Introduction
Marine invasions are recognized as one of the greatest threats to global biodiversity (Costello et al., Reference Costello, Coll, Danovaro, Halpin, Ojaveer and Miloslavich2010). Although the study of marine invasive species formally began in the 1970s, research on this field in southern Africa has only recently developed but is important due to the region's status as a biodiversity hotspot (Griffiths et al., Reference Griffiths, Mead and Robinson2009). In South Africa, at least 96 marine non-indigenous species (NIS) have been recorded; these include 55 that are invasive, and nine polychaete species (Miza et al., Reference Miza, Robinson, Peters, Majiedt, Jackson, Hampton, Sink, Sink, van der Bank, Majiedt, Harris, Atkinson, Kirkman and Karenyi2019). However, it is likely that the number of non-indigenous polychaete species identified in the region is underestimated, as evidenced by recent studies (e.g. Williams et al., Reference Williams, Karl, Rice and Simon2017; Malan et al., Reference Malan, Williams, Abe, Sato-Okoshi, Matthee and Simon2020; van Rensburg et al., Reference Van Rensburg, Matthee and Simon2020). Polychaetous annelids are frequent members of the non-indigenous marine fauna across the globe, likely due to factors that make them prone to introductions (Çinar, Reference Cinar2013), as well as biological mechanisms that enhance invasion success (Papacostas et al., Reference Papacostas, Rielly-Carroll, Georgian, Long, Princiotta, Quattrini, Reuter and Freestone2017). For example, many polychaetes exhibit r-selected life history strategies where they produce numerous planktonic larvae that can spend months in the water column, which in turn facilitates uptake into ballast water (Carlton & Gellar, Reference Carlton and Gellar1993). In addition, tubiculous (tube-dwelling) polychaetes (e.g. Spionidae and Serpulidae) can foul the hulls of ships, exist as symbionts of hull fouling organisms such as barnacles and sponges and bore into the shells of commercially reared molluscs, the latter of which is a major component of South Africa's aquaculture industry (Çinar, Reference Cinar2013; David et al., Reference David, Matthee and Simon2014).
Spionidae forms a significant component of the known non-indigenous polychaetes (Çinar, Reference Cinar2013), and many, especially members of the Polydora-complex (i.e. nine genera defined by an enlarged fifth chaetiger with modified spines (Walker, Reference Walker2011)), can be extremely destructive if they successfully invade a habitat. For example, the tube-dwelling spionid Boccardia proboscidea invaded a sewage outfall area in Argentina, producing biogenic reefs that excluded other organisms from the impacted area (Jaubet et al., Reference Jaubet, de los Angeles Sanchez, Rivero, Garaffo, Vallarino and Elias2011). Polychaetes such as B. proboscidea are considered examples of invasive ecosystem engineers (IEEs) that could have major ecological impacts through habitat modification (Guy-Haim et al., Reference Guy-Haim, Lyons, Kotta, Ojaveer, Queirós, Chatzinikolaou, Arvanitidis, Como, Magni, Blight and Orav-Kotta2018). It is therefore imperative that non-indigenous polychaetes be identified as early as possible to facilitate management before they become problematic, a major goal of rapid assessment surveys (e.g. David & Krick, Reference David and Krick2019; Pederson et al., Reference Pederson, Carlton, Bastidas, David, Grady, Green-Gavrielidis, Hobbs, Kennedy, Knack, McCuller, O'Brien, Osborne, Pankey and Trottin press). However, this is complicated by the fact that polychaetes are known to exhibit high levels of cryptic diversity (Carr et al., Reference Carr, Hardy, Brown, Macdonald and Herbert2011; Nygren, Reference Nygren2014; Nygren et al., Reference Nygren, Parapar, Pons, Meißner, Bakken, Kongsrud, Oug, Gaeva, Sikorski, Johansen and Hutchings2018; Malan et al., Reference Malan, Williams, Abe, Sato-Okoshi, Matthee and Simon2020) and many species that were once considered ‘cosmopolitan’ may actually be part of cryptic complexes (Hutchings & Kupriyanova, Reference Hutchings and Kupriyanova2018). Many purported species complexes result from limited investigations to distinguish species or poor taxonomic practices (i.e. they represent pseudo-cryptic species complexes; see Nygren, Reference Nygren2014). In contrast, true cryptic species are reproductively isolated but show strikingly similar morphologies and can only be distinguished genetically or through reproductive crosses (Rice et al., Reference Rice, Karl and Rice2008; Struck et al., Reference Struck, Feder, Bendiksby, Birkeland, Cerca, Gusarov, Kistenich, Larsson, Liow, Nowak and Stedje2018; Struck & Cerca De Oliveira, Reference Struck and Cerca De Oliveira2019).
