Comparative Analysis of Mollusc Assemblages from Di ﬀ erent Hard Bottom Habitats in the Central Tyrrhenian Sea

: Composition, trophic structure, and species-substrate relationships of molluscan assemblages inhabiting di ﬀ erent hard bottom habitats ( Sabellaria alveolata reef, photophilic bottoms, Phyllophora crispa sciaphilic assemblage, and coralligenous bioconstruction) were studied in two di ﬀ erent sites of the Tyrrhenian Sea. In particular, molluscs from the Sabellaria alvevolata (Linnaeus, 1767) reef and coralligenous concretion were investigated, testing the hypothesis that bioconstructions increase the diversity and abundance of associated biota compared to the surrounding habitats. A total of 3134 individuals belonging to the classes of Polyplacophora (5 species, 24 individuals), Bivalvia (39 sp., 2734 ind.), and Gastropoda (53 sp., 376 ind.) were identiﬁed. These three taxonomic groups showed di ﬀ erent distribution patterns in the studied habitats. Multivariate analyses revealed signiﬁcant inter-habitat di ﬀ erences in the composition of mollusc assemblages, especially between bioconstructions and the other habitats. S. alveolata and coralligenous host the highest rich molluscan fauna when compared to the neighboring hard bottom habitats characterized by photophilic and sciaphilic assemblages. The ﬁrst ones were dominated by bivalve suspension feeders, mainly represented by sessile and sedentary organisms, which act as bio-constructors, bio-eroders, or simply inhabit the several microhabitats provided by the bioconstructions, while the second ones host a rich molluscan fauna dominated by gastropod grazers and predators. The present study increases the comparative knowledge of molluscan assemblages inhabiting habitats of littoral plans of the Mediterranean Sea, providing pivotal information regarding biodiversity of coastal zones.


Introduction
In marine ecosystems, any structure built by living organisms that rises from the bottom toward the surface may be defined as bioconstruction, which strongly modifies the local environment through physical and ecological processes [1].
Although corals and calcareous algae are usually indicated as the builders of biotic reefs, barnacles, molluscs, encrusting bryozoans and tubeworms also play key roles in the creation of complex biogenic structures, especially in the Mediterranean Sea [2]. Ingrosso et al. [3] reviewed the most important biogenic habitats along the Italian coasts, and found that a total of seven different formations develop marine habitats and ecosystems is pivotal, especially in the light of the increasing seawater pollution of coastal zones.
The aim of this work is to analyze molluscan assemblages inhabiting five different Mediterranean habitats, assessing both molluscs' diversity and their functional trophic role in order to test the effect of habitat complexity.

Study Area and Sampling Methods
The study was carried out in two different sites in the central Tyrrhenian Sea (Figure 1). One is located along the Latium coast (Lido di Ostia, 41 •   At Lido di Ostia, samplings were performed in two habitats at 3 m depth: on a continuous S. alveolata bioconstruction (SABEL), developing parallel to the shoreline [7], and on the surrounding hard bottom (HARDB), represented by rocks very poorly covered by photophilic green and brown algae. The Sabellaria alveolata reef at Lido di Ostia reaches the greatest dimension among those reported in the Mediterranean Sea [44], and therefore can be used as a model for studies concerning the associated fauna.
At Punta del Lazzaretto, samplings were performed in three habitats, which characterize the substrate from 10 m to 40 m depth [14]. The rocky bottom at 10 m depth is characterized by shallow photophilous infralittoral algal assemblages (PHOTO), dominated by Padina pavonica (Linnaeus) Thivy, 1960 and Acetabularia acetabulum (Linnaeus) P.C. Silva, 1952, while at 30 m depth, the rocky bottom is covered by Phyllophora crispa sciaphilic assemblage (PHYLL) or by coralligenous bioconstructions (CORAL). The P. crispa turf was exclusively reported by Bianchi et al. [12] and Casoli et al. [14] along the coasts of Sardinian and Tuscany; coralligenous bioconstruction at Punta del Lazzaretto, in the proximity of the P. crispa turf, is extremely relevant in dimension ( [45] and references therein).
Samples were collected by SCUBA divers during late spring and summer of 2014. On SABEL, HARDB, PHOTO, and PHYLL, four replicate samples were collected by scraping the bottom within a 20 × 20 cm frame for a total thickness of 5 cm, using a hammer and chisel, and holding a 50 µm mesh bag below the sampling plot in order to minimize the loss of organisms. Sampling efforts were reduced on CORAL to four replicate samples with a surface of 10 × 10 cm and a thickness of 10 cm [14]. Then, samples were stored in 4% formaldehyde seawater solution to be transported to the laboratory, where the macrobenthos was sorted. All molluscs found were preserved in 70% alcohol solution and then identified to the lowest possible taxonomical level under a Leitz stereomicroscope.

