Introduction
The Santa Marta Sabrewing Campylopterus phainopeplus is the rarest and most poorly known of the 24 endemic bird species restricted to the Sierra Nevada de Santa Marta (SNSM) (BirdLife International 2024a), an isolated coastal mountain range in northern Colombia that is considered the most important continental centre of endemism in the world and Earth’s most irreplaceable nature reserve (Duran-Izquierdo and Olivero-Verbel Reference Duran-Izquierdo and Olivero-Verbel2021; Le Saout et al. Reference Le Saout, Hoffmann, Shi, Hughes, Bernard and Brooks2013). With only one documented record since 1946, this hummingbird is listed as “Critically Endangered” by the International Union for Conservation of Nature (IUCN) given its rarity and presumed very small population (<50 mature individuals; BirdLife International 2024b). The lack of confirmed records also resulted in Santa Marta Sabrewing being listed as a “lost” species (Rutt and Mittermeier Reference Rutt and Mittermeier2023) and included in the top 10 most wanted lost birds in 2021 by the Search for Lost Birds (https://www.searchforlostbirds.org).
The first specimens of Santa Marta Sabrewing were collected by English geographer and naturalist F.A.A. Simons during his expeditions to the south-eastern slope of the SNSM in 1878 (Simons Reference Simons1879, Reference Simons1881). The species was described by O. Salvin and F.D. Godman in 1879 based on Simons’ specimens (Salvin and Godman Reference Salvin and Godman1879). Simons spent seven months mainly in two localities named Atánquez and San José, now part of the Cesar Department, as well as in Alto de Macotama, now in La Guajira Department (Figure 1). Between 1945 and 1946, American ornithologist M.A. Carriker collected 11 additional specimens of Santa Marta Sabrewing for the National Museum of Natural History, Smithsonian Institution, Washington, DC. One of those specimens was collected in San José, the type locality, while the remaining 10 were collected in Chendúcua, located at a higher elevation along the Guatapurí basin (Figure 1). Since Carriker’s collection, there were no documented records of the sabrewing until 2010, when researchers took a photograph of a mist-netted male in the El Dorado private reserve, Magdalena Department, within the Important Bird and Biodiversity Area (IBA) of Cuchilla de San Lorenzo, located ~60–85 km west of the previously known localities (Figure 1).
(A) Historical records and confirmed locations of Santa Marta Sabrewing presence in the Sierra Nevada de Santa Marta, northern Colombia, together with locations of field surveys conducted in 2022–2023. Numbered localities are: (1) San Lorenzo; (2) Alto de Macotama; (3) Chendúcua; (4) San José; (5) Atánquez (~8 km from Chemesquemena). (B) The Guatapurí River basin has extensive agricultural areas mixed with riparian forests, which are more common on steep slopes such as along La Macana stream near Chemesquemena. (C) Santa Marta Sabrewing was found in riparian forest strips near watercourses with dense understoreys. (D and E) Male and female sabrewings observed at the study site. (Photographs by Elquin Toro, John C. Mittermeier, and Esteban Botero-Delgadillo © reproduced with permission.)
With such limited records, information on the distribution, ecology, and natural history of Santa Marta Sabrewing is sparse. The species has been suspected to be an elevational migrant, with seasonal movements occurring between the dry and wet seasons (BirdLife International 2024b), although this has never been confirmed. While inhabiting humid forest borders and mixed landscapes of forest and shade-grown coffee plantations at 1,200–1,800 m asl during the dry season (February–May), it has been supposed to migrate to the páramo and up to the snowline at 4,800 m asl during the wet season (July–October; reviewed in BirdLife International 2024b; Cárdenas-Ortiz and Cortés-Herrera Reference Cárdenas-Ortíz, Cortés-Herrera, Renjifo, Amaya-Villarreal, Burbano-Girón and Velásquez-Tibatá2016). Literature on its natural history suggest that it consumes nectar from flowers of Musa and Inga trees inside coffee plantations and Befaria and Palicourea plants in forest edges, and that courtship displays take place from April to June (see Cárdenas-Ortiz and Cortés-Herrera Reference Cárdenas-Ortíz, Cortés-Herrera, Renjifo, Amaya-Villarreal, Burbano-Girón and Velásquez-Tibatá2016). However, there are no direct observations or published evidence confirming foraging visits to coffee plantations or courtship behaviour. Most knowledge about the sabrewing’s ecology is based on anecdotal records, and the fact that it is regularly misidentified (see Ayerbe-Quiñones Reference Ayerbe-Quiñones2018; McMullan Reference McMullan2016) casts doubt on the validity of available information. For instance, males of the remarkably similar Lazuline Sabrewing Campylopterus falcatus and Black-throated Mango Anthracothorax nigricollis seem to be frequently misidentified as male Santa Marta Sabrewings, whereas females are confused with female White-vented Plumeleteers Chalybura buffonii. Some putative vocal recordings archived in acoustic online repositories have later been confirmed to be incorrectly assigned to the species (Salaman et al. Reference Salaman, Cortés-Diago, Luna, Borrero-Álvarez and Toro2022). Thus, it remains unclear whether the scarce information on Santa Marta Sabrewing is accurate and reliable.
