Faunal and biogeographic analysis of Syllidae ( Polychaeta ) from Rovinj ( Croatia , northern Adriatic Sea )

The study of hard-bottom Syllidae (Polychaeta) of the Rovinj area provides an updated measure of syllid species diversity in the northern Adriatic Sea. Faunal research in the coldest Mediterranean sectors may also help to address possible ecological and biogeographical consequences arising from climate change. Thirty-nine species were found, of which 13 are new for the northern Adriatic, increasing the species recorded from the area from 53 to 66. Some newly recorded species are dominant and typical of warmer areas. The lack of previous taxonomic updates is responsible for the high number of new findings. However, the northern Adriatic Sea is possibly undergoing long-term changes, with modifications of diversity due to the establishment of warm-water species. Syllid fauna of Sveti Ivan Island and its bioclimatic affinity are, in fact, unexpectedly more similar to the ones of some southern Mediterranean areas than to those traditionally reported for the northern Adriatic.


INTRODUCTION
The coldest Mediterranean sectors (i.e.Gulf of Lion, northern Adriatic and northern Aegean Seas) are strategic for monitoring faunal changes that may be caused by global warming (Boero et al., 2008).
The northern Adriatic Sea, bordered on the south by the line connecting Pescara in Italy and Zadar in Croatia (McKinney, 2007) is geomorphologically, hydrographically and biogeographically a peculiar Mediterranean region.The lower average temperatures of this northernmost Mediterranean region al-lowed the presence of a peculiar flora (e.g. the brown alga Fucus virsoides J. Agardh) and fauna [e.g. the Adriatic sturgeon Acipenser naccarii Bonaparte, 1836, the hydroid Tricyclusa singularis Schulze 1876 and the sprat Sprattus sprattus (L., 1758)].This supports its strong boreal affinity, as well as its ecological and biogeographical similarities with the North Atlantic (Bianchi et al., 2004;Boero and Bonsdorff, 2007).
Concerning polychaetes, the northern Adriatic Sea hosts the highest percentage of cold-temperate Syllidae species within the Mediterranean, but it also shows the lowest taxonomic updating index (Musco and Giangrande, 2005a).In fact, research on northern Adriatic polychaete fauna started in the second half of the 19th century and continued from the first half of the 20th (see Amoureux, 1983 and literature cited) to the present (Požar-Domac, 1994 and literature cited;Zahtila, 1997;Castelli et al., 1999;Mistri et al., 2002;Aleffi et al., 2003).However, most studies in the literature cover ecological aspects or simply report species lists (including records from grey literature or unpublished data), whereas taxonomic studies are scant.Moreover, most studies deal with soft bottoms, which are dominant on the western Italian coast, whereas very little is known about hard bottoms on the eastern, Istrian coast.
Syllidae is one of the most diverse polychaete families, common in many habitats but typical on hard bottoms.It is also the most diverse polychaete family in the Adriatic Sea, with approximately 100 species (Musco and Giangrande, 2005a).Syllid distribution seems to be a good indicator of local environmental conditions and might be effective in assessing large-scale ecological changes, even over relatively short periods (Giangrande et al., 2004;Musco and Giangrande, 2005a).In fact, some coldwater species previously reported in the northern Adriatic are no longer mentioned (Musco and Giangrande, 2005a).This might indicate an overall warming and tropicalization of the Mediterranean (Bianchi, 2007) causing local extinctions of some sensitive species, but it might also reflect lack of recent, updated taxonomic studies.
Taxonomic studies in the northern Adriatic Sea are, thus, particularly important in this period of global warming, as recently suggested by the CIESM (2008).Therefore, we explored the inshore hard bottoms of the Sveti Ivan Island (Rovinj) to increase knowledge of the northern Adriatic syllid diversity and biogeography.In addition to the inventory of the species found in the area, we also provide drawings, microphotographs and detailed descriptions of the most interesting findings, as an updated faunal background for future research in the area.

