These prokaryotic microorganisms are largely present within the endolithic assemblages responsible for the bioerosion of submerged stone artifacts. Many species of marine cyanobacteria have epilithic, endolithic or strictly euendolithic behavior. The species can grow inside the stone showing different patterns and levels of penetration, in relation to the microenvironmental features and their peculiar ecological needs. Although small in size (only a few micrometers), they form dense layers or large colonies (Figure 2A), with a considerable destructive action (Radtke and Golubic, 2005).

Endolithic cyanobacteria and their traces (i.e., holes, cavities and tunnels) (Figures 2B–D) were largely found in the submerged Archaeological Park of Baiae on statues, architectural and sculptural elements, mosaics and opus sectile pavements (Ricci, 2004; Ricci et al., 2009; Davidde et al., 2010; Ricci et al., 2013; Ricci et al., 2016a). This group was also observed on the marble statues of the decorative apparatus of the Grotta Azzurra of Capri (Sacco Perasso et al., 2015) and on several statues and stone artifacts recovered from the wreck of Antikythera (Davidde Petriaggi et al., 2017).

The ichnospecies and related biospecies found in the studied sites are listed below.

Scolecia filosa Radtke, 1991 is the trace of Plectonema terebrans Bornet and Flahault, 1889 (Ord. Oscillatoriales).

Scolecia filosa was observed on statues and architectural structures collected from the Underwater Park of Baiae (Ricci et al., 2013; Ricci et al., 2016a). This ichnospecies was mainly constituted by unbranched galleries penetrated into the stone up to one millimeter. The filaments, 1-1.5 µm in diameter, were long, flexuous, forming dense interwoven networks.

Eurygonum nodosum Schmidt, 1992 is the trace produced by the cyanobacterium Mastigocoleus testarum Lagerheim, 1886 (Ord. Stigonematales).

This ichnospecies showed curved or straight tunnels, 5-10 µm in diameter, with lateral branches and nodular swellings, 4-5 µm thick. It formed dense carpets with orthogonal branching, and lateral/terminal swellings, 6–10 µm in diameter, as traces of heterocysts. The network of tunnels was located orthogonally to the surface, up to 200- 300 µm in depth, decreasing in density in the inner parts of the substratum. Eurygonum nodosum was observed on statues and architectural structures from the underwater park of Baiae (Ricci et al., 2013; Ricci et al., 2016a).

Fascichnus dactylus Radtke, 1991 is the trace of Hyella caespitosa Bornet & Flahault, 1888 (Ord. Pleurocapsales).

The ichnospecies showed thick and straight branched tunnels, 5-8 µm in diameter, forming radiating bundles or expanded carpets (Figures 2E, F). The tunnels were slightly undulated with rounded ends. Fascichnus dactylus was observed on limestone tesserae collected from mosaic floors of the Villa dei Pisoni in Baiae (Ricci et al., 2016a).

This group of photosynthetic microorganisms played an important role in the bioerosion of stone artifacts. Several species dug tunnels and cavities up to the depth that allowed them to carry out photosynthesis. In many cases green algae bored deeper in the stone using the walls of cavities eroded by macroorganisms, such as sponges and bivalves.

The ichnospecies and related biospecies identified in the studied sites are listed below.

Ichnoreticulina elegans Radtke 1991 is produced by the siphonal green alga Ostreobium quekettii Bornet and Flahault, 1889 (Ord. Bryopsidales).

This ichnospecies showed several morphologies according to the different life stages of the biospecies O. quekettii. The first one was characterized by straight or winding filaments, 4-5 µm in diameter, narrowing gradually in distal direction. In the second stage, the filaments developed branches, 2-5 µm in diameter, forming zig-zag shaped networks (Figures 3A, B). The presence of numerous branched filaments produced dense horizontal networks of dorso-ventrally flattened borings (typical zig-zag pattern). Arch-like branches connected with the substrate surface were also observed. Several widenings and swellings with different shape and size, typical of the mature Ichnoreticulina, were frequently observed.

The ichnospecies was found in Lunense marble statues, in slabs of opus sectile floors (Cipollino, Penthelic, Paros, Docimio, Lunense, and Ancient Red Tenario marbles) collected from the site of Baiae (Ricci et al., 2009; Davidde et al., 2010; Ricci et al., 2013), and in limestone tesserae from Baiae (Ricci et al., 2016b). Its traces were also found in marble statues of the Grotta Azzurra, Capri (Sacco Perasso et al., 2015). Since Ichnoreticulina elegans is a key-ichnospecies for deep and dark habitats, its presence on the statues of the Grotta Azzura confirms the scarce lighting of the bottom of the cave in which artifacts were exposed

Fascichnus grandis Radtke, 1991 is the ichnospecies of Acetabularia acetabulum (Linnaeus) P.C. Silva, 1952 (Ord. Dasycladales).

Acetabularia acetabulum is a unicellular eukaryotic green alga which grows in sheltered environments on hard substrates. It lives in shallow waters in the Mediterranean Sea, Red Sea, eastern Atlantic, and the Indian Ocean (Guiry and Guiry, 2013). The life cycle has an adult stadium consisting of a rhizoidal part, with the nucleus in one of its branches, a thin stalk, and a reproductive cap, shaped like an umbrella, containing gametangial cysts (Van den Hoek et al., 1995).