A recent survey of intertidal fauna in the Knysna Estuary on the southern coast of South Africa (Figure 1A) revealed an unknown spionid worm inhabiting sediment tubes (Figure 1B). The species could not be identified using the taxonomic monograph of Day (Reference Day1967), but preliminary investigation using light microscopy tentatively identified the worm as Dipolydora cf. socialis based on the key provided by Blake (Reference Blake, Blake, Hilbig and Scott1996) (see Williams et al., Reference Williams, Karl, Rice and Simon2017). Dipolydora socialis (Schmarda, Reference Schmarda1861) inhabits sediment but can also be associated with colonial invertebrates such as sponges and molluscs (e.g. Sato-Okoshi & Takatsuka, Reference Sato-Okoshi and Takatsuka2001; Williams, Reference Williams2001; David & Williams, Reference David and Williams2012a), or as a borer in molluscan shells (Blake, Reference Blake1971). The species is considered cryptogenic i.e. ‘not demonstrably native or introduced’ (Carlton, Reference Carlton1996; Schwindt et al., Reference Schwindt, Carlton, Orensanz, Scarabino and Bortolus2020), along the Pacific coast of North America and Atlantic coast of South America and in Australia (Boyd et al., Reference Boyd, Mulligan and Shaughnessy2002; Orensanz et al., Reference Orensanz, Schwindt, Pastorino, Bortolus, Casas, Darrigran, Elias, Gappa, Obenat, Pascual, Penchaszadeh, Piriz, Scarabino, Spivak and Vallarino2002; Hayes et al., Reference Hayes, Sliwa, Migus, McEnnulty and Dunstan2005). The taxonomic history of D. socialis is complicated, with the species having undergone numerous taxonomic revisions, the most significant of which was a synonymization with another Dipolydora species, Dipolydora carunculata (Radashevsky, Reference Radashevsky1993) (Blake, Reference Blake, Blake, Hilbig and Scott1996), a designation which was later rejected due to insufficient evidence (Manchenko & Radashevsky, Reference Manchenko and Radashevsky2002). In this study we utilize morphological and molecular data to investigate the presence of D. socialis in southern Africa.
Fig. 1. (A) Map showing global distribution of Dipolydora socialis based on reports from the literature: 1 – Chile (type locality), 2 – Brazil, 3 – Trinidad & Tobago, 4 – Gulf of Mexico, 5 – Humboldt & Bodega Bay, California, 6 – Rhode Island, USA, 7 – British Columbia, Canada, 8 – Gulf of Alaska, 9 – Iberian Peninsula (Spain), 10 – Germany & Black Sea, 11 – Hong Kong, 12 – Taiwan, 13 – Philippines, 14 – New South Wales, Australia, 15 – New Zealand. ‘N’ represents the first southern African population of D. socialis found in the current study. (B) Colony of sand tubes constructed by D. socialis in the intertidal zone at the Knysna Estuary; inset photo shows a microscopic image of a live D. socialis with a recently regenerated anterior end (scale bar: 600 μm). (C) Microscopic image showing the mid-region of D. socialis with the visible gizzard identified (white arrow). (D) Microscopic image of the modified spines on the enlarged fifth chaetiger (scale bar: 100 μm).
