The Battle of the Aegates Islands marked the end of the First Punic War and took place on March 10th 241 BC, when the Carthaginian relieving fleet was defeated off of western Sicily.

The environmental context of the area off-shore of Trapani, Marsala, and Mazzara del Vallo (west and southwest of Sicily) that comprises the Aegates islands is a good example of a shelf affected by marine abrasion of a rocky, tectonized substratum, where scarce recent sediment deposits are mainly composed of bioclastic fragments (Colantoni et al., 1993).

The ram on which our study focuses work was discovered on the inner continental shelf that joins the islands of Favignana and Levanzo with the mainland (Figure 1). The shelf has a gentle slope with several shallow banks and minor islets that rise from the seafloor as erosional remnants. Depth minima occur on the shoals of Secca del Toro, 6 m; Secca dei Pesci, 21 m; and NW of Levanzo, 29 m (Colantoni et al., 1993).

A naval ram (rostrum in Latin) is a breakthrough thrusting weapon that was fitted to the bow of ancient galleys to break the hull framing of enemy ships. The rams found in the Aegates Islands weigh 170 kg on average and vary in thickness from about 1.5 cm in the laminar region to 5 cm in the front. The artifact examined in this study, referred to as ram 13, is a trident rostrum characterized of three sharp and blunt cuts (Figure 2A). This type of ram was common among the main Mediterranean warships beginning in the 4th century BC (Tusa and Royal, 2012). Ram 13 is of great importance because since it shows a Punic inscription on the upper sheath. This is the second rostrum with a Punic inscription discovered to date (the other was the Egates 3) (Tusa, 2020).

The discovery of these remains of significant scientific value occurred during several research campaigns as part of the Egadi Project. Exploration was carried out by the Soprintendenza del Mare della Regione Sicilia in collaboration with technical SCUBA divers from the organization GUE (Global Underwater Explorers). The project was conceived and directed through the cooperative efforts of Prof. Sebastiano Tusa and Jeffrey Royal and their institution, the Soprintendenza del Mare, Regione Siciliana, and the RPM Nautical Foundation, from 2005 to 2018. The ram in this study was recovered in October 2017 between 75 and 95 m depth, about 7 km north-west of the island of Levanzo, together with rostrum 12 and 10 bronze Montefortino helmets (Tusa, 2020).

The ram was restored in 2019 using metals and alloys by Stefano Ferrari and Antonella Di Giovanni of the ICR Restoration Laboratory. The first stage of restoration involved sampling and documentation of the sediment blocks and biogenic materials collected inside the inlet of the artifact. Images of the concretions and the material accumulated in the ram were collected using a digital camera. The sediment blocks compacted with Posidonia fibers and bioconcretions were separated from the sandy sediment. Every block was wetted with water, cleaned with a brush to remove the sediment, dried, and sieved with a 0.5 mm mesh. Subsequently, samples were observed under a stereomicroscope. All the biogenic fragmented remains, shells and tubes, were sorted by higher taxon and preserved in Petri dishes.

Benthic specimens collected inside the entire ram were identified to species level whenever possible, and all data were used to construct a binary presence/absence matrix. The species richness of total benthos and of mollusks, polychaetes, and bryozoans, respectively, was considered a measure of α-diversity. All faunal data were analyzed by means of multivariate ordination technique non-metric multidimensional scaling (nMDS) using the Jaccard index. A clustering analysis based on Ward’s minimum variance method compared the similarity between the faunal assemblage found in the ram and those of common shallow (infralittoral) and deep (circalittoral and bathyal) habitats. Analysis of similarities (ANOSIM) based on Bray–Curtis similarity matrix assessed significant differences between grouping of habitats. A non-parametric SIMPER (Similarity Percentage) test identified those species that contributed most to the distinction among groups of habitats. Published literature sources provided information on species composition in each natural habitat and functional traits of each species on the ram (Table 2).

