Distribution patterns of Syllidae ( Annelida : Polychaeta ) from seagrass ( Zostera marina and Z . noltei ) meadows in the Ensenada de O Grove ( Galicia , NW Spain )

This paper describes the distribution and composition of the syllid fauna inhabiting seagrass meadows in the Ensenada de O Grove (NW Spain). Samples were collected on muddy sediments colonized by either Zostera marina L., Zostera noltei Hornemann or by a mixed meadow with both species. Syllids were dominant (13340 individuals; 37% of total polychaete abundance), including 22 species (12 genera). The mixed meadows housed the highest number of species and the Z. noltei meadow had practically no syllids. The dominant species were Exogone naidina, Parapionosyllis elegans, Parexogone hebes and Prosphaerosyllis campoyi (>80% of total abundance). Carnivores (mainly species of Parapionosyllis, Amblyosyllis, and Streptosyllis) were dominant, especially in muddy sand with either Z. marina or Z. noltei and sandy mud with a mixed meadow. The most important abiotic variables for explaining the composition and distribution of the syllid fauna were bottom water salinity, sorting coefficient and carbonate content. The highest number of species was recorded at sites with a high salinity and carbonate content and the lowest at sites with a high sorting coefficient.


INTRODUCTION
Seagrass meadows are of great ecological importance in shallow-water environments, as their structures (leaves, rhizomes and roots) increase the habitat complexity.They harbour numerous epiphytic and epifaunal taxa (Orth and Heck 1980, Webster et al. 1998, Attrill et al. 2000), providing shelter and protection from predators (Heck and Thoman 1981) and a variety of food resources (e.g.seagrass, detritus and epiphytes) (Kitting et al. 1984, Hily et al. 2004, Fredriksen et al. 2005) to the associated faunal assemblages.Among them, the most diverse taxa are generally polychaetes, molluscs and crustaceans (Gambi et al. 1998, Nakaoka et al. 2001, Arroyo et al. 2006), which are often represented by small-sized, interstitial species that are also usually present in bare soft sediments (San Martín et al. 1985, Sardá 1985, Brito et al. 2005).In Zostera meadows, as in other seagrass meadows, these species may be favoured by the sediment retained by the seagrasses (Parapar et al. 1994, Fredriksen et al. 2010) and by the rhizomes and roots, which create spatial complexity within sediment and enable oxygenation (Tu Do et al. 2011 and references therein).
Meadows of the seagrasses Zostera marina L. and Zostera noltei Hornemann are typical of estuaries and shallow coastal areas in the northern hemisphere (Duffy and Harvilicz 2001), which are protected through the Habitat directive 92/43/EEC.On the Atlantic coast of the Iberian Peninsula, Z. marina and Z. noltei occur as extensive meadows in intertidal and shallow subtidal areas, particularly in the Galician rias (Laborda et al. 1997).In the sheltered, inner parts of the Ensenada de O Grove, highly dense meadows extend from the intertidal to the shallow subtidal (<20 m depth), providing macrofauna with protection against desiccation during low tide.
Studies exclusively focusing on the Syllidae from the Galician coasts are scarce (but see San Martin et al. 1985) and are most often included in wider benthic ecology studies (Moreira et al. 2006, Lourido et al. 2008) of both hard and soft substrata (San Martín et al. 1985, Parapar et al. 1994, Parapar et al. 1996a,b, Cacabelos et al. 2010).
As part of a broader project devoted to characterizing the soft-bottom benthic fauna of the Ensenada de O Grove (NW Spain) (Project XUGA30101A98), the main objective of this paper is to describe the syllid fauna inhabiting the seagrass meadows of the inlet in terms of composition, abundance, number of species and trophic structure.

