for species that present identification problems or when the main evidence is unsatisfactory, ambiguous or has uncertain availability.
We also updated the lists of species of probable and hypothetical occurrence in the state, following the criteria outlined in Bencke et al. (2010).In short, probable species are those for which the existing RGS records have distributional or biogeographic coherence but lack minimal ancillary information to allow a reliable taxonomic identification or an independent assessment of their validity; and hypothetical species are those known in RGS only from unsubstantiated records and whose occurrence in the state is not consistent with their distributional pattern.(del Hoyo et al., 2017b) and The Birds of North America (Rodewald, 2015) served as the main sources of information on geographic distribution and winter range of the species discussed in the text.

Additions.
Our compilation of published and unpublished new records resulted in the addition of 42 species to the main list of the birds of RGS.An additional species (Tringa inornata) was included as a result of taxonomic changes affecting the state list, totaling 43 new species compared to the previous list (Tab.I).Of these, 27 had records for RGS previously published in the scientific literature since the last update of the list, though in a few cases without accompanying diagnostic documentation.The remaining 16 species had only unconfirmed (taxa uplisted from the "probable" and "hypothetical" categories) or unpublished records for the state (Tab.I).Two taxa (Charadrius sp. and Catharus cf.swainsoni) could not be safely identified to species level from the available evidence, but undoubtedly belong to species not previously recorded for RGS.
We were able to find documentation for all but one of the newly incorporated taxa.Of these, 36 are documented with photographs, three with museum specimens and two with photographs complemented by voice recordings (Tab.I).One Nearctic migrant (Contopus virens) was added based on the re-identification of a photographic record that had been erroneously identified in the WikiAves online archive (see below).The inclusion of the only undocumented species (Pterodroma deserta) is based on information retrieved from data-loggers attached to living individuals tracked to oceanic waters within the EEZ of the state.
We herein publish or explicitly indicate supporting voucher documentation for the occurrence of 18 new species for which no published evidence was available in the bibliography (Tab. I;(17)(18)(19)(20), including one species (Pseudocolopteryx acutipennis) recently mentioned for the state without reference to any specific supporting documentation.With regard to T. inornata, although a specimen of its sister taxon T. semipalmata and photographs of both taxa had been previously mentioned for the state, no published photographs documenting its occurrence in RGS were available.We thus provide a photograph of an individual in nonbreeding (basic) plumage taken at the mouth of the Lagoa dos Patos estuary in 2014 (Fig. 4).We also indicate supplementary documentation for two other species that already had published records for RGS supported by material evidence (Picumnus cirratus and Catharus cf.swainsoni; Tab.I, Fig. 16).Nine species included in the main list were upgraded from the lists of species of probable and hypothetical occurrence owing to new documented records (Tab.I), including representatives of three taxonomic families not previously represented in the state list (Thinocoridae, Heliornithidae and Donacobiidae).In contrast, 13 of the newly incorporated taxa had not been previously reported from the state in the bibliography and constitute genuine additions, of which two are new for Brazil as well (Charadrius sp. and Sturnus vulgaris).
Taxonomic changes.Two species included in the previous list were substituted due to taxonomic splits.Thalassarche steadi Falla, 1933 is considered a species distinct from T. cauta (Gould, 1841) and replaces the latter in the list.In addition, the southern, yellow-billed form of gray seedeater previously considered a variant morph of Sporophila plumbea (Wied, 1830) has recently been described as a new species, endemic to the upland grasslands of southern Brazil, S. beltoni Repenning & Fontana, 2013, which substitutes the former on the main list.Documentation updates.We updated the documentation of 20 species previously included in the main list by compiling or providing higher-quality evidence for their occurrence in the state (Tab.II).Of these, eleven were previously known in RGS only from sight records, two were documented with voice recordings and seven with photographs.The new documentation includes eight museum specimens (in one case complemented by a voice recording), ten photographs (in one case complemented by a voice recording) and two audio recordings.Five updates were based on evidences previously published in bibliographic sources.We publish or indicate herein the new documentation for the remaining species (Tab.II, Figs 21-29).In one case (Fluvicola albiventer), we supplement the previous photographic documentation (VIREO archive d01/34/078) with a new photographic record because the former was only weakly connected to RGS.

DISCUSSION
Updated list of the birds of RGS.The changes implemented here increase the number of accepted species in the RGS checklist to 704 (Appendix 1), 6.5% more than in the previous assessment (i.e., Bencke et al., 2010).This increase was accompanied by a reduction in the number of species listed as probable and hypothetical in the state, from 10 and 16 to the current 7 and 14, respectively (Appendix 2 and 3), which signals an important gain in consistency towards the stability of the list.
