Digging the diversity of Iberian bait worms Marphysa (Annelida, Eunicidae)

During a visit to polychaete–rearing facilities in the vicinity of Bay of Cádiz (SW Iberian Peninsula, Atlantic Ocean), we sampled two populations of Marphysa (Annelida, Eunicidae) originally occurring at nearby intertidal soft bottoms, one being more than twice as long as the other at the same age. We analysed them using partial sequences of two mitochondrial genes, 16S rDNA and Cytochrome Oxidase I, and classical morphological observations. Our molecular results confirmed that the two populations corresponded to two different species, with PTP species delimitation values ranging from 0.973 (long–bodied species) to 0.999 (short–bodied species). Morphologically, the short–bodied species resembles the recently redescribed M. sanguinea (Montagu, 1813), but differs mainly in having some parapodia with two subacicular hooks (one bidentate and one unidentate) and three types of pectinate chaetae, Two isodont present all along the body, and one particularly large anodont asymmetric appearing only from mid–posterior parapodia. The long–bodied species resembles Marphysa aegypti Elgetany, El-Ghobashy, Ghoneim and Struck, 2018 both in size and in having very robust, unidentate subacicular hooks (single in most parapodia, two–both similar in size and form–in some posterior parapodia), but differs, among other features, in the maxillary formula, the number of acicula per parapodia and the number and shape of pectinate chaetae. Accordingly, we are here fully illustrating and formally describing the two Iberian populations as Marphysa gaditana sp. nov. (short–bodied) and Marphysa chirigota sp. nov. (long–bodied) and we are emending the description of M. aegypti based on our revision of the type material. Also, we discuss on the distribution of the species of the sanguinea–group and on the relevancy of taxonomically robust studies when dealing with species of commercial interest having the potential of being globally spread through human activities, as well as on the misunderstandings caused by the incorrect use of the “cosmopolitan species” concept.


Introduction
In addition to the intrinsic interest of the annelid polychaetes as ubiquitous and highly abundant members of virtually all marine benthic ecosystems, some of them are increasingly exploited commercially. There is a growing demand of these organisms as fishing baits and, thus, they are being harvested all around the world as an integral part of global coastal life [1,2], including Europe [3][4][5]. They have also been introduced in integrated polycultures to contribute managing organic matter and wastes produced by bivalves and fish [6], and are a natural source of proteins and omega-3 fatty acids so that they can be used as a nutritional resource and maturation diet in crustacean and fish aquaculture [1,7,8]. Therefore, in addition to the impacts of traditional bait harvesting [9], new mechanized methods of collection are being developed to increase efficiency, productivity and revenue [10,11]. Thus, we may only expect further impacts and native habitat deterioration, as well as a greater pressure on their wild stocks. Importing allochthonous species is a well-established alternative to exploiting local populations. However, this transfers the harvesting impact to the often remote areas where baits are collected. Also, this may lead to accidental introductions (even to invasions) whether some viable specimens manage to escape from fishing hooks or are directly released into the wild by anglers at the end of their fishing journey [12,13]. A more environmentally friendly alternative is rearing autochthonous species, although the number of feasible initiatives is still very low (e.g., [4,14]). Nevertheless, these activities may also entail the destruction of local habitats either when implementing the facilities or during the routine functioning activities. On the other hand, culturing allochthonous species must be disregarded and discouraged due to the implicit risks of accidental releasing of living specimens that would directly impact the wild surroundings.
So far, no commercial polychaete aquaculture initiatives have been successfully implemented in the Iberian Peninsula. There were some attempts by researchers from the "Grupo de Ecología" of the "Universidad de Cantabria", in cooperation with the company TEICAN Mediambiental SL. Also, the Institute of Marine Sciences of Andalucía (ICMAN), in cooperation with the private companies "Comercial de Cebos para la Pesca S.L" and SEAPRTNERS, attempted to develop cultures of supposedly Marphysa sanguinea (Montagu, 1813) [15] based on their relatively abundant, autochthonous populations in the Bay of Cádiz (SW Atlantic coasts of the Iberian Peninsula) [16][17][18]. The present work resulted from a visit of DM and JG to the facilities built during an initial phase of the project developed at the Bay of Cádiz.
During our visit to the polychaete-rearing facilities at Bay of Cádiz, we realised that the native populations of Marphysa originally collected from nearby intertidal soft-bottoms and designated by local fishermen as "sand" and "mud" Marphysa (according to their original habitats), clearly represented two different morphotypes. In spite of having being reared under the same environmental conditions, the adults of the former were more than twice longer and much more active than the latter at the same age.
