Symplegma (Ascidiacea: Styelidae), a non-indigenous genus spreading within the Mediterranean Sea: taxonomy, routes and vectors

Symplegma is a genus of compound ascidians (Fam. Styelidae) with warm water affinities and distribution in tropical and subtropical waters of the Pacific, Indian and Atlantic Oceans. The first record of this genus (as S. viride ) in the Mediterranean was from 1951 in the Levantine Sea, presumably entering the basin from the Red Sea through the Suez Canal. Subsequently, it has been expanding its distributional range northward along the Levantine Sea coast, probably following the prevailing surface current direction. Recently, Symplegma has colonized the Aegean, Ionian and Tyrrhenian Seas, where it is spreading quickly, most likely mediated by shipping (i.e


Introduction
The opening of the Suez Canal, the expansion and increase in intensity of maritime traffic, aquaculture and the marine aquarium species trade are the main vectors of introduction of non-indigenous taxa in the Mediterranean Sea, accelerated and favoured by climatic change (Zibrowius 1992;Bianchi and Morri 2003;Streftaris et al. 2005;Galil 2006;Occhipinti-Ambrogi 2007;Abdulla and Linden 2008;Zenetos et al. 2012;Ferrario et al. 2017). Among the world's seas, the Mediterranean Sea is the most invaded by non-indigenous species (NIS), at present hosting about 700 confirmed marine NIS (Zenetos et al. 2017;Galil et al. 2018). One genus of NIS Co-Editors' Note: This study was first presented at the 10 th International Conference on Marine Bioinvasions held in Puerto Madryn, Argentina, October 16-18, 2018 (http://www.marinebioinvasions.info). Since their inception in 1999, the ICMB meetings have provided a venue for the exchange of information on various aspects of biological invasions in marine ecosystems, including ecological research, education, management and policies tackling marine bioinvasions. ascidians found in the Mediterranean is Symplegma, which normally occurs in shallow and warm waters of tropical and subtropical seas around the world (Van Name 1945;Tokioka 1967;Monniot and Monniot 1997a;Kott 2004).
Amongst the Symplegma species, only S. brakenhielmi has been reported from the Mediterranean. The first record was made in 1951 off Cesarea (Israel) as S. viride by Pérès (1958a). Since then, the genus has exhibited a slow northwards expansion (Bitar and Kouli-Bitar 2001;Çinar et al. 2006;Izquierdo-Muñoz et al. 2009;Shenkar and Loya 2009). In 2008, the species was recorded in the Ionian Sea (present work), probably introduced via shipping, since it had not been previously recorded from the east coast of Greece and southern Italy, areas which would putatively have been colonised at an earlier stage considering the prevailing surface current pathway of the Eastern Mediterranean (Hamad et al. 2005). From 2014, new records of the genus have been made within the Aegean, Ionian and Tyrrhenian Seas (Ulman et al. 2017;Aydin-Onen 2018;Mastrototaro et al. 2019;present work), indicating its rapid expansion throughout the Mediterranean, and its striking colours have also attracted attention from SCUBA divers.
The aim of this study was to clarify the taxonomy and update the distribution of the Symplegma genus in the Mediterranean Sea. Additionally, we discuss possible temporal changes in the distribution of the genus and the likely introduction vectors for S. brakenhielmi in the Mediterranean Sea. A distribution of Symplegma spp. at the global level is also provided.

Materials and methods
A comprehensive literature search on the genus Symplegma was conducted, including available published scientific works, "grey literature" (scientific congresses, technical reports, student theses), and web databases (Ascidian World Database, Biodiversity Heritage Library, Google Scholar, Scopus, GBIF). Data from personal observations was also included, and first country Mediterranean records were cited according to the first actual collection/observation date, rather than the reporting date.
Colonies of Symplegma from different Mediterranean localities were observed and photographed at depths ranging from 0.5 to 17 m between 2005 and 2018 from natural and artificial substrates by snorkelling or SCUBA diving in the framework of different projects and studies carried out by the authors on benthic communities. Some of these colonies, collected from Egypt, Italy, Lebanon, Malta, Tunisia and Turkey (N = 14; Supplementary material Table S1) were dissected for subsequent histological studies in order to identify the specimens based on the morphological characters.
