First inventory of the shallow-water benthic hydrozoan assemblages of Gökçeada Island (northern Aegean Sea)

Abstract

The hydroid fauna of the Mediterranean Sea is considered one of the best known in the world, but the hydrozoans of the Aegean Sea remain poorly studied, hindering efforts to identify alien and invasive species in the region. The spatial and seasonal composition of the shallow-water (0–20 m depth) benthic hydrozoan assemblage from Gökçeada Island was investigated in summer 2012 and winter 2013. Overall, 48 hydrozoan taxa were identified, and their presence and ecological features are discussed herein. Twelve species are recorded for the first time in the Aegean Sea, and the same number for the Turkish coasts. Differences in species composition were detected between the northern and southern coasts of Gökçeada by cluster, multidimensional scaling (MDS) and permutational multivariate analysis of variance (PERMANOVA) analysis, whereas seasonal and vertical distribution patterns were not statistically significant. Differences in species richness and composition between the northern and southern coasts may be explained by the distinct geomorphological aspects of the shores, providing a spatial heterogeneity in the availability of substrates for the hydroid colonies. Observed differences are attributable to the occurrence and/or abundance of common species such as Sertularella ellisii, Aglaophenia tubiformis, Clytia hemisphaerica, Clytia linearis, Eudendrium racemosum, Plumularia obliqua, Eudendrium capillare, Turritopsis dorhnii and Dynamena disticha, rather than to the presence of rare, exclusive species.

Keywords:

• Hydrozoa
• biodiversity
• eastern Mediterranean Sea
• Aegean Sea
• Gökçeada Island

Introduction

The Mediterranean Sea, a biodiversity hotspot with approximately 17,000 marine species so far described, is renowned as “extremely well” studied (Coll et al. 2010), and its Hydrozoa fauna (siphonophores excluded) can be considered one of the best known in the world, with a total of 400 recorded species (Gravili et al. 2013). Because of their high reproductive potential (Boero et al. 2002), hydroids represent a key component of shallow-water communities, especially on hard bottoms. Nonetheless, knowledge of these organisms is not uniform throughout the entire Mediterranean Sea, because marine research efforts have historically been concentrated in the proximity of main oceanographic stations along the Western basin and the Adriatic Sea. Therefore, several areas of the Mediterranean, such as the African coasts and the eastern basin, still remain relatively undersampled (Boero et al. 1997). Recent studies have aimed to reduce the existing gap of information (e. g. Morri et al. 2009; Soto-Àngel & Peña-Cantero 2013), but much remains to be done in the Aegean Sea and, more generally, in the Levantine basin (Morri & Bianchi 1999).

The available information on the hydroids of the Aegean Sea and neighboring regions is summarized in the papers of Yamada (1965), Marinopoulos (1979), and Morri and Bianchi (1999). Other records of hydrozoans in the basin come from the works of Morri et al. (1999), Cocito et al. (2000) and Bianchi et al. (2011) in the hydrothermal vents of the Greek island of Milos, and of Rayyan et al. (2002) and Rayyan et al. (2004) in the northern (mainly Greek) part of the Aegean. The Turkish section of this sea, on the other hand, is relatively unexplored, and its hydrozoan fauna virtually unknown. Ninety hydrozoan species have been reported from the Turkish coasts so far, 63 of them from the Aegean Sea (Kocataş 1976; Marinopoulos 1979; Ünsal 1981; Marques et al. 2000; Gülşahin et al. 2013), and 42 from the Sea of Marmara (Demir 1954; Ünsal 1981; Marques et al. 2000; Isinibilir et al. 2010), with only 10 species of hydroids reported from Gökçeada Island (Ünsal 1981). Furthermore, an alien hydropolyp species (Macrorhynchia philippina) was recently recorded from the Levantine coasts of Turkey (Çinar et al. 2006).