Study area
The Ensenada de O Grove is located in the inner part of the Ría de Arousa (Galicia, NW Spain) between 42º41'N-42º28'N and 9º01'W-8º44'W (Fig. 1).The inlet has an area of 15 km 2 and is sheltered from wave action and dominant winds by the O Grove Peninsula.It receives freshwater inputs from rivers, both at the mouth and in the inner part.The inner and intertidal and shallow subtidal areas (<20 m) are soft bottoms largely colonized by Z. marina and Z. noltei.This inlet is of great socio-economic importance, especially with regard to mussel culture on rafts, bivalve collection (intertidal harvesting by hand and boat trawling) and fishing.Furthermore, the inlet is protected because of the seagrass meadows (Habitat Directive 92/43/CEE) and as a habitat for birds (ZEPAS, 1979 andRAM-SAR Convention, 1990).It is also a natural space of importance for the European Community, listed in the European Natura 2000 network.

Sample collection
The present study focuses on the inner soft bottoms of the Ensenada de O Grove colonized by Z. marina and Z. noltei.Ten sites were selected as representative of the different meadows (i.e.Z. marina, Z. noltei and mixed) and tidal conditions (intertidal vs subtidal).Sampling was done during October and November 1996 following the standard methodology for the XU-GA30101A98 project.We used a Van Veen grab with a sampling surface of 0.056 m 2 to collect five replicates per site (total area: 0.28 m 2 ).Samples were then sieved through a 0.5 mm mesh and all retained material was fixed in a 10% buffered formalin-sea water mixture.Additional sediment samples were used to determine particle-size composition and carbonate and organic matter contents, and single measurements of water temperature (°C), pH and salinity (practical salinity units, psu) and sediment pH and temperature (°C) were obtained in situ.

Laboratory analyses
Syllids were sorted out from the sediment under a stereomicroscope, identified to species level whenever possible, counted, and preserved in 70% ethanol.The names of species and higher taxonomic levels used follow the classification by Aguado and San Martín (2009), the MarBEF Data System (ww.marbef.org)and the WoRMS database (www.marinespecies.org).

Data analyses
The structure of the syllid assemblage was analysed using the PRIMER v 6.0 software package (Clarke and Warwick 1994).For each site, total abundance (N), total number of species (S), the Shannon-Wiener diversity index (H', log 2 ) and Pielou evenness (J') were determined using the DIVERSE routine.Affinities among sites were determined through non-parametric multivariate techniques (Field et al. 1982).Abundance data were fourth-root transformed (Currie and Small 2005, Bremec and Giberto 2006, Rueda et al. 2009) prior to constructing a matrix of similarities using the Bray-Curtis coefficient and calculating the centroids.Based on this matrix, the sampling sites were classified by cluster analysis (which was tested by the Simprof) and ordered through a non-metric multidimensional scaling (nMDS).These two analyses are complementary, so the graphic representation of the nMDS ordination includes the similarity levels derived from the cluster analysis.The SIMPER routine was used to identify the species most contributing to the dissimilarity among assemblages.Site 37 was excluded from the multivariate analyses (there was only one syllid).Also, the species in each group were classified according to the constancy and fidelity indexes (Glémarec 1964, Cabioch 1968), and those representing more than 4% of the total abundance per site or group were considered as dominant (Junoy 1996).The frequency × dominance (F×D) index was calculated to determine the numerical importance of species.The syllid species were assigned to one of the following guilds: carnivores, herbivores, detritivores, and omnivores (Rasmussen 1973, Fauchald and Jumars 1979, Gambi and Giangrande 1985a,b, Tena et al. 1993, 2000, Giangrande et al. 2000) (Table 1), and the importance of these guilds in the whole inlet and within the groups identified in the nMDS was analysed.
Correlations between assemblage descriptors and all measured environmental variables were determined through the non-parametric Spearman rank coefficient (SPSS 15 software package).Co-linearity (r>0.7) was also detected between some environmental variables and therefore only some of them were selected for the BIO-ENV routine (see Table 2).The rationale was the following: when two variables were highly correlated, that offering the most relevant information was selected for the BIO-ENV.For example, carbonate content was highly correlated with fine sand content (CARB-FS: 0.939).In this case, the latter was non-selected because other granulometric fractions that also provide information of sediment had already been included in the analysis.Variables expressed in percentages were previously log (x+1) transformed (Lourido et al. 2008, Sánchez Moyano andGarcía-Asencio 2009).

