Morphological and molecular systematic review of Marphysa Quatrefages, 1865 (Annelida: Eunicidae) species from South Africa

A vast polychaete fauna is hidden behind complexes of cryptic and pseudo-cryptic species, which has greatly hindered our understanding of species diversity in several regions worldwide. Among the eunicids, Marphysa sanguinea Montagu, 1813 is a typical example, recorded in three oceans and with various species considered its junior synonyms. In South Africa, specimens previously misidentified as M. sanguinea are now known as Marphysa elityeni Lewis & Karageorgopoulos, 2008. Of the six Marphysa Quatrefages, 1865a species recorded from the same region, three have their distributions restricted to South Africa while the others are considered to have worldwide distributions. Here, we evaluated the taxonomic status of the indigenous M. elityeni and investigated the presence of the widespread species Marphysa macintoshi Crossland, 1903 and Marphysa depressa Schmarda, 1861 in South Africa using morphological and molecular data. Our results reveal that M. elityeni is a junior synonym of Marphysa haemasoma, a species previously described from South Africa which is herein reinstated as a valid species. Both M. macintoshi and M. depressa are not present in South Africa and their status as being distributed worldwide deserves further investigation. Marphysa durbanensis Day, 1934 and the new species described here, M. sherlockae n. sp., had been misidentified as M. macintoshi and M. depressa respectively. Thus, the number of Marphysa species with distributions restricted to South Africa increased from three to five. This study reiterates the importance of implementing an integrated taxonomic framework to unravel local biodiversity.


INTRODUCTION
Studies implementing molecular and morphological tools in an integrated framework have found that a large portion of polychaete diversity has been hidden among complexes of cryptic and pseudo-cryptic species (Knowlton, 1993;Nygren, 2014;Hutchings & Kupriyanova, 2018). Thus, unravelling these species complexes can uncover patterns of distribution, regional biodiversity, and areas of endemism of previously overlooked polychaete species, which could have management and conservation implications (Bickford et al., 2007;Nygren, 2014).
Six valid species belonging to Marphysa are currently recognized as present in South Africa. Three have type localities in South Africa; Marphysa capensis (Schmarda, 1861), Marphysa posteriobranchia Day (1967), and Marphysa elityeni Lewis & Karageorgopoulos (2008). The latter is commonly known as the ''wonder worm'' by local fishermen, and is part of the global M. sanguinea species complex (Day, 1967;Lewis & Karageorgopoulos, 2008;Simon, Sato-Okoshi & Abe, 2019). The remaining three Marphysa species recorded in the region, namely M. corallina (Kinberg, 1865), M. depressa (Schmarda, 1861), and M. macintoshi Crossland, 1903 have type localities outside of South Africa and wide distributions (Day, 1967). Marphysa depressa's type locality is in Auckland, New Zealand (Schmarda, 1861), and has since been recorded in Hong Kong (Wang, Zhang & Qiu, 2018) and South African estuaries from Saldanha Bay to Durban Bay (Day, 1953;Day, 1967). Marphysa macintoshi was described from Zanzibar (Crossland, 1903) and has since been recorded from several localities including Australia, South Africa, Caribbean Sea, Mozambique, Red Sea, Trinidad and Tobago and China (Read & Fauchald, 2018). In South Africa, this species is supposedly present from Cape St. Francis to Durban Bay (Day, 1967). Interestingly, M. durbanensis (Day, 1934) described from KwaZulu-Natal in South Africa, is considered a junior synonym of M. macintoshi (Day, 1967). Similarly, M. haemasoma Quatrefages, 1866 was described from Table Bay in South Africa and is currently considered a junior synonym of M. sanguinea. Thus, both species probably represent valid indigenous species that were incorrectly synonymized.
In this study, we investigated whether M. depressa and M. macintoshi occur in South Africa and examined the taxonomic validity of M. haemasoma. These were achieved by conducting thorough taxonomic revisions and, where possible, molecular comparisons. We also provide redescriptions of M. haemasoma, M. durbanensis, and a description of M. sherlockae n. sp., a species new to science from South Africa.