The large Turkish island of Gökçeada is located at the core of the above-mentioned undersampled region of the northern Aegean Sea. The complex geomorphology of the island (Koral et al. 2009) is encircled by a coastal surface circulation driven by the anticyclone gyre, associated with the Dardanelles low-salinity water input (Olson et al. 2007). Gökçeada has a strategic position within the Aegean archipelago, mirroring the role of the Aegean Sea as a biogeographical crossroads between the Marmara Sea (and eventually the Black Sea) and the greater Mediterranean basin, as observed for a number of marine and terrestrial taxa (Ergüven et al. 1988; Topaloğlu 2001; Isinibilir & Tarkan 2002; Ateş et al. 2006; Karakulak et al. 2006; Aslan-Cihangir 2012). Given the above, the present study represents the first inventory of benthic hydrozoans (sub-class Hydroidolina, sensu Marques & Collins 2004), from shallow waters (0–20 m depth) at Gökçeada Island, with an assessment of spatial and seasonal ranges of distribution.

Material and methods

Study area

The island of Gökçeada, with a coastline of 92 km and a surface area of 279 km², is located about 20 km off the Gelibolu Peninsula, in the Aegean Sea, between 40° 05’ 12’’–40° 14’ 18’’ N and 25°40’06’’–26°02’05’’ E. On the northern coast of the island, the continental slope is steep and the Saroz Pit, a tectonic trench, is located immediately off the coast, creating a narrow continental shelf that rapidly gains depth and that is characterized mainly by a rocky substratum. The southern coast, on the other hand, has wide sandy beaches and a wider continental shelf with a slowly deepening, soft substratum. Particularly in summer, the wind circulation around Gökçeada is dominated by northeasterly and southwesterly winds (Kocataş & Bilecik 1992). Furthermore, due to the presence of numerous close islands, the water movements in the northern Aegean Sea in the vicinities of Gökçeada vary greatly; the Black Sea waters from the Dardanelles Strait, for example, have an orientation towards the north in winter and towards the south in summer (Kocataş & Bilecik 1992). The area is enriched by nutrients of the Meriç River and has a variety of pelagic fish feeding on the abundant phytoplankton (Ulutürk 1987). The water mass surrounding the island is affected by cooler and less salty water coming from the Sea of Marmara (Kocataş & Bilecik 1992).

Twelve sampling stations were selected around Gökçeada Island; eight (2-Kanyer [KY], 3-Kuzulimanı [KZ], 5-Şeytankale [ŞE], 6-Mavikoy [MA], 7-Yıldızkoy [YI], 10-Kaleköy [KL], 11-Kaşkaval [KŞ] and 12-Pirinç (Kuş Burnu) [KU]) were located on the northern side while four (1-Avlaka [AV], 4-Aydıncık [AY], 8-Laz koyu [LA] and 9-Uğurlu [UG]) were positioned on the southern side of the island (Figure 1). Stations 1, 11 and 12 are pristine sites with no access from land. Stations 6 and 7 are located within a Marine Protected Area, while 7 hosts thousands of tourists that reside elsewhere on the island but come to Yıldızkoy for clean beaches in the summer. Station 8 suffers similar tourism pressure while stations 2, 3, 4 and 9 are influenced by a somewhat lower (but still consistent) number of tourists in the same season. Additionally, station 7 suffers from sand dumping for recreational purposes by the local municipality before the summer season, and the sand is withdrawn into the sea in autumn, causing temporary turbidity and covering sessile fauna. Station 3 is located in the main harbor of the island and includes samples from both inside and outside the pier. Station 9 is located at a pier with low use by local fishermen. Station 10 is heavily influenced by the nearby human settlement of Kaleköy, and is the only location on the island where sewage discharge occurs.

Figure 1. Map of the study area. Stations 1–6 and 9–12 were sampled in July; stations 9–12 (in squares) were also sampled in February; 7 and 8 (in circles) were sampled only in February.