Remarkably, a population of Santa Marta Sabrewing was located in July 2022 by YV near the upper part of the Chemesquemena village (Cesar Department; ~8 km from Atánquez), not far from an area where Cristobal Navarro, a naturalist working on conservation projects with the Kogi indigenous people, may have seen the species in 2015 and 2019 (C. Navarro, personal communication, May 2022). In January 2023, CEL and AMC found isolated male individuals at a nearby locality in Chemesquemena and at San José (Figure 1). These findings were widely covered by national and international media and elicited a response from a consortium of research and conservation organisations to collect more information on the species. Here, we present the results of this effort including information collected over a 16-month field survey in four different localities in the SNSM (Figure 1). Specifically, we provide: (1) updated information on the species’ distribution from historical data and field expeditions conducted during 2022 and 2023; (2) preliminary projections of the species’ extent of occurrence (EOO) and area of occupancy (AOO) based on confirmed presence records; (3) new information on the species’ ecology and natural history based on a monthly monitoring of the population at Chemesquemena, including local rate of occupancy, seasonal movements, habitat associations, social and breeding behaviour, and a first characterisation of its vocal repertoire. We discuss our findings in the context of the information currently available for the species and evaluate the conservation implications.
Methods
Geographical distribution assessment and field expeditions
We conducted a thorough revision of the scientific literature, public databases, digital collections, and museum bird collections in search of confirmed records and specimens of Santa Marta Sabrewing. Public databases consulted included the Global Biodiversity Information Facility (GBIF) (www.gbif.org), eBird (www.ebird.org), xeno-canto (www.xeno-canto.org), and national-level data repositories, including SiB (www.sibcolombia.net). Bird collections consulted were those with good representation of SNSM specimens, including the Instituto Alexander von Humboldt (IAvH) and Instituto de Ciencias Naturales, Universidad Nacional de Colombia, in Colombia; the Carnegie Museum of Natural History, National Museum of Natural History (Smithsonian Institution), and the American Museum of Natural History in the USA; the Natural History Museum in the UK. We only considered evidence-based records including photographs, videos, audio recordings, and specimens.
In addition to fieldwork around Chemesquemena and nearby areas, we also conducted three field expeditions to search for Santa Marta Sabrewing at other localities where the species could occur (Figure 1). All searches for Santa Marta Sabrewing were conducted at elevations between 1,200 and 2,000 m. The first expedition was carried out during four days in December 2022 in Aracataca (Magdalena Department), a locality where the species could potentially occur according to recent distribution hypotheses. The second expedition was conducted during seven days in March and April 2023 in the San Lorenzo IBA (Magdalena Department), where the 2010 report was made. The third expedition was carried out during three days in May 2023 in the upper Rio Ancho basin in Dibulla (La Guajira Department), close to Alto de Macotama, where there have been recent unconfirmed sightings (C. Navarro, personal communication, May 2022). Searches consisted of ~6 hours/day, mainly between 07h00–11h00 and 15h30–17h30.
We built a minimum convex polygon (MCP) based on the confirmed records of Santa Marta Sabrewing to project the species’ EOO. Subsequently, we projected the species’ AOO by overlapping the EOO polygon with a geographical layer of suitable habitat (see IUCN 2022). A suitable habitat layer was generated by selecting areas of mature and secondary forests, riparian forests, and areas of perennial crops mixed with patches of native vegetation (see BirdLife International 2024b), delimited by the extent of the three biomes where the species’ presence was confirmed (see Results). We compared our EOO and AOO projections with the threshold values given under IUCN’s criteria B1 and B2 (IUCN 2022). Suitable habitat delimitation was conducted in ArcGIS Desktop 10 (ESRI 2011) using a vegetation layer from the CORINE Land Cover Methodology for Colombia, 2018, at 1:100,000 scale (IDEAM 2021), and a layer of biomes compiled in the public-access portal of the Land-use Planning Geographic Information System (SIG-OT) of the “Agustín Codazzi” Geographic Institute of Colombia (https://geoportal.igac.gov.co/contenido/datos-abiertos-geodesia). We present AOO metrics at both the scale of the analysis (i.e. 1 km2) and at a scale that is comparable with the Red List categories (i.e. 4 km2).