The colonies were anesthetised with menthol crystals, fixed with 4% formalin in seawater and preserved in 70º ethanol. Some zooids from each colony were dissected, stained with Masson's haemalum, and dehydrated in ethyl and butyl alcohols for mounting on permanent slides in Canada balsam. Taxonomic identification was made considering the original and subsequent descriptions of Symplegma species in literature (Van Name 1945;Tokioka 1961Tokioka , 1967Monniot 1972Monniot , 1983Monniot , 1988Kott 1985Kott , 2004Nishikawa 1991;Monniot and Monniot 1997a;Rocha and Costa 2005). The specimens examined in this study were deposited at the Marine Research Center of Santa Pola (CIMAR) of the University of Alicante (Spain) with the following identification codes: Sy.br-Eg01(from Egypt), Sy.br-It01 (from Italy), Sy.br-Le01-02 (from Lebanon), Sy.br-Mt01-02 (from Malta), Sy.br-Tu01-07 (from Tunisia) and Sy.br-Tk01 (from Turkey).

Symplegma brakenhielmi
Material examined: See Table S1. Description: Crusted and thin colonies, 1-2 mm thickness and up to 12 cm in diameter. Colonies were attached to ascidians, rock, seagrasses, mussels, sabellids, ropes and other artificial structures. Living colonies were lightcoloured (red, rose, yellow, orange, rose, brown, white, grey; Figure 1), that turned greyish or yellowish upon fixation. The oval zooids were embedded by the tunic, depressed dorso-ventrally (rarely erect), close together, and randomly arranged without forming apparent systems. They represented different sizes (max. 3 mm), the youngest intercalated between the adults. The thin test was rather tough in consistency and was transparent, allowing us to observe a "reticulated" pigmentation of the living zooids that concentrated in the branchial sac and the gut, although some colonies from Tunisia and Malta had a more opaque tunic. The siphons were similar, close and tubular and did not have coloured rings or a band around or between them. The thin and transparent mantle allowed the observation of the branchial sac and digestive tract. There were between 12 and 18 buccal simple tentacles, arranged in three orders, of which 6-9 were large and long and the others smaller ( Figure 2A). The peribranchial area was bordered by a narrow velum in a V-shape, where the oval dorsal tubercle, with a small aperture inside, was located ( Figure 2A); and the dorsal lamina was smooth. There were very small atrial tentacles.
Branchial sac was without folds, with four longitudinal vessels on each side of the body, with the first two on the left being incomplete and joining the dorsal lamina at the level of 6 th -7 th rows of stigmata ( Figure 2B). In mature zooids, about 11-13 rows of stigmata on the left and 10-12 rows on the right were present, separated by transverse vessels of uniform size, which do not meet the medium dorsal vessel exactly opposite each other. The number of stigmata per half row was between 22 and 25, and there were usually from 4 to 5 stigmata between the internal longitudinal vessels, except next to the dorsal lamina, where 6-8 stigmata were present. The stigmata formula at the 5 th row on the right side was the following: E 4-5 v 4 v 4 v 4 v 6-8 DL.
The gut occupied a third of the left part of the body and formed an open loop. The oesophagus was narrow and curved followed by a cylindrical or globular stomach. The stomach had between 9 and 13 well-marked longitudinal folds, and a short, curved and stout caecum, united by two connections to the intestine ( Figure 2C). The rectum bent forward and ended in a non-lobed anus at the 5 th -6 th row of stigmata.
One hermaphrodite gonad was located on each side, with two lobed testes and an ovary with 1-4 ovules of different sizes ( Figure 2D, E). Male and female gonads were present at the same time. The testes had 2-6 lobules, little or deeply divided, and the common spermiduct was narrow with variable length. In a particular colony (station MT-2), incubated eggs and 3-5 free larvae were concomitantly observed. The larvae presented a single organ (photolith) and three sharp papillae radially ranged ( Figure 2E), with a length of 0.9 mm, corresponding to a length of 320 μm up to the trunk.
Biology and ecology (Mediterranean Sea): The colonies sampled in summer (July-September) and early autumn (October) showed fully developed gonads and free fertilized ovules; the colonies from Monastir (Tunisia) sampled at the end of August contained larvae.
Global distribution: Table 1 shows the distribution of the Symplegma spp. throughout the world. Symplegma brakenhielmi represents the most widely-distributed species, both in tropical (Atlantic and Indo-Pacific) and in warm-temperate waters (NE, NW and SW Atlantic; NW Pacific; Mediterranean Sea; and Australasia).

Determination of Symplegma in the Mediterranean Sea
The genus Symplegma was created by Herdman in 1886 (p. 144) and was initially represented by the single species S. viride. Michaelsen (1904, p. 50) regarded this genus as nomen dubium due to Herdman's incomplete description and subsequently replaced it with the genus Diandrocarpa (Van Table 2. Records of Symplegma brakenhielmi (Michaelsen, 1904)

Region
Zone    Table 2), surface current direction (Hamad et al. 2005) and boundaries of the Mediterranean sub-basins (dashed lines) are also represented.
