Range extension of the invasive fish Xiphophorus maculatus ( Günther , 1866 ) ( Cyprinodontiformes : Poeciliidae ) in the upper Magdalena river basin , Colombia

This study reports the expansion of the known distribution of the invasive fish Xiphophorus maculatus (Günther, 1866) into the upper Magdalena river basin. The new record comes from a wetland in the Dry Tropical Forest biome in Tolima and represents the first collection of this species from the upper Magdalena river basin. An updated distribution for Colombia is provided.

The family Poeciliidae includes some of the smallest vertebrates, and the species are highly polymorphic and have ovoviviparous reproduction (Rosen & Bailey 1963).They show marked sexual dimorphism, with males exhibiting showy color patterns and anatomical modifications such as the gonopodium, which is used to transfer sperm to the reproductive tract of the female (Rosen & Bailey 1963, Rodríguez 1997, Stoops et al. 2013).This modification is used in their taxonomic classification because of its great variation in shape and size (Stoops et al. 2013).
The dimorphism of males of some species of Xiphophorus includes the prolongation of the lower caudal-fin rays (Rosen 1960, Rosen & Bailey 1963), and it is used as a diagnostic character for some of them.This modification has permitted the artificial segregation of the species in this genus into those that have the prolongation of the lower caudal-fin rays: Xiphophorus helleri (Heckel, 1848) and Xiphophorus montezu mae (Jordan & Snyder, 1899), among others, and those that have not such as Xiphophorus maculatus and Xiphophorus variatus (Meek, 1904) (Kallman et al. 2004, Kallman & Kazianis 2006, Gómez-Gonzales et al. 2014).
Poeciliids have successfully invaded diverse aquatic ecosystems on almost all of the continents (Lowe et al. 2000, de Brito et al. 2013, Holitzki et al. 2013).These invasions have been facilitated by their use as ornamental species, and as biological control for mosquitoes (Axelrod et al. 2007, Chandra et al. 2008, McDowall et al. 2010, Ghosh et al. 2010).
This study reports the expansion of the known distribution of X. maculatus into the upper Magdalena river basin, specifically from the natural wetlands of Dry Tropical Forest biome, and constitutes the first record based on collected specimens from that biome.
Sampling was done on 4 June 2015 in the Azuceno Wetland (04°01′09.5″N, 074°57′23.7″W; 322 m above sea level [a.s.l.]), which is located in the municipality of Guamo, Tolima Department, Colombia (Fig. 1).This natural wetland is located in the Tropical Dry Forest biome of the upper Magdalena river basin.It is surrounded by a matrix of pastures with a town nearby.
The area of this wetland is permanent and covers 18.8 ha, lacks surface outflows, and fills during the rainy season.The substrate is mud, but turbidity is low and depth does not exceed 2 m.Floating macrophytes and submerged grasses are the principal aquatic vegetation present.Shrubs and trees are scarce, and the scarce riparian vegetation forms small isolated patches (Fig. 2).
The pH, conductivity and temperature water were measured using a Schott Instruments Handylab Multi 12/SET.For the other water parameters, 2 L samples were collected in plastic containers and refrigerated for analysis by the Laboratorio de Servicios de Extensión en Analisis Quimico (LASEREX) of the University Tolima.The water quality parameters are given in Table 1.
Fish were collected using portable electrofishing gear (400 V, 1 A), and a small drag seine (2 × 1 × 0.05 m).Elec-trofishing was employed for 15 min in areas where the macrophytes allowed for its use, and was complemented with 5 seine pulls of 3 m length each, passing underneath the floating plants towards shore.Specimens were fixed in 10% formalin and taken in sealed plastic bags, with the corresponding field labels, to the Zoological Research Laboratory (LABINZO) of Tolima University.The research permit was provided by Corporación Autónoma Regional del Tolima (CORTOLIMA) in "Convenio de cooperación interinstitucional 030 del 31 de Julio de 2013", as part of the project: "Formulación del plan de manejo ambiental de los humedales en el departamento del Tolima: Caracterización ambiental de los humedales en zonas bajas del Tolima".Taxonomic identifications were made using keys and diagnoses provided by Rosen 1960 andRosen &Bailey 1963.Morphological measurements and counts were done following Miller (1948).All measurements were taken point-to-point with digital calipers on the left side of specimens to the nearest 0.1 mm and are expressed as percentages of standard (SL) and head length (HL).To recognize structures of the gonopodia, 3 adults were cleared and stained (C&S) using procedures outlined by Dingerkus & Uhler (1977).
The new distribution record was corroborated using the SiB Colombia platform, which includes records from the data bases of the following ichthyology collections: Colección de Ictiología, Universidad de Antioquia, Medellín (CIUA);   Xiphophorus maculatus (Günther, 1866) (Fig. 3)   We collected 72 specimens, 48 females and 24 males.The identification as X. maculatus was confirmed upon review and corroboration of the diagnostic characters listed by Rosen (1960).These include structures of the gonopodium: distal tip of fifth ray having a small rounded element; posterior region of fourth ray with distal serration well developed that includes four hooks; third ray with long thin hooks.As well as some counts: 22-25 lateral scales and 7-11 dorsal-fin rays (Fig. 4, Table 2).The fleshy palp covering the tip of the gonopodium usually present in species of Poecilia was not observed (Rosen & Bailey 1963) (Fig. 4).
