In recent years, many marine organisms have been introduced into new habitats by human-mediated transport, mostly by fouling of the hulls of ships or other submerged surfaces (e.g. anchors, buoys and nets; Carlton and Hodder 1995, Gollasch 2002, Floerl et al. 2009) and through transfer in ballast waters (Carlton 1985, Carlton and Geller 1993, Gollasch et al. 2000). To date, the maritime traffic across the oceans remains the most significant factor contributing to the current patterns of biogeographic distribution of many marine organisms worldwide (Carlton 1985, 1987). Many different taxa, such as algae, sponges, molluscs, polychaetes, crustaceans, bryozoans, ascidians and fish, have spread as part of the fouling community or as passengers in ballast water tanks due to their ability to survive for several weeks during transoceanic journeys (Gollasch et al. 2000).

Bryozoans are species that are particularly likely to be transported by such means because of their tendency to grow on a great variety of substrates, which makes them a major component of marine benthic hard-bottom communities worldwide (Woollacott and Zimmer 1977). In particular, cheilostome bryozoans are known to be easily dispersed as fouling organisms to non-native regions (Watts et al. 1998). Furthermore, some of them possess a planktotrophic larva that may have a potential planktonic life of up to several weeks, so their transport by ballast waters is also considered (Carlton 1985, Carlton and Geller 1993, Watts et al. 1998).

Celleporaria brunnea (Hincks, 1884) is an ascophoran bryozoan of northeastern Pacific origin, first described by Hincks (1884, as Cellepora brunnea) in British Columbia (Canada). The species is widely distributed in the Pacific Ocean; in fact it was reported in the vicinity of the Panama Canal, in the Galapagos Islands (Hastings 1929) and Ecuador (Osburn 1952, Soule et al. 1995). Furthermore, C. brunnea is one of the most abundant species in the Gulf of California (Soule 1961, Soule and Soule 1964). Recently, the species was also recorded in the Hawaiian Islands, where it may have arrived from the Californian coasts (Godwin 2003). In 2004, C. brunnea was discovered on aquaculture steel cages in southern Korea, where it was considered to have been introduced from the eastern Pacific Ocean (Seo and Min 2009). In the Mediterranean Sea, C. brunnea was recorded for the first time in 2004 near Alsancak harbour, in the inner part of Izmir Bay (Turkey, Aegean Sea), fouling on Mytilus galloprovincialis Lamarck, 1819 (Koçak 2007). In addition, Harmelin et al. (2009) reported the presence of the species in the coastal waters of Lebanon, on the basis of preserved material collected between the late 1990s and the early 2000s. Recently, a recent revision of non-indigenous bryozoans in the eastern Mediterranean Sea (Harmelin 2014) assigned to Celleporaria sp. aff. brunnea the specimens collected in Beirut (Lebanon) in 2003. In 2012, C. brunnea was detected in Cascais marina, located west of Estoril in the greater Lisbon area, Portugal, representing the first record of the species in the eastern Atlantic region (Canning-Clode et al. 2013). Furthermore, Harmelin (2014) assigned to C. brunnea specimens of Celleporaria aperta (Hincks, 1882) recorded by André et al. (2014) in the Arcachon Basin on the Atlantic coast of France.

Besides C. brunnea, six other species of Pacific origin belonging to the Celleporaria genus have been reported with scattered distribution in the Mediterranean Sea: Celleporaria aperta, Celleporaria fusca (Busk, 1854), Celleporaria labelligera Harmer, 1957, Celleporaria pilaefera (Canu and Bassler, 1929), Celleporaria sherryae Winston, 2005 and Celleporaria vermiformis (Waters, 1909). C. aperta was first recorded in the Mediterranean along the coasts of Israel, where it was considered to have been introduced through the Suez Canal (Powell 1969, d’Hondt 1988). In the mid-1970s, the species was reported, together with C. pilaefera, on fouling cages of an oyster farm on the island of Malta, where both species were most probably introduced via shipping (Agius et al. 1977). Recently, these latter two records were considered “questionable” by Sciberras and Schembri (2007) because there are no other records of these species from Malta in the literature and it is not known whether populations of both species still persist. C. fusca was reported for the first time in the Mediterranean at Haifa Bay (d’Hondt 1988), on the basis of preserved material collected in 1974, but no recent records have been reported (Zenetos et al. 2012). C. labelligera, C. sherryae and C. vermiformis were recently reported along the Lebanese coasts by Harmelin (2014), the former species being very common in Beirut and Tripoli harbours.