Name, 1902) containing the single species D. brakenhielmi. However, Van Name (1921, 1930, 1945 considered both species synonymous, giving priority to S. viride, stating that S. brakenhielmi cannot even be maintained as a subspecies of S. viride. This categorical statement was supported by subsequent authors until the 1980s, when Monniot (1983) established the differences between the aforementioned two species (S. viride and S. brakenhielmi), together with S. rubra, in Caribbean waters. Therefore, many pre-1983 publications on Symplegma in the Western and Eastern Atlantic, as well as in the Indian Ocean, adopted Van Name's description (1945) and considered these congeneric species as a single species: S. viride. Thus, one can infer that the first records of Symplegma (as S. viride; Herdman 1886) in the Suez Canal and Mediterranean Sea (Harant 1927;Pérès 1958a, b;Steiniz 1967;Por 1978;Koukouras et al. 1995), and Tropical Western Africa (Pérès 1949(Pérès , 1951Millar 1953;Monniot 1969;Lafargue and Wahl 1986) correspond to S. brakenhielmi. Symplegma brakenhielmi, S. rubra and S. bahraini present lobed testes in contrast to S. viride and S. reptans without lobed testes. However, in S. brakenhielmi, the tunic is transparent in living colonies with the pigments localised in the blood cells, thus imparting a reticulated aspect due to the concentration in the branchial sac and gut (S. rubra and S. baharaini have opaque tunics). Additionally, S. brakenhielmi does not have the characteristic red rings that encircle both siphons (a characteristic of S. rubra); although Herdman (1906) and Tokioka (1961) do indicate the presence of red circles around the siphons in S. brakenhielmi var. ceylonica and S. oceania (a synonym of S. brakenhielmi), respectively. Other characters, such as the number of testes lobes and the division between them (more numerous and deeper in S. rubra and S. bahraini) as well as the length and width of the common spermiduct are less diagnostic (Monniot 2002;Couto 2003). The simultaneous presence of testes and ovules in S. brakenhielmi and S. bahraini separates them from S. rubra (Monniot 1983;Monniot and Monniot 1997a), but this is not applicable for immature zooids . Rocha and Costa (2005) included the following other characters to distinguish S. brakenhielmi from S. rubra: the swollen dorsal tubercle of the former, with a small circular aperture inside; two incomplete dorsal left longitudinal vessels, reaching the dorsal lamina at the level of the 4 th and 7 th rows of stigmata; and only two tissue connections linking the caecum with the intestine. These characteristics have all been observed in the zooids of samples originating from the Mediterranean, and a molecular study conducted by Mastrototaro et al. (2019) corroborates the presence of S. brakenhielmi in the basin.
It has to be noticed that the pigmentation of the tunic and the deeply divided testicular follicles from the Maltese and Tunisian (i.e. Monastir) specimens resemble S. bahraini (Monniot and Monniot 1997a;Kott 2004). However, Monniot and Monniot (1987: 10) found colonies of S. oceania (synonymy of S. brakenhielmi) with a pigmented tunic; and Tokioka (1961, Figure 7) drew deeply divided testicular follicles for the same species. The presence of larvae in the Tunisian colonies with a single sensory organ (Millar 1953;Kott 1985;Mastrototaro et al. 2019) suggests that these colonies are more closely related to S. brakenhielmi, since S. bahraini has two sensory organs (Monniot and Monniot 1997a). Furthermore, the remaining morphological characters fall within the range of S. brakenhielmi and its synonyms (Monniot and Monniot 1997a;Monniot 2002;Kott 2004), such as its stomach folds, rows of stigmata, stigmata per row, pyloric caecum and its links with the intestine; longitudinal vessels on the left join the dorsal lamina, and the simultaneous presence of functional male and female gonads.
In the present study some colonies sampled in summer and autumn showed fully developed gonads or contained larvae. Larvae had also been recorded in June in Israel (Levantine Sea) by Shenkar and Loya (2009); and in July in North-eastern Sardinia (Tyrrhenian Sea) by Mastrototaro et al. (2019). Generally, data on reproductive period is scarce and scattered, e.g. from January to May in the Tropical Eastern Atlantic (Senegal : Pérès 1949;Monniot 1969;Ghana: Millar 1953), as well as immature colonies in January (Pérès 1951); in May in the Southern Indian Ocean (Mauritius: Vasseur 1967); from July to October in the Northern Indian Ocean (S-India: Renganathan 1985; Bahrain: Monniot and Monniot 1997a); and in November in the Tropical Pacific (W-Australia: Kott 1985).
