Introduction
Species of Clinostomum Leidy 1856 have an indirect life cycle involving a snail as the first intermediate host, fish or amphibians as the second intermediate hosts, and fish-eating birds, mammals, or occasionally reptiles as the definitive hosts (Kanev et al. Reference Kanev, Radev, Fried, Gibson, Jones and Bray2002). Until recently, species diversity of clinostomids remained largely unknown. This changed with the application of combined morphological and genetic tools, revealing significant species diversity within this family, particularly in the genus Clinostomum (e.g., Pérez-Ponce de León et al. Reference Pérez-Ponce de León, García-Varela, Pinacho-Pinacho, Sereno-Uribe and Poulin2016; Caffara et al. Reference Caffara, Locke, Echi, Halajian, Benini, Luus Powell, Tavakol and Fioravanti2017; Locke et al. Reference Locke, Caffara, Barčák, Sonko, Tedesco, Fioravanti and Li2019; Sereno-Uribe et al. Reference Sereno-Uribe, García-Varela, Pinacho-Pinacho and Pérez-Ponce de León2018; Briosio-Aguilar et al. Reference Briosio-Aguilar, Pinto, Rodríguez-Santiago, López-García, García-Varela and Pérez-Ponce de León2018; Montes et al. Reference Montes, Plaul, Croci, Waldbillig, Ferrari, Topa and Martorelli2020, Reference Montes, Barneche, Pagano, Ferrari, Martorelli and Pérez Ponce de León2021, Reference Montes, García, Paredes del Puerto, Barneche, Ibañez Shimabukuro, Reig Cardarella, Martorelli and Pérez Ponce de León2023). In Argentina, seven COI lineages of Clinostomum have been described or reported (Montes et al. Reference Montes, Plaul, Croci, Waldbillig, Ferrari, Topa and Martorelli2020, Reference Montes, Barneche, Pagano, Ferrari, Martorelli and Pérez Ponce de León2021, Reference Montes, García, Paredes del Puerto, Barneche, Ibañez Shimabukuro, Reig Cardarella, Martorelli and Pérez Ponce de León2023), and phylogenetic analysis supports the division of the genus into two clades: one from the Old World and another from the Americas (Locke et al. Reference Locke, Caffara, Marcogliese and Fioravanti2015).
In Argentina, a few metacercariae have been described or cited based solely on morphology, including from Hypostomus plecostomus (Linnaeus, 1758) from Córdoba province (Weyenbergh, Reference Weyenbergh1878) and from Neofundulus paraguayensis (Eigenmann & Kennedy, 1903) from Formosa province (Szidat, Reference Szidat1969). Additionally, metacercariae were reported in Trigonectes sp. (Myers) and Hoplosternum littorale (Hancock, 1828) from Salta province, but without molecular data (Davies et al. Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016). The records of metacercariae in Rivulidae by Szidat (Reference Szidat1969) and Davies et al. (Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016) are particularly important because of the high value of these fishes to aquarists and their significance as endangered species in the Neotropics (Alonso Reference Alonso2022; Alonso et al. Reference Alonso, Terán, Serra Alanís, Calviño, Montes, García and Casciotta2023, Reference Alonso, Terán, Calviño, Serra Alanís, Montes, García and Casciotta2024).
Seasonal killifish, members of the Neotropical Rivulidae and African Nothobranchiidae families within the order Cyprinodontiformes, inhabit temporary wetlands that periodically dry up. These fish lay desiccation-resistant eggs that undergo metabolic and developmental arrests known as diapause, regulated by environmental cues. The eggs hatch when ponds refill with rainwater, whereas adults do not survive the dry period (Podrabsky and Hand Reference Podrabsky and Hand2015; Furness Reference Furness2016). Seasonal killifish exhibit rapid growth and maturity within a short lifespan, thriving in harsh conditions (Costa Reference Costa, Malabarba, Reis, Vari, Lucena and Lucena1998; Berois et al. Reference Berois, Garcia and De Sá2015). Nearly 48% of killifish species in the Neotropical region are threatened due to their narrow geographic ranges and reliance on seasonal aquatic habitats highly impacted by human activities (Costa Reference Costa2016; Alonso et al. Reference Alonso, Terán, Calviño, García, Cardoso and García2018, Reference Alonso, Terán, Serra Alanís, Calviño, Montes, García and Casciotta2023, Reference Alonso, Terán, Calviño, Serra Alanís, Montes, García and Casciotta2024).