Population monitoring and occupancy
From September to December 2022, we conducted bird counts to estimate local occupancy and obtain a raw estimate of abundance along the course of La Macana stream (a tributary stream of the Guatapurí River) near Chemesquemena, where the sabrewing had been found by YV (hereafter referred to as the rediscovery site). Bird surveys were carried out using 20 georeferenced point-count stations (30 m radius), in which one observer (YV) searched passively (i.e. not using playback or other methods to attract the species) for Santa Marta Sabrewing for 15 minutes and documented every visual or acoustic detection. One count per station was conducted every month from 07h00 to 11h00, with visits to each site alternating between early and mid-morning across replicates. Count stations were located from 1,162 m to 1,819 m asl (mean: 1,615 m) in four land-cover types that were defined using the CORINE land-cover layer: secondary vegetation (thickets and early succession vegetation of at least 2 m height); early growth (shrubberies and abandoned pastures covered with bushes and ferns); mixed vegetation (mosaics of pastures, native vegetation, and productive lands such as agroforestry); pastures and open areas.
To estimate occupancy rate, we generated a detection history file consisting of all detection/non-detection data based on 80 surveys (four replicates at each of 20 point-count stations). Detection histories were analysed using the “occu” function in the unmarked package (Fiske and Chandler Reference Fiske and Chandler2011) in R 4.0.2 (R Core Team 2020). Given the limited number of detections (n = 23), we fitted an intercept-only model to reduce the risk of over-parameterisation. We then fitted two additional complex models to explore whether occupancy varied in response to elevation and/or land cover in the study area. One “saturated” model included as predictors the linear and quadratic effects of elevation and the proportion of native vegetation around each point count (100 m buffer around the point location), whereas the second one was an “elevation-only” model. The proportion of native vegetation was estimated using a buffer analysis that allowed us to estimate the percentage of native vs transformed vegetation around each point based on the CORINE land-cover layer. The detection probability was held constant for all three models.
Natural history and behavioural observations
We conducted ad libitum observations at the rediscovery site from July 2022 to October 2023 to describe different aspects of the species’ natural history. Over this period, YV made monthly 2–3-day visits to monitor individuals and territories, recording whether sabrewings were present along the course of La Macana stream (~1,150–1,850 m). We qualitatively described feeding habits and aspects of social and breeding behaviour, including territorial and lekking displays, vocal activity, and lek conformation.
Habitat associations were evaluated by calculating the proportion of detections of Santa Marta Sabrewing obtained in each of the four land-cover types registered during point counts (see above). We also investigated patterns of habitat selection, comparing the observed and expected frequencies of use for each land-cover type. We estimated the proportional extension of all vegetation types using remote images and the CORINE Land Cover Methodology for Colombia: secondary vegetation corresponded to 66% of all areas surveyed; early growth to 5%; mixed vegetation to 25%; pastures to 4%. Habitat use proportions and the confidence intervals were then compared with expected proportions through a Fisher’s exact goodness-of-fit test. Expected values by habitat were obtained by dividing the total number of detections of Santa Marta Sabrewing by the proportion of the area covered by each habitat (i.e. the combined area of all point counts and a 100 m buffer around them). Neu’s selection ratio (the ratio of observed to expected frequencies; Neu et al. Reference Neu, Byers and Peek1974) was also calculated for each land-cover type.
Vocal analyses
We conducted acoustic analyses to describe the sabrewing’s territorial call based on audio recordings of three males obtained near Chemesquemena. Given that these males were highly territorial and seemed to vocalise throughout the morning, and considering the distance between display sites, we assumed that these were different individuals. We selected and analysed a single high-quality cut per male of 20 seconds each. The first male (Male A) was recorded on 5 March 2023 at 09h00 at a small farm lot with coffee plantations intermixed with banana trees and patches of sugarcane (1,350 m asl, ~150 m distance to La Macana stream). This bird was perched ~4.5 m above the ground, and as no other individuals were present in the area, was likely the sole owner of its territory. The second male (Male B) was recorded on 21 May 2023 at 08h30 in a small secondary riparian forest patch and was perched ~2.5 m above the ground (1,480 m, ~40 m distance to La Macana stream). This individual was at a site where 3–5 males were frequently observed interacting throughout our study, in which we considered a lek. During the recording of Male B, another male was simultaneously vocalising nearby. The third male (Male C) was recorded on 27 May 2023 at 10h00 in a dense and relatively large patch of secondary premontane forest on a steep slope (1,780 m, ~200 m distance to La Macana stream). This bird was perched at ~3 m above the ground and while it was singing alone, a second male was observed ~50 m away. CEL obtained these recordings using a Zoom H6 recorder coupled with a Rode NTG4 phantom-powered shotgun microphone. The sampling rate was 48 kHz with a depth rate of 24 bits for Male A, and 44.1 kHz with a depth rate of 16 bits for Males B and C. We visualised and analysed the recordings using the Seewave package (Sueur et al. Reference Sueur, Aubin and Simonis2008) implemented in R.