Sampling and analysis

Sampling took place at the 12 selected stations in either summer (July 2012) or winter (February 2013), or both seasons, depending upon the prevailing climatic conditions (Figure 1). All sampling stations were characterized by a substrate composed of a mixture of rocks, sand and Posidonia meadows. At each station, benthic hydroid colonies from the bottom and fragments of diverse substrates supporting or susceptible to support hydrozoan colonies (e.g. algae, sea-grasses, invertebrates, etc.) were collected by SCUBA diving in each of four depth zones (0–5 m, 5–10 m, 10–15 m and 15–20 m) along a vertical transect. For the purposes of a faunistic inventory, a visually oriented collection was carried out (Piraino et al. 2013). Trained divers selectively picked up only hydroid colonies or their potential substrates from a single, larger homogeneous bottom belt (30 cm height × 1000 cm length = 3000 cm²) at one intermediate depth within each depth zone, whenever available substrates offered a sampling opportunity.

Specimens were sorted in the laboratory and anesthetized using 7% magnesium chloride solution in seawater before being observed alive under a binocular stereomicroscope for primary identification. Hydroids were fixed in 4% formaldehyde a few hours after sampling in order to preserve them for ultimate confirmation of their taxonomic identity under both low- and high-power microscopes. When possible, subsamples were fixed in 90% ethanol to keep a tissue bank for future genetic studies. Nematocysts were examined either on live specimens or from formalin-preserved material, with the aid of a light microscope. Biological samples were deposited at the Hydrozoa Collection of the University of Istanbul (HCUI). Due to the benthic target and adopted sampling, representatives of orders Actinulida, Limnomedusae, Narcomedusae and Trachymedusae (collectively grouped also in the subclass Trachylinae) were not included in our investigation.

Morphological description, updates on synonymy and distribution at the Mediterranean scale are not dealt with in detail since these issues have been recently covered elsewhere for all of the hydroid species found in this study (e. g. Peña-Cantero & García-Carrascosa 2002; Bouillon et al. 2004; Schuchert 2004, 2006, 2007, 2008a, 2008b, 2010).

There is still a lack of consensus concerning the higher classification of the class Hydrozoa and alternative names are still in use for suprafamily taxa, especially at subclass and order levels (see Bouillon et al. 2006; Cartwright & Collins 2007; Schuchert 2014a). The classification given in Bouillon et al. (2006) is mainly adopted here, although corresponding nomenclature is often provided as from the online World Hydrozoa Database (Schuchert 2014b).

Cluster and ordination analysis were carried out using the presence/absence data of taxa in order to identify spatiotemporal patterns of species composition in the study area. The selected similarity index was that of Sørensen (1948). The generated similarity matrix was used to obtain a cluster diagram through the hierarchical grouping method and a multidimensional scaling (MDS) representation in order to identify the existing relationships among sampling stations based on biological information. The SIMPROF (similarity profile) test was used in association with the cluster analysis (Clarke et al. 2008) in order to test the null hypothesis of no meaningful structure within groups of samples (average of 1000 permutations and 999 simulations, α = 0.05).

Differences in the hydrozoan assemblages according to the sampled region of the island (south vs. north), season (summer vs. winter) and depth of collection (0–5 m, 5–10 m, 10–15 m and 15–20 m) were tested using a permutational multivariate analysis of variance (PERMANOVA, Anderson 2001) in which each term was tested using 9999 random permutations, while a similarity percentage analysis (SIMPER) was performed to determine the contribution of each taxon to the dissimilarities between groups obtained from the cluster and MDS. Our experimental design included three factors: (1) “exposure” (fixed, two levels, north and south); (2) “season” (fixed, two levels, summer and winter); (3) “depth” (fixed, four levels, 0–5 m, 5–10 m, 10–15 m and 15–20 m). All of the above analyses were carried out with the Primer v. 6 software (Clarke & Gorley 2006).

Results and discussion

Forty-eight hydrozoan (Hydroidolina) taxa (Table I) were found along the coasts of Gökçeada, out of which 12 represented first records for the Aegean Sea, and 13 for the Turkish coastal area. Overall, we collected 19 representatives of the Athecata/Anthomedusae and 29 of the Thecata/Leptomedusae (also synonymized as orders Anthoathecata and Leptothecata, respectively, by Cornelius 1992). Twenty-seven species were collected from both the northern and southern coasts, whereas 21 species were recorded only from one of the two coasts, 15 from the northern shores and six from the southern shores of Gökçeada.