The species has a semi-endolithic habitus and settles by means of branched rhizoids which can penetrate carbonate substrates probably due to chemical actions, as for cyanobacteria (Garcia-Pichel, 2006; Garcia-Pichel et al., 2010). The endolithic penetration of rhizoids was well documented in works that described the bioerosive activity through the resin embedding-casting technique (Radtke et al., 1997; Perry and Macdonald, 2002; Radtke and Golubic, 2005; Radtke and Golubic, 2011; Casoli et al., 2015).

The boring role of A. acetabulum was observed on numerous artifacts in Baiae (Figure 3C) (Ricci et al., 2016a; Ricci et al., 2016b), underlining its ability to bore stones with different chemical composition. In particular, it was observed a semi-endolithic habitus of the rhizoidal apparatus (Figures 3D–F), in marble samples with a lower depth of penetration and a smaller number of branches than other calcareous stones, less hard like limestone. These results confirmed that the boring pattern is strongly influenced by the characteristics of the substrate, as already highlighted by Golubic et al. (1975).

Marine fungi are chemotrophic microorganisms that live independently of light. Their growth in the submerged lapideous artifacts is linked to the presence of organic matter inside the stone, as food source. For this reason, fungi are an euendolithic group able to penetrate mineral substrates (Golubic et al., 2005). Gleason et al. (2017) reported an exhaustive review on the role of endolithic fungi in bioerosion and disease in marine ecosystems. They described the importance of fungi in natural substrates as components of the marine calcium carbonate cycle and bioerosion phenomena, because of their contribution to the biodegradation of calcium carbonate shells/skeletons and calcareous geological substrates.

Fungi traces are produced by the thin exploratory hyphae, more or less uniform in diameter, quite similar to of the filaments of cyanobacteria, and bag-shaped swellings with spores dispersion function (Figure 4A), whose shape and dimension are useful for the identification (Wisshak et al., 2019).

Traces of euendolithic fungi were often found in stone artifacts in association with cavities produced by other microorganisms and boring sponges, which had provided a food source and had modified the petrographic features of the stone, favoring their settlement (Ricci et al., 2013; Sacco Perasso et al., 2015; Ricci et al., 2016a). Studies on microboring fungi showed that these microorganisms are generally slow in colonizing newly exposed substrata, becoming more abundant only after one year of exposure (Kiene et al., 1995; Gektidis, 1999; Casoli et al., 2015).

The ichnospecies and related biospecies identified in the studied sites are listed and described below.

Orthogonum fusiferum Radtke, 1991, is the ichnospecies produced by the fungus Ostracoblabe implexa Bornet and Flahault, 1891.

This ichnospecies is characterized by straight and branched filaments, more than 100 µm long, with diameter ranging from 1 to 4 µm, spreading parallel to the substrate surface, with bean-shaped swellings (Figures 4B, D). The ichnospecies was very frequently found in marble artifacts in Baiae and Capri (Ricci et al., 2013; Sacco Perasso et al., 2015; Ricci et al., 2016a). In particular, the study carried out on the Grotta Azzurra statues showed the presence of O. fusiferum in association with boring phototrophics (i.e., I. elegans) and sponges (Entobia isp.) (Sacco Perasso et al., 2015). The relevant presence of fungal traces was related to the micro environmental conditions present on the sea bottom of the grotto. The fungal colonization can be considered a phase successive to the colonization by other micro- and macro-organisms.

Saccomorpha clava Radtke, 1991, ichnospecies produced by the fungus Dodgella priscus Zebrowski, 1936.

This ichnospecies was observed in the ichnocoenosis observed on some marble statues of the Grotta Azzurra (Sacco Perasso et al., 2015). The traces were single club-shaped sacs, with diameter ranging from 25 to 40 µm, penetrated vertically into the material (Figure 4C). Globular swellings were connected to the surface by a narrow neck, isolated or together in small groups, referable to the fungal reproductive structures.

Saccomorpha sphaerula Radtke, 1991, ichnospecies of the biospecies Lithopythium gangliforme Bornet and Flahault, 1889.

This ichnospecies was frequently observed in association with Ichnoreticulina elegans in some marble statues of the Grotta Azzurra (Sacco Perasso et al., 2015). The traces were sphere/pear-shaped swellings, 7-9 µm wide, placed perpendicularly to the substrate surface. Several thin tubes, originating at the base, spread radially to interlink the sacs.

Sponges are among the most common and destructive boring organisms affecting archeological artifacts (Figures 5A–E). Any type of calcareous substrates immersed for long periods (i.e., archeological structures) can be colonized by sponges which typically characterize the late successional phase of borer communities (Chazottes et al., 1995). The damage caused by sponges is superficially extended but often produces deep cavities in the substratum thus affecting a large part of the archeological remains (Figures 5A–C) (e.g., La Russa et al., 2013; Davidde Petriaggi et al., 2017).