To infer possible colonization patterns of the artifact, we considered the main functional traits (Table 2) of the dominant species associated with the ram. We considered seven traits to describe the species’ niches: (1) reproductive mode (sexual, asexual), (2) development strategy (brooding, free spawning, eggs laid in capsules or masses), (3) larval type (planktotrophic or lecithotrophic, for species with pelagic larvae), (4) modularity (solitary, aggregation, aggregation of a few individuals, colonial), (5) adult motility (sessile, motile), and (6) engineering (primary constructor, binder, dweller), (7) size (large, medium, small, according the details described in the caption of Table 1). Information on the engineering role of each species was based on Fagerstrom (1988). Specifically, primary constructors refer to erect well-skeletonized builders that provide volume and rigidity to the concretion; binders are encrusters that expand and connect the organic structures; and dwellers are mostly motile (rarely sessile) organisms that are not strictly builders but inhabit cavities and crevices of the concretion. Moreover, the ecological affinity of each species was assigned by combining personal observations and literature data reported in Table 2.

The faunal assemblage in the ram included 114 species, including 58 species of mollusks (51%), 33 species of gastropods, 25 species of bivalves, 33 species of polychaetes (29%), and 23 species of bryozoans (20%). Table 1 provides a complete list of species identified from the ram and their functional traits and ecological affinities. Sessile species, i.e., bryozoans, serpulids, and a few bivalves, colonized the ram surface extensively and grew in epibiosis on calcareous surfaces of other organisms. Motile species, mostly mollusks, occupied the inside of the ram in large numbers together with the remains of sessile species (colony fragments, tubes) (Figures 2B–E). The percentage of the species sharing each of the functional traits is reported in Figure 3.

Sexual reproduction, pelagic spawning, and extended pelagic larval duration characterized most of the mollusks. Embryonic development in masses or capsules with short pelagic larval duration characterized about half of the gastropods, e.g., Alvania spp., Bittium reticulatum, Chauvetia giunchiorum, Jujubinus exasperates, and Turritella turbona. Almost all of the species were solitary and motile. A few bivalves were sessile and primarily contributed to the concretion, such as Chama gryphoides and Ostraeidae. In contrast, other sessile bivalves, i.e., Acar clathrata, Arca tetragona, Asperarca secreta, Pteria hirundo, and Striaca lactea, and all the motile mollusks were concretion dwellers. As for polychaetes, all species reproduced sexually, except for Filograna sp., Filogranula spp., and Josephella marenzelleri, which reproduced asexually. Most polychaetes were pelagic spawners with external fertilization and planktotrophic larvae and long pelagic duration, while only Filograna sp. and five Spirorbinae species were brooders that produced lecithotrophic, short-duration larvae. All polychaetes were sessile in the adult stage, and most were solitary. Among the serpulids, only Filograna sp., J. marenzelleri, and Spiraserpula massiliensis formed aggregations: Filograna sp. formed dense aggregations of small tubes in the ram. Spirorbid species occurred in a specific epibiothic association on mollusk shells, bryozoans, and tubes of other serpulids. Gregarious settlement occurred in the spirorbid Janua. Based on the tube size and their role in building the concretion, we considered serpulids important encrusting species that played the role of binders. This group included 20 species with medium and large tubes. We considered an additional 11 small-sized species from the genera Filogranula and Semivermilia and subfamily Spirorbinae as dwellers because they colonized crevices and interstices among calcareous surfaces and settled on top of shells and skeletons of other organisms. Bryozoans on the ram included species with sexual and asexual reproduction. Their colonies grew by asexual fragmentation and budding. Lecithotrophic larvae produced by sexual reproduction developed into new colonies. In the identified species, individuals placed eggs either into external brood chambers or retained them in the body cavity. Almost half of the bryozoan species grew on the substrate with erect-rigid colonies: Myriapora truncata, Pentapora fascialis, and Schizoretepora imperati with large colonies (more than 20 mm), and Adeonella pallasii, Cellaria fistulosa, Diporula verrucosa, Frondipora verrucosa, Hippellozoon mediterraneum, Hornera frondiculata, H. lichenoides, and Smittina cervicornis with medium-sized colonies (11–20 mm). These bryozoans formed new colonizable substrata and increased habitat heterogeneity with their three-dimensional structures. The other half of the bryozoan species formed monolaminar and multilaminar colonies encrusting the substratum, such as Cribilaria hinckinsii, Escharella variolosa, Onychocella marioni, and Palmiskenea skenei, firmly attached to the substratum with their large multilaminar colonies. Other species, e.g., Celleporina spp., formed cellepoliform colonies and, together with species with small-sized (less than 10 mm) monolaminar encrusting colonies, e.g., Cribellopora trichotoma and Escharoides mamillata, colonized mostly fissures and interstices.