Among the additions to the main list, nine (21%) are long-distance, non-pelagic boreal migrants that winter primarily in lower latitudes of the Neotropics or in the tropics and subtropics of the Old World and Australasia.Four of these (Buteo platypterus, Leucophaeus atricilla, Contopus virens and Catharus cf.swainsoni) are Nearctic breeders that have their normal wintering ranges well to the north of RGS and    Eight other additions (19%) are austral migrants that breed in the southern cone of South America and in some cases also northward along the Southern and Central  suggested that birds recorded in RGS may be latitudinal migrants in transit between breeding grounds in the Pampas of Argentina and wintering areas in coastal marshes of southeastern Brazil.The existence of a restricted breeding population of P. acutipennis in the lowlands of southeastern South America has only recently been revealed (Roesler, 2009).Therefore, these records are most likely attributable to pseudo-vagrancy (sensu Gilroy & Lees, 2003), i.e., extralimital records of seemingly vagrant individuals that are actually pioneering new migration routes or using previously unknown migratory flyways at low densities (see Dias et al., 2010).
Though not considered an austral migrant (sensu Chesser, 1994 andStotz et al., 1996), Phoenicoparrus jamesi moves from its breeding sites in the high Andean plateaus to lower elevation wetlands in the central plains of Argentina in the non-breeding season (Caziani et al., 2007).
Dias & Cardozo (2014) attributed the only RGS record of this species to vagrancy and speculated that it might have reached the Atlantic coast of the state following the more abundant and widespread Chilean Flamingo Phoenicopterus chilensis Molina, 1782, which breeds in north-central Argentina and regularly migrates to southern Brazil.
Six novel species (14%) are pelagic or coastal seabirds that disperse widely in southern oceans or throughout the Atlantic in non-breeding season.All but Phoebetria palpebrata are expected as regular (albeit scarce) transients or non-breeding visitors to waters off southern Brazil based on their known at-sea distribution and migration routes.Three are transequatorial migrants from the northern (P.deserta, Calonectris diomedea) or southern hemisphere (Stercorarius maccormickii).The latter is mostly pelagic in winter (Furness et al., 2017), while the coastal Sterna vittata may be overlooked because of identification problems and potential confusion with similar species present along the RGS coast at different times of the year (Bugoni & Vooren, 2005).Fregetta tropica has tentatively been regarded as rare off the coast of southern and southeastern Brazil, but nocturnal habits and difficulty of distinguishing it from F. grallaria (Vieillot, 1818) in pelagic censuses may be leading to an underestimation of its regional abundance (Petry et al., 2016).Records of P. deserta in particular point to the great potential of geolocation data to reveal the occurrence of rare or unexpected seabirds in pelagic waters within the Exclusive Economic Zone of the state, particularly species that are not usually attracted to fishing vessels, as is the case with most gadfly petrels (Pterodroma spp.).As demonstrated by such studies, Pterodroma arminjoniana (Giglioli & Salvadori, 1869) has a high probability of occurrence in RGS territorial waters during the breeding season (Krüger et al., 2016).This species has been recorded in both Uruguayan (Abreu et al., 2010) and Argentinean (Savigny et al., 2005) waters, and therefore is a strong candidate to have its occurrence in RGS confirmed in the near future.
Four additions (9%) -Melanerpes cactorum, Tarphonomus certhioides, Phacellodomus sibilatrix and Saltatricula multicolor -are non-migratory species primarily or entirely restricted to the Chaco region (Stotz et al., 1996), of which the Espinal Province (sensu Cabrera & Willink, 1973) can be considered an extension.The Espinal reaches RGS as the Vachellia-Prosopis espinilho parkland, typical of the westernmost tip of the state, where several Chaco species occur (Belton, 1984).Bird species above have range boundaries nearly abutting the western border of RGS in adjacent Argentina and Uruguay.Because range boundaries are dynamic and abundances tend to be lower near the edges of a species distribution (Brown et al., 1996), the presence of extralimital Chaco specialists in this part of the state, either as "transgressive" individuals or temporary colonizers, may simply reflect the spatial variation that is expected to occur over time at the periphery of their geographic distributions.The same may apply to records of P. cirratus, which presumably refer to the Chaco subspecies pilcomayensis, although in this case a recent range expansion may also be involved (Santos et al., 2015).Saltatricula multicolor is considered largely sedentary (Jaramillo, 2017), but data presented by Delhey & Scorolli (2002) indicate that it is apparently migratory at the southern end of its range in La Pampa and southwestern Buenos Aires provinces.Therefore, birds recorded in RGS in May (Bellagamba et al., 2013) could alternatively be thought of as being austral migrants originating in this region, as suspected by Di Giacomo (2005) for birds recorded mostly in fall and winter at El Bagual Reserve in the province of Formosa, Argentina.