Study area, sampling methods and taxonomic approach
Sveti Ivan Island is located close to Rovinj (Croatia, northern Adriatic Sea), and characterized by calcarenitic rocky plateaus extending from 0 to about 25 m depth, with a gentle-medium slope.Three stations were sampled in summer 2007 (Fig. 1) using scuba diving.At each station, three depths were selected (1.5 m, 5 m and 25 m), with three replicate samples at each depth by scraping the hard bottom benthos within a 10x10 cm metal frame, for a total of 27 observation units.According to our aims, the airlift sampler was not used in order to limit the mechanical damages and stress, which would induce autotomy in syllids and thus reduce the taxonomic accuracy.At each site temperature and salinity were measured; in situ observations and further analyses of underwater photographs were used to characterize the sessile macrobenthos (Table 1).
Samples were fixed in 8% formaldehyde seawater solution, sieved through 0.5 mm mesh and preserved in 70% ethanol.Syllids were sorted under a stereomicroscope.Diagnostic characters were analysed, drawn and photographed under an optical microscope.Specimen width at proventricle level, excluding parapodia, and length of prostomium plus first 10 chaetigers (H+10) were measured.The following abbreviations are used throughout the text: DC (dorsal cirrus/dorsal cirri), Chr (chaetiger/ chaetigers).In the taxonomic accounts, only the most recent literature on species description (also providing synonymies and references) is given.All specimens are deposited at the Center for Marine Research of the Ruđer Bošković Institute in Rovinj, Croatia.

Biogeographic and ecological analyses
Using the checklists of the coastlines of the northern Adriatic, Spain, northwest Italy, eastern Sicily, Israel, Cyprus, the Turkish and Greek Aegean, and the Ionian and Adriatic coastlines of the Salento Peninsula reported in the recent biogeographic revision of the Mediterranean Syllidae (Musco and Giangrande, 2005a), the Sveti Ivan syllid species inventory was compared with those of the above-mentioned coastlines both faunistically (i.e.comparing the syllid species lists based on a presence/absence matrix) and bioclimatically, assigning the species to 6 bioclimatic categories: cold, temperate-cold, temperate, temperate-warm, warm and eurythermic.Differences among inventories were assessed by means of the Bray Curtis similarity index and multidimensional scaling (MDS) analysis (PRIMER, Plymouth Marine Laboratory, UK).
In the Sveti Ivan syllid assemblages, differences in the spatial distribution of the individuals and the species richness were tested by analysis of variance (ANOVA).Differences in the spatial distribution of the species richness grouped in bioclimatic categories were tested by permutational multivariate analysis of variance (PERMANOVA) using the PER-MANOVA computer program (Anderson, 2005).

Diversity and biogeography
From a total of 39 species, 27 (266 specimens) were found at 1.5 m, 28 (320 specimens) at 5 m and 22 (367 specimens) at 25 m depth.The mean number of species per sample was always higher but also more variable at 5 m depth; the mean number of individuals was higher at 25 m depth at station B and C, while the minimum abundance was observed at the same depth at station A (Fig. 2A).The ANOVA (Table 2A) revealed significant differences in the local distribution of the species richness both vertically (among depths) and horizontally (among stations).Most species were locally rare (Fig. 2B), with 13 out of 39 being found only in one sample, and only Sphaerosyllis pirifera being common in the whole area (25 samples).Consequently, neither the accumulation curve, nor the estimator curves (Jacknife1, Jacknife2, Bootstrap) showed a tendency to reach the asymptote (Fig. 2C), suggesting a potentially large number of species for the area (i.e. from 44 to 63).Based on their abundance, twelve species are considered as the most representative of the area (Table 3).
From a bioclimatic point of view, about 38% of the species were temperate-warm, 28% eurythermic, 23% temperate, 7% temperate cold and 3% warm, while none of them was cold.PERMANOVA revealed significant differences in the horizontal distribution of the richness of species grouped in bioclimatic categories (among stations), but no differences among depths (Table 2B).(1) Test done using 4999 permutations of appropriate units, the p-value given in bold was obtained using 4999 Monte Carlo samples from the asymptotic permutation distribution.
Fig. 2. -Species distribution pattern in the sampled area.A, mean number (± SE, n = 3) of syllid individuals (bars) and species (lines) at the analyzed sites.B, distribution of species according to the number of replicates occupied.C, species area accumulation curve and estimator curves (J1, Jacknife1; J2, Jacknife2; B, Bootstrap; SO, Species observed).
The faunistic analysis (Fig. 3A) shows that Sveti Ivan is just opposite the northern Adriatic, whereas the other Mediterranean sectors are in an intermediate position, forming two groups (SP, NWI, GR, TU, CY, SA, SI and ES, IS).In turn, the bioclimatic analysis confirms the differences between the northern Adriatic and the remaining coasts, while Sveti Ivan most closely resembles the southern Mediterranean (Fig. 3B).
Distribution: Mediterranean Sea, W Atlantic and E Atlantic Ocean from the North Sea to South Africa, Indian coast of South Africa, Australia (San Martín, 2005).
Remarks: Exogone rostrata is easily distinguishable from the congenerics in the area due to the special spiniger-like chaeta from chaetiger 1, with a long blade and stout handle ending with a triangular ridge forming a cup-shaped structure.