Examined material
Fresh Marphysa depressa-like specimens were collected from rock crevices in the fringing intertidal zones from Strand (−34.116108, 18.821698) (n = 4) (Fig. 1). Fresh specimens of M. elityeni were collected from the fringing intertidal zone at low tide from burrows in gravely-sand type sediment under boulders in Kommetjie (n = 5) (−34.159709, 18.327851) (Fig. 1). Full collection data for both species can be found in the respective species accounts in the 'Results'. Live specimens were brought back to the laboratory where they were anesthetized with 7% MgCl 2 in distilled water, and photographed. Whole specimens from Strand were fixed in 96% ethanol. Posterior ends of the Kommetjie specimens were fixed in 96% ethanol, while the anterior ends were fixed in a 4% seawater-formalin solution.

Morphological examination
Species descriptions were produced based on the type material, but a variation section with all specimens reviewed was also included. The general structures such as the prostomium, peristomium, anterior region of the body, maxillary apparatus, branchiae, parapodia, chaetae, and pygidium were included in the descriptions. A dorsal incision was made in the specimen to extract and describe the maxillary apparatus, after which it was returned to its original position. The maxillary formula (MF) and measurements were taken according to Molina-Acevedo & Carrera-Parra (2015) and Molina-Acevedo & Carrera-Parra (2017). Six parapodia (three from the anterior region, two from the median, and one from the posterior region) were dissected to describe the morphology of the cirri and lobes, and simple and compound chaetae.
The chaetigers where branchiae and subacicular hooks start were indicated depending on the side where they began ('L' for Left, 'R' for Right) with the chaetiger number. In the region with the maximum number of branchial filaments, the long filaments are ≥4 times as long as dorsal cirri, whereas the short filaments are <4 times as long as dorsal cirri. The terminology used for the descriptions of the pectinate chaetae is according to the classification proposed by Molina-Acevedo & Carrera-Parra (2015), Molina-Acevedo &Carrera-Parra (2017) andZanol, Da Silva &Hutchings (2016). Herein, thin and thick refers to the thickness of the pectinate shaft; wide and narrow refers to the width of the pectinate blade; and anodont and isodont refer to the relative length of external teeth in relation to each other and internal teeth, e.g., thin, wide isodont with long and slender teeth.
The length through chaetiger 10 (L10) and the width of chaetiger 10 excluding parapodia (W10) were measured in the specimens as standard measures when the specimens were collected incomplete. Likewise, the total length (TL) and variations of the total number of chaetigers (TChae) were recorded. All descriptions were illustrated with a series of photos taken with Canon EOS T6i. These were then stacked using Helicon Focus R 6 (Method A) software to improve the depth of field, and the final editing was performed in Adobe Photoshop R 2020.
To understand patterns of intraspecific variation, linear regression analyses were conducted to evaluate the possible relationships between size (length of specimens using L10 measurement) and morphological features such as the chaetigers where branchiae or the subacicular hooks begin and the number of branchial filaments. The degree of predictability of variation in morphological features following size variation is given by R 2 (e.g., R 2 = 0.63, p = 0.05, n = 34). The electronic version of this article in Portable Document Format (PDF) will represent a published work according to the International Commission on Zoological Nomenclature (ICZN), and hence the new names contained in the electronic version are effectively published under that Code from the electronic edition alone. 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:C4C08B70-EC42-4AE1-9F9A-FDC717142D35. The online version of this work is archived and available from the following digital repositories: PeerJ, PubMed Central and CLOCKSS.

DNA extraction, amplification and sequencing
DNA was extracted from tissue samples using the ZR Genomic DNA Tissue MiniPrep Kit according to the standard manufacturer's protocol. The universal primer pair LCO1490 and HCO2198 (Folmer et al., 1994) was used to amplify a fragment of the mitochondrial gene cytochrome oxidase I (COI). PCR amplifications were carried out using 12.5 µl of OneTaq Quick-Load Master Mix (New England BioLabs), 9.5 µl of molecular biology grade water, 0.50 µl of forward and reverse primer (10 µM), 1 µl of 1% bovine serum albumin (BSA) and 1 µl of template DNA to make up a total reaction volume of 25 µl. Thermal cycling conditions were as follows for M. elityeni and M. sherlockae n. sp.: initial denaturation at 95 • C for 3 minutes, followed by 35 cycles of 94 • C for 20 seconds, 45 • C for 30 seconds and 72 • C for 1 minute, followed by a final extension time at 72 • C for 5 minutes. Amplicons were Sanger sequenced at the Central Analytical Facility at Stellenbosch University using just the forward primer (LCO1490). Quality control was performed on sequences to check for any sequencing errors using BioEdit (v7.2.6) (Hall, 1999).