The present study is part of a wider systematic survey aimed at studying non-indigenous species in Italian commercial harbours (Lodola 2013). It documents the occurrence of C. brunnea in the Italian ports of La Spezia, Olbia and Lampedusa, located in the Ligurian Sea, western-central Tyrrhenian Sea, and Sicilian Channel, respectively. These findings represent the first records of this species in Italian waters and its western- and northernmost occurrence within the Mediterranean Sea. A taxonomic comparison with previous records of C. brunnea in the Mediterranean Sea and the Atlantic and Pacific Oceans is also provided. Furthermore, we discuss the most likely vectors of introduction of C. brunnea into the western Mediterranean basin.

A systematic survey was carried out in two consecutive summers (2010 and 2011) in the Italian commercial harbours of La Spezia, Olbia and Lampedusa (Fig. 1, Table 1) to detect the presence of non-indigenous species in port environments. La Spezia and Olbia are two large harbours, both characterized by international trade shipping, recreational boating and aquaculture activities (mainly mariculture of the bivalves Mytilus galloprovincialis and Ruditapes philippinarum Adams and Reeve, 1850). La Spezia is located on the eastern coast of the Ligurian Sea, between Genoa and Leghorn, while Olbia is situated on the northeast coast of Sardinia and faces the western-central Tyrrhenian Sea. Lampedusa is a small island, belonging to the Pelagian Archipelago, located in the middle of the Sicilian Channel; its harbour hosts national trade ships, local fishermen and touristic boats, and military facilities. In each harbour (Fig. 2), five sampling sites were selected within a main dock: two sampling sites were identified for residential mooring (code “P”) and three for transit mooring (code “TR”). The biological sampling consisted of scraping the artificial hard substratum (the concrete dock walls) just under the low tide level, with a hand-held rigid net (1 mm of mesh size, surface 25×20 cm) using a semi-quantitative methodology, standardized for hard-bottom macrobenthos collections (Occhipinti-Ambrogi 2000, Marchini et al. 2004, Minchin 2007). For each sampling site, five replicates were collected in late spring/early summer, when Mediterranean benthic communities are known to exhibit their maximum development, and in late summer, just before the autumn crisis, when benthic communities decrease in biomass until the next recruitment season (Boero 1994). Because of the colonial nature of most components of the benthic fauna, the fouling community was evaluated using a semi-quantitative methodology, with a four class abundance index (ab. index) depending on coverage: +, presence of isolated colonies; ++, a few colonies; +++, well-established colonies, and ++++, overwhelming abundance (Occhipinti-Ambrogi 1991). This approach, already tested in other studies on the artificial hard-bottom communities (Marchini et al. 2004, Savini et al. 2006, Lodola et al. 2012) allowed us to compare the abundance of both vagile individual and colonial sessile components of the benthic fauna, whose quantitative estimation is not possible by enumeration of individuals.

Morphological measurements of some parts of zooids were made using both the microscope camera AxioCam ERc 5s of the AxioVision Carl Zeiss Software for image analysis (accuracy 0.01 μm), and the Tescan FESEM (Field Emission Scanning Electron Microscope) series Mira 3XMU for the SEM pictures, with increasing magnification, at 6-19 mm working distance, using an accelerating voltage of 10 kV, with graphite metallization and detection by secondary electrons. Specimens used for SEM pictures were previously bleached to remove all organic residues.

Celleporaria brunnea colonies found in our samples were well-developed, forming large coarse incrustations and irregular masses, bearing ovicells, thus confirming that the species was well-acclimated and reproducing. The presence in three distant localities of the western Mediterranean Sea of established populations, whose abundance increased during the two years of observation, indicates the probability of further records in this area.