We hypothesized initially that the short-bodied population corresponded to M. sanguinea, which has been widely reported in the Iberian Peninsula [37], while the other could represent an undescribed species. To resolve this question, we analysed specimens from both populations by using partial sequences of the mitochondrial genes 16S rDNA (hereafter 16S) and Cytochrome Oxidase I (COI), as well as classical morphological observations. As a result, both south Iberian morphotypes are here described as species new to science. Both species are compared with the most similar ones, including Marphysa aegypti Elgetany, El-Ghobashy, Ghoneim and Struck, 2018 [31], whose description is emended based on our revision of the type material. We finally discuss on the distribution of the species of the sanguinea-group and on the relevancy of taxonomically robust studies when dealing with species of commercial interest having the potential of being spread globally.

Collection
Samples were originally collected by professional fishermen in intertidal shores of the Natural Park of the Bay of Cádiz (SW Iberian Peninsula) and transported to isolated polychaete-rearing facilities located at the San Ramón saltworks (Chiclana de la Frontera, Spain). Within the frame of an agreement between "Comercial de Cebos para la Pesca" and the Institute of Marine Sciences of Andalucía (ICMAN-CSIC), the specimens studied herein were collected by hand digging at the rearing facilities on the 10 th of May 2011. For morphological observations, the specimens were gently relaxed prior to being fixed in a buffered 10% seawater/formaldehyde solution and then transferred to 70% ethanol. For molecular purposes, fragments of the specimens obtained by natural autotomy of posterior ends were directly preserved in absolute ethanol and kept in the dark at -20˚C.

DNA extraction, amplification and sequencing
Total DNA was extracted from small body wall pieces using REDExtract-N-Amp kit (Sigma Aldrich, www.sigma.com) and DNAeasy Tissue Kit (Qiagen) for 16S and COI genes, respectively, following the manufacturer's protocol. REDExtract-N-Amp kit DNA extractions were diluted (1/6) in ultrapure Millipore water before using them for PCR amplification of fragments of the mitochondrial gene 16S. We amplified 826 bp of 16S and 660-700 bp of COI. We used primers designed in this study with the software Primer3 v 0.4.0 [38] for the 16S: Mar_16SF 5' GTGAGCTGATCTTTACTTGC 3' and Mar_16SR 5' GCTCTGGAGGA AGATTAGTC 3'. For COI, we used the primers polyLCO 5' GAYTATWTTCAACAAATCA TAAAGATATTGG 3' and polyHCO 5' TAMACTTCWGGGTGACCAAARAATCA 3' [39]. For 16S, PCR amplification reactions were performed in a 20 μL total reaction volume with 10 μL of REDEXtract-N-ampl PCR reaction mix (Sigma Aldrich), 0.8 μL of each primer (10 μM), 7.4 μL of ultrapure water, and 1 μL of DNA diluted template in the case of 16S. For COI PCR reactions were performed in a 25 μL total reaction volume with 2.5 μL of NH 4 No MgCl 2 Bioline Reaction Buffer (10X), 0.5 μL of MgCl 2 (50 mM), 1 μL of nucleotide mix (10 mM each dNTP), 0.8 μL of each primer (10 μM), 0.15 μL of BIOTAQ DNA polymerase (5 U/ μL, Bioline), 1 μL of template DNA and 17.25 μL of nuclease-free water. The PCR temperature profile for 16S was as follows: a first step at 95˚C for 5 min, followed by 35 cycles at 94˚C for 1 min + 42˚C for 1 min + 72˚C for 1 min, and a final step of 72˚C for 5 min. For COI: a first step at 94˚C for 10 min, followed by 5 cycles at 94˚C for 40 sec + 44˚C for 40 sec + 72˚C for 1 min, 35 cycles at 94˚C for 40 sec + 51˚C for 40 sec + 72˚C for 1 min, and a final step of 72˚C for 5 min. Agarose gel electrophoresis were used to visualise PCR products and to confirm fragment amplifications. Successful amplifications were purified and sequenced in both directions (forward and reverse) by Macrogen, Inc. (Seoul, Korea) with the same primers used in amplifications.
Additional sequences belonging to other Marphysa species, together with other genera of Eunicidae and one Onuphidae, were obtained from GenBank ( Table 1). Sequences of 16S rDNA were edited using Geneious vs. R8 and aligned along with GenBank sequences using the Q-INS-I option of MAFFT v.7 [40] and manually adjusted. COI sequences were edited using BioEdit v. 7.0.5.3 software [41], translated into aminoacids and aligned by hand together with GenBank additional sequences in Mesquite v.3.6 [42].