At least nine species added (21%) are, or are likely to be, recent invaders from the north.These include both open-zone (Gampsonyx swainsoni, Todirostrum cinereum, Phaeomyias murina, Fluvicola nengeta, Xolmis velatus, Casiornis rufus, Campylorhynchus turdinus) and forestbased species (Pipra fasciicauda, Tangara ornata).Explicit evidences of recent range expansion in southeastern South America exist for T. cinereum, F. nengeta, C. turdinus and X. velatus (Piacentini et al., 2004;Straube et al., 2007;Bodrati et al., 2012;Meyer, 2013Meyer, , 2016)), and elsewhere for G. swainsoni and F. nengeta (Aguiar, 2010;Dort et al., 2010).With the exception of C. rufus, none of these species has so far been reported from Uruguay.Melanerpes cactorum is believed to be expanding its range in northeastern Buenos Aires province (Guerrero & Agnolin, 2016) and should perhaps be included among the newly colonizing species.Records of C. rufus in northwestern Uruguay are recent and suggest a range expansion directly from the Paraná-Paraguay river corridor or from the south, since the species is apparently absent along the Uruguay river in eastern Corrientes, but occasionally reach the gallery forests of the La Plata river (Montaldo & Roitman, 1999).A double invasion of the state by F. nengeta and X. velatus can be inferred from the WikiAves records.Both species appear to have entered RGS simultaneously from the north/northwest and via the southern littoral of Santa Catarina, thereby circumventing the highlands of the Planalto.Birds of open and semi open zones undoubtedly benefited from large-scale land-use changes in originally forested regions of southern Brazil, northeastern Argentina (Misiones) and southeastern Paraguay, which created favorable conditions and enabled the southward dispersion of these species.Significant positive trends in surface air temperature recorded in southern Brazil and elsewhere in southeastern South America over the last few decades may also be driving shifts in regional species ranges (Bencke, 2010;Guerrero & Agnolin, 2016).
Thlypopsis sordida and Donacobius atricapilla may be recent invaders from the north as well.The former is suspected to be expanding in southern Brazil (Arzua et al., 2001;Ghizoni-Jr & Silva, 2006), while the latter is known from unconfirmed records in RGS (Bencke, 2001).However, both have been recorded in southern Misiones and eastern Corrientes close to the RGS border (Chebez, 1996;Capllonch et al., 2005;Nores et al., 2005), either historically or in recent times, and may simply have gone unnoticed until recently.
The record of Heliornis fulica does not clearly fit into any of the previous cases.Instances of short-distance vagrancy involving this species have been recorded elsewhere (Trinidad, Bonaire and New Mexico, USA; Bertram & Kirwan, 2017).Nearby RGS, it is known from several localities in the provinces of Misiones and Corrientes (south to at least 30°S) and from an old record in northeastern Santa Catarina (Carbonell, 1987;Parera, 1987;Rosário, 1996).It may be regularly present in RGS in small numbers and should be looked for along the Ibicuí and other well preserved rivers of the north and west of the state.
Finally, the exotic S. vulgaris has not been listed among the non-native bird species introduced to Brazil (Fontoura et al., 2013).It is presumably spreading northwards from Uruguay and Argentina, perhaps via jump dispersal, since it is not known to occur in these countries close to the RGS border.However, its establishment in the state as a wild bird has yet to be confirmed.
Documentation.Following the documentation updates, the percentage of accepted species without documented occurrence in the state decreased from 1.8% in 2010 to 0.7% in the present review.Currently, only five RGS species still lack documentation.These are known exclusively from sight records -Notharchus swainsoni (Gray, 1846), Anodorhynchus glaucus (Vieillot, 1816) and Chloris chloris (Linnaeus, 1758) -and band/data-logger recoveries -P.deserta and Anas discors Linnaeus, 1766.Of the 699 species for which physical evidence is available, 590 are documented primarily by museum specimens (in one case complemented by an audio recording), three by osteological material, 95 by photographs (in three cases complemented by voice recordings), and 11 by audio recordings.
Identification and documentation accounts.The documentation indicated or provided in Tables I, II and in Figures 1-29 allows accurate identification of most species included in the main list based on unpublished information.Below we detail records and discuss diagnostic characteristics of some of the added or newly documented species whose identification is not straightforward.
Phoebetria palpebrata.One specimen found on the beach at Itapeva, Torres, was photographed and collected by the MUCIN staff on 14 May 2014, and its skeleton was preserved, together with tissue samples (MUCIN 591).A second specimen (field number AM 881) was found on the beach at Arroio do Sal on 5 October 2016; it was photographed (Fig. 1) and both the skeleton and tissue samples will be included in the same collection (M.Tavares, pers.comm.).Both specimens are clearly assignable to P. palpebrata rather than to the similar Phoebetria fusca (Hilsenberg, 1822) due to their light upper dorsum and neck.
Thalassarche steadi.This species was included in the main list based on an immature female collected in 2011 on the northern littoral of the state and identified using discriminant functions and molecular techniques (Pereira et al., 2016).The specimen (also immature) that served as the basis for the inclusion of Thalassarche cauta in previous lists, reported by Petry et al. (1991), has not been examined with the same methods and cannot be presently assigned to either species because immatures of T. cauta and T. steadi are indistinguishable on external characters (Pereira et al., 2016).Following this reasoning, we excluded the former from the list.
Pterodroma deserta.Ramírez et al. (2013) provided tracking data based on global location sensors (GLSs), or geolocators, deployed on breeding birds at Bugio Island, North Atlantic Ocean.At least five birds made intense use of waters within the EEZ of southeastern and southern Brazil.Some daily locations plotted on map of Fig. 1 in Ramírez et al. (2013:273) are within EEZ waters of RGS.Because some of these locations are clearly within 140 km of the coast of RGS and geolocators have a maximum error of 200 km (Phillips et al., 2004), even adding the maximum possible error of devices, locations are still within the 200 nautical miles (c.370 km) of the EEZ.