Table I. List of Hydrozoa taxa recorded in the present study on Gökçeada Island. Species marked with “ + ” are first records for the Aegean Sea, and those marked with “#” are first records for the Turkish coasts. Previous records of the species in the Turkish Aegean coasts, Greek Aegean Sea, Marmara Sea and the Levantine Basin are indicated with: TA, A, MS and L, respectively (references for these records are Demir 1954; Kocataş 1976; Marinopoulos 1979; Ünsal 1981; Svoboda & Cornelius 1991; Morri & Bianchi 1999; Marques et al. 2000; Morri et al. 2009). Acronyms of biogeographic affinity are as follows: C, cosmopolitan; ME, Mediterranean endemic; B, boreal; I, Indo-Pacific; CT, circumtropical; AM, Atlanto-Mediterranean; TA, tropical Atlantic. Acronyms of seasonality and reproductive status: ^R, fertile; J, June; F, February.

Species	Previously recorded from	Substratum	Depth (m)	Biogeographical affinity	Seasonality/reproductive status	Stations
Bougainvilliidae
Bougainvillia muscus (Allman, 1863)	MS; A	Bioconcretions, rocks	0–23 m	CT	J^R, F^R	1, 10, 11, 12
Bougainvilliidae sp.		Bioconcretions	15–20 m	–	J	10
Cytaedidae
Perarella schneideri (Motz-Kossowska, 1905)	+ #	Schizoporella longirostris	0–10 m	ME	J, F	2, 3, 6, 10, 11
Eudendriidae
Eudendrium album Nutting, 1898	+ # L	Bioconcretions, rocks	1–15 m	ME	J	1
Eudendrium armatum Tichomiroff, 1890	# A	Rocks	0–20 m	ME	J, F	9, 11
Eudendrium capillare Alder, 1856	MS; TA; A	Algae, bioconcretions, rocks	0–21 m	C	J, F	1, 4, 6, 9, 10, 11, 12
Eudendrium glomeratum Picard, 1952	TA; A; L	Bioconcretions, algae	10–25 m	CT	F	10
Eudendrium cf. merulum Watson, 1985	L	Rocks, bioconcretions, barnacles	0–20 m	CT	J, F	1, 2, 6, 10, 11, 12,
Eudendrium racemosum (Cavolini, 1785)	TA; A; L	Rocks, bioconcretions, coralline algae	0–25 m	C	J^R, F	1, 2, 4, 8, 9, 10, 11,
Eudendrium ramosum (Linnaeus, 1758)	TA; MS; A	Rocks, bioconcretions	15–20 m	C	J^R	4, 11
Hydractiniidae
Hydractinia aculeata (Wagner, 1833)	+ #	Gastropod shell	10–15m	ME	J^R	1
Stylactaria fucicola (M. Sars, 1857)	TA	Bioconcretions, barnacles	5–20m	AM	F	10, 11
Oceaniidae
Turritopsis dohrnii (Weissmann, 1883)	+ # L	Bioconcretions, polychaete tubes	0–20 m	C	J^R	1, 2, 6, 9, 10, 11, 12
Pandeidae
Amphinema rugosum (Mayer, 1900)	+ # L	Algae, barnacles	1–20 m	CT	J	1, 11
Cladocorynidae
Cladocoryne floccosa Rotch, 1871	TA; A	Algae, hydrorhiza of E. racemosum	0–25 m	CT	J^R, F	2, 9, 10, 11
Corynidae
Corynidae sp.		Mytilus sp.	0.5 m	–	J	5
Slabberia halterata Forbes, 1846	+ #	Sponge	15–20 m	CT	F	10, 12
Tubulariidae
Ectopleura larynx (Ellis & Solander, 1786)	MS; TA	Rocks	5–10 m	B	J	2
Zancleidae
Zanclea sessilis (Gosse, 1853)	+ #	Remains of bryozoans	3 m	AM	J^R	4
Aglaopheniidae
Aglaophenia elongata Meneghini, 1845	TA; A; L	Bioconcretions	20–25 m	AM	F	10
Aglaophenia harpago Schenck, 1965	+ #	Posidonia leaves	0–20 m	ME	J, F	2, 8, 9, 10, 12
Aglaophenia octodonta (Heller, 1868)	TA; A; L	Bioconcretions, Cystoseira	0–5 m	AM	J, F^R	4, 12
Aglaophenia picardi Svoboda, 1979	A; L	Algae, rocks	2–15 m	AM	J	1,2,3,6,