Phylogenetic and species delimitation methods
The COI sequences were edited, trimmed, and aligned with ClustalW (Thompson, Higgins & Gibson, 1994) using multiple alignment methods in BioEdit (v7.2.6). Several species belonging to the Marphysa genus were included in the analysis for comparison together with seven other species from different genera within the Eunicidae and one species from Onuphidae as they were used as outgroups to root the tree (see Table 1). DnaSP v5 (Librado & Rozas, 2009) was used to generate a nexus file for subsequent analysis. PAUP (Swofford, 2003) and MrModelTest v2.3 (Nylander, 2004) were used to calculate the best fit model of evolution for the data set using the Aikaike Information Criterion (AIC). Bayesian inference (BI) was used to reconstruct phylogenetic relationships using the best fit model SYM+G in MrBayes 3.1.2 (Ronquist et al., 2012). The trees were calculated using 4 Markov Chains of 5 million generations sampled simultaneously with every 1000th tree sampled. A 50% majority-rule consensus tree with posterior probability support was constructed by discarding the first 25% of trees as burn-in. Tracer v1.5 (Rambaut & Drummond, 2009) was used to investigate the convergence of runs by analysing the average standard deviation of split frequencies (≤ 0.01). The mixing quality of all parameters was verified by analyzing the plot of likelihood versus the sampled trees and the effective sample sizes (ESS >200), of which both criteria were satisfied. FigTree v1.4.4 (Rambaut, 2012) was used to visualize trees. A Maximum Likelihood tree was computed in MEGA X (Kumar et al., 2018) and was run for 500 bootstrap replicates using the best-fit model of evolution, GTR, that was calculated in the same program.
A Newick formatted phylogenetic tree generated using FigTree v1.4.4 from the previous analysis was used as input for the Bayesian implementation of the Poisson tree process (bPTP) (Zhang et al., 2013) model for species delimitation using the online webserver https://species.h-its.org/. The tree was rooted and run for 500,000 MCMC generations, with thinning set to 100 and burn-in and seed set to 0.1 and 123, respectively. The convergence of MCMC chains was visually checked on the maximum likelihood plot generated by the online server.
MEGA X was used to calculate the interspecific genetic distances between species using the Kimura 2-parameter (K2P) model with complete deletion of gaps.

RESULTS
Thorough morphological comparisons indicate that M. macintoshi and M. depressa do not occur in South Africa. Instead, M. durbanensis (type locality: South Africa), which was previously synonymized with M. macintoshi (type locality: Tanzania/Zanzibar) (Day, 1967) has been found to differ from the latter species with regards to the shape of the prostomium, anterior postchaetal lobes, pectinate chaetae, and the shape and distribution of branchiae throughout the length of the body. As a result, we herewith consider M. durbanensis as a valid species.
Moreover, specimens initially identified as M. depressa (type locality: New Zealand) in South Africa were a misidentified and instead represent a new species to science, whereby named M. sherlockae n. sp.. Morphological comparisons reveal that M. sherlockae n. sp. differs from M. depressa in the shape and distribution of compound chaetae, the shape of postchaetal lobes, and the maximum number of branchial filaments. COI sequences of M. depressa were not available from its type locality and could not be compared with sequences of M. sherlockae n. sp. Nonetheless, M. sherlockae n. sp. forms an independent phylogenetic clade with high posterior probability and bootstrap support (Fig. 2) and genetically differs from other Marphysa species included in the phylogenetic analysis by 18-25%, confirming that it is a separate species. Additionally, results from the bPTP analysis supported M. sherlockae n. sp. as a single independent species (BS>0.95) (S1,  supplementary information). M. sherlockae n. sp. is phylogenetically closest to Marphysa californica Moore, 1909, and Marphysa brevitentaculata, but the clade is poorly supported. Nonetheless, all three species genetically differ from each other by 18-20%.