Relation to the Substrate and Trophic Groups
Functional groups concerning molluscs' motility (sessile, sedentary or vagile) and feeding guilds were assigned to the identified species. According to Donnarumma et al. ([35] and references therein), the following categories were considered: Micro-and Macro-grazers (MG), feeding on both diatoms and/or algae growing on rocks, on shells or on large plants; Deposit feeders (DF), feeding on both diatoms and/or microalgae of sandy sediments; Predators (P), actively feeding on sedentary or motile animals; Scavengers (SC), feeding on dead animals; Ectoparasites (E), feeding on much larger animals on which they live during their life cycle; Suspension feeders (SF), feeding on the organic particles suspended in water, which mainly comprise of bacteria, diatoms, and algal tissue, or, in coastal habitats, nano-zooplankton.

Data Analysis
Species-area curves were produced for each habitat to evaluate the efficiency of sampling. Molluscan assemblages were analyzed using sinecological indices, such as the species richness (SR), the number of individuals (N) per 2 dm 3 , the Pielou's Evenness (J), and the Shannon-Weaver diversity (H'). The quantitative (DI, percentage of individuals of a given species upon the total of individuals) and qualitative (DQ, percentage of species of a given taxon upon the total of species) indices were also calculated. Further habitat comparisons were carried out by sample-based and individual-based interpolation (rarefaction) and extrapolation curves, following Colwell et al. [46] and Chao and Jost [47]. Differences on each sinecological index among habitats were assessed through a permutational analysis of variance (PERMANOVA), based on Euclidean distance and performed as univariate analysis [48]. A one-way model was used with habitat as a fixed factor. With the same design, PERMANOVA analysis based on Bray-Curtis similarity was used to test differences in the mollusc assemblages among the five habitats with four replicates. The pairwise test was performed to further highlight differences among habitats, and non-metric multidimensional scaling (n-MDS) ordination [49] was then plotted. Data were transformed using log (x+1) to reduce the effect of the dominant species in the samples [50], and 4999 permutations of the raw data units were computed to obtain p-values. A similarity percentages-species contribution analysis (SIMPER) was carried out to identify the species that mostly contribute to the similarity among habitats, as well as the species that mostly characterized each one [50].

Composition of Mollusc Assemblages
A total of 3134 individuals were collected from the five examined habitats. The taxonomic identification yielded 96 species, 62 of which were recorded in only one of the five habitats and were represented by one or a few individuals; only two species, the gastropod Bittium reticulatum and the bivalve Musculus costulatus, were widespread in all habitats (Table 1).      (Figure 2a). A high number of very recently settled juveniles of M. galloprovincialis were detected in SABEL (125 ind.) and in HARDB (201 ind.), even though they were not considered for statistical analyses in order to exclude the background noise. -----

Species-Substrate Relationship
Among the 96 species belonging to the three taxonomic groups, vagile species were 53.13%, sessile species 30.21% and sedentary species 16.67%.
On the contrary, gastropod assemblages were more greatly represented by vagile (96.23%) than sessile (3.77%) species. The first group, with 51 species and 367 individuals, was detected in all habitats, while the second, exclusively belonging to the Vermetidae family (Thylaeodus semisurrectus and Vermetus granulatus) with 9 individuals, was found only in CORAL habitats. As for polyplacophorans, they were exclusively sedentary and dominated the CORAL habitat.