As part of our studies on killifishes in Argentina, particularly in the western Chacoan region, we discovered Clinostomum metacercariae infecting Trigonectes aplocheiloides Huber 1995 and Titanolebias monstrosus (Huber, 1995). The objectives of this study were to analyse the Clinostomum metacercariae found in Rivulidae fish, describe their morphology, and report their COI sequences to evaluate their phylogenetic position.
Materials and methods
Collection of samples and morphological study
Nine Tr. aplocheiloides (TRI) and two Ti. monstrosus (TMO) were sampled using hand nets near Hickman on the route to Embarcación city (–23.215479, –63.706228, Salta province). The collection site is the same as the one reported by Davies et al. (Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016) where they found clinostomids parasitising H. littorale. It is also located approximately 100 km in a straight line from Bañados del Quinquincho, where those authors discovered infected Trigonectes sp. (Fig. 1).
The fish were transported alive in aerated, water-filled bags to a field laboratory. The hosts were euthanised with an overdose of eugenol anesthetic (30 mg/L) and subjected to necropsy. Internal organs (testes, ovaries, liver, intestine, and mesentery) were excised and fixed in 10% buffered formalin for light microscopic studies. The metacercariae of Clinostomum were removed from their cyst using needles, rinsed in 0.85% saline solution, and preserved in 96% ethanol for molecular analysis. Some parasites were fixed in formalin after excysting for morphological analysis.
Molecular analysis
DNA was extracted from whole specimens of metacercariae infecting Tr. aplocheiloides and Ti. monstrosus using PURO-Genomic DNA (Productos Bio-logicos SA) following the manufacturer’s protocol. A fragment of the partial COI-mtDNA gene was amplified using polymerase chain reaction on an Eppendorf Mastercycler thermal cycler with the forward primer DICE 1F (5′ –ATT AAC CCT CAC TAA ATT WCN TTR GAT CAT AAG- 3′) and the reverse primer DICE 14R (5′ –TAA TAC GAC TCA CTA TAC CHA CMR TAA ACA TAT GAT G- 3′) (Van Steenkiste et al. Reference Van Steenkiste, Locke, Castelin, Marcogliese and Abbott2015). The reaction was performed with GoTAQ Master Mix (Promega) following the manufacturer’s protocol. The thermocycling conditions followed Montes et al. (Reference Montes, García, Paredes del Puerto, Barneche, Ibañez Shimabukuro, Reig Cardarella, Martorelli and Pérez Ponce de León2023). polymerase chain reaction products were sequenced by Macrogen Inc. (Korea). Sequences were manually edited using the Geneious 11 platform. Nucleotide alignment was checked for pseudogenes by translating sequences into amino acids based on the invertebrate mitochondrial genetic code. Newly sequenced barcode fragments were aligned with COI sequences from GenBank using the MAFFT v.7 program (Katoh and Standley Reference Katoh and Standley2013).
Sequences of Euclinostomum heterostomum (Rudolphi 1809), Ithyoclinostomum yamagutii Rosser et al. 2020, and Odhneriotrema incomodum (Leidy 1850) were used as outgroups as previously done by Montes et al. (Reference Montes, García, Paredes del Puerto, Barneche, Ibañez Shimabukuro, Reig Cardarella, Martorelli and Pérez Ponce de León2023). Optimal partitioning schemes and substitution models for each DNA partition were determined using the Bayesian Information Criterion with the “greedy” search strategy in Partition Finder v. 1.1.1 (Lanfear et al. Reference Lanfear, Calcott, Ho and Guindon2012). The dataset encompassing barcode fragments was partitioned based on first-, second-, and third-codon positions, each employing the appropriate nucleotide substitution model. The first codon position used the Tamura-Nei model with estimates of invariant sites and gamma-distributed among-site variation (TrN+I+G). The second codon position utilised the Kimura 1981 model with unequal base frequencies (K81uf), and the third codon position was characterised by the general time-reversible model with gamma-distributed among-site variation (GTR + G). For the Bayesian Inference analyses, the implemented model was GTR for all three positions because the less complex TrN+I+G and K81uf are not implemented in Mr. Bayes. The first codon with invariant sites and gamma-distributed among-site variation (GTR+I+G) while the second codon position used a model with equal-distributed among-site variation (GTR).