Marphysa haemasoma is a valid species. The examination of type materials allowed us to confirm that M. haemasoma differs from M. sanguinea in the shape of the postchaetal lobe in anterior chaetigers and subacicular hooks, the maximum number of branchial filaments and in the distribution of the swollen base of ventral cirri. Furthermore, types of M. elityeni only differ from those of M. haemasoma in size-related features, such as the length of prostomial appendices, and where branchiae and ventral cirri with a swollen base start. For these reasons, and in view of the principle of priority (ICZN, 1999, Arts. 23), we consider Marphysa haemasoma a senior synonym of M. elityeni. Furthermore, M. haemasoma forms a well-supported phylogenetic clade independent of the M. sanguinea clade (Fig. 2). The species are genetically different from each other by 20%, with results from the bPTP analysis (S1 supplementary information), confirming their separation as independent species (BS>0.95). Thus, these species are not synonymous.
Pygidium with dorsal pair of anal cirri as long as last eight chaetigers; ventral pair short, as long as last two chaetigers. Variations. Material examined L10= 12-14 mm, W10= 3.6-4 mm, TChae= 322-380. Palps reaching middle of first or second peristomial ring; lateral antennae reaching middle of second peristomial ring or first chaetiger; median antenna reaching first chaetiger. The maxillary variations are MII 5-6+6-8, MIII 6, MIV 3-4+6-8. The proportion of maxillary apparatus varies as follows: MI are 3.1-3.2 times longer than maxillary carriers; MI are 4.6-5.3 times longer than closing system; MII are 3.5-3.6 times longer than length of cavity opening. Branchiae from chaetigers 28-32 to 10-13 chaetigers before pygidium. Maximum number of branchial filaments varied from 11 to 12. Postchaetal lobe well developed in the first 40 chaetigers. Ventral cirri with a swollen base from chaetigers 4-5 to 25 chaetigers before pygidium. Start of subacicular hooks in chaetigers 46-47. Habitat. Day (1934) does not provide information about the specific substrate, although he did clarify that the collection was between the tidemarks in Durban Bay and Umkomaas. Distribution. Day (1934) recorded this species from Durban Bay and Umkomaas in KwaZulu-Natal, South Africa. Remarks. The original description of Marphysa durbanensis provides a variation of the two specimens collected that matches with the specimens deposited in the BMNH. Day (1934) described almost colorless eyes, but they were not observed in this study. Possibly the color has faded due to the long-term preservation of the specimens. The best-preserved specimen is herein selected as a lectotype to fix the species definition (ICZN, 1999, Arts. 74.1, 74.7.3), whereas the other is considered a paralectotype (ICZN, 1999, Art. 74F). Day (1934) considered M. durbanensis different from morphologically similar species such as M. simplex Crossland, 1903 (= M. teretiuscula), and M. acicularum when he described the species. However, in his monograph of the polychaetes from South Africa, the author considered M. durbanensis a junior synonym of M. macintoshi without making any reference to this nomenclatural action (Day, 1967, page 378). Herein, apparent differences were found between the species. Marphysa durbanensis (L10: 14 mm) has a bilobed prostomium, the branchiae are pectinate and start from chaetigers 28-32, the postchaetal lobe is digitiform in first four chaetigers, and there are five types of pectinate chaetae; while in M. macintoshi (L10: 4.5 mm) the prostomium is unilobed with a shallow median sulcus at the anterior edge, the branchiae are palmate with a short button-shaped branchial stem and start from chaetiger 32-47, the postchaetal lobe is conical in the first four chaetigers, and there are only three types of pectinate chaetae. Due to these morphological differences, M. durbanensis is considered a valid species.
Marphysa durbanensis resembles M. haemasoma (see below) by the presence of compound spinigers distributed in all chaetigers; however, M. durbanensis has more teeth in MII (5-6+6-8), digitiform postchaetal lobes in first four chaetigers, five types of pectinate chaetae, and the subacicular hook with a continuous distribution even in bigger specimens. However, M. haemasoma has fewer teeth in MII (4+4). The postchaetal lobe is ovoid in the first four chaetigers. There are only four types of pectinate chaetae, and the subacicular hook has a discontinuous distribution in small specimens.