Endolithic sponges are often difficult to be detected in situ due to their cryptic habit (Marlow et al., 2019). On the surface, they produce papillae which represent the openings of their aquiferous system; through these circular holes the water enters carrying oxygen and food inside the sponge, and it exits transporting waste products outside. These papillae appear as small dots (often only a few mm in diameter), thus difficult to be detected, even though sometimes they can be vividly colored (yellow, red, purple or green), and more easily visible. This type of organization is generally called “alpha stage”, a typical growth organization of most of the Mediterranean boring species. Conversely, some species may also grow in beta and/or gamma and delta stages. In the beta stage, the sponge completely covers the substrate developing a thin ectosomal layer connecting papillae, while in the following gamma stage, the sponge completely erodes the substrate and grows freely like a non-endolithic sponge, with a massive shape; sponges in delta stage live partially buried in sediments and incorporate particles. Although the last kinds of organizations are mainly typical of tropical species, in the Mediterranean beta and gamma stages are characteristic of the sponges Cliona viridis (Schmidt, 1862) and C. celata Grant, 1826 (Cerrano et al., 2007; Evcen and Çinar, 2015).

In the Mediterranean Sea, the identification of boring sponges at the genus level can be relatively easy, just basing on their macroscleres (i.e. larger skeletal elements) (Figure 6A–D); for example, Siphonodictyon spp. present only oxeas, Cliona spp. only tylostyles, Pione spp. tylostyles and microxeas, etc. However, the identification at the species level is more problematic and additional features, such as microscleres (i.e., smaller skeletal elements), are mandatory for their recognition. Microscleres, in some cases, may be very rare and even if microscleres and macroscleres of boring species can be found in the etched chambers (Figures 6C, D), other not-boring species can exploit existing holes: in this case, the observed spicule set may actually belong to several different sponge species.

Bioerosion traces can be attributed to different perforating sponge taxa, based on their morphology and organization. Most of Porifera, being modular organisms, typically produce numerous serial erosion chambers (Figure 5D) connected to each other, creating a macro pattern of bioerosion, and to the surface of the substrate through the papillae. These features can be useful diagnostic elements since many species are characterized by cavities with a peculiar shape and/or size (Bromley, 1978; Calcinai et al., 2007). In fact, many erosion traces are attributed to different ichnospecies belonging to the ichnogenus Entobia (biogenus Cliona) (Figure 6B) considering the morphological characteristics of the cavity networks (Bromley and D’Alessandro, 1984).

Nevertheless, it is important to keep in mind that sponge erosion patterns may vary with the type of substrate (Bromley and D’Alessandro, 1984; Schönberg, 2000), the proximity of other boring organisms, or when the available substrate for the boring agent is spatially limited (Bromley and D’Alessandro, 1984). Morevover, the same erosion pattern may be produced by different species (Färber et al., 2016). For these reasons, the taxonomic identification of boring sponges should be conducted considering not only the erosion pattern but also the skeleton characteristics. On the contrary, the attribution of a cavity to the taxon Porifera is possible in light of the presence of the typical microbioerosion signs (pits) on the chamber walls (Figures 6A–D). These scars are exclusively produced by sponges and, in some cases, can be attributed to specific genera according to the different pattern (Calcinai et al., 2003; Calcinai et al., 2004). In this way, an erosion cavity can be traced back to the sponge action even if no sponge (and so no diagnostic spicules) are longer present inside the cavity, as happens in fossil structures.

The family Clionaidae is the taxon including most of the excavating sponges. Cliona celata and other species belonging to this genus are the most frequently reported sponges into archeological artifacts (Ricci et al., 2007a; Ricci et al., 2007b; Ricci et al., 2009; Ricci and Davidde, 2012; Ricci et al., 2016a). In Baiae, this species was responsible for intense and extensive damages on statues (Figure 5A) and other artifacts, destroying in many cases the original sculptural model (Ricci et al., 2009; Davidde et al., 2010; Aloise et al., 2014). Its growth was also observed on mosaic floors of the same site (Figure 5E). In addition, some studies carried out on tesserae showed that C. celata eroded the limestone, causing a loss up to 80% of the stone material (Ricci et al., 2007b).

Among the excavating taxon of Clionaidae, Dotona pulchella mediterranea Rosell and Uriz, 2002 (Figure 6C), Cliona janitrix Topsent. 1932, Cliona schmidtii (Ridley, 1881) (Figures 5D, 6D), Cliona vermifera Hancock, 1867, Pione vastifica (Hancock, 1848), and Spiroxya levispira (Topsent, 1898) were identified into archeological artifacts recovered from Grotta Azzurra and Antikythera shipwreck (Sacco Perasso et al., 2015; Davidde Petriaggi et al., 2017; Calcinai et al., 2019). These endolithic species were often found in association with other unidentified species belonging to the genera Alectona and Siphonodictyon in marble statues from Antikythera (Calcinai et al., 2019).