Materials and methods
Sediment samples were collected from the intertidal region within the Knysna Estuary, Western Cape Province, South Africa (34°3′57″S 23°3′17″E) on 21 January 2015. During collection, sandy tubes were processed individually, and all fragments per tube were interpreted as an individual undergoing architomic division. Samples were sorted, anaesthetized in 7% MgCl2 and either fixed in 4% formalin in seawater and then stored in 70% ethanol, or directly in 96% ethanol. Photographs of whole animals were taken through an Olympus SZ61 microscope using a Canon Powershot S3 IS camera. For morphological comparison, additional Dipolydora socialis specimens were obtained from Discovery Bay, Hong Kong and Rhode Island, USA (specimens collected by J. Williams). Voucher specimens from Chile (type locality) (USNM 1006390), Philippines (USNM 187534), Brazil (USNM 1022162) and Taiwan (USNM 1022160) from the National Museum of Natural History were also examined. Specimens were compared morphologically using both light microscopy (Leica Microsystems) and scanning electron microscopy (SEM) (FEI Quanta 450). For SEM preparations, worms were dehydrated in an ascending ethanol series (75%, 80%, 85%, 90%, 95%) for 10 min each and in 100% three times for 15 min each. Specimens were critically point dried over CO2 (Samdri-795 Critical Point Dryer), mounted on an aluminium stub with sticky tape and coated with gold (EMS-550 Sputter Coater).
For genetic analyses, genomic DNA from two South African specimens (referred to previously as Dipolydora cf. socialis) was extracted using the DNeasy blood and tissue DNA extraction kit (QIAGEN, Hilden, Germany), following the manufacturer's protocol. A ~800 bp fragment of the 18S rRNA gene and a ~710 bp fragment of cytochrome c oxidase I (COI) gene were amplified using the polymerase chain reaction (PCR) and the forward and reverse primer pairs from Teramoto et al. (Reference Teramoto, Sato-Okoshi, Abe, Nishitani and Endo2013) and the Dorid_COI.3F and Dorid_COI.3R primers from Williams et al. (Reference Williams, Karl, Rice and Simon2017). Cycling parameters for 18S rRNA included initial denaturation of 95°C for 4 min, followed by 40 cycles of 95°C for 30 s, annealing 55°C for 30 s, extension 72°C for 30 s, and final extension 72°C for 7 min. Cycling parameters for the COI gene followed the conditions outlined by Williams et al. (Reference Williams, Karl, Rice and Simon2017). Amplified PCR products were verified on a 2% agarose gel stained with ethidium bromide and purified using a gel clean-up kit (QIAGEN, Hilden, Germany). Amplicons were sequenced by GeneWiz (South Plainfield, NJ, USA) using the forward primers and Big Dye Terminator Cycle Sequencing. Sequence data generated were deposited into the GenBank database for archiving (accession nos. MT019828 & MT019829 and MT040509 & MT040510).
The generated sequence data were first compared with the GenBank database using the BLASTn tool to determine initial similarity indices. COI and 18S rRNA datasets were then compiled using archived polydorid sequences from GenBank. Dipolydora cf. socialis sequences from South Africa were obtained from Williams et al. (Reference Williams, Karl, Rice and Simon2017) based on a tentative identification prior to the present study (accession nos. KY677859, KY677899, KY002976), whereas the North American COI sequence was based on a single individual sampled from Massachusetts, USA (David & Krick, Reference David and Krick2019) (accession no. MK189200). The archived 18S rRNA gene sequences for Dipolydora cf. socialis (Williams et al., Reference Williams, Karl, Rice and Simon2017) were sampled from the same locality as those from the current study. Both datasets were aligned using the Clustal W alignment algorithm in BioEdit ver. 7.0.5.3 (Hall, Reference Hall1999) and edited by eye. Genetic relatedness among taxa was assessed by calculating pairwise uncorrected p-distances in MEGAX (Kumar et al., Reference Kumar, Stecher, Li, Knyaz and Tamura2018). In addition, a Maximum likelihood phylogenetic tree was also constructed in MEGAX with 1000 bootstrap to determine cladal support, using the General Time Reversible (GTR) evolutionary model as determined by AICc index in jModelTest2 (Darriba et al., Reference Darriba, Taboada, Doallo and Posada2012). All trees were edited and visualized in FigTree ver. 1.4.3 (Rambaut, Reference Rambaut2007). Voucher specimens were deposited at the Iziko South Africa Museum (SAM) and the National Museum of Natural History (USNM, Smithsonian Institution).