Species delimitation
To explore the potential clustering of our samples to other Marphysa species, we reconstructed phylogenetic trees for both markers separately, including sequences of Marphysa available in GenBank (NCBI), of published studies or thesis that authors had access to, and of six outgroup taxa (five of other genera of Eunicidae and one Onuphidae) ( Table 1). We used jMo-delTest 2 [71] as implemented in CIPRES Science Gateway V. 3.3 [72]. The most appropriate evolutionary models for our data determined by the Akaike Information Criterion (AIC) were GTR+I+G for 16S and HKY+I+G for COI. Bayesian Inference (BI) reconstructions were ran in MrBayes 3.2.6 [73] as implemented in CIPRES Science Gateway V. 3.3 [72], with two independent runs (each performed for four Markov-Chain Monte Carlo simulations) for 9 million generations for 16S and for COI analyses, sampled every 1,000 generations and initial 25% trees discarded as burning. We considered convergence of runs (average standard deviation � 0.01) and effective sample size of parameters (ESS � 200) calculated using Tracer v. 1.7.1 [74] to evaluate runs and accept results of the analyses. For both datasets, we calculated pairwise genetic distance using K2P model and partial gap deletion (cut-off 95%) in MEGAX [75].
The Poisson Tree Processes model (PTP, [76]) using BI rooted trees, 100,000 generations and removing all outgroups with exception of Paucibranchia was applied to infer putative species boundaries among our target samples and GenBank sequences using the webserver The Elexis Lab (https://sco.h-its.org/exelixis/web/software/PTP/index.html). We visually checked the convergence of MCMC runs in the maximum likelihood plot generated by the software.

Morphological study
To describe the diagnostic morphological features, we followed the terminology proposed by [52,67,77]. When necessary for descriptions or photography, relevant morphological structures (e.g., jaw apparatus, parapodia) were dissected and mounted on slides. Particularly, we dissected representative parapodia of the new species from anterior (5), median (40) and posterior (120-130) chaetigers to illustrate parapodial morphology and along-body variability.
Whole body pictures were taken with a PowerShot-SX710-HS digital camera. Light microscopy photos were taken with a CMEX 5 digital camera connected to a ZEISS Stemi CS-2000-C stereomicroscope and with a SP100 KAF1400 digital camera connected to a Zeiss Axioplan compound microscope. When necessary, dissected structures were stained with Methyl blue to highlight relevant characters. The same equipment was used to measure relevant morphological structures (with the help of the ISListen software, version 5.4(1) copyright by Tucsen Photonics Co. Ltd.), as well as to make the drawings (with the help of the Adobe Illustrator CC, version 2015.3.1, and Photoshop CC, version 2015.5.1, copyright by Adobe systems Inc.). For Scanning Electron Microscope (SEM) observations, specimens were prepared using standard SEM procedures [78]. SEM images were taken with a Hitachi TM3000 TABLETOP microscope at the SEM service of the CEAB-CSIC.
The type series of the Iberian populations are deposited at the Museo Nacional de Ciencias Naturales of Madrid (MNCN) and in the Natural History Museum Oslo (NHMO). The type material of M. aegypti was revised thanks to a kind loan of the NHMO.

Nomenclatural acts
The electronic edition of this article conforms to the requirements of the amended International Code of Zoological Nomenclature, and hence the new names contained herein are available under that Code from the electronic edition of this article. This published work and the nomenclatural acts it contains have been registered in ZooBank, the online registration system for the ICZN. The ZooBank LSIDs (Life Science Identifiers) can be resolved and the associated information viewed through any standard web browser by appending the LSID to the prefix http://zoobank.org/. The LSID for this publication is: urn:lsid:zoobank.org:pub:5053C03E-0822-4581-8F7A-3FE78C6BC4EA. The electronic edition of this work was published in a journal with an ISSN, and has been archived and is available from the following digital repositories: PubMed Central, LOCKSS, ResearchGate and DigitalCSIC.