Aglaophenia tubiformis Marktanner-Turneretscher, 1890	# A	Rocks, bioconcretions, algae, sponges, bryozoans, barnacles	0–23 m	AM	J^R, F^R	1, 6, 7, 8, 9, 10, 11, 12
Campanulinidae
Lafoeina tenuis Sars, 1874	+ # L	A. octodonta, E. racemosum, D. disticha	0.5 m and 15–25 m	B	J, F	4, 11, 12
Haleciidae
Halecium lankesteri (Bourne, 1890)	TA; A; L	Algae (Cystoseira sp.)	15–25 m	TA	J, F	10, 11
Halecium mediterraneum Weismann, 1883	TA; A	Rocks, algae	5–23 m	AM	J^R	2, 12
Halecium pusillum Sars, 1856	TA; A; L	Algae, remains of bryozoans	0–15 m	CT	J	2, 6, 9
Halecium tenellum Hincks, 1861	+ #	Mytilus sp.	0.5  m	C	J	5
Halopterididae
Antennella secundaria (Gmelin, 1791)	TA; A	Algae, polychaete tube, bioconcretions	15–25 m	C	J^R, F	2, 4, 6, 10, 11, 12
Halopteris diaphana (Heller, 1868)	TA; A; L	Algae, bioconcretions	0–20 m	CT	J^R, F	1, 10, 11
Hebellidae
Anthohebella parasitica (Ciamician, 1880)	TA; L	A. tubiformis, S. ellisi	0–10 m	CT	J^R, F	2, 11, 12
Kirchenpaueriidae
Kirchenpaueria pinnata (Linnaeus, 1758)	TA; A	Algae (including coralline algae), bioconcretions	5–21 m	AM	J, F	1, 4, 6, 9, 10
Kirchenpaueria halecioides (Alder, 1859)	MS; TA; L	Rocks, bioconcretions	1–10 m	CT	J^R	1, 10
Lafoeidae
Filellum cf. serpens (Hassall, 1848)	+ MS	Polychaete tubes, hydrorhiza of A. tubiformis, hydrorhiza and basal caulus of S. ellisii	5–32 m	C	J, F	1, 10, 11
Plumulariidae
Plumularia obliqua (Johnston, 1847)	TA; A	Algae (Cystoseira sp.), bioconcretions	0–20 m	I	J, F	7, 10, 11, 12
Plumularia setacea (Linnaeus, 1758)	TA; A; L	Bioconcretions, bivalve shell, coralline algae	0–5 m	C	J, F	2, 10
Sertulariidae
Dynamena disticha (Bosc, 1802)	TA; A	Algae, Posidonia leaves, bioconcretions	0–20 m	CT	J^R, F^R	1, 2, 7, 8, 10, 11, 12
Sertularella ellisii (Deshayes & Milne Edwards, 1836)	TA; A; L	Bioconcretions, algae, rocks, remains of bryozoans, cirripedes	0–25 m	AM	J^R, F^R	1, 2, 4, 6, 8, 9, 10, 11, 12
Sertularella polyzonias (Linnaeus, 1758)	MS; TA; A; L	Algae	15–21 m	C	J	4
Sertularia perpusilla Stechow, 1919	TA; A	Posidonia leaves	5–10 m	M	J	2, 10
Campanulariidae
Campanularia hincksii Alder, 1856	L	Bioconcretions, coralline algae, polychaete tubes	10–15 m	C	J	1, 4
Campanularia volubilis (Linnaeus, 1758)	+ #	Hydrorhiza of A. tubiformis, Posidonia leaves, algae	0–10 m	B	F	8, 9, 10
Clytia hemisphaerica (Linnaeus, 1767)	MS; TA; A; L	Bioconcretions, coralline algae, hydrorhiza of A. tubiformis, S. ellisii and E. racemosum, remains of bryozoans, polychaete tubes, barnacles	0–25 m	CT	J^R, F^R	2, 3, 4, 7, 9, 10, 11, 12
Clytia linearis (Thorneley, 1900)	TA; A;L	Bioconcretions, rocks, coralline and other algae, hydrorhiza of A. tubiformis, polychaete tubes, barnacles	0–25 m	CT	J^R, F	1, 3, 4, 9, 10, 11, 12
Clytia cf. paulensis (Vanhöffen, 1910)	TA; A	Hydrorhiza of T. dohrnii	5–8 m	CT	J	9
Obelia bidentata Clark, 1875	MS; TA; A	Rocks, bioconcretions, algae	0–32 m	CT	J^R, F^R	9, 10, 11
Obelia dichotoma (Linnaeus, 1758)	MS; A; L	Rocks, bioconcretions, algae	1–23 m	C	J^R, F^R	1, 4, 10