The nMDS ordination analysis (Figure 4A) highlighted differences between benthic assemblages in three groups. The ram assemblage most closely resembled the assemblages in coralligenous and detritic habitats in group 1. This group separated from shallow photophilic habitats and zoosteracean meadows (group 2) on one side of the nMDS plot and deeper mesophotic and bathyal habitats and caves (group 3) formed a second group on the other side of the nMDS plot (Figure 4A). The cluster analysis (Figure 4B) supported the nMDS ordination by yielded three main clusters. The assemblage on the ram grouped together with coralligenous and detritic habitats in group 1, while the shallower habitats and the deeper habitats in groups 2 and 3 formed separate clusters. The distinction between groups occurred at about 82% similarity. A total of 67 and 54 species on the ram, respectively, had affinities for coralligenous reefs (COR) and detritic bottoms (DET); meanwhile, bathyal habitats were less represented on the ram, with 30 species (Figure 4C).

A between-habitat ANOSIM comparison confirmed significant differences among these groups (Global R = 0.857, mean rank within group = 5.5, mean rank between groups = 17.5, and p < 0.005). Simper identified the species contributing most (>50%) to dissimilarity among habitats, including 22 species responsible for separating groups 1 and 2 (average dissimilarity 59%) and another 28 species that separated groups 1 and 3 (average dissimilarity 67.8%) (Table 3). A mostly sessile group of 12 species of bryozoans, eight species of serpulids, and two species of gastropods differentiated groups 1 and 2; bryozoans that form erect rigid colonies (66.7% of the species) contributed 54.5% to dissimilarity, with encrusting species comprising the remaining 33.3%. Large- or medium-sized serpulids (except S. crenata) accounted for 36% of dissimilarity. In contrast, 16 species of mollusks, all gastropods, were primarily responsible for the dissimilarity between groups 1 vs. 3 (57.2%). Different shell sizes and a wide range of ecological affinities characterized these gastropods, including sciaphilic circalittoral and photophilous infralittoral habitats (except for Volvarina mitrella, Alvania subareolata, and C. giunchiorum, which show specific preference for coralligenous and detritic habitats). Ten species of bryozoans contributed to the dissimilarity (35.7%), half of which form erect colonies and the other half form encrusting colonies.