Results
SYSTEMATICS
Order SPIONIDA sensu Rouse & Fauchald, 1997
Family SPIONIDAE Grube, 1850
Genus Dipolydora Verrill, 1881
Dipolydora socialis (Schmarda, Reference Schmarda1861)
Figures 1C, D, 2
Abbreviated synonomy (main taxonomic sources provided)
Fig. 2. Scanning electron micrograph (SEM) plates of Dipolydora socialis showing (A) lateral view of anterior chaetigers, (B) dorsal view with palps removed showing caruncle and nuchal organ, (C) modified fifth chaetiger with spines and (D) pygidium with tufts of cilia in the centre. Scale bars: A, 250 μm, B, 150 μm, C, 25 μm, D, 50 μm.
Leucodore socialis Schmarda, Reference Schmarda1861: 64, figures a–c, pl. 26, figure 209
Polydora caeca var. magna Berkeley, Reference Berkeley1927: 419; Pettibone, Reference Pettibone1967: 11.
Polydora magna Berkeley & Berkeley, Reference Berkeley and Berkeley1936: 473;, Reference Berkeley and Berkeley1952: 21.
Polydora socialis plena Berkeley & Berkeley, Reference Berkeley and Berkeley1936: 469; 1953: 20–21.
Polydora socialis Hartman, Reference Hartman1941: 310–311, pl. 48, figures 41–42; Reference Hartman1969: 147; Hartmann-Schröder, Reference Hartman-Schröder1962: 137–139, figures 167–168; Reference Hartman-Schröder, Hartmann-Schröder and Hartmann1965: 209–211, figures 200–203; Blake, Reference Blake1971: 20–23, figures 13–14; Reference Blake, Light, Smith and Carlton1975: 215, figures 237–238; Reference Blake1979: 607–609; Reference Blake1981: 950; Reference Blake1983: 264; Carrasco, Reference Carrasco1974: 194–196, figures 27–32; Light, Reference Light1977: 71; Reference Light1978: 179–181, figure 180; Sato-Okoshi & Okoshi, Reference Sato-Okoshi and Okoshi1997: 486; Blake & Kudenov, Reference Blake and Kudenov1978: 248–250, figure 38d–e; Johnson, Reference Johnson, Uebelacker and Johnson1984: 6–28 to 6–30, figures 6–19 and 6–20.
Polydora plena Foster, Reference Foster1971: 24–25, figures 22–29.
Polydora neocardalia Hartman, Reference Hartman1961: 96–98, pl. 14, figures 1–4; 1969: 141, 2 figures; Lissner et al., Reference Lissner, Phillips, Cadien, Smith, Bernstein, Cimberg, Kauwling and Anikouchine1986: appendix D; Steinhauer & Imamura, Reference Steinhauer and Imamura1990: figure 1.
Dipolydora socialis Blake, Reference Blake, Blake, Hilbig and Scott1996: 189–192, figure 4.34; Williams, Reference Williams2001: 442–445, figures 7–8; Walker, Reference Walker2009: 39, 133; 2011: 52; Schwindt et al., Reference Schwindt, Carlton, Orensanz, Scarabino and Bortolus2020: tables 1, S1.
Dipolydora cf. socialis Williams et al., Reference Williams, Karl, Rice and Simon2017: 107, 109–113, figures 1–5, tables 2, 3.
?Dipolydora socialis Abd Elnaby, Reference Abd Elnaby2019, figure 8.
Specimens examined
Knysna, Western Cape Province, South Africa (34°3′57″S 23°3′17″E) from sediment, January 2015, coll. C. Simon and F. Smith (8 specimens, destroyed for molecular analyses, 5 complete specimens SAM A089093 and 4 specimens on 2 SEM stubs USNM 1620898); Discovery Bay, Lantau Island, Hong Kong (22°18′0.74″N 114°01′0.84″E), from burrows in gastropod shells inhabited by hermit crabs, 8 June 2004, coll. J. Williams (2 specimens destroyed for molecular analyses, 3 specimens in ethanol USNM 1620899); Pettaquamscutt River, Narragansett, Rhode Island, USA (41°26′57.6″N 71°27′2.0″W), from burrows in Crepidula fornicata, 16 October 1998, coll. J. Williams (3 specimens destroyed for molecular analyses); Rowes Wharf, Boston, MA, USA from mudtubes on Ostrea edulis, July 2018, coll. J. Carlton & A. David (2 specimens); Sombrero Island, Philippines, July 1997 (USNM 187534, 10 specimens); Pontal do Sul, Parana, Brazil, August 1998, coll. V. Radashevsky (USNM 1022162, 11 specimens); Hsiangshan, Taiwan, July 1999, coll. V. Radashevsky (USNM 1022160, 3 specimens); Chiloe Island, Chile, February 1998, coll. W. Sato-Okoshi & M. Takatsuka (USNM 1006390, 4 specimens).