Charadrius mongolus/leschenaultii.There are four photographs of an unidentified Charadrius plover taken at the Parque Nacional da Lagoa do Peixe in the WikiAves website.Photographs WA1943013 and WA1944101 were taken by A. Briso on 5 December 2015, whereas WA1963959 and WA1963968 were obtained by F. Ronaldo on 29 December 2015.All photographs are presumably of the same individual, as inferred from the overall aspect of the birds and commentaries made by F. Ronaldo (who guided A. Briso and saw the bird on both occasions).In all photographs, the depicted individual is in basic plumage.The upperparts are dull gray-brown, with wing feathers displaying lighter margins.The forehead is white and the whitish supercilium broadens behind the eye.The underparts are white except for the large grayish-brown patches on the sides of the breast.The bill is black and relatively long, with mandibles gently tapering to a point at the tip.The maxillary unguis in A. Briso's photographs is approximately half of the total bill length.Legs are long (including the tibia), dark olive-gray in A. Briso's photographs and light gray in F. Ronaldo's.Differences in color reflect the influence of substrate (wet mud in A. Briso's photographs and dry soil in F. Ronaldo's).Toe joints appear to be the same color as the rest of the legs (visible in WA1963968).The tarsus-to-bill-length ratio ranged from 1.86-2.0 in WA1943013 and 1.94-2.04 in WA1944101 (variation of four different measurements of both tarsus and bill).The species is a mid-sized Charadrius, judging by comparisons with a Calidris fuscicollis (Vieillot, 1819) and two Charadrius semipalmatus Bonaparte, 1825 that appear in the background of WA1944101 and WA1963959, respectively.
A combination of characteristics, namely lack of a complete white neck collar, single incomplete breast-band, long, relatively large bill and long, dusky legs point to either the Greater Sand Plover Charadrius leschenaultii Lesson, 1826 or the Lesser Sand Plover Charadrius mongolus Pallas, 1776 (Hayman et al., 2011;O'Brien et al., 2006).Both species breed in Asia and spend the non-breeding season in more southern latitudes of the Old World (Hirschfeld et al., 2000;Wiersma et al., 2017a,b).Vagrancy is common, with both species being recorded in North America and C. mongolus also in South America (Abbott et al., 2001;Le Nevé & Manzione, 2011;Wiersma et al., 2017a,b).Clear separation of these species based on photographs of birds in basic plumage is complex because of the overlap in some diagnostic characteristics between the smaller, slenderbilled C. l. columbinus Wagler, 1829 and the "atrifrons" subspecies group of C. mongolus, which have slender and more pointed bills than the subspecies of the "mongolus" group (Hirschfeld et al., 2000).Identification of the RGS bird is further complicated because important diagnostic field marks of the wing and tail are not visible in the photographs (Hirschfeld et al., 2000).Leg color and tibia exposition point to C. leschenaultii (Hirschfeld et al., 2000).However, some C. mongolus may have greenish-grey legs and C. leschenaultii may have grayer legs (Hirschfeld et al., 2000).Judging the length of tibia exposition from photographs is further complicated because of distortions caused by the bird compressing or fluffing belly feathers in response to air temperature (Hirschfeld et al., 2000).The RGS bird not only has compressed body feathers but also has a concave abdomen, which may pass the impression that the tibia is more exposed than it really is.The concolor toe joints in relation to the rest of the legs points to C. mongolus, since C. leschenaultii often (but not always) has darker toe joints (Hirschfeld et al., 2000).The relatively large, pointed bill, which is clearly not blunt-ended, is reminiscent of C. leschenaultii (Hirschfeld et al., 2000).The maxillary unguis in A. Briso's photographs is approximately half the total bill length, which is within the range of C. l. columbinus (Hirschfeld et al., 2000).On the other hand, the tarsus-to-bill-length ratio proposed by Millington (1988) to separate these species puts the RGS bird within the range of C. mongolus.Although this formula seems to be accurate, the ratio is difficult to determine because the angle from which the photograph is taken must be considered (Hirschfeld et al., 2000).Nonetheless, all ratios we calculated were of birds in lateral views and they were far above the range of C. leschenaultii, which is 1.43-1.78(Millington, 1988).Differences in body molt schedules between both species, summarized in Hirschfeld et al. (2000), are not applicable because in December both species are not molting and should be in full basic plumage.However, upperparts are relatively paler and sandier -characteristic of C. leschenaultii -than in C. mongolus (Abbott et al., 2001).Although we are inclined to consider the bird as C. leschenaultii for presenting a larger set of characters within the range of this taxon, it shows a considerable overlap in some characters with C. mongolus.We thus consider the available evidence insufficient for an accurate identification and include the taxon in the RGS list as Charadrius sp.
Numenius hudsonicus Latham, 1790 and N. phaeopus.Sangster et al. (2015) recommended that the New World form hudsonicus (American Whimbrel) be treated as a species distinct from the Eurasian Whimbrel N. phaeopus based on previously published evidences and a new molecular phylogenetic analysis using available mtDNA datasets.The separation of these two taxa had already been proposed independently by Livezey (2010) based on a cladistic analysis of phenotypic characters.This treatment has recently been adopted by the CBRO (Piacentini et al., 2015).Bencke et  2010) provided photographic evidence for the occurrence of hudsonicus in RGS and attributed all documented records of whimbrels hitherto known from the state to this form (as N. phaeopus hudsonicus).Following the split of hudsonicus, this species should replace N. phaeopus in the RGS list.However, one individual of the Eurasian Whimbrel was recently photographed on the coast of RGS by a group of birdwatchers (Tab.I), thus justifying the maintenance of this species on the state list.The extensively white rump and uppertail-coverts of the bird rule out the commoner American Whimbrel (O'Brien et al., 2006).
