Specimen history

During offshore artisanal pole fishing in Rapa Nui (27º9’S, 109º26’W) (Fig. 1), an unidentified fish specimen was caught by local island fishermen. Their target catch was yellowfin tuna, but as an incidental catch they caught this specimen (^{Alverson et al. 1994}). The specimen was brought to shore dead, and discarded at the main port of Hanga Piko. None of the fishermen present at Hanga Piko were able to identify the specimen. The specimen was frozen and shipped to Santiago de Chile by plane for scientific determination. Upon arrival the specimen was frozen at -20 °C for storage, but during the trip the specimen lost the first dorsal spine.

Only one specimen was caught due to the limitations of pole fishing, and the artisanal boat infrastructure. The pole fishing lines for tuna were designed to reach an maximum depth of 100 m, limiting the maximum depth while fishing. Additionally, the artisanal boats in Rapa Nui are 7.1 m long and 1.89 m wide fiber glass boats that have a weight limitation of 50 tons (^{Acuña et al. 2018}), which limits the amount of weight they could safely carry on an expedition.

Morphology and meristic

To study the external and internal anatomy, the specimen was first thawed, and taken to a radiology center. Here, five x-ray photographs were taken to cover the whole length of the fish with an x-ray SUMMIT LX125V. All x-rays were assembled with Adobe® Illustrator CC v. 2017. The internal image of the specimen was used for a morphological and meristic analysis, following ^{Burhanuddin et al. (2002}) and ^{Nakamura & Parin (1993}). All measurements were made to the nearest 0.1 mm, and proportional measurements were rounded off to the first decimal place.

Sample and DNA extraction

A piece of muscle from the tail was cut and stored in 99% ethanol for molecular analysis, and the specimen was deposited in the ichthyology collection of the National Museum of Natural History of Chile (MNHNCL-ICT 7613). The DNA was extracted using the salting-out method (modified from ^{Jowett 1986}). A partial sequence of the mitochondrial gene encoding cytochrome oxidase subunit I (COI) was amplified using primers FishF2 and FishR2 described by ^{Ward et al. (2005}). The polymerase chain reaction (PCR) mixture had a total volume of 35 µl, and contained 1X PCR Buffer (200 mM Tris-HCL pH 8.4, 500 mM KCL), 2 mM MgCl₂, 0.2 mM of each dNTP, 0.1 µM of each forward and reverse primers, 0.3 U Taq DNA polymerase (Invitrogen™), and 70 ng of template DNA. The thermo-cycling conditions consisted of an initial denaturation step at 94 °C for 3 min, followed by 25 cycles of 94 °C for 30 s, 52 °C for 30 s, 72 °C for 1 min, and a final extension at 72 °C for 10 min. The PCR product was sequenced in an ABI PRISM 3100 Genetic Analyzer in the sequencing center of the Pontificia Universidad Católica de Chile. The sequence was aligned, and edited using BioEdit v7.0.5 (^{Hall 1999}) using the option ClustalW, and later a visual inspection was done.

Phylogenetic analysis

A phylogenetic analysis was done using maximum parsimony (MP), maximum likelihood (ML), and Bayesian inference (BI) methods in order to establish the phylogenetic position of the unidentified fish. The phylogenetic reconstruction was based on partial sequences of the COI gene of the Superfamily Trichiuroidea, with the following GenBank accession number: AP012507, AP012509 (^{Miya et al. 2013}), JN990842, JN990843, JN990845, JN990846, JN990862, JN990868 (^{Tzeng & Chiu 2012}), KU945006, and KU945019 (^{Chang et al. 2017}). The tree was rooted using species from the Family Scombridae as outgroup, GenBank accession number HM007708 (^{Cawthorn et al. 2011}), and a species of the Family Gempylidae as a sister group, GenBank accession number KU945005 (Chang et al. 2017). After obtaining all available sequences, they were aligned with the sequence of the unidentified fish (GenBank accession number MN052559) in BioEdit v7.05 (^{Hall 1999}), using the option ClustalW. The MP analysis was done in PAUP* version 4a (^{Swofford 2002}), using a heuristic search with the tree bisection-reconnection, and branch-swapping options. The ML analysis was done in RAxMLGUI v 1.5b1 (^{Silvestro & Michalak 2012}). For the statistical support of the nodes in the MP and ML analyses, a non-parametric bootstrap of 1000 pseudoreplicates was done. The BI analysis was done in Mr. Bayes v3.2.6 x64 (^{Ronquist et al. 2012}) with two independent runs of MCMC analyses with 10 million generation, each with 4 Markov chains. The MCMC analyses were sampled every 1000 generations, discarding the first 20% of trees. Convergence was considered checking the ESS (Effective Sample Size) and PSRF (Potential Scale Reduction Factor) values. For both ML and BI analysis, the GTR model (General Time Reversible) plus Invariant sites and Gamma parameter (GTR+I+G) was used, selected by JModelTest 2.0 (^{Darriba et al. 2012}) under both Akaike information criterion (AIC) and Bayesian information criterion (BIC).