Description of South African specimens
Whole specimens, 100–180 chaetigers in length. Prostomium bifid, extends as a caruncle to the end of chaetiger 3, surrounded by rows of sensory cilia (Figure 2A, B). Palps long and extend posteriorly to chaetigers 10–12 with ventral food groove lined by frontal cilia (other ciliary groups not observed because not fixed originally for SEM; see Worsaae, Reference Worsaae2001). Eyes and body pigmentation absent, brown pigmentation along feeding groove on palps present in some preserved specimens. Chaetiger 1 with noto- and neurochaetae, chaetigers 2–4 and 6 with two rows of capillary notochaetae and neurochaetae. Hooded hooks with curved shaft without constriction with an angle of 45° between apical tooth and main fang; no observable changes in angle in subsequent chaetigers. Hooded hooks begin on chaetiger 7; 3–4 per fascicle, accompanied by 1–2 capillary chaetae; companion chaetae not present from chaetiger 10. Branchiae small, beginning on chaetiger 8 for all specimens examined (Figure 2A). Branchiae decrease in length after chaetiger 10; absent on terminal chaetigers. Chaetiger 5 enlarged; approximately twice the size of preceding and succeeding chaetigers, with 4–5 simple stout spines (Figure 1D) accompanied by dorsal bundle of geniculate chaetae and ventral fascicle of capillaries (Figure 2C), spines falcate-shaped with subterminal protuberance and alternating row of companion chaetae. Posterior notopodial spines present as fine needle-like capillaries. Conspicuous gizzard beginning in chaetiger 16, extending for ~2 chaetigers with longitudinal muscles; only observable under light microscopy (Figure 1C). Pygidium disc-shaped with dorsal notch and tuft of cilia in the centre (Figure 2D).
Remarks
Morphologically, Dipolydora cf. socialis specimens from South Africa were almost indistinguishable from Dipolydora socialis reported from several geographic regions, including the type locality, except for slight variations such as caruncle extension and overall shape of the pygidium (see Table 1). During initial identification, regeneration of anterior segments was observed in 40% of worms collected (Figure 1B inset), with a maximum of eight anterior chaetigers regenerated; in addition, at least five individual fragments were regenerating both anterior and posterior ends simultaneously.
Table 1. Morphological comparison of Dipolydora socialis from selected geographic regions
PH, Philippines; BR, Brazil; HK, Hong Kong; TW, Taiwan; *CH, Chile; SA, South Africa.
a Type locality.
Genetic barcoding
DNA barcoding using the sequence data available from GenBank corroborated morphological analyses in identifying the specimens as Dipolydora socialis. Overall genetic distances of the in-group taxa ranged from 0.000–0.047 for the 18S rRNA marker and 0.000–0.224 for the COI marker. The maximum intraspecific distance for D. socialis specimens were 0.000 for the 18S rRNA marker and 0.015 for the COI marker with the species forming a distinct clade with high bootstrap support (Figures 3 & 4). For the 18S rRNA dataset, D. socialis also exhibited low genetic distances with D. cardalia (0.009) and D. carunculata (0.007).
Fig. 3. Maximum likelihood phylogenetic tree showing phylogenetic position of Dipolydora socialis relative to other polydorid taxa based on 18S rRNA barcode data. Codes adjacent to taxa represent GenBank accession codes and numbers above and below branch nodes represent bootstrap support based on 1000 replications. Nodes without bootstrap support represent clades where bootstrap values are less than 50. Sequences in bold font are those generated in the present study.
Fig. 4. Maximum likelihood phylogenetic tree showing phylogenetic position of Dipolydora socialis relative to other polydorid taxa based on COI barcode data. Codes adjacent to taxa represent GenBank accession codes and numbers above and below branch nodes represent bootstrap support based on 1000 replications. Nodes without bootstrap support represent clades where bootstrap values are less than 50. Sequences in bold font are those generated in the present study.