The phylogenetic trees were reconstructed using two parallel analyses of Metropolis-Coupled Markov Chain Monte Carlo for 20 × 106 generations each, to estimate the posterior probability distribution using Bayesian Inference through MrBayes v. 3.2.1 (Ronquist et al. Reference Ronquist, Teslenkovan, van der Mark, Ayres, Darling, Höhna, Larget, Liu, Suchard and Huelsenbeck2012). Topologies were sampled every 1,000 generations. The first 25% of the sampled trees were discarded as ‘burn in’. The consensus tree was visualised in FigTree 1.4.2 (Rambaut Reference Rambaut2014). The proportion (p) of absolute nucleotide sites (p-distance) (Nei and Kumar Reference Nei and Kumar2000) was obtained to compare the genetic distance among and between lineages, using MEGA 7, with 1,000 bootstrap replicates and a nucleotide substitution (transition + transversions) uniform rate. The obtained sequences were deposited in the GenBank database (http://www.ncbi.nlm.nih.gov) (Table 1).
Table 1. Information on clinostomids species/lineages used to construct the cytochrome c oxidase subunit I (COI) phylogenetic tree showed in Fig 1 (new sequences in bold)
Figure 1. Map of Argentina showing the sampling localities and previous reports of Clinostomum spp. Dots: Samples sites without genetic information. 1 Province of Cordoba (Weyenbergh Reference Weyenbergh1878). 2 Province of Formosa (Szidat Reference Szidat1969). 3 Uribelarrea city (Boero and Led 1971). 4 Pirané city, Formosa province (Lunaschi and Drago Reference Lunaschi and Drago2009). 5 Quinquincho Wetland, Salta province (Davies et al. Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016). 6 Hickman locality, Salta province (this study). 7, 8 Ibera Lagoon, Corrientes province (Montes et al. Reference Montes, Plaul, Croci, Waldbillig, Ferrari, Topa and Martorelli2020). 9 Concordia city, Entre Ríos Province (Montes et al. Reference Montes, Barneche, Pagano, Ferrari, Martorelli and Pérez Ponce de León2021). 10 Montecaseros, Corrientes province (Montes et al. Reference Montes, Barneche, Pagano, Ferrari, Martorelli and Pérez Ponce de León2021). 11 Santo Tome, Santa Fe Province (Montes et al. Reference Montes, Barneche, Pagano, Ferrari, Martorelli and Pérez Ponce de León2021). 12 Juan Blanco stream, Buenos Aires Province (Montes et al. Reference Montes, Barneche, Pagano, Ferrari, Martorelli and Pérez Ponce de León2021). 13 La Balandra, Buenos Aires Province (Sutton and Damborenea 2000). 14 Espinillo stream, Buenos Aires Province (Montes et al. Reference Montes, García, Paredes del Puerto, Barneche, Ibañez Shimabukuro, Reig Cardarella, Martorelli and Pérez Ponce de León2023). 15 El Palmar National Park, Corrientes Province (Montes et al. Reference Montes, García, Paredes del Puerto, Barneche, Ibañez Shimabukuro, Reig Cardarella, Martorelli and Pérez Ponce de León2023).
Morphological analysis
Specimens of Clinostomum TRI and TMO were stained in hydrochloric carmine, dehydrated through a series of ethanol concentrations, cleared, and mounted in Canada balsam (Pritchard and Kruse Reference Pritchard and Kruse1982). Specimens were photographed with an AmScope MU 1000 MP digital camera attached to an Olympus BX51 microscope and measured using ImageJ software (Schneider et al. Reference Schneider, Rasband and Eliceiri2012). Whole specimens were photographed with a Leica DMC 4500 digital camera attached to a Leica M205A stereomicroscope. Voucher specimens were deposited in the Invertebrate Collection of the Museo de La Plata, La Plata, Argentina under the accession numbers MLP-HE 8150 (Clinostomum TRI) and MLP-HE 8151 (Clinostomum TMO).