The benthic fauna associated with the ram included a highly diverse assemblage primarily dominated by mollusks and secondarily dominated by polychaete serpulids and bryozoans. The analysis of functional traits revealed that the majority of species in the assemblage were large, sessile, solitary, or colonial invertebrates, i.e., all the bryozoans, the serpulids, and only nine bivalves. Among them, the bryozoans with erect colonies played the role of primary constructors. This role was filled by very few species (H. mediterraneum, M. truncata, P. fascialis, and S. cervicornis), similar to what occurs on hard substrata in shallow photophilic and deeper habitats, as well as on coralligenous reefs, semi-dark caves, and mesophotic bioconstuctions. Bryozoans often act as frame builders with erect colonies resulting in distinct conditions, such as the facies of large arborescent bryozoans on coastal detritic bottoms or structures built by just a few species (Cocito, 2004; Bianchi, 2009). In contrast, in many bioconcretions, bryozoans play a role as secondary constructors (encrusters) and produce crusts rather than erect layers (Cocito, 2009). The ram exhibited this latter colonization pattern, where the bryozoans largely encrusted the substatum and played the role of binders. Species such as Escharina vulgaris, Onychecella marioni, E. variolosa, Cribrilaria setiformis, Cribrilaria hincksii, and P. skenei, as well as other small-sized colonies, e.g., Celleporina caminata, Celleporina lucida, and C. trichotoma bound together the concretion as well as detrital material. They expanded their mineralized colonies both on the ram (as primary substratum) and on organic calcareous surfaces (as secondary substrata), overgrowing colonies, shells, and tubes of other organisms. They cemented the underlying species and, in turn, were colonized by epibiontic organisms. Similarly, several medium- to large-sized serpulids (e.g., Serpula vermicularis, Serpula concharum, Protula tubularia, Spirobranchus triqueter, Vermilipsis labiata, Vemiliopisis infundibulum, Vermiliopsis monodiscus, and Placostegus tridentatus) contributed to build the basal framework of the concretion, settling with their calcareous tubes directly on the bronze substrate. Many serpulids grew as epibionts on calcareous skeletons and valves of other sessile organisms, thus acting as secondary constructors together with the bryozoans. In this way, serpulids and bryozoans enhanced the small-scale spatial heterogeneity and increased the three-dimensionality of the ram’s surface. They created new microhabitas, from laminar crusts to massive crusts rich in interstices, pits, and crevices. These microhabitats provided cryptic refuges that served as high-quality habitat for many dweller species.

The engineering category of dwellers constituted the majority of the ram assemblage, including the inhabitants of the interstitial spaces of the concretion and of the detritus collected inside the ram. Small-/medium-sized species formed a conspicuous component of this functional group because of their small dimensions, an effective adaptation to cryptic habitats. Serpulids (Semivermilia cribrata, Semivermilia crenata, Filogranula annulata, F. calyculata, F. gracilis, Janua eterostropha, and Pileolaria militaris) with their small tubes and bryozoans (Celleporina spp., C. trichotoma, and Escharoides mamillata) with small-sized colonies both settled in clusters on interstitial surfaces. Similarly, many small gastropods typically occur in shaphilic coralligenous (Alvania subaereolata, Stricteulima jeffreysiama, and Volvarina mitrella) and bathyal habitats (Anatoma umbilicata, Epitonium tiberii, Raphitoma pseudohystrix, and Talassia degueneti) and colonize interstices and crypts where they find refuge from currents, irradiance, and predation. In contrast, most mollusks of medium to large size colonized both the concretion and the detritus accumulated inside the ram. Here, many bathyal and cave-dwelling species (Asperarca nodulosa and Venus nux) co-occurred with other species typical of shaphilic detritic and coralligenous habitats (Abra prismatica, Astarte sulcata, Cardiomya costellata, Pecten jacobaeus, Mimachlamys varia, Palliolum incomparabile, Papillicardium papillosum, and S. lactea).

Our analysis of the reproductive and life-history traits of the species in the ram supports conclusions from previous studies on the colonization of artifacts. Most species (60%) reproduced sexually through pelagic fertilization and produced larvae with long pelagic durations that may undergo long-range dispersal in the intense deep currents in our study area (Suriano et al., 1992). Thus, we consider the ram an island-like habitat (Meyer et al., 2017) of solid substrate surrounded by a soft seabed. In fact, the ram interrupted the continuity of the soft seafloor and offered a surface suitable for larval settlement of hard-bottom species as well as adequate elevation above the bottom to expose organisms to stronger bottom currents and associated particulate food sources. Moreover, planktonic larvae advected by oceanographic currents were not the only dispersal mechanism for the ram fauna. The presence of a remarkably high percentage (40%) of brooding species on the ram (i.e., all the bryozoans, a few serpulids, and half of the gastropods with embryonic development inside attached capsules) suggested that recruitment on the ram also took place via short-dispersal planktonic propagules, small-sized adults, and fragments of colonies arriving from source populations in the area by passive drift. Moreover, we hypothesize the arrival of some species by migration from nearby habitats, particularly for large-sized motile gastropods from coralligenous habitats (e.g., Murexsul aradasii and Calliostoma conulus) and detritic bottoms (e.g., T. turbona and Bolma rugosa).

