Fishes of the Cusiana River (Meta River basin, Colombia), with an identification key to its species

Abstract The Cusiana River sub-basin has been identified as a priority conservation area in the Orinoco region in Colombia due to its high species diversity. This study presents an updated checklist and identification key for fishes of the Cusiana River sub-basin. The checklist was assembled through direct examination of specimens deposited in the main Colombian ichthyological collections. A total of 2020 lots from 167 different localities from the Cusiana River sub-basin were examined and ranged from 153 to 2970 m in elevation. The highest number of records were from the piedmont region (1091, 54.0 %), followed by the Llanos (878, 43.5 %) and Andean (51, 2.5 %). 241 species distributed in 9 orders, 40 families, and 158 genera were found. The fish species richness observed (241), represents 77.7 % of the 314 estimated species (95 % CI=276.1–394.8). The use of databases to develop lists of fish species is not entirely reliable; therefore taxonomic verification of specimens in collections is essential. The results will facilitate comparisons with other sub-basins of the Orinoquia, which are not categorized as areas of importance for conservation in Colombia.


Introduction
The Orinoco River, with an estimated richness of 1002 species of freshwater fishes, is the second most diverse drainage in the Neotropical region (Reis et al. 2016). Nonetheless, the basin has been exposed to increasing threats due to human activities that place the enormous fish diversity at risk (Barletta et al. 2010, Rybicki and Hanski 2013, Lasso et al. 2016. The systems draining the Andean region (western tributaries of the Orinoco) are considered the most threatened at basin scale (Rodríguez et al. 2007, Machado-Allison et al. 2010, Lasso et al. 2016. The rivers originating in the Andes are heavily exposed to threats like habitat fragmentation, contamination, deforestation, the introduction of non-native species and mining (Machado-Allison et al. 2010, Anderson and Maldonado-Ocampo 2011, Lasso et al. 2016. Additionally, large gaps regarding the basic knowledge of fish diversity of the Andean sub-basins are persistent, especially in Colombia , Machado-Allison et al. 2010, Lasso et al. 2016. Filling those gaps are essential to guide adequate conservation efforts for the freshwater ecosystems and therefore face the threats above mentioned.
The Meta River basin, with headwaters on the Eastern Cordillera of Colombia, is one of the major tributaries of the Orinoco River (Usma-Oviedo et al. 2016). Studies on its fish diversity (e.g., Urbano-Bonilla et al. 2009, and recent efforts (Zamudio et al. 2008 have advanced our understanding of the ecology of some species. The Cusiana River sub-basin is one of the best-known Andean tributaries of the Meta River basin; the first inventories of its fish diversity dated back to the 90's with the establishment of oil companies in the area. The Cusiana River sub-basin has been considered as a conservation priority area for the Orinoco biodiversity due to its high diversity in several groups , Trujillo et al. 2011, including fishes .
Here an updated checklist and an identification key are presented for the fishes of the Cusiana River sub-basin. We hope our results may establish a guideline that can be replicated in other basins of the Orinoco drainage.
The checklist was assembled by examining specimens deposited in Colombian ichthyological collections. Acronyms used in the text follow Sabaj-Pérez (2016) except uncatalogued material housed at Fundación Universitaria del Trópico Americano (UNITROPICO). The taxonomic list follows the classification system proposed by Reis et al. (2003) with recent modifications proposed by Oliveira et al. (2011) for characiform families, Betancur-R et al. (2016) at high-level groups for osteichthyans in general, and Thomaz et al. (2015) for genera of the Stevardiinae. Valid species names were confirmed through queries on the Catalog of Fishes of the California Academy of Sciences (Eschmeyer et al. 2017). Species were categorized as endemic (DoNascimiento et al. 2017), threatened (Mojica et al. 2012), migratory (Usma-Oviedo et al. 2013), and species subject of conservation (González et al. 2015).
Species richness interpolation and extrapolation was calculated following Chao et al. (2014) and using the package iNEXT 2.0.12 (Hsieh et al. 2016) for R v.3.4.0 (R Core Development Team 2017). The number of localities were obtained per Orinoco basin from the data set of the "Catalogue of the Freshwater Fishes of Colombia" (Do-Nascimiento et al. 2017).
To construct the key (for order, families and species), original descriptions of species, taxonomic revisions, and direct examination of specimens were used. Finally, in order to share the information produced herein, the dataset was uploaded to SiB Colombia's data portal (GBIF Colombia Node) in accordance with their protocols for species lists. For the latter, the complete dataset was structured and standardized to comply with the international biodiversity standard: Darwin Core standard (Wieczorek et al. 2012). After mounting the dataset on a Darwin Core spreadsheet template, it was uploaded to SiB Colombia's Integrated Publishing Tool for international visualization in their data portal. A DOI was provided by SiB Colombia for the shared dataset available at http://doi.org/10.15472/er3svl, all the results, discussion and quantities herein cited follow the version 1.8 of the published dataset.