Results
Geographical distribution
We found confirmed specimen records of Santa Marta Sabrewing from only four localities, all of them in or near the Guatapurí River basin: Atánquez, San José, and Chendúcua in Cesar Department, and Alto de Macotama in La Guajira Department (Figure 1 and Supplementary material Table S1). While Simons labelled his specimens as Atánquez, which is in the Badillo River basin, some or all might have been collected nearby Chemesquemena in the Guatapurí basin (1,200 m) (see Salvin and Godman Reference Salvin and Godman1879). In addition to specimen records, we reviewed one photographic record of a supposed male Santa Marta Sabrewing in the San Lorenzo IBA (Magdalena Department) from 2010 (see e.g. https://proaves.org/en/first-ever-photo-of-endangered-santa-marta-sabrewing/). This record has been widely accepted by the ornithological community for more than a decade, but in our review, it emerged as a clear outlier relative to the other confirmed records both in its geographical location and in its pattern of detection. Most importantly, the best-known field mark for distinguishing between Lazuline Sabrewing and Santa Marta Sabrewing (the bird’s tail) is not fully visible in the photograph. We concluded that this record should be treated as inconclusive, given the lack of additional supporting evidence.
Misidentified records have been excluded from public databases such as eBird or xeno-canto in previous years, but erroneous entries still need to be corrected or removed from other data sets. For instance, four female specimens collected in 2015 and deposited in the IAvH ornithological collection were examined and identified as female White-vented Plumeleteers (G. Bravo, personal communication, May 2023), but these records are still labelled as “Santa Marta Sabrewing” in the GBIF and SiB portals. Likewise, a sound recording that was originally attributed to a male Santa Marta Sabrewing (XC235605) from the San Lorenzo IBA in 2007, was an individual not seen well and hence not visually confirmed. We were able to confirm that those vocalisations corresponded to Black-throated Mango (https://xeno-canto.org/forum/topic/54177). There are also 16 listed occurrences dating from 2004 to 2017 in GBIF that are likely erroneous, being either linked to human observations without supporting documentation or to collected tissues that need to be examined, such as those described above.
We did not find Santa Marta Sabrewing during our expeditions to San Lorenzo, Rio Ancho or Aracataca in 2022–2023. This left the four historically reported localities along the Guatapurí basin as the only georeferenced records with unquestionable documentary evidence. Based on those data, we projected the species’ EOO to be ~28 km2.
Confirmed records of Santa Marta Sabrewing did not exceed elevations above 3,000 m, neither the exact coordinates nor a 2 km buffer around each point locality. Consequently, its documented presence was associated with three biomes: the Santa Marta and Macuira low orobiome, the Santa Marta orobiome, and the Santa Marta high orobiome. After intersecting areas of suitable habitat with these biomes within the species’ EOO, the AOO was projected to be ~23 km2 and restricted to the Guatapurí basin (Figure 2A). Following the Red List guidelines, the upscaled AOO value would correspond to ~24 km2. If unconfirmed localities mentioned by Simons were to be considered (i.e. San Miguel, La Guajira Department; San Sebastián de Rábago, Cesar Department), the projected AOO would possibly extend to ~164 km2 (Figure 2B).
Proposed area of occupancy (AOO) of Santa Marta Sabrewing delimited by the extension of suitable habitats (excluding the 2010 San Lorenzo record, see text). Distribution hypotheses consider (A) only confirmed records of the species, and (B) confirmed records and all observations on the southern and north-eastern slopes of the Sierra Nevada de Santa Marta described by F.A.A. Simons during his expedition (Salvin and Godman Reference Salvin and Godman1879).