Molecular analyses
In resulting trees, specimens from Cádiz formed two different well supported monophyletic groups (Figs 1 and 2 Results of PTP analyses supported that the two Iberian populations correspond to two different species, with 16S and COI species delimitation support values ranging, respectively,
Prostomial appendages arranged in semicircle (median and lateral antennae in about the same line, palps a little more anterior), extending beyond prostomium by ca. 2/5 their length (Figs 3A, 3B and 4A-4C). Median antenna, about as long as lateral antennae, all of them directed anteriorly, reaching from middle of chaetiger 2 to posterior border of chaetiger 3 when folded back. Palps about 1/5 shorter than antennae, directed anteriorly reaching from anterior border of chaetiger 2 to middle of chaetiger 3 when folded back. Ceratostyles and palpostyles tapering, lacking peduncle, style basally thicker, non-articulated (Figs 3B and 4A-4C). Ceratophores and palpophores all ring shaped, slightly wider than bases of ceratostyles and palpostyles, almost 12 times shorter than styles length ( Fig 3B).
Remarks. In addition to the marked molecular differences found in our analyses (Figs 1 and 2) and the distinct biogeographical origin [8] [30], whose subacicular hooks start after chaetiger 60 (instead of before).
The five following species, normally included or associated with the sanguinea-group and mainly described from European waters, have been discarded due to incomplete original descriptions and lack of redescriptions, which prevents a comparison: Leodice opalina Savigny in Lamarck, 1818 [85] [29,35,36], but differs, among other characters, in having some parapodia with two subacicular hooks, the second one being unidentate (instead of only one, bidentate in M. sanguinea), in having three types of pectinate chaetae in posterior parapodia (instead of only two, with Type 1 lacking, in M. sanguinea) and in the shape of the anodont pectinate chaetae from posterior chaetigers, which are very large and have 5-10 teeth with filiform tips (instead of normal size, 6-14 teeth, lacking filiform tips in M. sanguinea) (Fig 10A-10H).
Despite the absence of reliable evidences, it has been suggested that the presence of a secondary subacicular hook in some parapodia in the species of Marphysa could represent a replacement for the main one [81]. However, the fact that, when present, the secondary hook in M. gaditana sp. nov. is unidentate and lacks guards, while the main one is bidentate and has a pair of guards, casts some doubts on this replacement hypothesis. The presence of subacicular hooks seems to be a variable character within a given specimen, as they may also be absent from some parapodia (after first appearing). Therefore, we strongly recommend to consider this variability as a relevant character in species description.
Marphysa gaditana sp. nov. differs from M. chirigota sp. nov. and M. aegypti in having bidentate subacicular hooks with guards (unidentate in the other two species). All molecular species delimitation methods used herein grouped the COI sequences of M. gaditana sp. nov. with those in GenBank from Cap de la Hague (France), Sado Estuary (Portugal), and Florida and Virginia (USA), all them in Atlantic waters. It is feasible that the first two localities fall within the natural species distribution area (particularly the second one), but the records from the USA are certainly surprising. This wide disjoint distribution is uncommon for the family and deserves further investigation.
Etymology. The specific epithet refers to Gadir, the Fenician name of the oldest settlement of the city of Càdiz; "gaditana" means "from Gadir" and it is the Spanish epithet (feminine) for Cádiz inhabitants.
Distribution. Type materials collected at the Salina de San Ramón; however, according to ICZN Article 76.1.1 [88], the type locality must be the nearby intertidal muddy shores of the Natural Park of the Bay of Cádiz (approx. 36 Description. Holotype complete, very long, with ca. 370 chaetigers (26.5 cm long, 7.9 mm wide at mid-body, without parapodia), slightly widest all along median region, progressively tapering at regenerating posterior end (Figs 3C and 11A-11G); round in cross-section until around chaetiger 20-25, then dorsoventrally flattened (Fig 3C). Chaetigers more than 13 times wider than longer at widest body region (Figs 11D and 11E).
Prostomial appendages arranged in semicircle (median and lateral antennae in the same line, palps a little more anterior), extending beyond prostomium between half and 2/3 their length (Figs 3D and 11A). Median antenna, about as long as lateral antennae, all of them directed anteriorly, reaching from middle of chaetiger 1 to posterior border of chaetiger 3 when folded back. Palps about 1/5 shorter than antennae, directed anteriorly, reaching from middle of chaetiger 1 to middle of chaetiger 3 when folded back. Ceratostyles and palpostyles tapering, lacking peduncle, style basally thicker, non-articulated (Figs 3D and 11A). Ceratophores and palpophores all ring shaped, slightly wider than bases of ceratostyles and palpostyles, almost 13 times shorter than styles length (Fig 3D).