Tringa inornata.The separation of T. inornata and T. semipalmata in basic plumage is based mostly on structural features (O'Brien, 2006;O'Brien et al., 2006).Tringa inornata is larger and more elegant, with a more elongated body, longer legs and neck, and longer, slimmer, finer-tipped bill.The angle from the bill to the forehead is steeper, and the crown is sometimes distinctively tall or puffy (O'Brien, 2006;O'Brien et al., 2006).The bird photographed at São José do Norte (Fig. 4) is readily identified as T. inornata based on bill and head features, and is more diagnosable than other individuals mentioned in Martínez-Curci et al. (2014).
Calidris pugnax.C. B. Andretti photographed a single bird along the Talha Mar road, Parque Nacional da Lagoa do Peixe, Tavares, on 18 January 2017 (WA2466615).G. R. Peres later photographed what appears to be the same individual on the same locality on 6 February 2017 (WA2462329, WA2477328, WA2477334).The bird is noticeably larger than C. fuscicollis and about the same size as or slightly smaller than Pluvialis dominica (Statius Muller, 1776), which appear with it in WA2477334.It has a small head, a rounded body with a prominent belly and a humped back.The upperparts look scaled due to distinctive pale fringing on the grayish brown scapulars, wing coverts and tertials, most of which are centered or barred black.Most of the head, hindneck and mantle has a slight sandy-buff wash.The belly, flanks and undertail coverts are white, and the throat and the area around the base of the bill are whitish.
The bill is dark and droops slightly toward the tip.The legs are long and yellow-orange.Based on comparisons with C. fuscicollis and P. dominica, the bird measured approximately 190-230 mm, which is within the range of females or rare faeder males, which measure 200-250 mm (regular males measure 260-320 mm) (Karlionova et al., 2007;Hayman et al., 2011).This bird is also in definitive plumage, as inferred from the lack of a buff wash on the underparts and of buff fringes on scapulars, wing coverts and tertials (Hayman et al., 2011).In addition, juveniles have dull yellowish-brown or greenish legs, never yellow-orange as the RGS bird (Hayman et al., 2011).M. Alievi photographed a second individual along the same road on 15 February 2017 (WA2513429; Figs 23, 24).This bird is similar to the first one but its bill is marked orange-yellow and the upper breast has vitiligolike white blotches, thus suggesting a male.Structural and morphological characteristics visible in photographs separate the birds from any other shorebird occurring in RGS and the somewhat similar Tringa totanus (Linnaeus, 1758) and Calidris acuminata Horsfield, 1821 from the Old World (Hayman et al., 2011).
Nannochordeiles pusillus.The combination of rufoustinged upperparts, white throat, barred underparts, white bar on primaries, and wings about the same length as the tail when resting are distinctive (Cleere, 1998;Ridgely et al., 2015).The Itaqui bird (Fig. 27) appears to be a female, as indicated by the small white marks on the primaries, duller and buffier tips to secondaries, and lack of white tips on the rectrices (Cleere, 1998).The somewhat similar Chordeiles minor (Forster, 1771) is larger, less rufous, has wing tips projecting beyond the tail when resting and the white band on the primaries is larger and closer to the base of the wing (Cleere, 1998).The very similar Chordeiles acutipennis (Hermann, 1783), which has not been found in RGS, is also larger, with no rufous above, and has the white band on the primaries closer to the base of the wing (Cleere, 1998).As noted by Cleere et al. (2017), the race of the birds recently found in northeastern Argentina and RGS has yet to be determined.Pseudocolopteryx acutipennis.Although Bornschein et al. (2017) mention three records of P. acutipennis for RGS, there are actually only two.One individual was photographed by P. Fenalti in a wetland on the margin of Lagoa da Custódia, Tramandaí, on 9 December 2012 (WA828480) and another was photographed by P. Fenalti (Fig. 8; WA1972873), V. E. Florencio (WA1979557) and O. A. Fenalti (WA1967150) on the eastern edge of Lagoa dos Patos at Porto do Barquinho, Mostardas, on 28 December 2015 (P.Fenalti, pers.comm.).Records for "Mostardas" and "Lagoa do Peixe" on Table 1 of Bornschein et al. (2017) thus refer to the same bird and are mislocated.Diagnostic features visible in the photographs include the yellowish-olive hue to the upperparts, the dusky lores and cheeks, the olive crown concolor with the back, and the bright yellow underparts (Ridgely & Tudor, 2009).The similar Pseudocolopteryx flaviventris (d'Orbigny & Lafresnaye, 1837) is duller, has brownish upperparts and a rufescent crown (Ridgely & Tudor, 2009).Pseudocolopteryx dinelliana Lillo, 1905, which has not been found in the state (Bencke et al., 2010), has a rufescent tone on crown, lores and cheeks (Ridgely & Tudor, 2009).