Local occupancy rate and seasonal movements
We detected Santa Marta Sabrewing in 9 of the 20 point-count stations surveyed, resulting in a naive occupancy estimate (sD/s) of 0.45. The intercept-only model indicated that the species’ occupancy rate and detection probability were 0.46 (SE = 0.11) and 0.63 (0.09), respectively (see Supplementary material Table S2). The “saturated” occupancy model was not conclusive regarding a potential effect of elevation on Santa Marta Sabrewing occupancy (Supplementary material Table S2). The “elevation-only” model fitted the data better and suggested a very small increase in occupancy with elevation (Supplementary material Figure S1), although this effect had weak statistical support (Supplementary material Table S2).
The raw abundance of Santa Marta Sabrewing was similar between the lower (1,150–1,550 m asl) and higher (1,550–1,900 m asl) elevational ranges of the study area, with an average of 0.39 and 0.44 birds/count. While only 0.2 birds/count were recorded in stations that were mainly surrounded by transformed areas, 0.6 birds/count were detected in stations where land cover was dominated by native vegetation (i.e. secondary vegetation and early growth, as defined previously). The highest raw abundance corresponded to point-count stations surrounded mostly by native vegetation below 1,550 m, with an average of 1.4 birds/count.
We confirmed the presence of Santa Marta Sabrewing at the rediscovery site (~1,150–1,850 m) for 16 consecutive months between July 2022 and October 2023. Males and a few females were recorded during every visit, with males always detected in the same spots. We found no evidence of birds migrating.
Habitat associations, foraging, and behaviour
Of the 23 sabrewing point-count records, 78% were in secondary vegetation and 22% in mixed vegetation. No records were in point counts in early growth areas or open pastures. Neu’s habitat selection index suggested that secondary vegetation was preferred (1.2) to the remaining habitats (early growth: 0; mixed: 0.9; open: 0). However, neither habitat was used more frequently than expected given its areal extent (Fisher’s exact test P = 0.74). We obtained 25 behavioural records from ad libitum observations of the sabrewing. Away from the point counts, sabrewings were mainly observed in remnant riparian forest strips, all of which had been heavily disturbed, on steep slopes along La Macana stream and directly at the Guatapurí River above Guatapurí village. Individual hummingbirds were observed visiting agroforestry systems from nearby patches of secondary vegetation and early growth. We grouped all sabrewing observations into three basic behavioural categories: foraging (12% of records); flying (36%); vocalising in perches (52%). Most foraging observations were obtained in mixed habitat, particularly in agroforestry systems with shade-grown coffee, banana, and sugarcane. Females and males took nectar from flowering banana trees (Musa sp.) and from 7–14 m tall Calliandra and Erythrina trees. One foraging event involved a male taking nectar from the flowers of Malvaviscus concinnus 1.7 m above the ground in a riparian forest. Most observations of “flying” and “vocalising in perches” involved males, which vocalised non-stop during the morning (~07h30–12h00) while defending their territories.
Males would defend territories alone or aggregated in loose leks composed of up to four individuals located ~9–15 m apart (n = 2 leks). Two leks were found inside 30–60 m wide and 12–18 m tall patches of riparian forest at elevations of 1,551 and 1,746 m, respectively. Territorial behaviour was observed year-round. Males used perches 3–7 m above the ground to vocalise throughout the morning and defended their territory from conspecifics and other hummingbird species by aggressively chasing intruders in flight (Supplementary material Appendix 2). We also observed displays that consisted of males slowly approaching an intruder with a threatening posture, hovering forward with tail feathers and wings extended to appear larger (Supplementary material Appendix 2). In November 2022, a different display-like manoeuvre was observed: two males engaged in a vertical ascent display (15–20 m), after which they turned around and performed a display dive with their tail feathers open.
We obtained only one observation that could indirectly suggest breeding activity in the population of sabrewings at La Macana stream. This was a female collecting spider web with her bill in May 2023 that could be an individual gathering material for nest building.
Vocalisations
Male territorial vocalisations were composed of single (one-type) notes of short duration. The number of notes produced per minute (i.e. calling rate) was variable and may be associated with conspecific local density (e.g. isolated vs lekking males). Male A, which was observed alone, had the lowest calling rate (42 notes/min), whereas Male B that was observed at a “crowded” lek site had the highest calling rate (93 notes/min), and Male C had an intermediate calling rate (78 notes/min; see Figure 3). The mean note duration was relatively similar among males (Male A: 0.024 seconds; B: 0.026 seconds; C: 0.036 seconds), but the duration of silent intervals showed more variation (Male A: 1.177 seconds; B: 0.608 seconds; C: 0.696 seconds). The mean spectral frequencies were 4.257 kHz (SD = 0.459 kHz) for Male A, 3.570 kHz (0.385) for Male B, and 3.577 kHz (0.437) for Male C. Peak of frequency spectrum was 4.50, 3.78, and 3.87 kHz for males A, B, and C, respectively. Male A, which had the lowest calling rate, had the highest mean and peak frequencies (Figure 3).