Sphaerosyllis pirifera
Remarks: Sphaerosyllis pirifera was the most abundant species at all depths and stations.Moreover, it is the largest Mediterranean Sphaerosyllis species and has never been previously mentioned in the area.
Distribution: Mediterranean Sea, E Atlantic Ocean (San Martín, 2003).New for the N Adriatic Sea.

Distribution:
The species is considered cosmopolitan in warm-temperate areas.However, its distributional range should possibly be limited to the E Atlantic Ocean and the Mediterranean Sea (San Martín, 2003).Distribution: The species is considered cosmopolitan in warm-temperate areas (San Martín, 2003).

Remarks:
The specimen resembles Paraehlersia (?) dionisi (Núñez and San Martín, 1991) but differs in having shorter dorsal cirri at proventricular level, shorter long-bladed chaetae (up to 58 µm vs 80 µm in the original description) and particularly in having a longer peristomium (as long as Chr 1 vs very short, covered by the Chr 1 in the original description).This species is included in Paraehlersia since the antennae, tentacular and anterior cirri are articulated, and it has spiniger-like compound chaetae; however, its systematic position remains unclear, being thus indicated as P. (?) dionisi (San Martín et al., 2009).Distribution: E Atlantic Ocean, Mediterranean Sea (Spanish and Turkish coast) and Antarctica.It was found in Mediterranean from 43 to 470 m depth (San Martín, 2003).Providing our specimen actually belongs to P. (?) dionisi, this would be the first record for the Adriatic Sea and the shallowest one, since it was found at 25 m depth.Distribution: Mediterranean Sea, Atlantic Ocean (San Martín, 2003).New for the N Adriatic Sea.
Remarks: Our specimens correspond to the descriptions of San Martín (1984;2003).However, the dark spots on antennae and cirri were not reported in the holotype either by Gravier (1900) or by San Martín et al. (2008a).Other species having dark spots on cirri and similar chaetae are B. thylacine San Martín et al., 2008a,b, B. cirropunctata Michel, 1909and B. maculata (Imajima, 1966).The first two, however, differ from our specimens in the shape of anterior falcigers, while B. maculata differs in having longer antennae and DC.Branchiosyllis exilis is considered a species complex and would require a detailed revision.Our specimens, as well as the Iberian ones, might belong to a different species (San Martín et al., 2008a).
Distribution: The B. exilis species complex, as well as the genus Branchiosyllis, has a circumtropical distribution, including the warmer areas of the Mediterranean Sea (San Martín et al., 2008a).This is the northernmost record in the Mediterranean area.
Distribution: Mediterranean and Red Sea, Australia, E Atlantic from the North Sea to the Canary Islands and W Atlantic Ocean from North Carolina to the Gulf of México (San Martín et al., 2008a).
Remarks: Our specimens agree with San Martín (2003) but differ slightly from the holotype redescription (Licher, 1999) in having somewhat shorter lateral antennae and mid-body pseudospinigers.Like S. rosea and S. garciai (see descriptions below), S. beneliahuae is characterized by pseudospinigers, as is S. cornuta (Rathke, 1843), a species commonly reported from both hard and soft bottoms in the N Adriatic (Zahtila, 1995 and references cited).However, according to Licher (1999) and San Martín (2003), this species should not be present in the Mediterranean Sea, being typical for colder regions (neotype locality Trondheimsfjorden, Norway).We herein hypothesize that the previous records of S. cornuta in the area, par-ticularly as far as hard bottom findings are concerned, may possibly belong to S. beneliahuae, S. rosea or/ and S. garciai, but also to S. parapari (see San Martín and López, 2000), which were not previously recorded in the N Adriatic.The descriptions of the first three species are herein given, to facilitate the identification in future research in the area.Similarly, soft bottom reports of S. cornuta in the N Adriatic Sea should be carefully reconsidered.In fact, the analysis of the N Adriatic soft bottom material from the Rovinj Center for Marine Research collection (unpublished data) revealed that the specimens assigned to S. cornuta actually belong to S. parapari San Martín andLópez, 2000, but also to S. garciai andS. cf. alosae San Martín, 1992.This confirms that previous findings of S. cornuta in the Mediterranean should probably be attributed to S. parapari, whose presence, however, is not formally confirmed for the Mediterranean (San Martín, 2003).Distribution: Mediterranean Sea, E Atlantic Ocean and Caribbean coast (Licher, 1999).New for the N Adriatic Sea.Verrill, 1900 Typosyllis corallicola, Licher, 1999: 116 Syllis corallicola, San Martín, 2003: 439 Material: A5a, 1 ind.; A5b, 2 ind.; A5c, 4 ind.; B5a, 3 ind.;B5b, 4 ind.; 5c, 9 ind.; C1.5a, 1 ind.; C1.5b, 8 ind.; C5a, 2 ind.; C5b, 2 ind.; C5c, 1 ind.; C25a, 1 ind.; C25b, 1 ind.; C25c, 1 ind.