In conclusion, morphological studies on different color morphs (red, yellow, white, brown, pink) with transparent and opaque tunics led us to the same species in the Mediterranean Sea: S. brakenhielmi. In this regard, Mastrototaro et al. (2019) have not found significant morphological or genetic differences between red and yellow colonies of S. brakenhielmi from the Western (Sardinia) and Central (Taranto) Mediterranean Sea, thus highlighting the possible synonymy between S. brakenhielmi and S. rubra by Automatic Barcode Gap Discovery (ABGD) species delimitation analysis. However, to confirm this, it is necessary to perform genetic analysis of our specimens and compare them with S. bahraini and S. rubra. Michaelsen (1918b) recorded S. brakenhielmi (as S. viride f. stuhlmanni) in the Gulf of Suez (Red Sea) in 1914. Later in 1924, Harant (1927 found it in the Suez Canal at El Katera, 46 km away from Port Said. Pérès (1958a) reported the species (as S. viride) in 1951 from the Mediterranean coast of Israel, together with Ascidia cannelata Oken, 1920, Phallusia nigra Savigny, 1816 and Hermania momus Savigny, 1816, solitary ascidians of Indo-Pacific origin. The same author (Pérès 1958b) also argued how these new findings were probably recent introductions from the Red Sea via the Suez Canal, since he had only found them in Israel, and not in Syria, Greece or Turkey.

Possible causes of spread in the Mediterranean Sea: routes and vectors
In the Levantine Sea, S. brakenhielmi has progressively spread northwards, following the trajectory of the prevailing surface coastal current (Figure 3; Hamad et al. 2005), and has become a common species on both artificial and natural habitats in this area (Bitar and Kouli-Bitar 2001;Çinar et al. 2006;Izquierdo-Muñoz et al. 2006Bitar et al. 2007;Shenkar and Loya 2009). In Cyprus it was found on artificial substrata in harbours in 2016 (Savva and Kleitou 2017;Ulman 2018;Ulman et al. 2019), although a study conducted in the same country in Larnaca Bay in June 2007 (UNEP-MAP RAC/SPA 2007) did not find the species; this may suggest a recent introduction mediated by shipping. Regarding the westward spread further along the coasts of Egypt, Halim and Abdel Messeih (2016) recorded the species in Alexandria in 1988-89 andin El Alamein, west of Alexandria, in 2013 (YRS pers. obs., present study), both within ports. All living colonies observed were the reddish or pink colour morphs ( Figure 1A).
The recent appearance of S. brakenhielmi along the Aegean coasts of Turkey (in 2015-2016 at Kiyikislacik andAkbük;MEÇ pers. obs. and Aydin-Onen 2018) and Greece (Rhodes and Heraklion;Ulman 2018;Ulman et al. 2019) might suggest a further expansion of the species within the Aegean Sea through the Eastern Mediterranean surface current. However, these observed colonies had white, brown and greenish-grey morphs (not red; Figure 1C-D) and were found in harbours and rocks near fish-farm cages, which could suggest either aquaculture and/or shipping as the main introduction vectors.
In the Ionian Sea, the first record was made in October 2008 (Monastir Marina, Tunisia; AR pers. obs.), followed by new records in October 2014 (present study). The species was observed on rocks and seagrass meadows (Cymodocea nodosa and Halophila stipulacea) around the harbour with white and red colour morphs. Studies carried out in 2005 and 2006 in Hammamet, 70 km north of Monastir (Chabbi et al. 2010(Chabbi et al. ), and 2009(Chabbi et al. -2010 in the Gulf of Gabes (Zarzis and Sfax harbours, Ramos-Esplá et al. 2011), did not find this species. In December 2014, the merchant-ship "Rochelle" sailing from Ghana to Turkey was stranded near the Kuriat Islands in front of Monastir, and was colonized by abundant colonies of S. brakenhielmi with different colours (red, pink, yellow, orange, cream, white; Figure 1) on the hull. Later, the species was recorded in other Tunisian localities further north (YRS pers. obs.), including the port of El Kantaoui (October 2015) and Goulette Canal (November 2015), all with red color morph. Monastir is an important marina for boat traffic, thus possibly representing a propagule seeding hub for adjacent ports. International maritime traffic by commercial vessels could be considered the most likely vector of introduction of S. brakenhielmi, e.g. specifically for Marsaxlokk (Malta), which represents one of the main cargo trans-shipment ports of the Mediterranean Sea; recreational boats could also be a vector (Ulman et al. 2017;Ulman 2018;Mastrototaro et al. 2019). Symplegma brakenhielmi was frequently recorded within several sampled marinas, and it was also found on recreational boat hulls in Turkey (Ulman et al. 2017), suggesting that recreational boating may be an important secondary vector in its rapid expansion through the intensely sailed Mediterranean Sea (Cappato 2011).
