First detection of an ocellate octopus in the Revillagigedos ecoregion, a biodiversity hotspot located in the Tropical East Pacific Province

Abstract The biodiversity of mollusks, particularly cephalopods, has not been exhaustively determined in the Revillagigedos ecoregion, which is a biodiversity hotspot for several marine groups located in the Tropical East Pacific Province. In our study, we detected and examined ocellate octopuses from Socorro and Clarion Islands, and determined their identity using morphological criteria and molecular data from two mitochondrial genes (COIII and COI). The taxon identified was Octopus oculifer, a species considered endemic to the Galapagos Archipelago. In addition, according to our analyses, O. mimus, O. hubbsorum and O. oculifer are very closely related and may represent a species complex comprised of three morphotypes. We found that the evolutionary relationships among octopuses are not determined by the presence of ocelli. This study is the first to report a clade represented by ocellate and non-ocellate species, in addition, the identity of cephalopods in the Revillagigedos was determined with analytical support.

Ocelli are considered an important diagnostic trait within octopodids, and are defined as false eye-spots in the form of round or ovoid conglomerations of chromatophores that may possess an outer concentric dark or light ring and an iridescent blue, purple, gold, or green inner ring (Packard and Hochberg 1977;Jereb et al. 2016). For octopodids, diagnostic features are highly valuable and needed, mainly due to the increased number of taxonomic confusions that derive from overlapped morphological characters among species (Norman and Hochberg 2005). Perhaps the most outstanding example of this problematic aspect is O. vulgaris Cuvier, 1797, which is considered a complex species that is being disentangled into different species (e.g., Octopus insularis Leite & Haimovici, 2008) and morphotypes by using morphological and molecular approaches (see Gleadall (2016), González-Gómez et al. (2018) and Amor et al. (2019)). Octopodids from the northeastern Pacific are no exception, for instance, Pliego-Cardenas et al. (2014) suggested that O. mimus Gould, 1852and O. hubbsorum Berry, 1983could be conspecific, and Díaz-Santana-Iturrios et al. (2019 confirmed that O. californicus Berry, 1911 andO. alecto Berry, 1953 should be reassigned into new genera. Determining the biodiversity of octopodids is relevant given that several species constitute fishery resources (4.8 million tons extracted during 2005-2014) (Sauer et al. 2019) or present aquaculture potential (Baltazar et al. 2000;Iglesias et al. 2000;Chapela et al. 2006), thus, it is important to implement species-specific conservation and management measurements, especially in poorly known areas such as islands. Many insular systems are biodiversity hotspots (Hazen et al. 2013), often difficult to access, which hinders the characterization of biodiversity. The Revillagigedo Archipelago is an insular system declared Biosphere Reserve since 1994, and later, in 2016, the World Heritage Committee included it in the World Heritage List of the United Nations Educational, Scientific and Cultural Organization; more recently, in 2017, it was declared National Park (CONANP 2004;SEMARNAT 2017;UNESCO 2019). The Archipelago is located in the Tropical East Pacific Province and the Revillagigedos ecoregion, according to Spalding et al. (2007). Revillagigedos is considered a biodiversity hotspot, at least for sea turtles, sharks, whales and giant manta ray (Dutton et al. 2014;Muntaner 2016;SEMARNAT 2017). Moreover, in Mexican waters, Clarion represents one of the five islands with the greatest marine species richness, and Socorro is one of the nine islands with the highest number of strict endemism (Koleff et al. 2009). Determining the identity of taxa in biodiversity hotspots can be informative to understand and monitor global biodiversity patterns, especially in groups that respond markedly to the current climate change, as is the case for cephalopods (Renema et al. 2008;Robinson et al. 2009;Rodhouse et al. 2014). We visited Revillagigedos during 2018 with the purpose of characterizing the coastal malacofauna of the Archipelago and found ocellated octopuses; thus, in this study, the objective was to identify these octopodids collected in Clarion and Socorro Islands through morphological comparisons and partial COIII and COI gene sequence analyses.