The high number of new records in this study is mainly due to the remarkable scarcity of studies in the region. Two species (Bougainvillia muscus and Filellum cf. serpens) were previously recorded in the Marmara Sea, but there were no further records from the Turkish Aegean coasts. In addition, Clytia linearis is a lessepsian species that was recorded for the first time in Turkey in the present work. Thus, the present record of 48 species, all previously known in the western Mediterranean, constitutes a significant increase in the number of hydroid species recorded in the Aegean Sea. Previously, Yamada (1965) had recorded 12 hydrozoan species from the vicinity of Athens, while Kocataş (1976) found 10 species in the Gulf of İzmir, Morri and Bianchi (1999) observed 31 in two southern Aegean islands (Kos and Milos), and Ünsal (1981) recorded another 23 in the Turkish Aegean coasts. Thus, this study provides the basis for the increase of the number of hydroid species known from the Aegean Sea to a total of 88.

The hydroid fauna of Gökçeada Island is dominated by circumtropical species (15 species, 33%). This proportion is similar to that reported from the Levantine Sea (Morri et al. 2009) and the Maltese islands (Soto-Àngel & Peña-Cantero 2013), the general warm-water affinity of the Gökçeada hydrozoan fauna being evident. Overall, the present collection showed a relatively low proportion of Atlantic–Mediterranean elements; these data, however, must be taken with caution, since the total number of Atlantic–Mediterranean species may increase, provided that more studies are conducted in the region.

As for substrate preference, most species are epibionts, being found mainly on hard surfaces such as algal bioconcretions, bryozoans and polychaete tubes. Although some species (e.g. Aglaophenia elongata, Kirchenpaueria halecioides) were found to be only epilithic in our samples (eight species), they have been observed growing on other substrate types elsewhere (see for example Boero 1981; Svoboda & Cornelius 1991). Twenty-one species in our samples were both epilithic and epibionts; three other species were exclusively found on other hydroids (Lafoeina tenuis, Anthohebella parasitica and Clytia cf. paulensis), and five species could utilize both hydroids and other substrates. Clytia cf. paulensis, however, was observed only once, so the information on its preferred substrate is by no means conclusive. Only four species were found on P. oceanica meadows, and two of them, namely Aglaophenia harpago and Sertularia perpusilla, are obligate and specialized epibionts restricted to this sea grass species. Posidonia oceanica meadows exhibit a complex biotic community and have a very high productivity and diversity of invertebrates (Pergent et al. 1994; Borg et al. 2006). Hydrozoans are one of the most remarkable taxa inhabiting these meadows, as pointed out in several studies (e. g. Boero 1981, 1987). Hydrozoan diversity may, however, be underestimated in the Gökçeada Posidonia meadows, because of our sampling approach which was not focused on the leaves of this species. Notwithstanding this, the present collection of species suggests that substrate diversity at Gökçeada Island directly influences hydrozoan diversity in the region.