As previously mentioned, together with S. brakenhielmi, Pérès (1958a) also identified three other solitary Lessepsian ascidians: A. cannelata, P. nigra and H. momus. While the former ascidian remains localised to the south-eastern portion of the Levantine Sea (Izquierdo-Muñoz et al. 2009;Shenkar and Loya 2009), P. nigra and H. momus have formed proliferating populations further northward, following the same colonization dispersal pattern promoted by surface currents through the Levantine and Aegean Seas (Çinar et al. 2006;Kondylatos et al. 2010;Koutsogiannopoulos et al. 2012;Gerovasileiou and Issaris 2014;Kondylatos and Corsini-Foka 2017). At present, only H. momus has reached the Ionian Sea in Marsaxlokk, Malta (Evans et al. 2013), which is very close to the sampling site of S. brakenhielmi recorded in this study and was most likely introduced through maritime traffic.
Therefore, the colonial S. brakenhielmi appears to be spreading faster than these three solitary ascidians. The possible causes of this postulated expansion could be traced to the characteristics of its colonial strategy and larval development (Millar 1971). The rapid growth (as encrusting colonies) and short life cycle of this species are, in fact, competitive strategies for space (Jackson 1977). Symplegma brakenhielmi exhibits two-dimensional growth, and the larva develops inside the zooids (Berrill 1940), which is advantageous for the colonization of artificial structures within fouling communities and for growth as an epibiont (Green et al. 1983;Ramos-Esplá and Ros 1990;Rocha 1991;Dijkstra et al. 2007), whereas P. nigra and H. momus are oviparous ascidians, with less-specialised larva (Millar 1971).
Ascidians are one of the most frequent NIS in artificial habitats due to their rapid spread and population outbreaks (Lambert 2007;Zhan et al. 2015). Generally, ascidians can be associated with multiple vectors of introduction, including ballast water, rafting, hull fouling and aquaculture (Zhan et al. 2015). While the role of ballast water in the transport of ascidians remains unclear, several studies provide evidence for the colonisation success of ascidians on recreational boat hulls as biofouling components (e.g. Darbyson et al. 2009;Lambert 2002;Ulman et al. 2017), rafting colonies living attached to seagrass leaves (Worcester 1994), and in association with shell-fish aquaculture facilities (e.g. Davis et al. 2007;Ordóñez et al. 2015). Among the Symplegma genus, the three species recorded in two or more distant biogeographic regions (i.e., S. brakenhielmi, S. rubra and S. reptans; Table 1) are likely to have been associated with the biofouling and aquaculture vectors of introduction and spread (Rocha et al. 2009;Carman et al. 2011;Mastrototaro et al. 2019;Ulman et al. 2019). In addition, colonies of S. rubra have also been observed on free-floating algae and other drifting materials (Dias et al. 2006), thus could have been transported by rafting. Similarly, Symplegma species recorded in closer biogeographic regions (i.e. S. viride, S. bahraini and S. japonica; Table 1) may have locally spread through shipping or by natural means (Tamilselvi et al. 2011;Zhan et al. 2015).
Currently, regulations addressing the control of NIS only relate to aquaculture (EC 2007) and ballast water (Ballast Water Management Convention of the International Maritime Organization, IMO) vectors, while management of the biofouling vector is still unregulated and insufficiently assessed in the Mediterranean Sea. Specific management guidelines for this vector were previously formulated (IMO 2011), and an ad-hoc project (GEF-UNDP-IMO GloFouling Project) was launched in December 2018 in order to address the management of biofouling. The continuous monitoring programmes in high-risk areas of NIS introduction (e.g. ports and aquaculture facilities) and the use of preventive measures to limit NIS spread are recommended best practices against marine bioinvasions (Lehtiniemi et al. 2015). The Mediterranean Sea brings together many national jurisdictions, placing the onus for the effective management of new introductions and their related vectors on a basin-wide early detection and warning system. Finally, the dissemination of taxonomic knowledge and the elucidation of ambiguous taxonomy are essential for an accurate species identification, particularly urgent for the class Ascidiacea (Shenkar and Swalla 2011;Zhan et al. 2015).