Morphological and meristic analyses

Based on the external characteristics of the specimen found in Rapa Nui, it is suggested that it belongs to a cutlassfish of the genus Evoxymetopon. Until now, there are only four described species of the genus Evoxymetopon, and little information is known about their biology (^{Nakamura & Parin 1993}, ^{Chakraborty et al. 2006}, ^{Fricke et al. 2014}). Nevertheless, as other species of the Family Trichiuridae, the species of the genus Evoxymetopon are described as benthopelagic species, inhabiting continental shelfs and slopes, and are described as abundant in sea mounts (Nakamura & Parin 1997). Since there are only two described species of the genus Evoxymetopon that have the first dorsal spine elongated, all morphological and meristic comparisons were done with Evoxymetopon poeyi Günther, 1887, and Evoxymetopon macrophthalmus^{Chakraborty, Yoshino & Iwatsuki, 2006.}

The morphological and meristic values obtained for the specimen from Rapa Nui are shown in Table 1. Values obtained for the counts of dorsal fin elements, the origin of the first anal spine, the origin of pelvic fins, and the total number of vertebrates in the Rapa Nui specimen, correlate with the values of other E. poeyi from Japan and Taiwan. The only differences observed between the Rapa Nui specimen and the other E. poeyi, are in head length and eye diameter, with these measurements being slightly bigger in the Rapa Nui specimen Table 1. All of these measurements and counts were done on the compilation of the 5 x-rays. This x-ray compilation shows that specimen was captured with an empty stomach.

Table 1 Morphometrics and meristic data of E. poeyi from Rapa Nui, with comparative material from other E. poeyi and E. macrophthalmus from Chakraborty et al . (2006) and Koeda & Ho (2017) / Morfometría y mirística de E. poeyi de Rapa Nui, con material comparativo de otros E. poeyi y E. macrophthalmus de Chakraborty et al. (2006) and Koeda & Ho (2017) 

Phylogenetic analyses

The total alignment of the COI gen used for the phylogenetic analyses had 651 pb. The MP analysis retained a total of 3 trees, with a tree length of 458 steps, and the consistency index was of 0.6135, the homoplasy index 0.3865, the retention index 0.6856, and the rescaled consistency index 0.4206. The ML analysis obtained a final ML optimization Likelihood of -3,003.08. The BI analyses resulted in one consensus tree, with statistical values of ESS above 200, and PSRF values with an average of one, suggesting that the analysis reached convergence.

The topology of all three trees were congruent, therefore, only the BI consensus tree is shown, with the statistical support of each node (Fig. 2). This phylogenetic reconstruction, groups the specimen found in Rapa Nui with the species E. poeyi in one clade, presenting only two variable sites between them. Furthermore, the genus Evoxymetopon is paraphyletic, resulting in a taxonomic and systematics issue that has not been addressed by other authors. In this analysis, there was no possibility of including all four species of the genus Evoxymetopon, since no genetic data is available.

Figure 2 Consensus tree obtained using a partial sequence of the cytochrome oxidase I (COI) mitochondrial marker with a Bayesian inference analysis. The maximum likelihood of the tree is -3,003.08. Values above the nodes are, from left to right, the bootstrap value of maximum parsimony, maximum likelihood (above 70%) and posterior probability of the Bayesian inference (above 0.95) / Árbol de consenso del análisis bayesiano del gen citocromo oxidasa I (COI) marcador mitocondrial. El valor de máxima verosimilitud del árbol es -3.003,08. Valores arriba de los nodos son, desde izquierda a derecha, valor de bootstrap de máxima parsimonia, máxima verosimilitud (arriba de 70%) y el valor de probabilidad posterior obtenida del análisis bayesiana (arriba de 0,95) 

The morphological, meristic and molecular analyses done, suggest that the specimen found in Rapa Nui, belongs to an E. poeyi, commonly known as Poey’s scabbardfish. This fish species was first described for the Indian Ocean in the Mauritius and in the Western North Pacific Ocean, near Japan territory (^{Nakamura & Parin 1993}), and it has been occasionally reported in the Reunion Island (^{Letourneur et al. 2004}), near Taiwan (^{Koeda & Ho 2017}), and Cook Islands (^{McCormack 2007}). Most of these occasional reports, state that these fish have been caught as incidental catch, or have shown up dead while pole fishing in the open ocean. As an outcome, there are only 7 individuals catalogized in museums, with published information (^{Chakraborty et al. 2006}, ^{Koeda & Ho 2017}), and only 4 COI sequences in GenBank, including the one found in Rapa Nui, confirming, the rareness of this species.

This is the first record of the rare Poey’s scabbardfish for the remote island Rapa Nui. It is possible that this species has been off record until now, because most offshore fishing activities focus on yellowfin tuna, that are fished at depths of 100 m, and E. poeyi is described inhabiting below 200 m from the surface. Additionally, most marine biodiversity expeditions at EIE are done near shore. Only recently, have there been expeditions that explore the marine fauna offshore at EIE (^{Parin et al. 1997}, ^{Friedlander et al. 2013}, ^{Easton et al. 2017}, 2018). These expeditions are difficult, but crucial for the study of deep marine biodiversity, especially near Rapa Nui, because of its pristine seamount trails, and isolation. To the east of Rapa Nui, there is the Salas y Gómez ridge, to the southwest there is the Rano Kau ridge, that has the Apolo peak, and to the west, there is the Moai seamount and Pukao Seamount (^{Haase et al. 1997}) (Fig. 1). Of these seamount trails, pole fishing by natives is sporadically reported in the Pukao seamount, and the Apolo surface peak (^{Mecho et al. 2019}).

The presence of E. poeyi in Rapa Nui, is a sign of how unexplored seamounts are in general and in the Pacific Ocean. Additionally, the geographical distribution of E. poeyi is extended from the Indian Ocean to the whole Pacific Ocean, including the North and Southern hemisphere. So, it is suggested that E. poeyi inhabits seamounts, and that it is a possible mid-water migrator, since the specimen found had no stomach content. Therefore, the report of the rare E. poeyi in Rapa Nui, enlighten new information about the biology, ecology, habitat, and distribution of this species.