Head and body covered with large scales.Dorsal margin of head straight to slightly convex from anterior margin of eye to a vertical above opercle.From there to dorsal-fin origin convex, then gently declined to base of last dorsal-fin ray and then slightly convex to base of caudal fin.Mouth superior, protrusible and not reaching anterior margin of eyes.Eyes large (33.9-39.72%HL) (Table 2).Ventral profile from mouth to pelvic-fin insertions slightly curved, then from base of last anal-fin ray to caudal fin it is slightly concave.
In life, X. maculatus is silvery on flanks with black and iridescent blue patches on head, and fins are reddish from base to tips (Fig. 3).In alcohol, males may have melanophores on flanks, when if present are concentrated on caudal peduncle, sometimes covering the bases of caudal-fin rays.Dorsal fin has small dark spots, mostly located on interradial membranes, which are concentrated on distal thirds of branched rays.Caudal fin with melanophores on interradial membranes to point where rays begin to branch; caudal-fin margins lines with thin line of black spots, more evident in larger specimens, almost imperceptible in smaller ones.Pectoral fins with melanophores along edges of rays and membranes.Pelvic fins with melano-   (Steindachner, 1880), Synbranchus marmoratus (Bloch, 1795) and Andinoacara latifrons (Steindachner, 1878).This new distribution record enlarges the known distribution of X. maculatus in Colombia, increasing from 6 (Gutiérrez et al. 2012) to 7 the number of departments where it is present.Furthermore, it is the first record of this species from a natural wetland in the Dry Tropical Forest biome of the upper Magdalena river basin.It is the first record with associated museum specimens because the previous reports from this region are bibliographic (Alvarado & Gutiérrez 2002, Gutiérrez et al. 2010, Gutiérrez et al. 2012, Sánchez-Duarte 2016 pers. comm).
In aquatic environments, introduction of species is frequently associated with human activities (Ortega 2015), and this seems to be true for exotic poeciliids that are used as biological control agents for diseases vectored by mosquitoes as well as ornamental aquarium fish (Axelrod et al. 2007, Chandra et al. 2008, McDowall et al. 2010, Ghosh et al. 2010, Mousavi-Sabet & Eadgeri 2014).
The cause of the introduction of X. maculatus to the Azuceno Wetland is unknown, but since it is an ornamental species, one can infer that it may have been from a local aquarist (Gutiérrez et al. 2010(Gutiérrez et al. , 2012)).
Since invasive species adversely impact the natural community (Mack et al. 2000), such as the alteration of the trophic network and community structure and structural changes to the habitat (Mills et al. 2004, Gutiérrez 2006, Lasso et al. 2014, Magalhães & Jacobi, 2017), some poeciliids have become models for the study of these impacts (MacDonald et al. 2012, Holliztki et al. 2013).Species like P. reticulata, X. helleri, Gambusia holbrooki (Girard, 1859) and Gambusia affinis (Baird & Girard, 1853) modify the macroinvertebrate community through predation, which triggers an increase in the density of the periphyton community and so alters the trophic network (Pyke 2008, Maddern et al. 2011, Stockwell & Henkanaththegedara 2011, Holliztki et al. 2013, Walton et al. 2016).These invasive poeciliids also directly affect native fishes by eating their eggs and juveniles (Ivantsoff & Aam 1999, Rincón et al. 2005, Stockwell & Henkanaththegedara 2011, Schumann et al. 2015), by segregation that results from territoriality (Warburton & Madden 2003) and hybridization (de Brito et al. 2013).Also, it has been shown that aquatic ecosystems suffering human impacts with fish communities with poor species richness are more likely to be colonized by invasive species (Gutiérrez 2006).Those conditions have been successfully exploited by poeciliids since they are favored by their generalist feeding habits (Maddern et al. 2011, Stockwell & Henkanaththegedara 2011).
We infer that the Azuceno Wetland conforms to this scenario because its fish fauna has low richness (5 species), and it is exposed to loss of natural vegetation as a result of human activities in the area.Taking into account the proposal of Gutiérrez et al. (2010) to establish the risk level for invasive species in Colombia, X. maculatus is classified as a high risk species because it has been recognized as invasive at a global, regional and national scale.Likewise, there exists information of its life history that allows one to predict negative effects as well as the impossibility of its management once it has established in natural communities.For this reason, we consider it necessary to study the impact of X. maculatus not only on the local fish community, but also on the periphyton and macroinvertebrates present in the Azuceno Wetland, with the objective of adequately evaluating the level of the impact it causes on the structure and dynamics of the aquatic ecosystem.Such studies would provide information useful for understanding the ecological implications of this introduction to a wetland in the Dry Tropical Forest biome.

ACKNOWLEDGEMENTS
Financial support was provided by CORTOLIMA.We are grateful with all the members of the Grupo de Investigación en Zoología (GIZ) who helped us with the fieldwork.Thanks to J. E. García-Melo for the photos.Special thanks to D. Taphorn for translating to English version and for his suggestions on the manuscript.This manuscript was enriched and improved by the helpful comments from G. Echevarría and the anonymous reviewers.

Table 3 .
Records of X. maculatus in Colombia.Asterisk = new report, double asterisk = bibliographic report with approximate coordinates.