Syllis corallicola
Remarks: Our specimens agree with the descriptions of Licher (1999) and San Martín (2003) except in having ventral cirri as long as parapodial lobes vs shorter in Licher (1999), while anteriorly longer and posteriorly shorter than parapodial lobes in San Martín (2003).The proventricle is one third longer than the pharynx, with 26-30 muscle cell rows vs as long as the pharynx with 34-40 muscle cell rows, according to the above mentioned authors.
Remarks: Our specimens agree with the description of San Martín (1984) of S. truncata cryptica, which was originally described from the Gulf of Elat, Red Sea (Ben-Eliahu, 1977) and subsequently commonly recorded in the Mediterranean Sea.However, this sub-species was considered as a synonym of S. gerlachi by Licher (1999), who described S. gerlachi by combining the characters of an incomplete syntype from the Red Sea (lacking the last segments) and those listed in the original description (Hartmann-Schröder, 1960).Our specimens differ from the original description in having posterior DC with 12-15 articles, compared with 6-8 in Hartmann-Schröder (1960) and in having a pharynx/proventricle length ratio of 3, vs 1.4 and 1.6 respectively in the original descriptions of S. gerlachi and S. truncata cryptica.In Licher (1999), the pharynx is even shorter than the proventricle.Although we consider these differences as intraspecific, further analyses are needed, particularly of alive specimens from the Red Sea type locality.In fact, irregularly arranged faintly marked orange dorsal spots, which completely disappear after preservation, were observed in the live specimens of S. gerlachi from the N Ionian Sea (LM, personal observation).
Distribution: Mediterranean and Red Sea, Indian Ocean, Caribbean coast, Gulf of México (Licher, 1999).New for the N Adriatic Sea.
Distribution: Mediterranean and North Sea, N Atlantic and Central-East Atlantic Ocean; records outside the above mentioned areas are considered doubtful (Licher, 1999).