Sacco Perasso et al. (2015) and Calcinai et al. (2019) documented how erosion chambers in archeological material from Antikythera and the Grotta Azzurra have been secondarily occupied by non-boring sponges and the assignment of the agents responsible for the erosion of the artifact may be uncertain. When the artifact crosses several phases of cover-up, often the chambers and cavities may be empty, devoid of sponge tissue and so devoid of diagnostic characters; it happens when the original sponge died and the tissue with the spicules has been washed away, depriving us of important taxonomic characters. Another problem compromising species identification is related to the cultural value of the material which prevents us from sampling large quantities. In these cases, the species identification can be impossible and only the observations of the micro-erosion pattern (type of pits) and of the spicule set directly under the SEM (bypassing the traditional method for spicule preparation) may support the identification, but often only at the genus level (Calcinai et al., 2019). For perforating sponges, in fact not only the microscopic pattern, but also the macroscopic boring pattern (i.e., shape and organization of the erosion chambers) represent additional characters useful for taxonomy (Figure 5D, Figure 6B).

For this reason, the embedding casting procedure is useful in sponge erosion studies, allowing for the accurately reproducing of the erosion traces (Figure 6B), also because they can be observed using SEM, highlighting the superficial ornamentation produced by the sponges themelves.

Mollusca is one of the most species-rich and widely distributed phylum in the Mediterranean Sea, encompassing around 2000 species (Coll et al., 2010): such biodiversity reflects several adaptations to different environmental features and food sources. Bivalves, which among Mollusca are represented mainly by sessile or sedentary benthic organisms, show a variety of relationships with the hard substrates colonized. For example, they adhere to the rock as epilithics, actively perforate the substrate as euendolithics, creep into pre-existing cavities produced by other organisms or vacated by boring bivalves as nestlers (such as Coralliophaga lithophagella, Lamarck 1819, Barbatia barbata, Linnaeus 1758) not acting as borers (Morton, 1980; Morton, 1982; Morton, 2014).

Endolithic bivalves have adapted to boring into hard substrata such as consolidated clay, calcareous rocks, and wood (Clapp and Kenk, 1963; Turner, 1971; Perry, 1998; Devescovi, 2009; Bagur et al., 2013) and play a significant role in the degradation of the submerged archeological artifacts (Ricci et al., 2015). Calcareous rocks are highly susceptible to boring bivalves’ colonization, although lithotype (limestone, calcitic or dolomitic marble, travertine), surface inclination, morphology of the rocky bottom, and depth reflect the different degree of substrate aggression (Kleemann, 1973a; Kleemann, 1974a; Fanelli et al., 1994). Furthermore, both abiotic and biotic factors affect colonization patterns and growth of boring bivalves, such as substrate composition, hydrodynamic conditions, physical features of the habitat, nutrient concentration, and intraspecific competition for food and space (Kleemann, 1973a; Kleemann, 1973b; Kleemann, 1974a; Kleemann, 1974b; Devescovi and Ivesa, 2008; Devescovi, 2009).

Bivalves are characterized by slow growth rates and long lifespan (Montero-Serra et al., 2018) and dominate the final stages of the ecological succession of endolithic communities (Hutchings, 1986; Chazottes et al., 1995; Casoli et al., 2016).

The highest number of boring bivalves has been reported in tropical waters, where the bioerosion processes and borers’ biodiversity greatly influence the morphology and functioning of carbonate coral reefs; on the contrary, few species occur in temperate regions (Valentich-Scott and Dinesen, 2004; Valentich-Scott and Tongkerd, 2008). Boring bivalves mostly belong to Mytilidae, Pholadidae, Gastrochaenidae, and Petricolidae. In the Mediterranean Sea, the most widespread species of boring bivalves are Lithophaga lithophaga Linnaeus, 1758 (Colletti et al., 2020), Petricola lithophaga Retzius, 1888 (Donnarumma et al., 2018; Casoli et al., 2019a), Pholas dactylus Linnaeus, 1758 (Voultsiadou et al., 2010; Danovaro et al., 2018), and Rocellaria dubia Pennant, 1777 (Schiaparelli et al., 2005; Fava et al., 2016; Casoli et al., 2019a). These species share chemical and mechanical erosion processes through which they penetrate the rock by secreting acids and rotating the shells (Kühnelt, 1933; Kleemann, 1990; Kleemann, 1996).

Some of the species mentioned above colonized underwater archeological calcareous artifacts (mosaic floors, statues, and architectural elements) and were defined as responsible for the biological degradation (Ricci et al., 2015; Ricci et al., 2016a; Casoli et al., 2016). In addition, boring bivalves are important biological sea level indicators. The boreholes of boring bivalves in Puteoli, Naples (e.g., Serapis temple - Macellum) indicate the ancient sea levels and give information on the bradyseism phenomena and related movements that affected the area (Morhange et al., 2001; Morhange et al., 2006).

Lithophaga lithophaga is protected by the EU Habitats Directive and several international directives and conventions in all Mediterranean countries. It is distributed along the whole Mediterranean coastline (Fisher et al., 1987), throughout the Atlantic Ocean from Portugal to the northern coast of Angola (Gonzalez et al., 2000), and it is also reported on the coast of Mozambique (Macnae and Kalk, 1958), albeit Indo-Pacific records of the species should be possibly carefully revised (Huber, 2010). It is an edible endolithic bivalve illegally harvested by breaking the rocks that it colonizes: because of such exploitation, the rocky reefs turn into a biological desert, with dramatic biodiversity and habitat loss (Colletti et al., 2020 and references therein; Bevilacqua et al., 2021).