Based on the ecological affinity of the species, the benthic assemblage on the ram differed in similarity from those in infralittoral, circalittoral, and bathyal habitats. All these habitats used as comparisons have been reported in the waters of the Aegadian archipelago. The Aegates Islands seabeds represent an area of great ecological value, with several endemic habitats protected by EU regulations and effectively managed through the designation of the largest Italian MPA (Aegates Islands MPA. Nevertheless, knowledge gaps remain regarding the distribution, biodiversity, and ecological status of benthic communities in the area. Uniformly distributed Posidonia oceanica meadows in the infralittoral zone cover the sandy seabed to depths over 30 m. Conversely, algae-dominated (i.e., brown algae belonging to the genus Cystoseira) belts develop at shallow depths and characterize photophilic rocky floors (Catra et al., 2006). The presence of the Cystoseira spp. thalli shapes environmental features and creates suitable conditions for invertebrate settlement (Sanfilippo et al., 2017). Coralligenous reefs occur mainly on steep walls, hard bottoms below 40 m depth, and shoals around the whole archipelago and represent the most attractive seascape for diving tourists (Cocito et al., 2014). Among the 200 Rhodophyta species reported in the algal checklist of the area (Catra et al., 2006 and the references therein), widespread distributions across all islands of the archipelago characterize the Corallinales and Peyssonneliales that form the basal layer of coralligenous reefs. Recent distribution patterns of encrusting red algae likely reflect the spread of coralligenous reefs in the study area. The carbonatic lithology of the islands supports several submerged caves that originated through karstic phenomena (Gerovasileiou and Bianchi, 2021). Furthermore, organogenic detritus composed of animal debris enriches the carbonate sandy bottom sediments of the Aegadian archipelago, likely stemming from the erosion of coralligenous reefs and marine cave communities. Recent work described deep and cold-water coral assemblages in seamounts and banks surrounding the Aegetes islands between 240 and 300 m depth (Bo et al., 2014; Angiolillo et al., 2021). Although this previous work reported Porifera and Cnidaria as the most frequent taxa, our study did not find them. Angiolillo et al. (2021) highlighted the importance of hard bottoms (sparse boulders and wrecks) for the richness and diversity of the megabenthic assemblages, resulting in high ecological value and conservation interest for the whole area.

All of these factors play a key role in understanding the composition of the ram assemblage. Indeed, the distribution and bathymetric range of the benthic communities in the region explain the higher affinity in species composition of the ram assemblage with coralligenous reefs and detritic bottoms, which form the group 1 in the nMDS plot and in the cluster analysis. On the one hand, both coralligenous and detritic habitats thus represent the main source populations that would have provided the larval supply necessary for colonization of the ram; in fact, pelagic spawning species with long-lived pelagic larvae dominated. On the other hand, the strong regional hydrodynamic regime, which presumably promoted the transport of both propagules and fragments or mineralized remains inside the ram, can partially the presence of species whose affinity links shallow habitats, such as Posidonia meadows. In support of this hypothesis, the facies dominated by Laminaria rodriguezii reported in the region (Suriano et al., 1992; Araújo et al., 2016) indicate the presence of high-speed bottom currents on the surrounding seabed.

These results offer insights regarding the expected timeframe for a submerged wreck to match natural habitats. Previous research showed that the benthic invertebrate communities on shipwrecks up to more than a century old do not match the background community (Perkol-Finkel and Benayahu, 2004; Perkol-Finkel et al., 2005, 2006). Indeed, wrecks generally have lower functional diversity than natural habitats and are dominated by species with long pelagic larval duration and/or asexual reproduction by fission. Our study showed that a ram that has accumulated biota over many centuries hosts a community with high functional diversity and species that occur in a range of surrounding natural habitats.