Notopodial cirri triangular, tapering (almost three times longer than wide at basis), decreasing in length towards posterior end (0.97 mm at chaetiger 5, 0.48 mm at chaetiger 40, 0.25 mm at chaetiger 120), longer than post-chaetal lobes in anterior chaetigers, as long as in median chaetigers and shorter in posterior ones (Fig 12A-12F), but longer than post-chaetal lobes in posterior-most chaetigers (Figs 11F and 11G). Lateral sense organs as single ciliated bump, located below the notopodial cirri ( Fig 7B). Ventral cirri thumb-shaped with round wide tips; from chaetiger 5 with inflated bases along most of the body, except about last 20, being round to triangular, with distinct round tips (Fig 12A-12F). Ventral cirri about half as long as notopodial cirri in anterior-most chaetigers (0.59 mm at chaetiger 5), similar in length in maximum branchial development region (0.54 at chaetiger 40) and decreasing in length at end of branchial region (0.43 mm at chaetiger 120).
Pygidium longer on ventral side, with two pairs of tapering pygidial cirri on ventral side; dorsal pygidial cirri ca.14 times longer than ventral ones (Fig 11H).
Marphysa chirigota sp. nov. most closely resembles the recently described M. aegypti in overall body size and in having very robust, unidentate subacicular hooks (single in most parapodia, two-both similar in size and form-in some posterior parapodia). However both species differ in numerous morphological characters: the ratio chaetiger width/length, the ratio prostomium/peristomium length, the presence of peduncle in cerato-and palpostyles, the posterior end of muscularised pharynx, the maxillary formula, the shape of notopodial cirri, the length and proportions of branchial filaments, the number and colour of neuropodial acicula, and the shape and number of pectinate chaetae (Table 2, Fig 14). Despite the numerous differences, distinguishing the two species requires a careful observation of key characters. These subtle differences are also reflected in the low genetic differentiation between both species (2.9-3.74%), which are borderline between intraspecific and interspecific for polychaete species [92,93], suggesting a recent speciation event.
Etymology. The specific epithet "chirigota" is a tribute to the "Chirigotas", a genre of choral folksong typical of Cádiz province (SW Iberian Peninsula), the type locality of the species. The chirigotas are satirical-humoristic songs performed predominantly in the streets by costumed performers during Carnival, and reflect much of the refined local sense of humour and hospitality that the first two authors had the chance to enjoy during the collection trip. By selecting this species name, all authors aim to contribute promoting the amazing heritage (natural, cultural and human) of the whole region of the Gulf of Cádiz.  [88], the type locality must be considered as the nearby intertidal sandy shores of the Natural Park of the Bay of Cádiz (approx. 36.349˚N, 6.181˚W), Chiclana de la Frontera, Cádiz (Iberian Peninsula), from), where the specimens or their progenitors were originally native. The species is also likely to be present in south Portugal, namely in Parchal (Algarve). This area is characterised by intertidal sand banks in marine sheltered waters (similar to the type locality) and harbours a population of Marphysa with very long specimens, known by Portuguese anglers as "ganso do Parchal" (Nuno Lopes, pers. comm. 28 April 2019). inflated basis along whole body) (Fig 15C and 15D); (2) three black aciculae per parapodium (instead of four, three black and one yellow) (Figs 15E, 15F and 16A); (3) one (occasionally absent, occasionally two in posterior chaetigers) unidentate subacicular hook (instead of subacicular hooks absent, reported as "yellow acicula" in the original description) (Figs 15E, 15F, 16B and 16C), present from chaetiger 32 (paratype NHMO C6966), 40 (holotype, paratype NHMO C6964), 48 (paratype NHMO C6965); (4) two pairs of pygidial cirri, ventral ones absent but scars indicating presence (instead of lacking ventral pair of pygidial cirri); (5) pectinate chaetae of four types; Type 1 in anterior body region, progressively replaced at midbody by Type 2; Types 3 and 4 in posterior and posterior-most chaetigers, respectively (instead of three types, with Type 1 in anterior and midbody segments and Types 2 and 3 in posterior segments); (6) Type 1 thin, symmetrical, isodont with external teeth markedly differing in length, with ca. 10-15 teeth; Type 2 thin, asymmetrical, isodont, with > 25 teeth; Type 3 relatively thick, asymmetrical, isodont, with about 15 teeth, coarse, with pointed tips; Type 4 thick, asymmetrical, anodont, with about five long and coarse teeth (instead of Type 1 isodont, with about 19 teeth, Type 2 with ca. 9 teeth and Type 3 with six teeth); and (7) Type 4 with teeth length vs. tip width ratio of 1 and teeth length vs. width ratio of 4 (not mentioned in original description) (Figs 14E-14H and 16D-16G).