Sonograms of Santa Marta Sabrewing’s calls for three different individuals (Males A, B, and C; see main text). The horizontal axis indicates time (in seconds), and the vertical axis indicates frequency (in kHz). Audio files for the three males are deposited at the Macaulay Library scientific archive (https://macaulaylibrary.org) with catalogue numbers ML615927194, ML615927198, and ML615927203.
Discussion
This is the first study to combine historical and reliable recent information on Santa Marta Sabrewing to describe aspects of its distribution and natural history, while contributing to an increased understanding of its conservation status. From a conservation perspective, our results highlight the need for more information to understand its distribution, habitat preferences, and population size. Given the erroneous or unsupported sabrewing records that we uncovered during our study, we argue that it is essential for new records of Santa Marta Sabrewing to be rigorously documented and verified to avoid false positives resulting from misidentifications. This particularly applies to records for previously unreported localities. Our study shows that a local population of Santa Marta Sabrewing can be found throughout the year in an area that coincides with the type locality and nearby areas where the species was first observed by Simons almost 150 years ago.
Historical distribution and seasonal movements
The assumption that Santa Marta Sabrewing had a wide distributional range, with individuals commonly observed up to the snowline, seems to originate from Salvin and Godman’s (Reference Salvin and Godman1879) transcriptions of Simons’ comments during his eight-month journey across the SNSM in 1878 (Simons Reference Simons1879, Reference Simons1881). In his notes, Simons mentions that Santa Marta Sabrewing was found as high as ~4,500 m and stated that sabrewings were commonly observed in June and July at ~3,300 m, in addition to being found in San Miguel at ~1,800 m and San Sebastián de Rábago at ~2,050 m (Salvin and Godman Reference Salvin and Godman1879). In contrast to these comments, however, most of Simons’ specimens were collected at elevations below 2,500 m (Salvin and Godman Reference Salvin and Godman1879; Todd and Carriker Reference Todd and Carriker1922), and neither the sabrewing specimens he collected nor the ones collected by Carriker in 1945 and 1946 came from localities above 2,800–3,000 m. Apart from these notes by Simons, no one else has observed the sabrewing at higher elevations. Thus, it is difficult to establish with certainty whether the species was indeed present at higher elevations, as has been suggested for decades. Simons was helped by indigenous people while crossing the snowline via the headwaters of the Guatapurí River (>4,000 m), and there is a chance that some of his observations of Santa Marta Sabrewing occurred during that journey. Revisiting this area would help resolve this uncertainty.
Similar to the assumptions about distribution, the supposition that Santa Marta Sabrewing is an elevational migrant is based solely on Simons’ notes (see Salvin and Goodman Reference Salvin and Godman1879). Our review found no empirical data that corroborate Simons’ idea, and in our fieldwork, we observed Santa Marta Sabrewings, likely the same individuals, over 16 consecutive months at La Macana stream, including two wet seasons. Considering current knowledge on the ecology of Campylopterus sabrewings (Ayerbe-Quiñones Reference Ayerbe-Quiñones2018; McMullan Reference McMullan2016; Winkler et al. Reference Winkler, Billerman, Lovette, Billerman, Keeney, Rodewald and Schulenberg2020) and our data, we propose three hypotheses regarding the distribution and possible migratory movements of the species. First, the sabrewing is a partial migrant with a portion of the population seasonally migrating to the páramo, while the remaining individuals occupy their territories at mid-elevation habitats year-round. Second, the species is a facultative migrant whose migrations depend on inter- and intra-annual fluctuations in precipitation and resource availability. Third, the species is a resident that does not migrate, and records outside the core area of distribution in the Guatapurí basin are vagrants. While the data remain sparse, our results favour the third hypothesis.