Results
A total of 72 digeneans were collected from Tr. aplocheiloides, with a prevalence of 78%, an abundance of eight, and an intensity of 14.4. Both Ti. monstrosus individuals were infected with more than 100 metacercariae. The parasites in both hosts were attached to different organs and musculature, and at various stages of development, with only those showing more mature genitalia being measured.
Clinostomidae Lühe, 1901
Clinostomum Leidy, 1856
Clinostomum TRI (Fig. 2a), TMO (Fig. 2b)
Description (based on five specimens found on Tr. aplocheiloides [TRI] and eight specimens on Ti. monstrosus [TMO], measurements Table 2). Body elongated, devoid of spines, flattened anterior end with oral collar. Oral sucker subterminal, rounded, smaller than ventral sucker. Developed prepharynx. Short pharynx. Intestinal caeca lateral to ventral sucker and genital primordium extending to posterior end. Diverticulated (TRI) or slightly diverticulated (TMO) intestinal wall. Ventral sucker 2–3 times larger than oral sucker, rounded, almost triangular opening. Genital complex posterior to body middle plane (TRI) or in posterior end (TMO). Testes slightly triangular, points rounded, apex of triangle wide but smaller than elongated base with an almost irregular margin. Concave base of testes facing the ootype. Posterior testis transversely elongated (in TMO). Kidney-shaped cirrus sac in right margin of anterior testis. Small, oval ovary, intertesticular and dextrally located. Tubular uterine sac not observed (TRI), tubular uterine sac, long, between genital complex and ventral sucker (TMO).
Table 2. Measurements of Clinostomum sp. infecting Rivulidae host from Embarcacion (Salta, Argentina) compared with those reported for Davies et al. (Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016)
In bold are measurements of the new specimens. Measurements are shown in μm with the mean followed by the range.
Abbreviations: AT, anterior testicle; B, body; Co, collar; Cs, Cirrus sac; Dbs, distance between suckers; Dbt, distance between testicles; Fb, forebody; Hb, hindbody; L, length; n/c, not calculated; O, ovary; Os, oral sucker; P, pharynx; PT, posterior testicle; Vs, ventral sucker; W, width.
* Calculated from the measurements provided by Davies et al. (Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016).
Figure 2. Clinostomid metacercaria from A) Trigonectes aplocheiloides; B) Titanolebias monstrosus. Abbreviations: At, anterior testis; C, caecum; Cs, cirrus sac; O, ovary; Oc, oral collar; Oo, ootype; Os, oral sucker; P, pharynx; Pp, Prepharynx; Pt, posterior testis; Vs, ventral sucker; Us, uterus. Scale bars = 500 μm.
Remarks
Despite minor morphological differences (such as pharynx size and cirrus sac length), both metacercariae found in Tr. aplocheiloides and T. monstrosus constitute a single biological entity and may be identical to those described by Davies et al. (Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016). The metacercariae described by these authors were obtained from the same host (Tr. aplocheiloides) and H. littorale, near the same sampling site as in the present study. The range of measurements and features is quite similar. The main difference observed in the measurements is the body length/body weight ratio, which is smaller in Trigonectes sp. compared to those reported here or in H. littorale by Davies et al. (Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016). Although the oral sucker width falls within the reported range, it is larger in both hosts found by Davies et al. (Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016). This measurement affects the ratios of oral sucker width/body width and ventral sucker width/oral sucker width, which are larger and smaller, respectively, in Davies et al. (Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016). The distance between the suckers is smaller in both hosts reported by Davies et al. (Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016). The anterior and posterior testicle width/length ratios are larger in Trigonectes sp. from Davies et al. (Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016), where the testicles are wider than they are long. The distance between the testes is also greater in Trigonectes sp. from Davies et al. (Reference Davies, Ostrowski de Nuñez, Ramallo and Nieva2016) compared to that found in Tr. aplocheiloides in this study. These observed differences may reflect variation in the maturation stage of the metacercariae.