Inter-Habitat Comparison of the Molluscan Assemblages
Species accumulation curves (supplementary material, Figure S1) showed that the molluscan assemblages were well-represented by the four replicate samples in the studied habitats. The mean values among replicates of sinecological indices ( Table 2) were higher in the bioconstructions (SABEL and CORAL) when compared to the neighboring habitats, except equitability (J) at Punta del Lazzaretto. Sample-based and individual-based interpolation (rarefaction) and extrapolation curves (supplementary material, Figure S2) corroborate the results of habitat comparisons. The PERMANOVA test on the sinecological indices displayed significant differences among habitats (Table 2). A significant difference was also detected when analyzing mollusc composition (PERMANOVA test: df = 4, F = 9.9816, p = 0.0002; Pairwise comparison, Table 3). Their spatial differences were highlighted by the n-MDS plot (Figure 4), which showed a clear separation between all the five habitats, while the four replicates of each habitat were consistent each other. The similarity percentages-species contribution analysis (SIMPER) indicated a similarity among habitats of 22.08%, due to 2 bivalves and 2 gastropods (Table 4), while for each habitat the similarity of dominant species among replicates ranged from 37.17% to 68.95% (Table 5).

Discussion
The present paper provides quantitative data on molluscan assemblages from relevant coastal habitats which are widespread in the Mediterranean Sea. Sabellaria and coralligenous bioconstructions hosted richer molluscan fauna, dominated by suspension feeder bivalves, when compared to the neighboring hard bottom habitats characterized by photophilic and sciaphilic assemblages, dominated by grazer and predator gastropods.
S. alveolata bioconstruction revealed a huge variety of associated bivalves that shed light on the complex ecological system of the reef. This grows at a very shallow depth (from the mesolittoral to around 3 m) where high-water movement contributes to the transport of sediment and suspension of organic nutrients, promoting the occurrence of associated suspension feeders which are able to remove micro-particulates from the water column [51]. The suspension feeders M. galloprovincialis, H. arctica, and S. lactea were the main components of mollusc assemblage. M. galloprovincialis is well-known as the food competitor of S. alveolata, contributing to the regression of the bioconstruction [9], while the high abundance of H. arctica and the occurrence of other species of suspension feeders, such as Parvicardium spp. and Venerupsis corrugata, are indicators of the softness of the reef that is made of poorly adherent sand grains. Indeed, H. arctica is a soft rock borer, while the latter two are a soft-bottom burrower species [52], accounting for the presence of loose sandy sediment on the reef. Furthermore, the occurrence of the rock-boring bivalve R. dubia indirectly suggests the presence of carbonate fragments, probably shells or coarse gravels, on the S. alveolata bioconstruction. Concerning gastropods, the high occurrence of micrograzers belonging to the genus Bittium have been also detected by Gravina et al. [53] for Sabellaria spinulosa (Leuckart, 1849) reefs [54,55].
Overall, molluscs abundance and species richness found in the investigated reef were markedly higher than those previously reported for other Mediterranean and Atlantic Sabellaria reefs [9,11,53,56].
As for the hard bottoms surrounding the S. alveolata reef, characterized by photophilic algae, the highest dominance of M. galloprovincialis influenced the dynamic of associated faunal assemblages. There is evidence in literature that mussels alter the bottom topography, causing both positive and negative effects on the associated biota. They offer suitable substrate for the colonization of both algae and invertebrates, but at the same time, preempt the space preventing the settlement for other kinds of sessile suspension feeder species [57][58][59][60]. In particular, the dominant species which were detected here, also associated with the suspension feeder M. galloprovincialis, were its direct predator, the shell-boring gastropod O. edwardsii [61], and the micrograzer gastropod B. reticulatum. According to Huston [62], the dominance of very few species in mollusc assemblages suggests a remarkable species competition, which, in this case, increases from the Sabellaria reef to the surrounding hard bottom, together with a decrease of species number and evenness (Sinecological indices, Table 2).