Syllis rosea
Remarks: Our specimens differ from San Martín (2003), particularly in having dorsal and ventral tentacular cirri somewhat longer.See also remarks for S. beneliahuae.Distribution: Mediterranean Sea, Madeira and Canary Islands, NW Pacific Ocean and Solomon Islands (Licher, 1999).New for the N Adriatic Sea.
Remarks: Our specimens agree with the holotype redescription by Licher (1999).The type locality (Krk Island) is close to our research area.As in Licher (1999), our specimens were generally dark, mostly lacking the typical "eyeglasses-shaped" colour pattern, possibly due to preservation.However, the specimens from Sveti Ivan Island bear stout, short, fusiform mid-posterior DC, resembling S. ferrani, while Licher (1999) reports slim middleposterior DC, scarcely attenuating towards the tip.Also, the simple dorsal chaeta is straight and distally serrated, while slightly curved and smooth in Licher (1999).Thus, our specimens better correspond to the Iberian specimens described by San Martín (2003).All above differences, as well as those between Licher (1999) and San Martín (2003) (i.e.number, length and dorso-ventral gradation of anterior chaetae, number of proventricular muscle cell rows, length of anterior dorsal cirri and ventral cirri), are herein considered as intra-specific variations.
Distribution: Mediterranean Sea, Atlantic Ocean and Pacific Marquesas Islands (San Martín, 2003).Trypanosyllis zebra (Grube, 1860) Remarks: Reports of this species beyond the type locality should be re-examined, as they may represent other species (San Martín et al., 2008a).The European specimens present intra-specific chaetal variation, having either unidentate blades only or both unidentate and bidentate blades (Campoy, 1982).Our specimens agree with San Martín (2003), in having ventral-most unidentate and dorsal-most bidentate blades.