In the Mediterranean Sea and Atlantic Ocean, C. brunnea was reported from harbours and marinas, showing a preference for both natural (e.g. mussel beds and seaweeds; Koçak 2007) and artificial hard substrates (e.g. docks and PVC plates; Canning-Clode et al. 2013, present study). Port systems are known to be highly disturbed environments subject to high propagule pressure of non-indigenous species, due to the presence of a variety of anthropogenic pathways (Smith et al. 1999, Ruiz et al. 2000, Occhipinti-Ambrogi and Savini 2003). Amongst these, shipping activity, for both commercial and tourism purposes, facilitates the introduction in new regions and the subsequent spread of non-indigenous organisms fouling on the hulls of ships or transported in ballast waters or sea chests (Carlton 1985, 1987, Carlton and Geller 1993). Like many marine sessile organisms, bryozoans are capable of using a variety of potential dispersal mechanisms, as they demonstrate great catholicity in their exploitation of substrates. Man-made installations located in major shipping ports and centres for pleasure craft can provide a wide and extensive range of hard surfaces for attachment that include breakwaters, pontoons and pilings, which constitute a suitable habitat for sessile organisms such as bryozoans (Glasby and Connell 1999, Connell 2000, 2001). C. brunnea was found in the mooring area of the Italian harbour of La Spezia, Olbia and Lampedusa, in the Portuguese marina of Cascais (Canning-Clode et al. 2013) and on artificial substrates in Turkey (Koçak 2007) and Lebanon (Harmelin 2014). Ship fouling, in particular, could provide a very rapid, reliable means of dispersal and is likely to be an effective vector for both local and global scales (Watts et al. 1998). Thus, hull fouling is the most likely vector of introduction and spread of C. brunnea in the Mediterranean Sea and the Atlantic Ocean. Similarly, Godwin (2003) suggested that the species may have been secondarily dispersed to the Hawaiian Islands via hull fouling of ships travelling from the Gulf of California, where C. brunnea was already known to occur. Although such dispersal will mostly be restricted to ports, other transport mechanisms may also increase the geographic range of bryozoan species beyond these points of entry, including rafting on floating materials (e.g. marine debris of anthropogenic origin, such as floating plastics) and drifting seaweeds (Watts et al. 1998, Barnes 2002); in recent years, marine debris has increased significantly, and this floating material is used as a ‘hitchhiking’ substrate for mature individuals of different taxa, bryozoans included (Barnes 2002, Aliani and Molcard 2003).

In addition, transport by ballast water has to be considered as a potential dispersal mechanism, since bryozoans are able to foul the inside of ballast tanks (Carlton 1985). In fact, larval forms and development traits are crucial in contributing to dispersal and habitat exploitation of marine invertebrates (Scheltema 1986a,b, Roughgarden et al. 1988). For instance, larvae of the cheilostome bryozoan Membranipora membranacea (Linnaeus, 1767) were found in ballast water (Carlton and Geller 1993), but the dispersal possibilities of this bryozoan is enhanced due to the long pelagic duration of its planktotrophic (cyphonautes) larva (Ryland 1964, Yoshioka 1982, Stricker et al. 1988). Conversely, C. brunnea has a lecithotrophic (coronate) larva with a wide area of coronal cells for ciliary propulsion and a short pelagic duration (Santagata and Zimmer 2000, Santagata 2008a,b). The low dispersal potential of C. brunnea larvae is thus not very effective in long-range transportation of the species, e.g. in ship ballast waters. The short pelagic duration of C. brunnea larvae, which can settle within a few hours of their release, means that they may not be able to disperse over long distances and their transfer in ballast waters can be considered effective just for short-range transportation. Furthermore, C. brunnea was also found on aquaculture steel cages in Korea (Seo and Min 2009); this vector has to be taken into account in the light of the records from Arcachon Basin (André et al. 2014), Olbia and La Spezia (present work), all of which are intense mariculture activity sites. The transfer as non-target species associated with mussels and clams for aquaculture purposes cannot be excluded as a potential vector of introduction, albeit a less effective one.

Current distribution and record dates of C. brunnea in the Mediterranean Sea might suggest a step-by-step anthropogenic route of invasion: firstly, long-distance introduction from the Pacific, most probably through shipping via the Suez Canal, to the Levantine basin, including the Lebanese and Turkish findings, followed by a stepping stone occurrence in the Sicilian Channel (Lampedusa harbour). This fact might have contributed to the further spread of the species northwards, up to the Tyrrhenian (Olbia harbour) and the Ligurian Sea (La Spezia harbour), as well as eastwards towards the Atlantic coast of Portugal (Cascais marina) and France (Arcachon Basin). The presence of C. brunnea in the polluted inner part of Izmir Bay near Alsancak harbour (Turkey) suggests that the most likely pathway of the species’ introduction in the eastern Mediterranean Sea is the hull fouling community of ships (Koçak 2007) originating from the Pacific through the Suez Canal. This is also the most likely explanation for C. brunnea occurrence along the Lebanese coast (Harmelin et al. 2009), as the location of the Levantine Basin “downstream” of the opening of the Suez Canal facilitates Indo-Pacific biota in becoming established first along the eastern Mediterranean coasts (Galil 2009). From there, C. brunnea might have been subsequently transferred and spread throughout the Mediterranean by carriers of different types (e.g. vessels and cruise liners). These operate all over the basin as far east as the Aegean and Levantine Seas, as far north as the Tyrrhenian and the Ligurian Sea and eastwards up to the Alboran Sea towards the northern Atlantic Ocean (www.biodivalue.eu; Accessed October 2014).