In addition to the continuous presence of the species at La Macana stream, the supposed photograph from San Lorenzo IBA is the only documented record of Santa Marta Sabrewing outside the Guatapurí basin. As stated above, we view this photograph as inconclusive. Moreover, regardless of whether the San Lorenzo photograph is a Santa Marta Sabrewing, it is very unlikely that it reflects the presence of a resident population of the species or a consistent pattern of migration. There have been extensive ornithological surveys and birding effort in San Lorenzo before and since the observation and our data, like Simons’ and Carriker’s, suggest that sabrewings have a relatively high detectability when present (see below). Furthermore, some reconstructions of the historical distribution of Santa Marta Sabrewing by other authors are remarkably similar to ours, circumscribing the species’ presence to the south-eastern slope of the SNSM (Ayerbe-Quiñones Reference Ayerbe-Quiñones2018; Hilty and Brown Reference Hilty and Brown1986; Salaman et al. Reference Salaman, Renjifo, López-Lanús, Renjifo, Franco-Maya, Amaya-Espinel, Kattan and López-Lanús2002). In summary, evidence points to Santa Marta Sabrewing as a case of microendemism, probably restricted to the Guatapurí basin and nearby river systems. Further research may reveal the species to occur in other parts of the SNSM, such as remnant forests of the northern flank (e.g. San Miguel and Río Ancho, La Guajira Department), or west of the Guatapurí River (e.g. San Sebastian de Rábago; Figure 2), but our results indicate that the burden of proof should be on confirming the species’ speculated migratory behaviour and presence outside the Guatapurí basin rather than vice versa.
Natural history and habitat associations
We found sabrewings to be highly localised in the rediscovery site, and conspicuous within this area. The species was detected with a probability of 0.6 in our surveys at the site and the birds vocalised repetitively during the morning, year-round. Simons also commented on sabrewings being common around Atánquez, and that their presence could be noticed with ease when vocalising (Salvin and Godman Reference Salvin and Godman1879). The fact that the species can be reliably detected when present points to very low numbers and a limited distribution, rather than difficulty of detection, being the reason that it has been reported so rarely.
Our observations and those of Simons suggest that this species is more abundant in mid-elevation forests, possibly in the vicinity of watercourses. We found no quantitative evidence of Santa Marta Sabrewing selecting forests over disturbed habitats, but the presence of territories and leks in riparian strips points to forests as an important habitat. Although the species uses cultivated areas, we hypothesise that its presence in transformed landscapes might be conditioned by the presence of native vegetation in association with watercourses. Thus far, we have found leks in dense understorey in riverine forest strips, and the presence of riverbanks near leks may indicate that female nesting sites could be nearby (see Esteves Lopes et al. Reference Esteves Lopes, Nogueira and Miranda2020). Nest sites in the vicinity of streams have been reported in other Campylopterus species (Hayes Reference Hayes, Hayes and Temple2002; Hayes et al. Reference Hayes, Trimm and ffrench2000; Marín Reference Marín2001).
Information on Santa Marta Sabrewing diet and breeding phenology remains extremely limited. From our review and field observations, the species has been seen consuming nectar from plants of five families, including Ericaceae (Befaria), Fabaceae (Calliandra, Erythrina, and Inga), Malvaceae (Malvaviscus), Musaceae (Musa), and Rubiaceae (Palicourea). This list can help inform conservation actions aimed at increasing suitable habitats in productive landscapes. Regarding the species’ breeding phenology, our only observation of what could be breeding behaviour occurred in May, coinciding with previous reports of courtship activity during April–June (reviewed in Cárdenas-Ortiz and Cortés-Herrera Reference Cárdenas-Ortíz, Cortés-Herrera, Renjifo, Amaya-Villarreal, Burbano-Girón and Velásquez-Tibatá2016).
Conservation, threats, and future work
The EOO projected in this study was considerably lower than the values reported in previous extinction risk assessments (2,900 km2 in BirdLife International 2024b; 7,702 km2 in Renjifo et al. Reference Renjifo, Amaya-Villarreal, Burbano-Girón and Velásquez-Tibatá2016; 4,000 km2 in Renjifo et al. Reference Renjifo, Franco-Maya, Amaya-Espinel, Kattan and López-Lanús2002). Similarly, our AOO projection was lower than the inferred remnant habitat (4,004 km2 in Renjifo et al. Reference Renjifo, Amaya-Villarreal, Burbano-Girón and Velásquez-Tibatá2016; 790 km2 in Renjifo et al. Reference Renjifo, Franco-Maya, Amaya-Espinel, Kattan and López-Lanús2002). It is worth noting that the latest extinction risk assessment carried out by BirdLife International lists Santa Marta Sabrewing as “Endangered” based on criterion B1 and “Critically Endangered” based on criterion C2 (inferred small and declining population size). If more liberal EOO and AOO projections were to be generated by considering all documented historical presences, including San Lorenzo, their extent would correspond to ~1,580 km2 and ~564 km2, respectively (Supplementary material Figure S2). This species is likely to remain listed as Critically Endangered until conclusive evidence suggests otherwise.