Tyrannus tyrannus.Photographs taken at Mostardas (Figs 9, 10) show a molting bird with faded and worn flight feathers (except outermost tertials, which are new).It has slaty upperparts and white underparts, with a sharp contrast between the dark head and the white throat.The bill and tail are black, the latter with an evident white tip.Characteristics visible in photographs, especially the white tip to the square tail, separate it from Tyrannus savana Daudin, 1802 and T. dominicensis (Gmelin, 1788), the latter not found in RGS (Ridgely & Tudor, 2009;Bencke et al., 2010).The similar T. caudifasciatus D'Orbigny, 1839 is resident in the West Indies and has a larger head, a stronger bill and a grayer back.
Contopus virens.We reidentified a photograph originally assigned to the Tropical Pewee Contopus cinereus (Spix, 1825) in the WikiAves website (WA2362350) as being of this species and obtained new photographs of the same individual that confirm the identification .The bird was photographed in a small clearing inside coastal restinga woodland during a birdwatching tour to the Parque Nacional da Lagoa do Peixe at Tavares (R. Fortes, pers.comm.).The Wood-Pewees (C.virens and C. sordidulus Sclater, 1859) can be distinguished from the local (nominate) subspecies of the Tropical Pewee by their longer primary extension, more distinct wing bars, heavier bill, and less uniform underparts .Tropical is also grayer and darker overall than either, especially on the crown and underparts (Schulenberg et al., 2007;Ridgely & Tudor, 2009;Farnsworth & Lebbin, 2017a,b;Farnsworth et al., 2017).In addition, this species is restricted to the dense forests of the extreme north in RGS, along the Pelotas and Uruguay rivers (Belton, 1994;Bencke et al., 2003).Therefore, its occurrence in the mostly open areas around Lagoa do Peixe is highly unlikely.
Separation of Eastern C. virens and Western C. sordidulus Wood-Pewees is far more problematic and they are often distinguished with certainty only by voice (McCarty, 1996;Bemis & Rising, 1999;Ridgely & Tudor, 2009).Differences in plumage, structure and bare part colors are subtle and much overlap exists (see review in Lee et al., 2008).Consequently, no identification can be regarded as conclusive unless based on a combination of characters.Western tends to be darker (less gray and more dusky-olive) overall, with a more extensive "vest" (broad chest band often not interrupted medially and typically extending down to flanks).Its upper wing bar is duller and fainter than the lower wing bar, showing less contrast with the dark wing, and its lower mandible is duskier, with the pale color usually restricted to the base and seldom extending nearly to the tip.Eastern in turn is generally paler throughout (typically olive-gray above and more extensively whitish below), shows two equally contrasting wing bars and has a proportionately longer tail (because tail extension, or the distance from wing tip to tail tip, is greater than primary extension in this species).The extent of light color on the lower mandible is also greater than in Western and adults often have the bill mostly or entirely orange-yellow underneath (Hilty & Brown, 1986;Stiles & Skutch, 1989;Fjeldså & Krabbe, 1990;McCarty, 1996;Bemis & Rising, 1999;Lee et al., 2008;Farnsworth & Lebbin, 2017a,b).
We assume the Tavares bird as being C. virens based on its relatively pale overall coloration, rather restricted "vest" (not extending down to flanks) and extensively orange-yellow lower mandible .Although the primary extension is not directly visible in any of the available photographs, the length of the tail extension also suggests this species instead of C. sordidulus.Evaluation of the relative contrast between the upper and lower wing bars in the Tavares bird is hampered by the high degree of wear of the median secondary coverts.However, although reduced, the upper wing bar appears to be as bright as the lower one, again suggesting C. virens.Finally, C. sordidulus is thought to winter primarily on the Andean slopes from 400 to 1700 m a.s.l., while C. virens is more widespread in the adjacent lowlands, especially in western Amazonia (Stotz et al., 1992;Ridgely & Tudor, 2009).Hence, the latter is the most likely species to be expected in eastern South America.
The RGS record of C. virens is apparently only the third for the country outside Amazonia.Previous known occurrences in eastern Brazil were at Chapada do Araripe, Ceará (Teixeira et al., 1993) and Ubatuba, northern littoral of São Paulo (Silveira & Uezu, 2011).Lagoa do Peixe also appears to be a new southernmost locality for the species within its wintering range, since outside the Brazilian territory there seem to be no published records south of Orán, Salta, in northwestern Argentina (Hoy, 1981;Ridgely & Tudor, 2009).