According to Gili and Coma [63], suspension feeders play a pivotal role in coralligenous reefs. Indeed, coralligenous, like in Sabellaria reef and the surrounding hard bottom, reveals a rich habitat mainly characterized by high bivalves abundance and diversity. This richness is due to species which contribute both to the biological accretion and erosion of coralligenous itself. The sessile and byssate species detected, such as Barbatia barbata and S. lactea, and cemented ones, such as Spondylus gaederopus, T. semisurrectus, and V. granulatus, act as secondary builders, due to their abundance and shell structure [64]. On the other hand, boring bivalves such as L. lithophaga, P. lithophaga, and R. dubia settled in this calcareous formation, bore the calcareous substrate by their glandular secretions and act as bio-eroders [65][66][67]. The occurrence of sand-burrowing bivalve species, such as Venus casina and Papillicardium papillosum, accounts for the heterogeneity of bioconstruction, well-known as a promoter of biodiversity [35,[68][69][70].
The presence of sandy sediments occurring in the coralligenous habitat favors the accumulation of organic materials and provides optimal condition for gastropod deposit feeders, such as Alvania spinosa, Alvania tenera, S. costata, and the most abundant species, C. semistriata. According to Terlizzi et al. [71], this latter species also characterizes the coralligenous reefs along Apulian coasts. In addition, the herbivore polyplacophorans were found almost exclusively on coralligenous habitats, due to the availability of their food source, such as diatoms and red algae [72,73]. Moreover, when compared with other habitats, the coralligenous also shows a high species richness belonging to several trophic groups, such as micro-and macro-grazers, carnivores, predators, and ectoparasites, sustaining complex food webs [15,74,75].
The two algal-dominated hard bottom habitats sampled at Giglio Island greatly differ from the others (Figure 4), as vagile gastropods were the main components of mollusc assemblages instead of bivalves.
From the literature, the turf of the sciaphilic red alga P. crispa provides a complex habitat due to the algal architecture, supporting several algal and animal epiphytes as food of herbivorous and carnivorous gastropods, respectively [4,14]. Zaitsev [76] and Bonifazi et al. [4] reported a rich and heterogeneous community characterized by more than 140 animal species in the Black Sea and in the Tyrrhenian Sea, respectively. Our results are consistent with these data, stressing the importance of the micrograzer B. latreillii and the predator Pusia savignyi as the main gastropods of the Phyllophora algal turf. Here, although less abundant, byssate bivalves were mainly represented by the sessile S. lactea and M. costulatus, occurring on substrate where the Phyllophora alga was attached, and the sedentary F. hyalinus, settling on the algal tallus [77].
The photophilic habitat dominated by algae P. pavonica and A. acetabulum, when compared with all the other studied habitats, shows the lowest values of both species richness and abundance of mollusc assemblages. This is consistent with some literature [37,78,79], although Donnarumma et al. [40] recently described photophilic hard bottoms from different Italian Marine Protected Areas as the richest habitat of coastal environments in terms of gastropod species and feeding guilds. This might probably be due to the different sampling procedure adopted in the field (air-lift pump rather than classical method of scraping hard bottom). Several authors reported how the complexity of algal architecture contributes to diversifying molluscan assemblages [30,80,81]. In this case, the very simple shape of tallus provided by P. pavonica and A. acetabulum affects the low abundance of molluscan fauna, mainly represented by the two congeneric micrograzers, B. latreillii and B. reticulatum and very few other species.

Conclusions
In conclusion, our findings clearly highlight that food availability and high habitat complexity due to biological structures growing on substrates may represent the key components determining mollusc assemblages, occurring from the infralittoral to the circalittoral plans of the Mediterranean Sea. Micro-topography and the tridimensional shape of bioconstructions (S. alveolata reef and coralligenous) promote the settlement surface and sheltered microhabitats which are mainly exploited by sessile bivalve-dominated assemblages, as well as suspension-feeding behavior. Instead, algal and their epiphytes, dominating the photophilic hard bottom, represent key factors of gastropod-dominated assemblages, mainly characterized by vagile micro-and macro-grazers.