In general terms, a pattern of variation between the samples collected on the northern coasts of Gökçeada and those obtained from the southern shoreline was identified by the classification analysis (Figure 2) and the PERMANOVA (Table II), and to a lesser extent also by the MDS (Figure 3).

Figure 2. Similarity dendrogram (Sørensen index) of the samples based on the hydroid species composition of 12 stations along the coasts of Gökçeada (Aegean Sea). The codes identifying the samples are as follows: number of the station; station acronym; “su” for summer or “wi” for winter; depth of the sample. Thick lines connect samples which differ significantly (similarity profile, SIMPROF P < 0.05).

Figure 3. Tridimensional representation of the multidimensional scaling (MDS) of the samples based on the hydroid species composition of 12 stations along the coasts of Gökçeada (Aegean Sea). Acronyms identifying the samples are omitted for clarity, except for the most distinctive Stations 3 and 5 (see text).

Table II. Permutational multivariate analysis of variance (PERMANOVA) of hydrozoan assemblages (48 taxa) of Gökçeada Island, (Aegean Sea). Statistically significant results are highlighted within a rectangle. Sources of variation are as follow: Se = season; Ex = exposure (north or south); De = depth; df = degrees of freedom; SS = sum of squares; MS = mean square; Pseudo-F = Pseudo-F statistic; P (perm) = probability after the permutations.

Source	df	SS	MS	Pseudo-F	P(perm)	Unique permutations
Season	1	5031.4	5031.4	1.4828	0.125	997
Exposure	1	3418.1	3418.1	1.3888	0.047	996
Depth	3	14138	4712.6	1.0073	0.446	996
Se × Ex	1	3525.6	3525.6	1.039	0.445	999
Se × De	3	11980	3993.3	1.1769	0.236	999
Ex × De	3	11504	3834.6	1.1301	0.289	997
Se × Ex × De	1	4102.8	4102.8	1.2091	0.29	998
Residual	29	98404	3393.2
Total	42	1.63 × 10^5

The PERMANOVA comparisons revealed that no significant difference can be attributed to the season or the depth of sampling, whereas the (southern or northern) exposure of the collection sites was significant in determining the observed differences between samples (Tables I and II). The main difference between the northern and the southern coasts of the island is related to their geomorphology. The southern coast has a much more gradual bottom slope and soft substrates are much more abundant than in the northern area, where it is common to find vertical or sub-vertical rocky cliffs. This difference has therefore influenced the number of sampling stations in the northern area (eight), double the number of hydroid samplings available in the southern coast (four).

Soft bottoms indeed offer a limited selection of natural substrates and a reduced diversity of benthic species which may represent potentially available substrates for epibiotic hydrozoans. The nature of the substrate significantly influences species composition, diversity, abundance and distribution of benthic organisms, including hydrozoans, and hydroid assemblages are generally more diverse and abundant on hard substrates (Calder 1991). It is worth noting that the differences observed between the northern and southern samples seem to be related to the presence of common, locally abundant species (Sertularella ellisii, Aglaophenia tubiformis, Clytia hemisphaerica, Clytia linearis, Eudendrium racemosum, Plumularia obliqua, Eudendrium capillare, Turritopsis dorhnii and Dynamena disticha) rather than rare, exclusive species as shown in the SIMPER analysis (Table III). Common, dominant species have often been held responsible for richness and diversity patterns in hydrozoan assemblages elsewhere in Mediterranean habitats, and, in fact, several spatial patterns in richness reported in the literature appear to be caused by relatively few, common species (Megina et al. 2013). Besides C. linearis, no new alien species were detected, providing evidence that Gökçeada may represent a still-unaffected hotspot of Eastern Mediterranean hydrozoan biodiversity.