Discussion
Thus far, seven Dipolydora species have been recorded in South Africa: D. capensis (Day, 1955), D. normalis (Day, 1957), D. keulderae (Simon, Reference Simon2011), D. cf. giardi (Mesnil, 1896), D. armata (Langerhans, 1880), D. flava (Claparède, 1870) and D. caeca (Oersted, 1843), with only the first three species native to the region (Day, Reference Day1967; Simon, Reference Simon2011). The status of the remaining species as native or non-native has not been confirmed. In this study, we show that the specimens identified tentatively as Dipolydora cf. socialis in South Africa by Williams et al. (Reference Williams, Karl, Rice and Simon2017) and those from this study exhibit extensive genetic and morphological similarities with the supposedly cosmopolitan spionid, D. socialis. It should be noted that although we confirmed reciprocal monophyly of specimens from the east coast of North America and South Africa, sequence data from type locality (Chile) were not available for inclusion in the analysis. The samples examined do share extensive similarities in traditional taxonomically informative traits across different populations, most notably the visible gizzard at chaetigers 15–22, a distinctive trait of the species (Blake, Reference Blake, Blake, Hilbig and Scott1996; Walker, Reference Walker2009, Table 1). Despite these similarities, some variability does exist. For example, South African specimens had a maximum of 180 chaetigers whereas Blake (Reference Blake, Blake, Hilbig and Scott1996) reported worms from California having up to 400 chaetigers. Although intraspecific variation in size may reflect environmental differences and is a common feature across marine invertebrate taxa (e.g. Zakas & Rockman, Reference Zakas and Rockman2014), this should be investigated further, as pseudo-cryptic species complexes are common among polychaetes (Nygren, Reference Nygren2014). This makes it particularly difficult to determine the taxonomic importance of size in this group. Furthermore, other variations detected, such as length of the caruncle, may be linked either to intraspecific variation in size, or reflect interspecific variation (Simon et al., Reference Simon, Sato-Okoshi and Abe2019a). Without genetic data from a wider distribution, we cannot confirm whether D. socialis is (1) a single widespread species (e.g. as shown for Polydora hoplura, Sato-Okoshi et al., Reference Sato-Okoshi, Abe, Nishitani and Simon2017) or (2) a member of a complex of cryptic species of which one or more may be widespread (as suggested for Spirobranchus kraussii, Simon et al., Reference Simon, van Niekerk, Burghardt, ten Hove and Kupriyanova2019b and Ficopomatus enigmaticus, Yee et al., Reference Yee, Mackie and Pernet2019). A recent study by Abe & Sato-Okoshi (Reference Abe and Sato-Okoshi2021) focusing on specimens in north-eastern Japan lends support to the latter hypothesis as the authors also morphologically confirmed D. socialis there, but the18S rRNA sequence exhibited relatively high genetic divergence from D. socialis sequences from South Africa (including those analysed in the present study). Therefore, until sequence data are obtained from morphologically confirmed specimens from additional sites, the most parsimonious option for referring to the South African specimens is using the nominal name, D. socialis.
We found interspecific genetic variation between D. socialis and D. carunculata to be an order of magnitude lower than that of the other pairwise comparisons. Both species also showed similarities in taxonomically informative traits based on Radashevsky's (Reference Radashevsky1993) description of D. carunculata and a museum specimen of D. carunculata that was analysed separately (David unpublished data). Blake (Reference Blake, Blake, Hilbig and Scott1996) synonymized D. carunculata as D. socialis, regarding these differences as intraspecific variation and extending the apparent distribution of D. socialis to the Sea of Japan. However, this synonymy was later rejected by Manchenko & Radashevsky (Reference Manchenko and Radashevsky2002) who argued that Blake (Reference Blake, Blake, Hilbig and Scott1996) did not consider morphological variation in relation to the ecology of each population. Considering the results from our study, we suggest that the status of D. carunculata with D. socialis as distinct species be re-evaluated with additional morphological and molecular data or reproductive crosses (e.g. Rice et al., Reference Rice, Karl and Rice2008).