Marphysa durbanensis resembles M. victori Lavesque et al., 2017, M. hongkongensa Wang, Zhang & Qiu, 2018, M. leidii Quatrefages, 1866, M. parishii Baird, 1869 by the presence of five types of pectinate chaetae; however, M. durbanensis has a digitiform postchaetal lobe in the first four chaetigers, and the subacicular hook is amber, while M. teretiuscula has an ovoid postchaetal lobe in the first four chaetigers, and the subacicular hook is reddish basally and translucent in the distal region. Also, M. leidii has a conical postchaetal lobe in the first chaetigers. Otherwise, M. durbanensis has long branchial filaments, and the branchiae are pectinate; while for M. hongkongensa, the branchial filaments are short, and the branchiae are pectinate and palmate with a short buttonshaped branchial stem in some regions of the body. On the other hand, in M. durbanensis (L10: 14 mm), the eyes are present, and the branchiae start in chaetigers 28-32; while M. victori (L10: 6.3-7.9 mm) lacks eyes, and the branchiae start in chaetiger 36. Finally, M. durbanensis has up to 11-12 branchial filaments while M. leidii (L10: 10.7-17 mm) and M. parishii (L10: 17.2 mm) only have 4 to 6 filaments.
Peristomium (2.8 mm long, 6.3 mm wide) wider than prostomium; first ring three times as long as second ring, separation between rings distinct only dorsally and ventrally (Figs.  6A-6C). Ventral region of the first ring with a slight central depression in anterior edge (Fig. 6B).
Branchiae pectinate with up to six long filaments for around 20-54% of the body, present from chaetigers 26L-27R to 308L-311R (Figs. 6I-6J). First two and last 13 pairs with one filament; with six filaments in chaetigers 79L to 173L (Fig. 4B). Branchial filaments longer than dorsal cirri except in first two and last branchiae.
Pygidium with dorsal pair of anal cirri broken; ventral pair as long as last chaetiger.
Habitat. Very common in the boulder fields at the lower intertidal zones of sheltered bays, and in rock pools. Worms can be found under rocks in sand burrows up to 1 m deep. Distribution. Table Bay to Buffels Bay, Cape Point, Western Cape South Africa (Quatrefages, 1866;Lewis & Karageorgopoulos, 2008). Branch et al. (2016) recorded this species to occur from Namibia in southwest Africa to East London in South Africa. Simon et al. (unpublished data) recorded this species from Melkbosstrand to Knysna in the Western Cape and therefore falls within the currently accepted distribution range of this species according to Branch et al. (2016). However, the records from Namibia have not been verified and may also represent an overlooked indigenous species of that region and therefore should be revised. Marphysa haemasoma (L10: 9.3-18.5 mm) is considered a different species from M. sanguinea (L10:11.5-20.4) because the former has up to 10 branchial filaments, and ovoid postchaetal lobes in anterior chaetigers; whereas the latter has 9-18 branchial filaments, and digitiform postchaetal lobes in anterior chaetigers. Moreover, in M. haemasoma the swollen base of the ventral cirri continues until the last chaetigers, and the subacicular hook is translucent; while in M. sanguinea the swollen base of the ventral cirri ends between 8-18 chaetigers before the pygidium, and the subacicular hook is reddish basally and translucent distally.
Regression analyses indicated that there are no correlations between the start of the branchiae (R 2 = 0.0702, p = 0.26, n = 11, Fig. 10), the maximum number of branchial filaments (R 2 = 0.000, p = 0.00 n = 11, Fig. 10) or the start of the subacicular hooks (R 2 = 0.1307, p = 0.35, n = 11, Fig. 10) with the length to chaetiger 10 for this species. The chaetiger where the branchiae start does not follow a pattern regarding their growth but starts to emerge from chaetiger 20 to 30 (Fig. 8, blue points). This same situation is repeated with emergence of subacicular hooks, starts between chaetiger 30 and 40 (Fig. 10, orange points). However, the number of filaments (two filaments) seems to be fixed regardless of the size of the organism, a contrasting pattern with other Marphysa species in which the number of filaments appears to increase with the length of the specimen.