Table III. Similarity percentage analysis (SIMPER) analysis of hydrozoan contributions to dissimilarities between northern and southern samples from Gökçeada Island, Northern Aegean Sea. δ = Average dissimilarity between northern and southern samples; SD = standard deviation of δ; Contrib.% = percentage contribution of each species to the overall dissimilarity; Cum.% = cumulative percentage of contribution; only main contributors (up to a cumulative 40%) are shown. Species in bold contributed the most to the similarity of southern stations. Species with * contributed the most to the similarity of northern stations.

Species	δ	δ/SD	Contrib.%	Cum.%
Sertularella ellisii*	5.71	0.94	6.47	6.47
Aglaophenia tubiformis	5.56	0.78	6.31	12.77
Clytia hemisphaerica	4.64	0.70	5.26	18.04
Clytia linearis*	4.27	0.81	4.84	22.88
Eudendrium racemosum	4.15	0.82	4.71	27.58
Plumularia obliqua	4.08	0.75	4.62	32.21
Eudendrium capillare*	3.14	0.54	3.56	35.77
Turritopsis dorhnii*	3.10	0.55	3.52	39.29
Dynamena disticha*	2.69	0.46	3.05	42.33
Average dissimilarity between southern and northern stations = 88.23
Average similarity within northern samples = 60.86
Average similarity within southern samples = 50.36

The cluster diagram (Figure 2) and MDS representation (Figure 3) showed that stations 3 (Kuzulimanı) and 5 (Şeytankale) were highly different from the other sampling stations, mainly due to their low number of species and unique species composition. Both stations present very particular conditions, which most probably are responsible for their singularity in terms of the hydrozoan fauna. Station 3 (Kuzulimanı) is the main harbor of the island, being subjected to a high anthropogenic pressure that seems to condition the presence of a reduced assemblage of hydrozoans. Only some widespread and opportunistic species (e.g. the well-known Clytia hemisphaerica, Clytia linearis and Aglaophenia picardi) are able to establish permanent populations in this disturbed environment. These results are consistent with the findings of Megina et al. (2013), who observed harbor hydroid assemblages to be significantly different from natural ones mainly due to their qualitative composition.

An interesting finding related also to the fauna of the harbor and its vicinity in Gökceada is that of Ectopleura larynx, which is considered a species found very often in or around harbors in temperate and cold waters around the world, and that may sometimes cause problems to aquaculture and port operations (Guenther et al. 2009, 2010). We found this species exclusively in station 2 (Kanyar), very close to the harbor of Kuzulimanı (station 3). This vicinity suggests that the species may actually be present in the harbor but was missed in the only sampling conducted at this station. The abundance of E. larynx in our samples was not very high, however, and more studies are needed to establish the extent of its populations in the study area.

Species inventories and distribution records represent key tools to biodiversity research. The diversity of benthic hydrozoans recorded in the last decade in the Mediterranean Sea is restricted to 156 species (Gravili et al. 2013) and, out of them, 48 taxa were collected by this study on the Gökçeada coast. In spite of the limited temporal samplings (1 week in summer; 1 week in winter), and the depth limitation (0–20 m) due to diving operations, nearly one third of the contemporary (i.e. recently collected) Mediterranean species of hydroids were detected in Gökçeada. Therefore, the northern Aegean Sea appears to be a representative, high-diversity area of the Eastern basin, probably acting as a biogeographic crossroads across the different ecological and hydrological features of the Black Sea, the Aegean Sea and the Ionian Sea. The near absence of hydrozoan aliens in the present faunistic inventory qualifies Gökçeada as a hotspot for future biodiversity studies and for early monitoring of the impact of climate change in the Eastern Mediterranean.

Acknowledgements

The authors are grateful to the scientists and technicians of Gökçeada Marine Research Center for their help during samplings and analyses. This study was supported by the Istanbul University (project number IRP-20587).