This species is commonly called Mediterranean date mussel because of its shell morphology and color. The shell is elongate-elliptical, and it can exceed 90 mm in length (Kefi et al., 2014; Peharda et al., 2015). Lithophaga lithophaga generally inhabits galleries bored in calcareous rocks by attaching its antero-ventral shell margin to the inner wall of the holes with byssus (Owada, 2007). Mechanical boring mechanisms are excluded due to the absence of erosion marks on the shell; neutral mucoproteins with calcium-binding ability, secreted by pallial glands, chemically mediate the calcareous excavation (Jaccarini et al., 1968). Date mussels start boring rocks after about 5-10 years from the larval settlement (Pierotti et al., 1966; sub Lithodomus lithophagus), producing deep club-shaped cavities that accommodate the shell, showing dumbbell-shaped openings at the rock surface corresponding to inhalant and exhalant siphons of the organism (Kázmér and Taborosi, 2012).

This species exhibits a slow growth rate and long lifespan: the application of von Bertalanffy growth function shows one of the lowest growth rates among bivalves (k = 0.03 year ^-1, Peharda et al., 2015). Shell length of 10 mm is reached at three years of age (Kleemann, 1973b; Galinou-Mitsoudi and Sinis, 1995; Kleemann, 1996), while a specimen measuring 30-50 mm has about 21-35 years, and regarded 80 mm are related to the age of 54-80 years (Kleemann, 1973b; Galinou-Mitsoudi and Sinis, 1995; Devescovi and Ivesa, 2008).

Lithophaga lithophaga was reported in most of the archeological artifacts (marble statues and architectural elements) submerged in Baiae, Capri, Antikythera, and Epidaurus (Figure 7A) and described by Ricci et al. (2015) and Davidde Petriaggi et al. (2019). Holes were round and oval, ranging from 0.5 to 3 cm in diameter. In many cases, two confluent holes forming the dumbbell-shaped elongate aperture were observed, as well as three or more holes confluent were also reported (Ricci et al., 2015; Ricci et al., 2017). The cavity orientation, indicating the shell position, can help the hypothetical reconstruction of the artifact lying conditions on the sea bottom (Figures 7B, C).

Rocellaria dubia colonizes both infralittoral and circalittoral hard bottoms and has been reported as a borer of the skeleton of living bivalves and echinoderm lying on soft sediments (El-Menif et al., 2005; La Perna, 2005; Morton et al., 2011; Belaustegui et al., 2013; Donnarumma et al., 2018; Casoli et al., 2019a; Mikac et al., 2021).

Rocellaria dubia is distributed in the Mediterranean Sea, Black Sea, British islands, and the eastern Atlantic Ocean (Tebble, 1966), preferentially boring horizontal or slightly inclined substrates and being able to cope with high sedimentation rates thanks to its long siphons that rise from the bottom, allowing feeding and avoiding suffocation (Kleemann, 1974b; Schiaparelli et al., 2003). Larval settlement occurs in holes and crevices in the substrate where they metamorphose into a juvenile and start to penetrate the substrate. The boring activity was described as a combination of chemical and mechanical processes due to the acid secretion produced by pallial glands (Morton and Tseng, 1982) and the contraction of pedal retraction muscles (Carter, 1978; Morton et al., 2011). It shows a variable boring pattern, changing the perforation direction to avoid competition or direct contact with other endoliths, or selecting the most suitable parts of the rock (Schiaparelli et al., 2003; Morton et al., 2011). This bivalve secretes an aragonitic chimney encircling siphonal tube that appears as an external “8” shaped borehole (Figures 7D, E), well distinguishable through the observation of substrates (Morton et al., 2011; Casoli et al., 2016; Mikac et al., 2021). Schiaparelli et al. (2005) defined the correlation between the size of the “8” shaped siphonal tube and the shell dimensions: the equation they provided represents an effective method to quantify the bioerosive activity of this species by measuring the volume of the chamber.

Although R. dubia was found in different artificial substrates (concrete, travertine tiles, and limestone) in the Mediterranean Sea (Nicoletti et al., 2007; Spagnolo et al., 2014; Fava et al., 2016), it was reported for the first time by Davidde et al. (2010) and later by Ricci et al. (2009 sub Gastrochaena dubia; Ricci et al., 2015; Ricci et al., 2016a) as a component of the endolithic assemblages on archeological artifacts recovered in Baiae and the Gulf of Puteoli. Dense colonization of living specimens was observed in situ on mosaic floors (Ricci et al., 2016a) in the Underwater Archaeological Park of Baiae. The shells were found inside the limestone, developing vertical and sub-horizontal cavities, and affected adjacent tiles: the specimens ranged from 11 to 13 mm in length. Casoli et al. (2016) described the colonization dynamic highlighting the role of the bivalve in the bioerosion process of artificial limestone panels submerged in the Underwater Archaeological Park of Baiae. The results showed that the colonization was fast, and the bioerosion rate depended on the interactions with epibenthic assemblage. The presence of elliptical/spherical-shaped holes indicated that the specimens have settled on the encrusting layer formed by epilithics (i.e., barnacles and bryozoans), partially reducing the amount of eroded calcareous material.