Environmental variables
Sampling sites were characterized by moderate to high silt/clay contents (6%-62%).Sand content was generally greater at subtidal sites and sediment ranged from muddy sand to mud (Table 2).Water salinity was lower than 33 psu, particularly at sites 34 and 37 (close to the river, 20 psu).Carbonate content ranged from 5% to 10% and organic matter content ranged from low at subtidal sites (1.3%) to high (10.7-15.5%)at intertidal inner sites.Site 37 also showed the highest organic matter content and the lowest carbonate content.

Description of faunal assemblages
Subgroup A1 included subtidal sites with Z. marina, mostly with muddy sand, a moderate selection, and a low organic matter content.N was high (mean ± sd: 5018±2657 ind.m -2 ) and total S was the highest (18), ranging from 7 to 14 per site.The group was dominated by E. naidina, P. elegans, P. hebes, P. minuta, P. campoyi and S. clavata.H' ranged from 1.29 to 2.30 and J' from 0.54 to 0.78.Carnivores and omnivores were dominant in abundance (36% and 28%, respectively) and number of species (33% and 33%) (Table 7).
As stated above, group B only included site 34 (an intertidal sandy mud area with a mixed meadow), and was characterized by low salinity, carbonate content, S (6), N (86 ind.m -2 ) and H' (1.34).The syllid assemblage, characterized by X. scabra, B. pusilla, P. tetralix, O. gibba and S. limbata, was clearly different from that in A (Table 7).