Catharus cf.swainsoni.Meyer et al. (2014) reported the first record of a migrant North American Catharus in RGS from Erechim, where one bird identified as C. swainsoni was sighted with the aid of playback on 15 February 2013.Two poor-quality audio recordings documenting this record were deposited in the WikiAves (WA888672, WA914300) and Xeno-Canto (XC127370, XC127519) websites (date in XC127519 mistakenly given as 15 March).Another RGS record of a Catharus, also identified as C. swainsoni, occurred on 18 January 2016 at Marques de Souza (Fig. 16; WA2261365, WA2261370).The photographs that document this record are of poor quality but allow the recognition of a Catharus thrush with distinct buffy eye-ring and loral stripe ("spectacles"), white underparts and densely spotted breast.The face, lower throat and upper breast are washed with buff (color saturation artificially enhanced by flash), the throat is bordered on each side with a dark stripe and the mandible is extensively pale.In neither case did the authors of records mention the diagnostic characteristics on which the field identification was based, nor did they refer the taxonomic source followed to treat C. swainsoni as a species separate from C. ustulatus (Nuttall, 1840), which presumably was the latest CBRO list available at the time of each record.These taxa are considered only subspecifically distinct in most recent sources (e.g., AOU, 1998 and supplements) and differ in subtle vocal and morphological features.They also have nearly completely segregated winter ranges, with russet-backed birds (ustulatus group) wintering in Mexico and Central America and olive-backed birds (swainsoni group) wintering primarily in Panama and western South America, south to northern Argentina (Collar & Christie, 2017).In Brazil, swainsoni has been treated as a full species by the CBRO since 2010 and is the only taxon with accepted occurrence (Piacentini et al., 2015).However, there are very few specimen records for the country (e.g., Pinto, 1944;Willis et al., 1993) and undoubtedly, most modern Brazilian records of a "spectacled" Catharus have been automatically assigned to swainsoni based solely on the assumption that this is the only likely taxon in Brazil.In our opinion, the available documentation for the RGS records does not contain sufficient information to allow a reliable identification based on diagnostic characters.Consequently, the identity of the birds as C. swainsoni can be assumed only on biogeographical grounds.The voice of the individual recorded at Erechim, although reminiscent of that of C. ustulatus/swainsoni, is barely audible on the recordings, which makes any comparison inconclusive.Likewise, the photographs taken at Marques de Souza do not show the upperparts of the bird, where lies the main difference between the two taxa (Mack & Yong, 2000;Collar & Christie, 2017).Therefore, we chose to cite these records as Catharus cf.swainsoni on the main list until better evidence is available.
Sporophila hypochroma.This seedeater was listed as probable in Bencke et al. (2010) because of the possibility of confusion with phenotypically similar individuals of Sporophila hypoxantha Cabanis, 1851 and plumage variants of S. cinnamomea (Lafresnaye, 1839) when the identification cannot be confirmed by voice, which was the case of the records hitherto known from the state.An audio recording made by M. Repenning at Uruguaiana on 1 February 2015 (WA1645476; Tab.I) is referred to S. hypochroma and confirms the occurrence of this species in the west of RGS.A photograph of the same individual (as inferred from information issued by the authors and the date and time of the record) taken by S. Messias is available at the WikiAves website (WA1647035).
Sturnus vulgaris.This Eurasian species became established around Buenos Aires, Argentina, in 1987 and has since been expanding its range in southeastern South America (Peris et al., 2005).The first Uruguayan records were in 2008 at Montevideo (Mazzulla, 2013).On 29 October 2014, and again on 23 September 2015, an adult male (Fig. 20) was photographed by V. Souza at his property in Lavras do Sul (Tab.I).These are the first records of this highly invasive species in Brazil.
Species of probable and hypothetical occurrence.A photograph of a Sporophila seedeater of the capuchino group made by E. Mandarino at Barra do Quaraí (30°12'S; 57°33'W) on 17 November 2013 (WA1155820) matches the color pattern of the recently described S. iberaensis from the Iberá marshes in northeastern Argentina and adjacent areas of Paraguay and southwestern Brazil.However, since there is considerable plumage variation among members of this group, and full diagnosis relies on voice characteristics, we consider it as probable in the state until further evidence is available.
Similarly, an atypical martin photographed north of Coxilha (28°07'S; 52°18'W) (Fig. 30), identified as a Greybreasted Martin Progne chalybea (Gmelin, 1789), closely resembles the rare and poorly known Sinaloa Martin Progne sinaloae of northwestern Mexico.The bird is glossy steel-blue above, with a deeply forked tail.The flanks, sides of throat and most of the upper breast are also steel-blue, with the lower breast, abdomen and undertail-coverts contrastingly white.The center of the throat is grey and the dark band on the breast is mottled or coarsely scaled whitish to greyish.The Sinaloa Martin is an endemic migrant breeder to the pine-oak woodlands of the Sierra Madre Occidental and Central Volcanic Belt in Mexico (Lethaby & King, 2010).Like the closely related and very similar Caribbean Martin Progne dominicensis (Gmelin, 1789), it is suspected to winter primarily in South America, but its wintering range remains unknown (Turner & Sharpe, 2017).It starts departing breeding grounds in late July and is absent from Mexico between early September and mid-February (Lethaby & King, 2010).The RGS record is consistent with this migration schedule and an occurrence of P. sinaloae in the state, at least as a vagrant, seems plausible.On the other hand, the species would perhaps be expected to winter much closer to its breeding area and well to the north of southern Brazil.Moreover, both the extent of plumage variation in nonbreeding Progne martins and the identification criteria used to distinguish the various species are still poorly known (Tobias et al., 2006).For these reasons, we regard the identification of the RGS bird as tentative and list the species as hypothetical in the state (Appendix 3).