Petricola lithophaga is a small bivalve living in shallow holes in calcareous rocks, likely excavated by acid secretion (Morton and Scott, 1988). This rock-borer bivalve is distributed in the Boreo-Atlantic littoral zone: the Atlantic coasts of Europe and North Africa, from Great Britain in the north to Morocco in the south, as well as in the Mediterranean, Aegean, Marmara, and Black seas (Kovalyova, 2017). Despite the cryptic mode of life, studies carried out in the Black Sea showed that P. lithophaga is widely distributed in limestone rocks, where it is present as dominant species showing a high abundance and biomas (Kovalyova, 2012; Kovalyova, 2015). Nevertheless, it is the lesser-known and studied among the boring bivalves found in the Mediterranean Sea. The shell is rounded or elongated-oval; shell length ranged from 4 mm to 24 mm. The valves are gray to white in color, thin, with c. 60 radial ridges on the surface (Kovalyova, 2015).

Petricola lithophaga was rarely reported as a bioeroder of Cultural Heritage. Davidde et al. (2010) and Ricci et al. (2015) found several specimens inside marble statues and architectural elements recovered from Baiae and the Port of Puteoli (Figure 7F). The dimension of the valves ranged from 11 to 19 mm, and the external holes had a diameter varying from 4 to 13 mm. As for the conservation of submerged archeological remains, P. lithophaga seems to play a minor role as abioerosion agent due to the smaller shell size and less frequency of occurrence on the artifacts if compared to L. lithophaga and R. dubia.

Polychaetes contribute greatly to internal bioerosion through their ability to bore calcareous substrates by digging cavities, chambers, and tunnels, via either through secretion of acid compounds, mechanical actions or combination of both. Chemical dissolution of carbonate is due to secretion of acid mucopolysaccharides that are produced by the ventral epithelium and segmental mucus glands; meanwhile the substratum is physically eroded using hard structures, such as jaw teeth or specialized chaetae (Blake, 1969; Hutchings, 1986; Hutchings, 2008).