Molecular analysis
Four partial COI mtDNA fragments were recovered (Table 1), two from Tr. aplocheiloides and two from Ti. monstrosus. The sequences measured 612 bp for Clinostomum TRI and 542-606 bp for Clinostomum MON. The final COI alignment was 624-bp long and consisted of 63 terminals, including the newly sequenced Clinostomum specimens and outgroups. Metacercariae from both rivulid hosts fell within the New World clade in the phylogenetic tree (Fig. 3), clustering with a p-distance of 0.52% (Supplementary material 1). Phylogenetic analysis showed this lineage as a sister species to the node formed by Clinostomum sp. 7, Clinostomum L1, and Clinostomum CRA. The smallest p-distances were with Clinostomum arquus García-Varela, Pinacho-Pinacho & Pérez-Ponce de León, 2018 and Clinostomum sp. 3 (10.91%–11.17%), Clinostomum L3 (12.50-12.76%), Clinostomum GBA (13.12%–13.38%), and Clinostomum caffarae Sereno-Uribe, García-Varela, Pinacho-Pinacho & Pérez-Ponce de León, 2018 (13.25%–13.77%).
Figure 3. Phylogenetic tree inferred using Bayesian Inference derived from cytochrome c oxidase subunit I (COI) gene dataset. Numbers in the nodes represent posterior probability (<0.9 are not shown). The Clinostomum species observed in this study are in bold. Reports from Argentina are indicated with a strong black bar, whereas the turquoise bars represent lineages or species recognised in the New World.
Discussion
The metacercariae found in Tr. aplocheiloides and Ti. monstrosus belong to the same entity and belong to the ClinostomumgGenus.
Previous findings (Montes et al. Reference Montes, Plaul, Croci, Waldbillig, Ferrari, Topa and Martorelli2020, Reference Montes, Barneche, Pagano, Ferrari, Martorelli and Pérez Ponce de León2021, 2023) revealed a wide range of second intermediate hosts of clinostomids in Argentina, including Characidae, Cichlidae, Crenuchiidae, Heptapteridae, and Lebiasinidae. This study is the first to sequence a Clinostomum lineage from the host belonging to the Rivulidae. The Clinostomum found here represents the eighth lineage identified in Argentina, with only one formally described as a new species so far (Clinostomum fergalliari Montes, Barneche, Pagano, Ferrari, Martorelli & Pérez-Ponce de León, 2021). Some metacercariae found are specific to their fish hosts such as Clinostomum ASC on Australoheros scitulus Říčan & Kullander 2003, Clinostomum GBA on Gymnogeophagus balzanii (Perugia, 1891), and Clinostomum PAU on Pyrrhulina australis (Eigenmann and Kennedy, 1903), all of them from Argentina. Others have a broader range of fish hosts, such as Clinostomum CRA in Characidium rachovii Regan 1913 and Psalidodon anisitsi 1907 (Eigenmann), Clinostomum heluans Braun, 1899, on Australoheros sp., and Cichlasoma dimerus (Heckel, 1840) (in Argentina), and Clinostomum L3 on Gobiomorus maculatus (Günther, 1859), Rhamdia guatamensis (Günther, 1864), Rhamdia laticauda (Kner, 1858) and Sicydium salvini Ogilvie-Grant, 1884. This last lineage, until the analysis of Montes et al. (Reference Montes, García, Paredes del Puerto, Barneche, Ibañez Shimabukuro, Reig Cardarella, Martorelli and Pérez Ponce de León2023), was considered separate from others found in Argentina, like Clinostomum CVI on Crenicichla vittata Heckel, 1840, C. dimerus (CDIM); and Clinostomum PLA on Pimellodella laticeps (Eigenmann, 1917). However, genetic distance among these and other metacercariae (Clinostomum sp1 and Clinostomum sp2) from the Americas indicates they belong to the same lineage but exhibit different morphologies. The conspecificity among Clinostomum L3, Clinostomum sp1, and Clinostomum sp2 was previously established by Perez Ponce de Leon et al. (Reference Pérez-Ponce de León, García-Varela, Pinacho-Pinacho, Sereno-Uribe and Poulin2016).