DISCUSSION
Syllids are common members of benthic assemblages associated with seagrass meadows (Çinar 2003), including those formed by Zosteraceae (Hutchings 1981), and the meadows at the Ensenada de O Grove were no exception (37% of total abundance and 24% of polychaete species richness; Quintas 2005).This contrasts with lower abundances found in other quantitative studies using the same sampling methodology on soft bottoms (coarse sand to mud; see Moreira 2003, Moreira et al. 2006, Cacabelos et al. 2008, Lourido et al. 2008, Lourido 2009, Cacabelos et al. 2010), which may be partially explained by the presence of a dense seagrass meadow in the present study rather than differences in granulometric composition or organic matter content (Table 8).In general, seagrass meadows reduce physical stress, trap sediment, reduce suspension, protect small invertebrates from predators, and enhance food availability, also adding complexity to the habitat (Orth et al. 1984).In the case of Cymodocea nodosa and Zostera noltei, syllids are among the most abundant polychaete taxa in the foliar and rhizome layers (Giangrande and Gambi 1986, Gambi et al. 1998, Brito et al. 2005).In fact, the tridimensional structure provided by those seagrasses and especially the rhizome structure make available a variety of microhabitats for small-sized taxa.Syllids are mostly interstitial animals and therefore the availability of small spaces along the rhizomes could favour their presence (Giangrande 1985, Somaschini and Gravina 1994, Brito et al. 2005).In fact, syllids require spatial structures at microhabitat rather than at macrohabitat level (Abbiati et al. 1987, Giangrande 1988), while the interactions among syllids and with other macrofaunal species have also been suggested as factors controlling the abundance and, partially, the variability of syllid assemblages (Musco 2012).
In the studied seagrass meadows, abundance, number of species, and diversity differed among sites, resulting in two distinct faunal assemblages: (1) the muddy sand with Z. marina and the intertidal muddy sand or sandy mud with Z. noltei or mixed meadows (with high values of S and H'), and (2) the intertidal sandy mud with mixed meadows and the intertidal mud flat with Z. noltei (low values of S and H').These differences may be partially explained by the sediment characteristics (carbonate and silt/clay content), the proximity of a river, and the dominance of Z. marina, Z. noltei or both seagrass species; the latter determines, in turn, the availability of microhabitats (size and shape of the leaves and rhizomes), food and amount of sediment retained by the rhizomes.The overall composition of the syllid assemblage in the meadows from the Ensenada de O Grove is similar to those reported from other seagrass meadows.Streptosyllis websteri occurred in mud, muddy sand and shallow muddy gravel bottoms with Zostera in the Ría de Ferrol (Parapar et al. 1994).At the island of Ischia (Tyrrhenian Sea, Italy) 33 syllid species (mostly Exogoninae: Exogone spp., Sphaerosyllis spp., Parapionosyllis spp.) were associated with C. nodosa and Z. noltei meadows, with E. naidina and P. elegans being positively correlated with the foliar substrate (Gambi et al. 1998).On the other hand, E. verugera, E. naidina, S. hystrix, B. pusilla, and S. clavata are cosmopolitan and ubiquitous species that are common in other habitats including bare bottoms (soft and hard substrata) (Sardá 1985).Exogone naidina, P. hebes, P. tetralix, P. campoyi, P. elegans, S. websteri were previously recorded in intertidal bare and soft sediments near to the seagrass meadows studied here (San Martín et al. 1985).Fredriksen et al. (2010) reported higher abundance of P. hebes and S. hystrix in Z. marina meadows than in bare soft sediments in Norway.Prosphaerosyllis campoyi is also abundant in C. nodosa and P. oceanica meadows (San Martín 2003) and bare intertidal (San Martín et al. 1985) and subtidal sedimentary substrata (Parapar et al. 1994).Similarly, Syllis garciai is common in C. nodosa and Z. noltei meadows (Gambi et al. 1998) and bare muddy sand (Parapar et al. 1996b, Lourido et al. 2008).In the Ensenada de O Grove, E. belizensis showed a noteworthy presence in subtidal muddy sand with Z. marina and intertidal sandy mud with a mixed seagrass meadow.This species has been reported from warm and tropical seas, including those of the Iberian Peninsula (López and San Martín 1997, Olano et al. 1998, San Martín 2003) and also in low densities in muddy sediments of the Ensenada de San Simón, Galicia (Cacabelos et al. 2010).In the Ensenada de O Grove, specimens were similar to those from the western Atlantic according to morphological characters.However, it has not yet been elucidated whether they have been accidentally introduced by human activities or have a true amphiatlantic distribution.Some warm-water species (mainly Mediterranean molluscs) have been previously collected in O Grove (Rolán et al. 1985, Rolán 1992, Quintas 2005, Quintas et al. 2005).The introduction of these species in this area has been attributed to commercial activities such as oyster importation (Rolán et al. 1985).In some cases, these accidental introductions may result in significant alterations in the composition of assemblages and biotic interactions (Grall and Hall-Spencer 2003).There is, however, no evidence of oyster importation being the cause of the presence of S. belizensis in the study area and direct dispersion should not be discarded as an alternative hypothesis.
Syllids were less abundant in the intertidal sandy mud with a mixed meadow at site 34, and, particularly, they were nearly absent in the nearby muddy sediment with Z. noltei at site 37, in the inner part of the inlet.This absence can be explained by the vicinity of the river mouth, which is associated with regular freshwater inputs, high silt/clay content and low salinity.In particular, the salinity and/or resulting horizontal stratification of waters is a key factor structuring the macrozoobenthic communities (mainly infauna and small, slowly motile epifauna) of Z. noltei meadows in the inner part of Arcachon Bay, France (Blanchet et al. 2004).Our results also agree with those of Cacabelos et al. (2010), who found that syllids were also scarce in Z. noltei meadows subjected to environmental conditions similar to those in the Ensenada de O Grove.
The wide spectrum of feeding habitats among syllids allows them to find a variety of suitable feeding resources on seagrass meadows.In the Ensenada de O Grove, carnivorous syllids were mostly represented by species of Parapionosyllis, Amblyosyllis and Streptosyllis.Omnivores such as P. hebes have been reported in other studies of leaf development of the plant (Gambi et al. 1995).S. hystrix and S. garciai have been considered as herbivorous and omnivorous, respectively (Giangrande et al. 2000, Sánchez Moyano andGarcía-Asencio 2009).In the studied meadows, surface and sub-surface deposit feeding polychaetes were numerically dominant (Quintas 2005), as has been found in other seagrass meadows (Jacobs et al. 1983, Kiting et al. 1984, Thayer et al. 1984, Junoy 1996).However, in this study, the collection of a high number of syllids, including carnivores, herbivores, detritivores and omnivores, shows the importance of syllids for understanding the trophic structure of these habitats.
In conclusion, the present study shows that Syllidae were well represented and had a high diversity in the Z. marina and Z. noltei meadows in the Ensenada de O Grove compared with bare soft bottoms of the inlet (Quintas 2005).This information suggests that seagrass meadows are biodiversity preservation hot spots.This is the first quantitative and systematic study based on the Syllidae family associated with the seagrass meadows of the inlet.Therefore, this paper can be considered as a baseline study for future monitoring and environmental management studies aimed at increasing the protection of seagrass meadows of the inlet.However, detailed long-term studies considering separately different spatial scales or microhabitats in the plant (blades, rhizomes) and the sediment are needed to better understand the temporal dynamics of syllid assemblages and the environmental factors governing them in the studied meadows.