Distribution range of Marphysa sanguinea
Robust taxonomic literature strongly supports not only that M. sanguinea fails to be a cosmopolitan species, but also that its distribution seems to be surprisingly restricted to areas surrounding its type locality [29,30,36]. Its current confirmed distribution ranges from Arcachon (Bay of Biscay, Atlantic coast of France) as southern limit, to the Southern Bight region (North Sea) as north-eastern limit [35,36], including both shores of the English Channel (type locality). The species has also been recorded in nearby regions, such as the western Irish coast [94], the Irish and Welsh shores of the Celtic Sea [95][96][97][98], and the Bristol Channel and Severn Estuary (e.g., [95,[97][98][99] and references herein), in what could represent its northwestern limit. Moreover, the Natural History Museum of UK (NHM) holds in its collections specimens identified as M. sanguinea from the Bristol Channel (NHMUK 1954.1.1.127, from Woody Bay; NHMUK 1970.7, from Porlock Weir). The species was also reported as "common" at Sully and Lavernock Point (Welsh coast of inner Bristol Channel) [99] and as "relatively rare" [98] or "occasional" [96] at Dale (Pembrokeshire, Welsh coasts of Celtic Sea), where it was not found in a 1988 survey [100]. Although we have not revised any of these records, many of them refer to intertidal specimens associated to hard substrates, so that at least some of them may feasibly represent the north-western limit of distribution for M. sanguinea.
The species has been recently reported as "introduced" in the southern North Sea, namely in southwestern Dutch shores [35,[101][102][103], where it is listed as an alien species. However, this statement seems to lack supporting evidence. Although there are not many published reports, M. sanguinea was previously known from the English coasts of the North Sea, at Whitstable, Kent [104] and Skipper's Island, Essex [105] and, more recently, from a shipwreck off Oostende, Belgium [106]. Furhermore, the NHM holds in its collections specimens identified as M. sanguinea from Norfolk (NHMUK 1903.5.16.1), Burnham-on-Crouch, Essex (NHMUK 1966.10.4), and Margate Beach, Kent (NHMUK 1966.4.21). Although we have not revised the identity of these specimens, all them support the existence of populations of Marphysa in the Southern Bight region for more than one century. However, the scarce records in the Southern Bight and their absence in the northern regions of the North Sea (with one exception, see below) seem to indicate that the genus is, if not accidental, at least rare or occurring there at low densities. This supports the Southern Bight as being the north-eastern limit of distribution for the genus (including M. sanguinea). Consequently, we suggest that the relatively rare presence of M. sanguinea in the Netherlands is more likely connected with the natural limit of its distribution, rather than to a human mediated introduction [35], with the recent findings in the area more probably resulting from achieved newly routine and more exhaustive monitoring programs.
The single report of Marphysa in the northern regions of the North Sea corresponds to a specimen described (but not named) from the Swedish west coast (near Uddevalla, Skagerrak) [107], a region under climatic and oceanographic conditions very different from those prevailing in the English Channel and Southern Bight. This specimen differs clearly from the sanguinea-group by the presence of tridentate composite falcigers in the anterior 15 chaetigers (against the single presence of composite spinigers in the sanguinea-group), a very uncommon feature, if not unique, in the genus. Regardless whether it may represent a undescribed species, it clearly differs from all the species targeted in our work.
The low records of M. sanguinea in north European waters above the Southern Bight or the Celtic Sea seem to indicate temperature as a key limiting factor for the northern distribution of this intertidal species. This is in accordance with the distribution of other intertidal invertebrates along the English Channel coasts, which diminish their densities eastwards in the direction of the Strait of Dover), were not only the average water temperature is lower, but also the severe winter cold results in a reduced fitness for southern species, which are replaced by northern taxa [108]. Regarding the Celtic Sea and Bristol Channel side, local episodes of high mortality of M. sanguinea were registered at certain areas of south Wales after the 1962-63 severe winter [98]).
As for the southern limit of distribution of M. sanguinea, it still needs further investigation along the Atlantic coasts of France, Spain and Portugal. The present descriptions of M. gaditana sp. nov. and M. chirigota sp. nov. from the Bay of Cádiz place that southern limit somewhere between Arcachon and Cádiz. Very likely, it may be north of the Sado Estuary (Portugal), as our analyses place the GenBank sequence of one specimen from that locality within M. gaditana sp. nov.