Boring polychaetes belong to the families Spionidae, Cirratulidae, Eunicidae, Dorvilleidae, and Sabellidae and their bioerosive activity is known in different types of calcareous substrates, such as rocks, corals reefs, calcareous algae, shells, and experimental substrates (Hutchings et al., 1992; Martin and Britayev, 1998; Pari et al., 1998; Pari et al., 2002; Hutchings, 2008; Gibson, 2017; Çinar and Dagli, 2021). Boring polychaetes has been studied mainly in tropical coral reefs (Hutchings and Murray, 1982; Hutchings and Peyrot-Clausade, 2002; Osorno et al., 2005; Hutchings, 2008), where eunicids, spionids, cirratulids, and sabellids actively perforate corals (Bromley, 1978; Hutchings, 1986; Le Grand and Fabricius, 2011). Boring polychaetes belonging to spionids (Polydora and related species) are particularly investigated in commercially valued shellfish species because of the consequence of such infestation on the economy (Blake and Evans, 1973; Diez et al., 2011; Diez et al., 2016; Waser et al., 2020). In contrast, compared to tropical areas, little data is available from the Mediterranean Sea particularly with respect to archeological artifacts. Despite the limitations imposed by the sampling of archeological remains, the micritic limestone tesserae collected from the submerged Roman mosaic floor of Portus Julius represent a unique substrate exposed for centuries to marine organism colonization and bioerosion processes (Antonelli et al., 2015). Here the cirratulid Dodecaceria concharum Örsted, 1843 (Figure 8A) dominates the endobenthic polychaete assemblage having a substantial role in bioerosion, while the spionids Polydora ciliata (Johnston, 1838) (Figure 8B) and Pseudopolydora antennata (Claparède, 1869) have a secondary role; in addition, the eunicid Lysidice unicornis (Grube, 1840) is found as nestler and possible borer (Gravina et al., 2019). The mechanisms of perforating and type of borings reflect the morphology and ecology of the various species. Spionids use their modified chaetae of the fifth chaetiger, which are assisted by the acid mucus produced by ventral glands (Dorsett, 1961), then enter a variety of substrates, including limestone and bivalve shells, as well as tropical coral skeletons. The burrows of the Polydora species are V- or U-shaped tunnels with small diameter entrances; here, the worms are folded in half and emerge from the openings with the head and the pygidium. By protruding the pair of long tentacles through the openings, Polydora individuals collect food particles from both the substrate and water column. As the individuals grow by elongating, the burrows are branched and result in a complex boring system of well-developed sinuous, cylindrical tunnels. In cultured bivalve species, such as oysters, scallops, mussels (Blake and Evans, 1973; Diez et al., 2011) and gastropods (Guţu and Marinescu, 1976), Polydora worms settle between the mantle and shell and induce the oyster to produce a mud blister around the worm’s body, as a mechanism to isolate the worm (Zottoli and Carriker, 1974; Lauckner, 1983). Consequently, depending on the intensity of infestation, the commercial value of the damaged oysters is reduced. The cirratulid D. concharum, in the Mediterranean Sea, together with D. fimbriata (Verrill, 1879) in northern Europe coast, burrows both in natural calcareous substrates, mollusc shells, corals, algae (Lithothamnion, Lithophyllum), and in man-made artifacts (Hartman, 1951; Gibson, 2008; Gravina et al., 2019). Dodecaceria species use their spoon-shaped chaetae present along the body to perforate the carbonate (Gibson, 2017), but chemical dissolution is an additional method to enter limestone substrates (Hutchings, 2008). Such mechanism is confirmed by the characteristic behavior of the worms, which exhibit close contact with the substrate during forward and backward movements and stop for periods of rest preceding the boring activity. The burrows of Dodecaceria are pouch-like, flattened in cross-section, distally enlarged and characterized by an 8-shaped single entrance, where the individuals are folded on themselves with the head and anus facing the opening. As deposit-feeders, these worms protrude their long palps through the opening for collecting sediment particles. Sabellids have no hard structures, thus their eroding ability is attributable to acid content of the mucus, which is produced by the glands that open at the base of the branchial crown (Chughtau and Wyn Knight-Jones, 1988). Unfortunately, this mechanism has not been demonstrated, although sabellids are well known to produce mucous material (Giangrande et al., 2014). Sabellids produce narrow, unbranched tunnels of circular section; here the worms remain with the pygidium at the blind end and keep their branchial crown expanded through the opening, thus ensuring respiration and filter-feeding. The tunnels are considerably longer than the animals and are lined with a chitinous coating that isolates the worms from the cavity wall. Potamilla sp. is the only sabellid that is known to bore in the Mediterranean Sea (Çinar and Dagli, 2021), but no sabellids have so far been found responsible for boring archeological artifacts (Gravina et al., 2019). Eunicids produce intricate branching tunnels of circular and oval sections of large diameter, where the animals remain folded in on themselves. The burrows differentiate from those of sabellids because they have unlined walls (Hutchings, 2008). Moreover, the tunnel of eunicids show evident teeth marks as signs of the bioerosive activity especially on corals, due to developed jaws equipped with mandible and maxillary toothed plates, as an efficient mechanical erosive system. Acid secretions are also hypothesized to occur. Some large eunicids, Eunice dubitata Fauchald, 1974, E. floridana (Pourtalès, 1867), and Leodice torquata (Quatrefages, 1866) live associated among the scleractinian corallites and the valve of the deep-oyster Neopycnodonte cochlear in the mesophotic zone of the Mediterranean Sea; although they are not clearly identified as opportunistic nestlers or true bioeroders (Gravina et al., 2021b). The same bioerosion mechanism is reported for the boring species of the related family Dorvillaeidae (Hutchings, 2008).

Concerning the archeological artifacts, the tesserae of the Roman mosaic floor are largely bored on their surfaces, showing holes and furrows different in arrangement, shape and extension (Gravina et al., 2019). Single or paired furrows with U-shaped and sinuous arrangement highlight that polychaete galleries develop both in the tesserae and in the space between the tesserae and the epilithic layer encrusting them (Figures 8C, D). At the same time, circular and elliptical holes, isolated or paired in an 8 shape, represent the entry and exit points of the galleries (Gravina et al., 2019). Similar morphologies of traces are noticeable on the experimental panels submerged in the Archaeological Marine Park of Baiae and confirmed that the previous cover by encrusting organisms favors the settlement and the subsequent bioerosive action of polychaetes (Casoli et al., 2019b). Due to its dominance, D. concharum appears to be a secondary borer, together with L. unicornis. Meanwhile, P. ciliata and P. antennata, present in low abundance, are consistent with the hypothesis of early colonization of these species. Such spionids continue to settle as long as new substratum remains available, but with reduced density. Experimental studies have been based on artificial calcareous panels placed in the same archeological site of the Roman mosaic floor of Portus Julius and allowed repeated observations on different time scales (Casoli et al., 2015; Casoli et al., 2019b). The endolithic polychaete assemblage that occurred in the panels confirmed the findings reported for the tesserae of the Roman mosaic floor (Gravina et al., 2019), where P. ciliata and P. antennata are the first colonizers and are dominant species after 1 year of underwater exposure. Then, their density declines in the following 2-3 years, in correspondence with the arrival of D. concharum, which thrives in the mature community; also, the eunicid L. unicornis is a component of this phase. Interactions between the encrusting and boring organisms are other relevant information supported by these studies. They highlight that encrusting epibenthos provides a new substratum available for the boring polychaetes, where the worms can complete their life cycles and find their feeding resources (Casoli et al., 2019b). Consequently, the worms bore the encrusting stratum and dig different holes and furrows in the surface of the panels, which remain as a trace of the interaction between encrusting and boring organisms (Casoli et al., 2019b). The same sequence of settlement has been highlighted on the valves of sea scallops, where spionids are the main colonizers in the early stage, while D. concharum is the secondary borer which settles in previously formed holes (Evans, 1969). Other data from tropical coral reefs show Polydora spp., dorvilleids (Schistomeringos) and sabellids (Notaulux) to be the early colonizers, while Dodecaceria spp. and some eunicids the component of the mature boring community (Hutchings, 2008).