In this study, we also confirm architomy in D. socialis by observing fragmenting worms along with regenerating anterior (Figure 1B inset) and posterior ends of smaller fragments, prior to anaesthetization. Although architomic division has been reported before in D. socialis based on lab experiments (Stock, Reference Stock1964) this is the first field observation of the phenomenon in the species. Asexual reproduction via architomy has thus far only been recorded from six spionids, including D. socialis (Blake, Reference Blake, Rouse and Pleijel2006; David & Williams, Reference David and Williams2012b; Whitford & Williams, Reference Whitford and Williams2016). Stock (Reference Stock1964) found that D. socialis can regenerate a maximum of eight chaetigers which was also confirmed from field observations in this study. Such a reproductive strategy in a non-indigenous species has important management implications since it may facilitate rapid establishment of the species in the introduced range (David & Williams, Reference David and Williams2012b). Dipolydora socialis also offers a potentially effective model system to study anterior and posterior regeneration in a single species that inhabits both soft and calcareous substrates. Investigating whether regeneration rates vary across these habitats (calcareous substrate vs sediment) would provide valuable insights into the polymorphisms of spioniform worms.
Dipolydora socialis may have been transported to southern Africa, and Knysna Estuary in particular, via multiple vectors (see review in Papacostas et al., Reference Papacostas, Rielly-Carroll, Georgian, Long, Princiotta, Quattrini, Reuter and Freestone2017). Firstly, the species is known to produce planktotrophic larvae that could survive for months in the water column (Blake & Arnofsky, Reference Blake and Arnofsky1999) and could therefore have been transported to South Africa in ballast water. Secondly, there are multiple harbours and marinas in Knysna Estuary (Claassens et al., Reference Claassens, Barnes, Wasserman, Lamberth, Miranda, van Niekerk and Adams2020), so the species could have arrived as an epibiont of hull fouling invertebrates such as sponges from an unidentified population elsewhere in the country. Finally, for many years, Knysna Estuary was a hub for oyster farming that relied on regular importation of oyster spat and movement of oysters (Haupt et al., Reference Haupt, Griffiths, Robinson and Tonin2010; Williams et al., Reference Williams, Matthee and Simon2016). Given that the species has been reported as a shell borer (Blake, Reference Blake1971) and shell fouler (Sato-Okoshi & Takatsuka, Reference Sato-Okoshi and Takatsuka2001), it may have arrived as a hitchhiker, even though it has not been reported on farmed molluscs in South Africa. Another record of D. socialis on the African continent was recently provided by Ab Elnaby (Reference Abd Elnaby2019); however, the description lacked critical taxonomic details and the images provided were not of sufficient quality to allow for accurate species delineation. As that study also lacked genetic data, we regard the Egyptian report as dubious.
In Australia, Walker (Reference Walker2009) reported D. socialis producing thick tube colonies which may cause geophysical alterations in the environment that could negatively affect other organisms. Similarly, in Bodega Harbor, California, historical colonization of sediment by D. socialis resulted in the extirpation of native fauna and the establishment of an ‘alternative’ community dominated by D. socialis and B. proboscidea (Bowles, Reference Bowles2013). Thus, D. socialis provides another example of a spionid that could be considered an ecosystem engineer, one with potential negative impacts in non-native regions. Finally, if D. socialis does have shell-boring capabilities, it could become established on shellfish farms which are known to serve as potential source populations for invasive polychaetes (David, Reference David2015; Williams et al., Reference Williams, Matthee and Simon2016). Shell borers are known to negatively impact molluscs, including causing an energetic burden, reducing shell strength and lowering growth rates of hosts (Nel et al., Reference Nel, Coetzee and Van Niekerk1996; Clements et al., Reference Clements, Bourque, McLaughlin, Stephenson and Comeau2018; Spencer et al., Reference Spencer, Martinelli, King, Crim, Blake, Lopes and Wood2020). Consequently, we recommend that future monitoring on the South African coast include explicit searches for this species as this will provide much-needed information on the extent of its distribution and its natural history in the region.
Acknowledgements
CAS thanks Frances Smith for first alerting her to the polychaetes in these sandy tube communities and the Knysna Basin Project for use of their laboratory facilities. Comments from three anonymous reviewers were helpful in improving the quality of the manuscript.
Financial support
Funding provided to CAS by the National Research Foundation Incentive funding, grant number 77747; funding provided to JDW by the National Science Foundation (DBI-1337525).