On the other hand, M. sherlockae n. sp. has similar characteristics to other species of Marphysa where the presence of compound chaetae is size-dependent (Aiyar, 1931;Pillai, 1958;Salazar-Vallejo & Carrera-Parra, 1998;Molina-Acevedo & Carrera-Parra, 2017;Molina-Acevedo, 2018). Marphysa sherlockae n. sp. specimens with L10 ≤ 6 mm possess compound falcigers to the last chaetiger. In this group of individuals, the number of falcigers per chaetiger decreased from median to posterior region, which was more noticeable in specimens with L10 close to 6 mm. Additionally, specimens with L10 >6 mm do not have falcigers in the posterior region. This condition indicates that in the largest specimens of M. sherlockae n. sp. falcigers will be lost, and only compound spinigers will be observed, as demonstrated in M. gravelyi Southern, 1921, M. borradailei Pillai, 1958and M. brevitentaculata Treadwell, 1921 Etymology: The species is named after Emma Sherlock, in recognition of her valuable work on the polychaete collections of BHNM. DNA barcode: Type region: Strand, False Bay, Western Cape, South Africa (Museum number: SAMC-A089090) (GenBank accession number: MT840249). 577 bp fragment isolated with universal mitochondrial cytochrome oxidase subunit 1 gene, primer pair: LCO1490, HCO2198 (Folmer et al., 1994). Habitat. Fringing rocky zones at low tide in sheltered bays. Worms can be found in rock crevices. Type locality. Langebaan Lagoon, South Africa. Distribution. Day (1953), Day (1967) andBranch et al. (2016) recorded this species to occur in rocky coasts and estuaries from Saldanha Bay in the Western Cape to Durban in KwaZulu-Natal, South Africa. Remarks. Day (1953) studied the material collected by himself and other members of the Zoology Department at the University of Cape Town during ecological surveys of the rocky coasts and estuaries in South Africa. The author identified some specimens as Marphysa depressa collected from localities such as East London, Bushman's Estuary, Still Bay, Cape Agulhas, and Langebaan Lagoon due to the presence of compound spinigers and falcigers in the same chaetiger which is similar to the New Zealand species. As a result, this was the first record of the species in South Africa. Additionally, Day compared his material with a specimen collected from New Zealand by Ehlers (1904), most likely to confirm his identification. However, thorough taxonomic revisions revealed marked differences between the material from South African and New Zealand and led us to conclude that the South African specimens belong to a new species named herein as Marphysa sherlockae n. sp.
Marphysa sherlockae n. sp. differs from M. depressa in the chaetal distribution. For example, the former has compound spinigers in all chaetigers, and compound falcigers restricted to the median and posterior chaetigers; whereas in M. depressa, the compound falciger is present in all chaetigers, but the spinigers are only present in the anterior region. Also, M. sherlockae n. sp. has a triangular postchaetal lobe, while M. depressa has a digitiform postchaetal lobe. Furthermore, M. sherlockae n. sp. (L10: 5.7-6.6 mm) has only two branchial filaments, while M. depressa (L10: 9.5 mm) has up to four filaments.

DISCUSSION
This study revealed that M. macintoshi and M. depressa recorded for the region actually represent (1) an incorrectly synonymized species, i.e., M. durbanensis that was reinstated herein, and (2) a new indigenous species that was previously overlooked and herein described, i.e., M. sherlockae n. sp., respectively. We also confirm the notion addressed by Lewis & Karageorgopoulos (2008), that M. sanguinea is not present along the South African coast. However, the local species should be named M. haemasoma Quatrefages, 1866 andnot M. elityeni Lewis &Karageorgopoulos, 2008, since the latter is a junior synonym of the former.
Marphysa depressa and M. macintoshi were first recorded along the South African coast by Day (1953) and Day (1967) with summary descriptions and general illustrations. The recurrent identification of M. macintoshi and M. depressa along the South African coast (e.g., Branch, Charles & King, 2016) reflects the overlooking of detailed characteristics and the use of traditional and conspicuous diagnostic features considered enough to define Marphysa species, such as the color and shape of the subacicular hook, distribution of compound chaetae throughout the body, the shape and distribution of branchiae, and the number of branchial filaments (Quatrefages, 1866;Grube, 1878;McIntosh, 1910;Hartman, 1944;Fauchald, 1970, among others). The sole use of distinctive conspicuous features in the identification may lead to spurious records of cosmopolitanism in species (Hutchings & Kupriyanova, 2018), and also to the proliferation of misleading species records and synonymization.