State of the art after type species resdescription
Since M. sanguinea redescription [29], two new species of the sanguinea-group were described from European and nearby locations, M. victori from the Bay of Biscay [3] and M. aegypti from the eastern Mediterranean, Suez Canal and Gulf of Suez [31], while we are here describing two more from the Bay of Cádiz. Similar situations occurred in other well studied coasts, such as Australia [29,52], South Africa [30], the Grand Caribbean [77], China [33,34] and Hong-Kong [51], where the presence of M. sanguinea proved to result from misidentifications or from a wrong use of the "cosmopolitan species" concept. Accordingly, several species (many new, some recovered from synonymies) have been reported, while others are still waiting to be reanalysed, likely to have their status removed from synonymy. Among [19,23,91,109]. As a result, the so-called sanguinea-group [50] or Group-B [110] currently comprises 32 species or subspecies of Marphysa, including one recently described from Philippines [8] and the two new ones described herin. Certainly, many more wait to be discovered in the near future.
Another important aspect allowing recognising and delineating the species of Marphysa is the habitat. Marphysa sanguinea seems to be virtually always present in association with hard substrates [25,111,112], while most species of the sanguinea-group occur in different, likely species-specific substrates [36], usually soft. In the case of the two new species here described, this habitat specificity also applies, as M. gaditana sp. nov. was found in muddy substrates, while M. chirigota sp. nov. was associated to sediments with higher contents of the sand fraction.
In addition to highlighting habitat specificity, we would also like to stress two other main aspects that emerge as useful issues in providing solutions to this species group. First, the obvious growing molecular standards. In our case, we were able to state, with different degrees of certitude, that both M. gaditana sp. nov and M. chirigota sp. nov. differed from all previously sequenced species of the genus, but also that the former apparently occurs in a number of different locations at both sides of the Atlantic, namely France, Portugal and USA. Second, the increasingly careful and detailed observations leading to highlight the presence of clearly discriminatory characters, many of them previously overlooked. In the case of Marphysa, in addition to traditional morphological traits (e.g., the shape of dorsal cirri and pre-chaetal, chaetal and post-chaetal lobes, the starting chaetiger of brachiae or subacicuar hooks, or the type of compound chaetae [77,81]), the shape of pectinate chaetae became a key argument, as first postulated for the type species [35] and later for some Chinese species of the sanguinea-group [34]. Classifying pectinate chaetae in a formulaic way may be tricky [8], but our results confirm the relevancy of carefully observing their morphology, number and presence, as well as the variations along the whole body. We also provide additional morphometric support based on the use of width/length ratios for the teeth of these chaetae in distinguishing among species, which turned to be particularly relevant when comparing M. aegypti and M. chirigota sp. nov.

Commercial interests and associated risks
Many polychaetes have a great commercial interest, and the species of Marphysa are not an exception. Most of them are being widely used as fishing bait by anglers all around the world, which is particularly favoured by two facts: 1) many of them occur intertidally or shallow subtidally, often in sheltered coasts, being thus easily collectable by hand or by digging the sediment, and 2) their relatively big size and robust muscular body is particularly adequate to be used as fishing bait. Their stiff bodies enable a tight fixation to fishing hooks, from which they cannot be easily detached. They most often remain intact if bitten by small fishes, being thus available for a bigger catch, which improves size selection and capture outcomes (Nuno Lopes, personal communication, 28 April 2019). Some anglers also sustain that the specimens of Marphysa may be bioluminescent, which would make them particularly attractive for night fishing (Nuno Lopes, personal communication, 28 April 2019). However, this statement still requires scientific confirmation.
Endurance once in the hook and catch selectivity have thus made the species of Marphysa sought and popular fishing baits everywhere in the world for such a long time. Scientific records of this particular use are known from Australia [29,35], China [33,34], Egypt [31], England and the English Channel [25,113], the French Atlantic [3] and Mediterranean coasts [21], Japan [23,24,26,27], Malaysia [32,114], South Africa [30], Sri Lanka [28], or Zanzibar [22]. However, our results emphasize the importance of knowing how many species are being currently traded under the name "M. sanguinea", not only in the Iberian Peninsula, but also in Europe and all around the world [1,13]. The fact that the south Iberian "M. sanguinea" turned to be two different, new species, as well as their differences in size, behaviour and habitat, indicate that they may have different life cycles, which probably also differ from those of the genuine M. sanguinea and any other species within the sanguinea-group. This may have obvious consequences for any commercial initiative (e.g., aquaculture, fishing baits), as well as for management programs of exploited natural populations, which may be extrapolated to all species within the group.