Fig. 1 .
Fig. 1. -Location of the Ensenada de O Grove (Galicia, Spain) showing the distribution of sampling sites with seagrass meadows.

Fig. 2 .
Fig. 2. -Spatial distribution and density of the numerically dominant syllid species in the seagrass meadows at the Ensenada de O Grove.

Table 1 .
-Systematic list of syllid species identified in the study.Sampling sites, trophic guild (TG; C: carnivores, H: herbivores, D: detritivores, O: omnivores) and abundance values (Abund.)per m 2 (considering all sites) are indicated.The species acronym used in the nMDS ordination is also listed.

Table 3 .
-Total abundance (N, individuals per m 2 ), number of species (S), Shannon-Wiener's diversity index (H', log 2 ) and Pielou evenness (J') for each sampling site in the Ensenada de O Grove.Values: mean±standard deviation.

Table 5 .
-Results of SIMPER analysis showing the main taxa contributing to the dissimilarity among subgroups determined from cluster analysis.Average abundance (Av.Ab.), average dissimilarity (Av.Disim.),ratio value (dissimilarity/standard deviation, Dis./SD) and percentage of cumulative dissimilarity (Cum.Disim.)were also included.

Table 6 .
-Best combinations of variables obtained by the BIO-ENV routine.SBW, salinity of bottom water; GR, gravel; CS, coarse sand content; MS, medium sand content; CARB, carbonate content; S o , sorting coefficient.ρW, Spearman's rank correlation.
Fig.4.-Non-metric multidimensional scaling (nMDS) ordination plot showing the syllid species ordination with a numerical dominance >4% at any given site in the study area.See Table1for species acronyms.

Table 7 .
-Summary of biotic and physical characteristics of the three assemblages derived from multivariate analysis (values: mean ± standard deviation).First ten constant species are listed including their fidelity (ELE, elective; PRE, preferential; ACE, accessory) and frequency x dominance values (in brackets).N, total number of individuals per m 2 ; S, total number of species; H', Shannon-Wiener diversity; J', Pielou evenness; C, carnivores; H, herbivores; D, detritivores; O, omnivores; TC, tidal condition; SSW, salinity of surface water; SBW, salinity of bottom water; OM, organic matter content; CARB, carbonate content; Q 50 , median grain size.N & S of trophic categories expressed in %.

Table 8 .
-Summary of biotic and physical characteristics of three studies carried out on Galician Rias with similar sampling methodology (values: mean ± standard deviation).S. type, sedimentary type (GR, gravel; CS, coarse sand; MeS, medium sand; FS, fine sand; MS, muddy sand, SM, sandy mud, M, mud); OM, Organic matter content (%); CARB, carbonate content (%); TC, tidal condition (S, Subtidal; I, Intertidal); N polychaetes, total number of polychaetes; N syllids, total number of syllids; S syllids, total number of species; Dom Species of Syllids, dominant species of syllids for each groups of sites.