DISCUSSION
Similarly to the polychaete assemblages of Otranto in the extreme south Adriatic Sea (Giangrande et al., 2003), the Sveti Ivan syllid assemblages are variable in terms of both species abundance and richness among depths and their horizontal distribution.At the considered spatial scale (hundreds of metres), in fact, this spatial variability is a common trend in benthic assemblages, and has already been reported for the Syllidae (Musco et al., 2009).
From a faunistic point of view, 13 species are newly recorded for the northern Adriatic Sea, increasing the number from 53 (Castelli et al., 2008) to 66.The number of species found in this study (39) is close to those from other Mediterranean hard bottoms (about 40 to 60 species) (Çinar, 2003;Giangrande et al., 2003;Corriero et al., 2004;López and Gallego, 2006;Musco et al., 2009), although our sampled area was considerably smaller and some specimens were possibly lost since the airlift sam-pler was not used.Moreover, the estimated number of species in our study area (from 43 to 63) suggests that its potential diversity is among the highest recently reported for the Mediterranean.Therefore, the northern Adriatic syllid assemblages appear to be complex and highly diverse, thus calling for regular faunal updating in the area (Musco and Giangrande, 2005a).Many species were, in fact, rare and the analyses of distribution patterns indicate that additional sampling would probably yield more species.Since one third of the records are new, we may assume that more new species will be further found in the northern Adriatic.Particularly, some newly recorded species were among the most abundant or frequent ones (e.g.Syllis rosea, S. gerlachi and S. corallicola), with Sphaerosyllis pirifera being the dominant one.Species dominating the Sveti Ivan syllid assemblages are also common and abundant in other Mediterranean areas, such as the northern Cyprus coast (Çinar, 2003) and the southern Adriatic coast (Giangrande et al., 2003).
Obviously, new northern Adriatic records represent the northward widening of the species distribution in the Mediterranean, particularly for the typically warmer (e.g.Syllis corallicola, S. gerlachi, S. garciai and S. westheidei) or endemic (e.g. S. ferrani and Exogone rostrata) ones.The presence of Branchiosyllis exilis is particularly relevant, since the whole genus is typically circumtropical.
Our study draws a scenario differing from the traditionally reported northern Adriatic one.When the species compositions are considered (faunistic analysis), Sveti Ivan is the farthest sector of the northern Adriatic.The relative positions of the Mediterranean sectors (excluding Sveti Ivan) resemble the pattern reported in Musco and Giangrande (2005a), who suggested an influence of both the different environmental features and, especially, of the taxonomic updating of the respective inventories (particularly for NA, ES and IS).Therefore, we suggest that the position of Sveti Ivan in the bi-plot might be due to its recent taxonomic updating.However, the area is represented by only one sampling time, so the potential finding of additional species in further surveys would possibly influence its degree of faunal similarity with the other sectors.The bioclimatic approach reveals that the species array at Sveti Ivan Island resembles the more southern Mediterranean syllid inventories rather than the previously reported northern Adriatic ones (dominated by cold-temperate species).
The high number of new records of warm-water species in our study might reflect a "meridionalization" of the northern Adriatic fauna (Bianchi, 2007;Boero et al., 2008), at least for the considered depth range, since no clear differences in the bioclimatic composition among depths were observed.However, the new records might also reflect incorrect characterization of species' distributional ranges, due to lack of recent taxonomic research in the area (Musco and Giangrande, 2005a).Some species, such as Syllis beneliahuae, S. ferrani, S. garciai and S. westheidei, may have not been recorded previously, as former taxonomic studies of hard bottoms in the northern Adriatic Sea (Banse, 1959;Amoureux and Katzmann, 1971;Katzmann, 1971Katzmann, , 1972;;Požar, 1972;Amoureux, 1975) were carried out before those species had been scientifically described.However, the same reason cannot justify the absence of other newly reported but well-known species, such as Myrianida convoluta, M. quindecimdentata, Exogone rostrata and, particularly, the abundant Sphaerosyllis pirifera, Syllis gerlachi, S. corallicola and S. rosea.Our findings might suggest an actual faunal change in the area.
Misidentification, most probably caused by the fact that taxonomic information currently available was not complete at the time of the former observations, is another possible reason for the lack of previous records of some species (see remarks for S. beneliahuae).
Whatever is the reason for the high number of new records (lack of taxonomic studies, meridionalization, misidentification), regular updates of the northern Adriatic Sea biota are important for a better understanding of the dynamics of this part of the Mediterranean, presumably highly sensitive to the influence of global warming (CIESM, 2008).The greatest problem in trying to relate changes in marine communities to climate changes is the lack of long-term data (Southward, 1995), concerning both the biota and the environmental parameters.Longterm monitoring studies of benthic fauna associated with measurements of the environmental variables in the northern Adriatic Sea, such as temperature, salinity, chlorophyll-a concentration and pH, are necessary in order to correlate climate changes with those of the polychaete fauna and thus support the postulated meridionalization.Further studies might, in fact, clarify whether the northern Adriatic is going through species enrichment due to the establishment of warm-water species, whose high abundance might contribute to a loss of the peculiar biogeographical features of this area.
Ircinia sp.) and the algae P. squamaria and EA, in the lower layer, and D. dichotoma, P. pavonica, F. petiolata and Jania sp.(and in low numbers Amphiroa sp. and Cladophora sp.) in the upper layer.bottom layer; dominant algae in the upper layer are P. pavonica and D. dichotoma (with Laurentia sp. and Amphiroa sp.present in low numbers).Dominance of P. squamaria in the bottom layer and P. C5(a,b,c) pavonica, D. dichotoma, Galaxaura oblongata, Jania sp. and encrusting algae in the upper layer.

Table 2 .
-Univariate and multivariate analysis of variance testing differences in the spatial distribution of the syllid assemblages [Factor 1: station (St), three levels, random; Factor 2: depth (De), three levels, orthogonal, fixed; Three replicates].A, differences in spatial distribution of individuals and species richness (ANOVA, untransformed data).B, differences in spatial distribution of species richness grouped in bioclimatic categories (PERMANOVA based on Bray-Curtis dissimilarity, untransformed data).DF, degrees of freedom; MS, mean square; F, F-ratio; pu, permutable units; MS denom , denominator mean square.Significant p-values are in italics.

Table 3 .
-Mean number of individuals (mean) per sampling unit (0.01 m 2 ) plus standard deviation (SD) of the 12 most abundant syllid species at 1.5, 5 and 25 m deep.Total: number of collected individuals; *: new for the area .