Methods
The Revillagigedo Archipelago is located approximately 390 km southwest of the southern tip of the Baja California peninsula and 890 km west from Manzanillo harbor, between 17°39'19" and 20°00'31"N, and 110°04'41" and 115°28'17"W. The Archipelago is comprised of four volcanic islands: Roca Partida, San Benedicto, Clarion and Socorro (CONANP 2004;SEMARNAT 2017).The octopuses evaluated in this study were fished for self-consumption by crew of the Mexican Navy in the military bases of Clarion and Socorro in accordance with the regulations stated by SEMARNAT (2017) for the Revillagigedo National Park and the permit to develop scientific research on the malacofauna from the Archipelago (official document number: F00.1.DRPBCPN.DIR.PNR.-001/2018) during June, 2018. In this convenience sampling, a total of 49 individuals were collected in Socorro (n = 8) and Clarion (n = 41) islands (Table 1) (Fig. 1). Due to the sampling particularities (convenience sampling), we were not able to transfer whole specimens from the Archipelago to the laboratory, instead, the morphological identification and morphometric measurements were conducted in situ and a small piece of tissue of each specimen was preserved in vials with 96% ethanol. However, in order to account for a type specimen of our sampling, a whole specimen was frozen and transferred to the laboratory facilities, preserved in 96% ethanol and deposited as ICML-EMU-12678 in the Regional Collection of Marine Invertebrates (ICML-EMU), Instituto de Ciencias de Mar y Limnología, Unidad Mazatlán, UNAM, in Mazatlán, México.
In Socorro Island, specimens were captured in each location by free and scuba diving using a hook; and in Clarion Island, octopuses were collected with a hook in the rocky intertidal during the lowest tide of each sampling site and day. Octopuses were sacrificed right after fishing. Individuals were sexed according to the presence (male) or absence (female) of a hectocotylized arm. Maturity stages for males and females were Octopuses were identified at the genus level using the morphological characteristics described in Jereb et al. (2016). For identification at the species level, the diagnoses of taxa of the genus Octopus that were reported for the Revillagigedo Islands by CONANP (2018)  naughey 1949;Berry 1953). The coloration patterns observed in the individuals were photographed with a Canon PowerShot D30 subaquatic camera. Images of morphological features of the octopus were recorded using a Canon EOS rebel T5 coupled to a stereoscopic microscope (Iroscope ES-24). The dorsal mantle length (DML) and total weight (TW) were recorded to determine the length-weight relationship (LWR). Class intervals were determined following the Sturge's rule. The following potential equation was employed to evaluate LWR: where: TW = dependent variable (total weight), a = coefficient of proportionality, DML = independent variable (dorsal mantle length), b = allometry coefficient (weight per unit of length).
The type of growth was estimated based on Student's t-test for the "b" values obtained from the model and compared to a theoretical value of b=3 which represents isometric growth.
For the molecular approach, sequences deposited in GenBank (Table 2) of the octopus species reported for the Revillagigedo Islands, ocellate octopuses of the world and species of various octopod genera were compared to our sequences. Separate analyses of partial COIII and COI sequences were performed to support the morphological All partial COIII and COI sequences were assembled and edited using BIOEDIT 7.2.6 software (Hall 1999). Edited sequences were deposited in the GenBank (Accession Numbers: MN259099-MN259105) ( Table 2). All partial sequences of each gene were aligned using MUSCLE (Edgar 2004) in MEGA 7 software (Kumar et al. 2016). The phylogenetic relationships among octopodids were reconstructed using Bayesian inference in MR. BAYES v3 (Huelsenbeck and Ronquist 2001) with the GTR+G+I model (Tavaré 1986) (selected by BIC in MEGA 7 software). The analysis was conducted with four default heated chains, running 1 million MCMC iterations and saving at every 1000 th generation. The first 1000 trees were discarded as burn-in. Inter-specific genetic distances were estimated by Kimura-2-parameter model (Kimura 1980) in MEGA 7 software.

Results
The individuals analyzed belonged to the genus Octopus Cuvier, 1797; these presented an ink sac and suckers in a two-row arrangement. The specimens presented ocelli (Fig. 2) and were identified as Octopus oculifer according to the morphological characteristics specified in its original description (Table 3).

Diagnosis of the specimens collected in the Revillagigedo Archipelago
The morphological features observed in the octopuses evaluated in this study are shown in Fig. 3A-G. Arm length ranged 3.5 to 4.4 times mantle length. Arm formula 3>4>2>1. Each arm with 230 to 280 suckers. Enlarged suckers on arms II and III. Gills with 10 lamellae per demibranch. Funnel organ W-shaped. Radula with 9 ele-ments, 7 rows of teeth plus marginal plates. Right third arm of males hectocotylized (with 180 suckers). Ligula tiny, 0.3% of hectocotylized arm. Calamus small, 0.25% of hectocotylized arm. Upper beak: rostral tip blunt and thick; rostral curvature well-defined and extended anteroventrally; the hood extends in posterodorsal direction; crest curved dorsoposteriorly; dorsal portion of the lateral wall sharply angled towards the tip of crest; wing and shoulder compressed posteriorly into an almost vertical position; jaw angle and edge concave ventrally. Lower beak: pointed rostral tip; jaw edge extends in moderate slope posteroventrally; wing fold slightly angled; wing extended in dorsoposterior direction; the dorsal edge raised in the central por-    tion; wall moderately curved posteroventrally towards the tip; free corner of the wall blunt; ventral edge of the wall curved. Color: five coloration patterns were observed in live individuals (Fig. 4), from pale and rugose body with few reddish-brown spots to completely dark-red smooth body without ovals. False-eye spots (ocelli) present as purplish black spot with a small pale central spot; ocelli are bound in an outer pale   ring. The individuals analyzed in this study presented isometric growth (b = 2.62; t = 2.38; p = 0.07) (Fig. 5).