As for C. cyaneus, the reddish legs (visible in Fig. 31) readily separate this species from Cyanerpes caeruleus (Linnaeus, 1758), which is not found in eastern and southeastern Brazil (Ridgely & Tudor, 2009).Furthermore, females of the latter have a blue malar streak and distinctly streaked underparts (Ridgely & Tudor, 2009).However, C. cyaneus is not known to occur south of São Paulo (Ridgely et al., 2015) and we do not rule out the possibility that the bird photographed at Santo Antônio da Patrulha (29°50'S; 50°31'W) was an escapee or had been intentionally released into the wild.For this reason, we prefer to include it among the species of hypothetical occurrence (Appendix 3) until better evidence is available.
Sicalis citrina is likewise listed as hypothetical because the voice recording that documents the only RGS record (Meyer et al., 2014) is considered not diagnosable.Although the species is known to occur in the upland grasslands of southeastern Santa Catarina (Rosário, 1996; WikiAves archive) and is therefore likely to occur in adjacent areas of RGS, the fact that the bird was not seen (Meyer et al., 2014) prevents a circumstantiated evaluation of this record.Moreover, we maintain Laterallus exilis as hypothetical despite a new evidence (voice recording; WA2053723), which in our opinion also does not allow a precise identification.
Concluding remarks.The rate of new birds added to the RGS list shows no signs of slowing.The addition of 43 species since the last update corresponds to an average increase of over six species per year, which is nearly 30% higher than that recorded for the previous period (44 additions in 2001-2010;Bencke et al., 2010).Of the newly incorporated taxa, only 10 to 14 (~23-32%) are known or suspected to be undergoing a recent range expansion in southeastern South America and are thus confirmed or potential newcomers to RGS (nine northerly species plus S. vulgaris, and perhaps P. cirratus, M. cactorum, D. atricapilla and T. sordida).The remaining are mostly migratory birds recorded as vagrants or irregular visitors (22 species), along with previously overlooked migratory or resident taxa present at low densities or with marginal occurrence in the state (7-11 species).This points to a better coverage -both spatial and temporal -of the state by amateur and professional birders in recent years, leading to an increased detection of unusual species ("rarities").
This greater observer coverage is due in large part to the increasing contribution of citizen science to knowledge on bird occurrence and distribution in RGS.Nearly 60% of the new occurrences reported here resulted from the activity of amateur bird watchers and photographers.This contribution is reflected in the way records are publicized, documented and geographically distributed.Of the 43 additions to the main list, 36 (84%) are based on records documented exclusively by photographs, of which no less than 28 first appeared in the WikiAves website.Common destinations of bird photographers, such as the Lagoa do Peixe and surroundings, account for a relatively high proportion of the new records for the state.In contrast, only three additions were substantiated by the collection of whole or partial specimens.Similarly, half of the documentation updates implemented here is based on photographic records.As a direct consequence of this trend, the proportion of accepted species documented by museum specimens in the state has decreased steadily over time, from 91% in 2001 to 88% in 2010 and 84% in the present review.This leads to a timely reflection: while citizen science is highly welcome and undoubtedly contribute a considerable amount of valuable data to science, it must be remembered that sound science is not only grounded on an adequate amount of information, but gathers strength and reliability from the thorough documentation of facts.Museum specimens allow direct, objective comparison of morphological traits and provide access to genetic resources for molecular studies and biological tissues for chemical analyses, among other things (Wiley et al., 2017).The recent record of C. diomedea for RGS (Oliveira et al., 2017) provides a good example of how important and effective specimens can be for accurate and reliable identification.Investment in scientific research and collection of voucher specimens should keep pace with the increasing interest birds arouse among people.We thus urge professional ornithologists, academic community and governmental organizations to document the biodiversity in RGS as rigorously as possible through responsible and carefully planned collections.In doing so we will be honoring and continuing Belton's landmark work on the birds of RGS, without which the task of updating this list would be at a much earlier stage.
casual or accidental vagrants to RGS and South America, since both records involved migrating birds that obviously strayed well off their normal routes.The occurrence of the latter is presumably linked to the transatlantic vagrancy of Palearctic migrants headed to Africa from northwestern Europe.Such birds may continue heading south after arriving in equatorial South America or the Caribbean and may eventually reach higher latitudes along the Brazilian coast.The only previous records of N. phaeopus in Brazil are from Fernando de Noronha Archipelago, a well-known vagrant trap for Old World species(Silva e Silva & Olmos, 2006).Concerning Charadrius sp., a central Asian species, two distinct routes by which displaced or misoriented migrants can reach South America are perhaps equally likely: one crossing the Atlantic ocean from Africa and the other via the islands off western Alaska and the Pacific coast of North America.

Figs
Figs 11-14.Contopus virens, Tavares, 8 November 2016 (J.L. J. Ribeiro).Note the whitish throat, center breast and belly contrasting with the darker breast-sides (giving a "vested" appearance), relatively long, pointed wings, distinct wing bars, and rather heavy-looking bill, which exclude the sympatric C. cinereus.The combination of a pale overall coloration (especially underneath), extensively orange-yellow lower mandible and relatively long tail extension are consistent with C. virens and not C. sordidulus.
Four decades after Belton: a review of records and evidences...