The detailed study of the traditional conspicuous features, the discovery of new unique characters as well as the examination of type specimens, as carried out here, has improved the morphological delimitation of Marphysa species, and the understanding of the diversity within the genus (e.g., Glasby & Hutchings, 2010;Molina-Acevedo & Carrera-Parra, 2015;Molina-Acevedo & Carrera-Parra, 2017). Therefore, recent studies on Marphysa have focused on detecting unique characters or in the re-assessment of those forgotten features, such as the shapes of dorsal cirri, postchaetal lobes, and pectinate chaetae, and the first appearance of the ventral cirrus with a swollen base. For instance, Miura (1986) and Molina-Acevedo & Carrera-Parra (2015) have shown that the distribution of the number of filaments and the region where the maximum number is reached can be informative in species delimitation. Here, the distribution of branchial filaments is different in each analyzed species (Fig. 4). Thus, whenever possible, it should be incorporated in future descriptions of Marphysa species. The main challenge of using ''new'' features in taxonomic investigations is the lack of this information in older descriptions preventing comparison. Thus, the examination of type material deposited in museums or examining newly collected material from the type locality in cases where no types were deposited previously is an essential step towards improving the taxonomy and recognition of new or inappropriate synonyms as in the case of M. haemasoma.
Molecular data provide an additional source of information that improves our knowledge on species boundaries and aids in recognition of intraspecific variation (e.g., Lewis & Karageorgopoulos, 2008;Zanol, Da Silva & Hutchings, 2016;Zanol, Da Silva & Hutchings, 2017;Lavesque et al., 2017;Elgetany et al., 2018;Lavesque et al., 2019;Glasby et al., 2019;Abe et al., 2019;Martin et al., 2020). The phylogenetic tree revealed two distinct South African monophyletic clades, belonging to the new species M. sherlockae n. sp., and the other to M. haemasoma. The molecular analyses reinforced the re-establishment of M. haemasoma as a valid species by confirming its distinction from M. sanguinea, which concurs with previous findings from the region (Lewis & Karageorgopoulos, 2008). Furthermore, for the first time, this study provided COI sequences of M. haemasoma, from South Africa.
A total of nine Marphysa species have been newly proposed or redescribed under an integrative taxonomic framework since 2003 (Zanol, Da Silva & Hutchings, 2016;Zanol, Da Silva & Hutchings, 2017;Lavesque et al., 2017;Elgetany et al., 2018;Lavesque et al., 2019;Glasby et al., 2019;Abe et al., 2019;Martin et al., 2020;present study), thus, increasing the number of publicly available sequences of Marphysa species globally. This framework, in turn, provides a starting point from which other studies can address more complex hypotheses, such as resolving the phylogenetic placements of species within the genus.
This study has confirmed that the indigenous diversity of Marphysa in South Africa was indeed previously underestimated and thus increases the number of described indigenous species from three to five (Day, 1967;Lewis & Karageorgopoulos, 2008) and reduces the number of putative cosmopolitan species to one (i.e., Marphysa corallina). Similarly, studies by Lewis & Karageorgopoulos (2008), Clarke et al. (2010), Kara, Macdonald & Simon (2018) and Simon, Sato-Okoshi & Abe (2019) provide additional evidence that many cosmopolitan species reported in the Day (1967) polychaete monograph for this region are actually incorrect assignments. Undoubtedly, the polychaete monograph by Day (1967) is an invaluable resource for polychaete descriptions and distributions. However, it is widely used by researchers from many disciplines, including those working outside of the region (Hutchings & Kupriyanova, 2018). Thus, biologists locally and internationally should take cognizance of this fact and use the monograph with caution, especially concerning species considered ''cosmopolitan''.
Using information from Day (1967) and Awad, Griffiths & Turpie (2002) determined that only 20% of polychaete species in South Africa are endemic to the region. Thus, if only half the remaining 80% prove to be misidentifications of indigenous species, our understanding of diversity, biogeography, and endemism of polychaete worms in South Africa has been severely underestimated, and priority conservation areas may need to be reviewed. Furthermore, the resolution of taxonomically confusing species, such as those belonging to Marphysa, and development of realistic diversity estimates will be improved if voucher specimens are deposited in museums for taxonomic and molecular investigations. . depressa has changed the composition of endemic and cosmopolitan species. As such, gaining a better understanding of our true local biodiversity may help us to understand the extent of biodiversity loss in the face of climate change and make better decisions regarding the designation of marine protected areas.