However, the taxonomic comparison between our morphometric measurements and literature data allowed us to assert a greater similarity with the Pacific specimens of C. brunnea, rather than with the Mediterranean Sea or Atlantic Ocean ones. Given these circumstances, an alternative hypothesis for the introduction route of this bryozoan to the western Mediterranean Sea appears to be more likely than the first one. Firstly, C. brunnea might have reached La Spezia harbour by means of cargo ships which travel regularly from the Pacific to the Ligurian Sea and back (www.porto.laspezia.it; accessed October 2014). In this case, La Spezia harbour would represent the first event of introduction into the western Mediterranean basin and the primary source for further events of secondary spread within the whole basin. While big tankers and merchant ships originating from the Pacific could have introduced the first propagules of C. brunnea into La Spezia waters, coaster and pleasure boats could have contributed to the secondary spread of the species southwards to Olbia harbour in the western-central Tyrrhenian Sea and to Lampedusa Island in the Sicilian Channel. Indeed, these localities are situated in areas with extremely heavy tourist traffic and feature busy marinas and moorings for pleasure boats and fast ferries; recreational boating was recently accepted as an important secondary transmission vector of non-indigenous species worldwide (Johnson et al. 2001, Floerl and Inglis 2003, Davidson et al. 2010, Clarke Murray et al. 2011). Because of the morphometric differences of the eastern Mediterranean Sea and Atlantic Ocean specimens, the occurrence of C. brunnea there has to be considered the result of multiple introduction events, possibly unrelated to the events of primary and secondary introduction in the western Mediterranean basin.

In La Spezia harbour, C. brunnea is the most frequent and abundant bryozoan, along with other bryozoan species (Table 4): the cosmopolitan Bugula neritina (Linnaeus, 1758), Bowerbankia gracilis Leidy, 1855, Cryptosula pallasiana (Moll, 1803), Watersipora sp. Neviani, 1896, the pseudo-indigenous (sensu Carlton 2009) Zoobotryon verticillatum (Delle Chiaje, 1822) and the non-indigenous Tricellaria inopinata d’Hondt and Occhipinti-Ambrogi, 1985. Table 4 shows that, when there was a high abundance of C. brunnea in Olbia harbour, other encrusting bryozoans such as Conopeum seurati (Canu, 1928) and C. pallasiana were less abundant or absent, and vice versa. In Lampedusa harbour, the most common assemblage of bryozoans included B. neritina, Cradoscrupocellaria bertholletii (Audouin, 1826), Watersipora sp. and Z. verticillatum together with C. brunnea, which generally displayed higher abundances in the summer of 2011 than in the summer of 2010. Furthermore, C. brunnea prevailed in samples collected at the end of the summer season in the three sampling locations during both years, while it was less abundant and frequent in samples collected in spring (Table 4). Unfortunately, the lack of previous bryozoan collection in the three Italian harbours makes it difficult to precisely assess the invasion history of C. brunnea in the western Mediterranean Sea and to make inferences about its effects on the native bryozoan community. The time of introduction of C. brunnea to La Spezia, Olbia and Lampedusa is unknown and the species might have remained unnoticed for a long time. Notwithstanding this, C. brunnea has been found to be widespread and abundant in the three investigated harbours since 2010, therefore suggesting that its introduction and acclimation in these areas can be definitely backdated.

The differences in taxonomic characteristics of specimens found in different areas may indicate multiple introduction events, and not just a single event of primary introduction followed by secondary spread. However, only a taxonomic review of the Celleporaria genus and genetic and taxonomic analysis of C. brunnea specimens, together with a comparison with Hincks’s type (Hincks 1884), would allow us to reconstruct the historical introduction events of C. brunnea, and to determine whether the morphological differences between specimens are to be assigned to a complex of species or a variation within the C. brunnea species.

The species belonging to the Celleporaria genus are known to be highly opportunistic and invasive (Dunstan and Johnson 2004), and their occurrence in harbours or marinas may indicate a tolerance capacity to thrive in polluted water conditions (Harmelin 2014). Thus, C. brunnea and the other congeneric species may be expected to occur in other localities throughout the Mediterranean Sea in the years to come.