Discussion
In this study we analyzed octopuses from the Revillagigedo Archipelago in an attempt to increase knowledge concerning cephalopods in this geographic area. We identified the octopuses to the species level primarily, according to their morphological attributes, and secondarily, using partial sequences of COIII and COI genes following Vecchione et al. (2017). The individuals were identified as Octopus oculifer (Hoyle, 1904) based on morphological and molecular examinations; however, the overlap of characters among the species reviewed in literature, especially between O. oculifer and O. hubbsorum, and the slight variation of arm formula in regard to original description (i.e., 3>4>2>1 instead of 3>2>4>1), explained why Jereb et al. (2016) pointed out that these species are a confusing complex that needs to be carefully re-evaluated (Table 3). Unfortunately, there is no holotype designated for O. oculifer, however, the original description is well-delineated and it is the only official conduit that endorses the identity of the species. In our study, there was no reason to suspect that the octopuses evaluated belonged to a new (undescribed) ocellated species or to any ocellated octopus other than O. oculifer.
For octopodids, particularly for the species evaluated in this study, a great deal of the taxonomic confusion is related to the fact that the morphological attributes are not standardized among species and that the diagnoses of octopodids from the northeastern Pacific had not been updated since the species descriptions, except for the validation of O. bimaculatus and O. bimaculoides within the genus Octopus performed by Norman and Hochberg (2005), and the recent evaluation performed by Díaz-Santana-Iturrios et al. (2019) for eight species of the genus Octopus. Thus, in this study, we provided a detailed characterization of the specimens collected in the Revillagigedo Archipelago and described attributes that were not included in the diagnosis of O. oculifer such as images of anatomic parts and coloration patterns, description of beaks, and lengthweight relationship. We suggest that such procedures should be addressed in further research concerning octopods from the Eastern Pacific in general.
The endemism and geographic distribution of O. oculifer restricted to the Galapagos Islands was well documented (Hoyle 1904;Edgar et al. 2004;Molina et al. 2004;Jereb et al. 2016); in fact, this area is the known distribution of the species to date. Contrastingly, in our research we found O. oculifer in the Revillagigedo Archipelago, Mexico, which is located approximately 3242 km northwest from the Galapagos Islands, Ecuador. Hence, with this information, in this study we report an increase of the distribution range of O. oculifer, which could be related to ocean-current patterns putatively used as dispersal mechanism during the paralarval stage, as was detected for larvae of the lobe coral Porites lobate Dana, 1846 in that same area (Galapagos and Revillagigedo Archipelago) (Reyes-Bonilla et al. 1999). Another explanation for this distribution expansion could be associated with changes in climatic conditions and similarities between original and new environments (Arkhipkin and Laptikhovski 2008;Stewart et al. 2014;Ramos et al. 2015), given that the Revillagigedos and the Galapagos Islands belong to the same realm (Tropical Eastern Pacific) according to Spalding et al. (2007). Thus, it is likely that the presence of O. oculifer in the Revillagigedo islands is related to the incipient climate change, which is beneficial for the abundance and distribution of cephalopods (Doubleday et al. 2016). It is worth noting, however, that the Revillagigedo Archipelago is a remote protected area, which limits human access, and therefore, the cephalopod diversity has been determined incidentally and with unreliable observations. As a result, octopods are only included as part of the mollusks and macroinvertebrate fauna of the Revillagigedo Archipelago in generalized taxonomic lists (i.e., González-Nakawaga and Sánchez-Nava 1986;Holguín-Quiñones et al. 1992;Ortega and Castellanos 1994;Ortega et al. 1995;Cabrera-Mancilla and Bautista-Moreno 2002;CONANP 2004;Friscione-Carrascosa 2005;Bedolla-Guzmán 2007;CONANP 2018) that do not specify the identification criteria or that employ taxonomic keys (i.e., Keen 1971; Abbott and Dance 1998) that were not developed by specialized cephalopod taxonomists. These lists include identifications at the genus level (Octopus sp.), tentative determinations (Octopus cf. bimaculatus) and/or erroneous identifications, such as Callistoctopus macropus (Risso, 1826) (formerly O. macropus), from the Mediterranean and northwestern Africa (Jereb et al. 2016). Thus, it is more probable that the presence O. oculifer in the Revillagigedo Archipelago occurred much earlier in geological time due to different ocean current patterns putatively used as dispersal mechanism during the paralarval stage, as explained earlier, than during the current climate change, and that its presence was not noticed until this detailed revision.
In addition, the molecular analyses of partial COIII and COI sequences strongly evidenced that O. oculifer, O. hubbsorum, and O. mimus are very closely related (interspecific distance lower than 1%) and it is highly likely that these taxa are conspecific and represent a species complex comprised by three morphotypes. However, our finding should be further confirmed with type material (when available) and complete re-descriptions must be performed to support that these taxa are the same species. Moreover, the closer evolutionary relationships found between ocellated and non-ocellated octopuses compared to the relationships among non-ocellated octopuses indicate that the presence of ocelli is not a determinant character in octopodid classification and therefore, it should not be considered a diagnostic attribute.

Conclusions
In this research, we conclude that according to our integrative species identification, the specimens collected in the Revillagigedo Archipelago are Octopus oculifer. According to our molecular analyses the non-ocellate O. hubbsorum and O. mimus and the ocellate O. oculifer are very closely related and might constitute a single species comprised of three morphotypes. In addition, ocelli should not be considered a diagnostic attribute for octopodids but rather a supplemental character.