Just a few years before its rediscovery in 2022, Santa Marta Sabrewing was feared to have become extinct in the wild (Salaman et al. Reference Salaman, Cortés-Diago, Luna, Borrero-Álvarez and Toro2022). We have found what seems to be a resident but likely small population along La Macana stream in Chemesquemena, and this raises hope for the survival of the species. Most of the species’ hypothesised distribution lies within the Sierra Nevada National Park (Supplementary material Figure S3), and the national government recently agreed with four indigenous groups on a long-term management plan and the expansion of the park. A strategy to strengthen governance is required to accompany this new arrangement given that the protection of natural ecosystems within the park has not been effective (Duran-Izquierdo and Olivero-Verbel Reference Duran-Izquierdo and Olivero-Verbel2021). The focal population near Chemesquemena and at San José is currently unprotected, and despite being located within Kankuamo and Kogi indigenous reserves, land-use practices (e.g. fires and forest clearing) in the Guatapurí basin may pose a threat. Sugarcane cultivation by indigenous peoples has been documented since the nineteenth century (Salvin and Godman Reference Salvin and Godman1879) and is still widely practised across the region. This practice usually leads to deforestation of strips of native riparian vegetation, which buffer river floods and protect local family lands and other crops (Luke et al. Reference Luke, Slade, Gray and Annammala2019). Using agroforestry systems instead of sugarcane can help conserve the population of Santa Marta Sabrewing along La Macana stream and ensure environmental sustainability and resilience of local peoples’ agricultural productivity.
Santa Marta Sabrewing has been labelled as perhaps the rarest and most threatened bird species in Colombia (Salaman et al. Reference Salaman, Cortés-Diago, Luna, Borrero-Álvarez and Toro2022), and continued studies are urgently needed to design effective conservation action for the species. First, further field surveys are necessary to produce geographical data so that models can be used to inform a more accurate distribution hypothesis for Santa Marta Sabrewing. Once it is feasible to use distribution models, other relevant measurements (e.g. EOO and AOO) will be available to better assess the species’ risk of extinction. Second, conducting a longer and more widespread study of the population at La Macana stream is necessary to determine whether Santa Marta Sabrewing occupancy and abundance vary with elevation and other relevant environmental factors. This will also make it possible to assess whether the individuals we monitored remained at mid-elevations during our study given the constantly humid conditions due to the past La Niña event, which spanned approximately three years (2020–2023). Finally, further efforts are needed to conduct nest searches during April–June along La Macana stream and its tributaries to characterise the species’ habitat requirements during this sensitive period of the annual cycle. We hope that the characterisation of the Santa Marta Sabrewing vocal repertoire and behavioural displays will facilitate the correct identification of this species, and in doing so help to increase information on its biology, ecology, and distribution.
The sabrewing populations we studied along La Macana stream occur inside a private reserve of the Kankuamo indigenous community. Any further study and observation require the active participation of the indigenous landowners and collaboration in developing conservation measures. We extend this call to the general public and actors involved in the birdwatching tourism industry, who can help the local community by being patient and by accompanying the development of a sustainable ecotourism strategy that benefits the Kankuamo people and the region’s endemic and threatened avifauna.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0959270925000103.
Acknowledgements
We are thankful to all those who allowed us to visit the study area, particularly the Cabildo Indígena del Resguardo Kankuamo and the Chemesquemena community, as well as the Kogi people of San José. We also thank Luz Eneida Ochoa and Jaider Carrillo for their support and Elquin Toro for his valuable help in the field and for facilitating photographs used in this manuscript. Special thanks to Cristobal Navarro for his advice and Gustavo Bravo for checking specimens at the IAvH ornithological collection. We also thank Andrés Gómez, Leonardo Lemus, María Camila Machado, Gina Gómez Junco, and Daniel J. Lebbin. This work was funded by the American Bird Conservancy and the Search for Lost Birds (Grant No. 20154F), World Parrot Trust, SELVA, ProCAT Colombia, and Universidad Nacional de Colombia. No birds were captured or manipulated during this study. Fieldwork protocols were approved under Resolution No. 179-2015 from Parques Nacionales Naturales de Colombia and Resolution 0597-2014 and 0189-2016 from Autoridad Nacional de Licencias Ambientales issued to SELVA. Data analysed in this study and supplementary media files are available from the Zenodo open data repository at https://doi.org/10.5281/zenodo.10651684.