Our results also contribute to highlight the relevance and necessity of accurate taxonomic studies dealing with species of commercial interest. Exploited saleable species used to be distributed locally, while at present, the growing global marked establishes the potential of being distributed worldwide [1,2,13]. The usage of incorrect identifications favours careless practices, and enables impunity in trading, transporting and (possibly) releasing living allochthonous species into the wild. As they are being officially traded under the same scientific name than autochthonous species, no legal actions can be taken if a, let us say, "M. sanguinea" from a remote region of the globe is released, accidentally or not, in a Mediterranean area were "M. sanguinea" also occurs. This way, different exotic species may be legally introduced to areas where they are non-native. This may be the case of M. gaditana sp. nov. in some of the locations here reported (particularly in the most remote ones). It must be highlighted that, at least at Cap de la Hague (France), this species lives in sympatry with M. sanguinea (although they occupy different habitats, i.e., soft and hard bottoms, respectively). Sympatry has been also reported for other species of the sanguinea group in Philippines, Zanzibar, the Florida Keys, Australia and Cádiz Bay ( [8], present work). The fact is that exotic species may perfectly survive to establish viable populations outside their native habitats after being released to the wild by anglers. This not only represents a risk due to the own presence of the introduced polychaete, but also may favour introductions of potentially dangerous associated organisms. A recent example has been reported for the beachworm Perinereis linea (Treadwell, 1936) [115], a traded fishing bait native from the NW Pacific that has a well-established populations in the Mar Menor lagoon (Mediterranean coast of the Iberian Peninsula) [116]. The reproductive females of this exotic species carry gelatinous egg masses where the embryos are attacked by symbiotic ciliate protozoans, thus keeping the potential of acting as carriers of diseases for the native beachworms [116].
An interesting additional question rises on whether all known species of the sanguinea group are native or introduced in the areas from where they were first described. Solving this question, however, would require having samples from all around the world to undertake a complex molecular analyses, which is completely out of the scope of our paper. Despite this, knowing the real identity of commercially exploited species may certainly contribute to recognise the risks and, thus, to control them by promoting the implementation of good habits among traders, but also among the possible final users (e.g., sport anglers). As introduced exotic species may always have the potential of becoming invasive, the consequences of lacking these controls for local species and habitats would be unpredictable, but certainly point to overall changes in biodiversity that would further affect food webs, ecosystem functioning, and the provision of ecosystem services [117].

The collapsing "cosmopolitan species" concept
The use of polychaetes as model organisms in many different types of studies, from biogeochemistry, biology and physiology to ecology and genetics, as well as their commercial interest and increasing trade market, combines with old incomplete descriptions and inadequate diagnostic features to generate a considerable number of worldwide citations for certain species that obviously lack a rigorous taxonomic support. Marphysa sanguinea is a perfect example of this problem. This species has been reported in many studies from locations as diverse as Japan, China, Hong-Kong, South Korea, Australia, USA, Morocco, South Africa, India, New Caledonia, Gulf of Aden, Persian Gulf, Gulf of Thailand and many more (e.g., [36,[118][119][120][121] and references therein), with examples of misidentifications all around the world existing mainly, but not exclusively, in faunistic or ecological papers. Misidentifications certainly include European waters, where specimens of supposedly M. sanguinea have been used to trace heavy metals and for diet tests based on survival and growth of juveniles in Portugal [122,123], to monitor life cycles in the Venice Lagoon [124], or as target of potential interest for aquaculture in the Bay of Cádiz [16,17]. As for other world citations, our results strongly support that many of these European reports do not refer to M. sanguinea, but to different species within the genus.
As a final remark, we would like to highlight that proving the incorrectness of the "cosmopolitan" character traditionally attributed to many polychaete species should no longer be considered a surprise or even an added value increasing the interest of a given research or publication. Virtually all "cosmopolitan" polychaete species that have been confronted with careful, detailed studies (morphological, molecular or, ideally, both combined) have shown geographically restricted distributions, habitat specificity and/or the existence of hidden species complexes [125]. As a result, most species having broad worldwide distributions would very probably be those that have been secondarily spread (i.e., introduced) by human activities.
Supporting information S1 File. Species delimitation. 16S fragment: PTP results: based on the Maximum Likelihood and Bayesian inference reconstructions. 16S fragment: PTP results: based on the Maximum Likelihood reconstruction. Species described in this paper highlighted in red. (DOCX)