Concerning threatened species, five are currently categorized as Vulnerable (Potamotrygon motoro, Brachyplatystoma vaillantii, Pseudoplatystoma metaense, P. orinocoense, and Zungaro zungaro), and two as Nearly Threatened (Potamotrygon orbignyi and Sorubim lima). There are 34 species endemic to the Orinoco drainage, 20 are migratory, and 8 are subjects of conservation. A total of 50 species are new records for the Cusiana River sub-basin, while Cetopsorhamdia shermani and Rhamdia muelleri are also new records for Colombia. Some species from the genera Andinoacara, Astroblepus, Ceratobranchia, Cetopsorhamdia, Chaetostoma, Characidium, Corydoras, Creagrutus, Curimatopsis, Hypostomus, Imparfinis, Microglanis, Parodon, Parotocinclus, Pimelodella, Poecilia, Spatuloricaria, and Trichomycterus, require further revision by specialists.    Base of posterior premaxillary tooth (6th) approximately equal to adjacent one (5th); ectopterygoid teeth unicuspid, generally 5 or fewer .  -Oviedo et al. 2016); it is a higher number of species than reported herein because re-identification and taxonomic updating processes of specimens excluded 59 of the 258 nominal species reported by Usma-Oviedo et al. (2016). For example, most of the undetermined (e.g., Ancistrus sp., Aphyocharax sp, Microglanis sp., and Odontostilbe sp.) and erroneous records (e.g., Hemibrycon cristiani, Pyrrhulina brevis, Schultzichthys gracilis, and Steindachnerina guentheri) originally counted as independent species were merged with other recorded species after our verification of the data. This is not surprising, taxonomy proceeds at a faster pace than the institutional ability to maintain updated records. For this reason, there is a need to account for validation of species identification when regional checklists are assembled from multiple secondary sources in order to avoid errors due to outdated or unverified data. Extrapolation suggests that the drainage could have a richness of roughly 314 species, indicating that the number of species found in the present study represents 77.7% of the expected richness in the area (Table 1). However, this estimate represents a rough estimate because sampling efforts have not been uniform across the drainage. A historical sampling-specific bias is not expected in the Cusiana River sub-basin, with the possible exception of an elevational bias (as is true for the whole Orinoco drainage in Colombia). Given the non-uniform nature of sampling in comparable river systems, we suggest that our extrapolations of species richness may be useful for comparisons among drainages using collection records as the input for rough estimation. This is particularly important since most checklist studies compare observed and not estimated richness, with the latter a more appropriate measure because it incorporates (and even overestimates) uncertainty from the samples into the estimation, and it also serves accounts for sampling effort among drainages.
Among similar Orinoco Andean tributaries, the Cusiana is one of the best-sampled sub-basins, exceeding in species richness other recently well-sampled sub-basins such as Orotoy (113 spp.; Ramírez-Gil et al. 2011) and Pauto (182 spp.;Maldonado-Ocampo et al. 2013). In fact, the Cusiana River sub-basin represents the quantile 0.83 among Orinocoan drainages, indicating that 83 % of the other drainages had fewer than 74 localities represented in the collections. The importance of the Cusiana River sub-basin is not only determined by its fish richness, but also because of its diverse and extensive aquatic ecosystem richness (rivers, streams, lagoons, estuaries, palm swamps, riparian forests, and flooded savannas) that provide important areas for fish reproduction, shelter and food. Because aquatic ecosystems have dynamic ecological and environmental processes (Teresa et al. 2015, Ribeiro et al. 2016, Toussaint et al. 2016, management and conservation projects of sub-basins should be addressed at regional (sub-basin) scales.
The documentation of the ichthyofauna in cis-Andean Colombian sub-basins has been increasing during the last decade, but new records and species can likely still be found in areas previously thought to be well-sampled (e.g., Ballen 2011, Vanegas-Ríos et al. 2015, Ballen et al. 2016a, 2016b, Burns et al. 2017, García-Alzate et al. 2017). Most of the sampling effort has been carried out in the piedmont and lowland areas in the Cusiana as well as in other sub-basins, and exploration of High Andean areas could lead to the discovery of local endemic species at the basin scale that usually are underestimated (Carvajal-Quintero et al. 2015).
Sub-basins adjacent to the Cusiana draining along the eastern slope of the Eastern Cordillera in the Orinoco region of Colombia (e.g., Guachiría, Casanare, Upía, Túa, and Cravo Sur) have not been well sampled and their richness is surely underestimated (Urbano-Bonilla et al. 2014). Continuous efforts are still to be carried out in order to document the fish fauna present along this region; this information is crucial to better understand how different anthropogenic activities (mining, oil extraction, agricultural, and livestock practices) are affecting the environmental conditions of these areas and as a consequence, the fish that live therein. Combination of this kind of information and further environmental data is a necessary step in order to generate freshwater conservation strategies using different approaches and therefore go further toward effective protection initiatives for species subject of conservation in the region.