The new Clinostomum lineage appears to be less host-specific, parasitising different hosts from the same locality and environment. This lineage seems endemic to the Western Chacoan region, particularly in temporary ponds. The parasite appears adapted to the hydrology of these ponds, filling with water in summer from rains and drying in autumn and winter (Alonso et al. Reference Alonso, Calviño, Terán and García2016). In addition, this shallow and confined environment might allow cercariae to parasitise secondary hosts more efficiently.
We found Clinostomum metacercariae displaying different degrees of development, from short individuals with underdeveloped genitalia to larger ones with developed testes, particularly in Tr. aplocheiloides (Supplementary material 1). Variations in developmental stages could lead to misidentification of distinct lineages without genetic analysis. Accurate description of metacercariae requires ‘mature’ specimens.
The genetic study has reduced the number of morphological-only descriptions (or citations without genetic information) of Clinostomum in Argentine fishes, to those found by Weyenbergh (Reference Weyenbergh1878) in H. plecostomus from Córdoba, which is not a species present in Argentina (Mirande and Koerber Reference Mirande and Koerber2020, Bogan et al. Reference Bogan, Méttola, Terán and Cardoso2024), and that may actually be Hypostomus commersoni Valencienne 1836 or Hypostomus cordovae (Günther, 1880); and by Szidat (Reference Szidat1969) in N. paraguayensis; and juveniles found in bird throats by Lunaschi and Drago (Reference Lunaschi and Drago2009) and Lunaschi et al. (Reference Lunaschi, Cremonte and Drago2007). Without genetic data, it is impossible to determine if these represent previously reported lineages.
The importance of this new metacercariae relies on several aspects. Parasites play vital roles in ecosystems as part of biodiversity, controlling host populations, and participating in energy flow within food chains (Lafferty et al. Reference Lafferty, Dobson and Kuris2006; Timi and Poulin, Reference Timi and Poulin2020). This study of parasites, particularly this lineage, is crucial. Many aquarists value rivulids captured in nature, and the ‘yellow grub’ disease (caused by several Clinostomum species) could pose a significant problem if they also collect snails for their biotopes. We have observed that some fishermen use rivulids as bait, which could lead to the translocation of their parasites. Conservation issues related to Rivulidae could result in biodiversity loss in temporary ponds and affect hidden diversity within fish. Changes in rainfall, temperature (Allen and Ingram Reference Allen and Ingram2002; Karl and Trenberth Reference Karl and Trenberth2003) and habitat surroundings, such as large soy plantations and agriculture, could endanger these pools. These habitats are susceptible to fire, agrochemical contamination, and other ecosystem changes (Alonso et al. Reference Alonso, Terán, Calviño, García, Cardoso and García2018). Limited knowledge of parasite biodiversity and life cycles is crucial for protecting host fishes and ecosystems. Parasites can detect environmental changes, showing more sensitive responses than their hosts and indicating biogeographical shifts (Parmar et al. Reference Parmar, Rawtani and Agrawal2016). This underscores the importance of describing this and other parasite species or lineages. Finally, a very interesting line of investigation on the coevolution of those parasites and their seasonal killifish host inhabiting temporary wetlands seems to be a promising avenue for future research.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1017/S0022149X24000658.
Ethics approval
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional guides on the care and use of laboratory animals.
Consent to participate
All the authors give their consent to participate in this work
Consent for publication
All the authors give their consent to the publication of this work
Availability of data and material
All the material will be deposited in Museums and the sequences deposited on GenBank
Code availability
Not applicable
Funding
This work was partially supported by the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET PIP 1713) and Agencia Nacional de Promoción Científica y Técnica (PICT 2020 SERIE A-01531) to MMM. Grants: PIBAA 2872021010-0128CO from CONICET and “Fondos complementarios para la investigación con impacto en el territorio argentino: Peces de la cuenca del Río Bermejo: biodiversidad, distribución espacio temporal, efectos de las actividades antrópicas y conservación” by Fundación Williams to FA.
Competing interest
On behalf of all authors, the corresponding author states that there is no conflict of interest.
Acknowledgements
We are grateful to M. Marcia Montes for the line drawings, to the Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET) and CEPAVE for the provision of facilities and equipment, and to the Province of Salta for the permits for sample collection.