Internal bioerosion has been studied with various techniques used to detect the type and distribution of burrows, recognize their burrowers, and reconstruct the bioerosion pattern. In natural substrates, as bivalve shells, the X-ray radiography has been used (Evans, 1969); while the micro-computed tomography has been employed for both the investigation on fossil bioerosion traces in Cretaceous belemnite (Wisshak et al., 2017) and traces present in experimental blocks (Färber et al., 2016). The internal bioerosion of archeological artifacts, as the Roman mosaic floor of the Underwater Archaeological Park of Baiae, has been studied with the technique of silicone casts, which highlighted six different ichnospecies (Gravina et al., 2019). The present borings clearly match to the fossil traces belonging to both ichnogenera Caulostrepsis and Meandropolydora, described by Bromley and D’Alessandro (1983; 1987), Costa de Almeida (2007), Voigt (1965; 1971; 1975), and Wisshak (2006). The traces associated to the ichnospecies Caulostrepsis taeniola and C. cretacea are identified by the tongue- and ribbon-shape of borings and differ in the presence of a vane interconnecting the limbs (Figure 8E). The traces ascribed to the ichnospecies Maeandropolydora barocca, M. decipiens, M. elegans, M. sulcans are characterized by a more complex system of tunnels, which differ in the arrangement of single, parallel U-shaped or spiral furrows (Figure 8F), as well as in the shape of the external holes. The complex arrangement of polychaete boring traces occurring in the mosaic tesserae results in the high ichnospecies’ diversity. In accordance with the dominance of their inhabitants in the artifacts and the succession pattern on experimental substrates (Casoli et al., 2019b), these variations show that the traces correspond to various bioerosive stages. They all correspond to the action of D. concharum: Caulostrepsis spp. corresponds to its primary bioerosion activity, while Maeandropolydora spp. corresponds to its secondary boring activity, when D. concharum uses and modifies primary borings made by spionids (Gravina et al., 2019).

The phylum Sipuncula comprises species adapted to live in a wide range of temperatures and depths, often nestling inside vacant shells or polychaetes tubes or burrowing sandy bottoms (Pancucci-Papadopoulou et al., 2014). Many literature works report about the ability of some Sipuncula species (belonging to the genera Phascolosoma and Aspidosiphon) of actively boring hard carbonatic substrates (e.g., corals, shells, rocks, and bones) mainly by mechanical action of the caudal shield (Rice, 1969; Rice and Macintyre, 1972; Williams and Margolis, 1974). However, evidences proved that the epidermal glands distributed on the worm’s body produce substances, like chelating agents or acids, able to chemically modify calcareous substrata (Williams and Margolis, 1974). Once burrowed a flask shaped tunnel, the organism settles inside it, with the anterior end directed toward the opening, and periodically extends the introvert outside the burrow scraping the surface of the substrate to collect food. The presence of hooks on the surface of the introvert and its movement along the surface can also contribute to the superficial erosion of the material (Rice, 1969).

The bioerosive role of sipunculids on archeological artifacts has been reported for the first time by Antonelli et al. (2015). Two specimens belonging to the species Aspidosiphon muelleri Diesing, 1851 were observed in erratic mosaic fragments recovered from the Underwater Archaeological Park of Baiae. This is the most widespread species of the genus Aspidosiphon, distributed between 5 and 2900 m in depth, from the north-eastern Atlantic to Chile, from Japan to the Mediterranean Sea (Cutler, 1994; Hong and Fenglu, 1998; Dean, 2001; Ferrero-Vicente et al., 2011; Pancucci-Papadopoulou et al., 2014). Sipunculids belonging to this species are characterized by an elliptical dark brown anal shield, a caudal shield with papillae separated by tangential furrows, and an introvert with compressed uni- or bi-dentate hooks arranged in rings in the distal part and pyramidal, uni-dentate hooks randomly arranged in the proximal part (Stephen and Edmonds, 1972; Cutler and Cutler, 1989; Pancucci-Papadopoulou et al., 2014). The two reported specimens had penetrated the mosaic boring both limestone tesserae and mortar, producing tunnels up to 5 mm in diameter (Figure 9A). After this report several A. muelleri specimens were observed in the mosaics of Portus Julius (Underwater Archaeological Park of Baiae; Figure 9B). As for the previous report, the organisms were observed both inside the tesserae and mortar and had produced tunnels up to 4 mm in diameter.

Differently from all the other bioeroder groups discussed in the present paper, echinoids belong to the vagile epifauna and their boring action is usually associated with their grazing activity, leading to the formation of superficial scars on rocks and other hard substrates. However, some species developed a rock burrowing behavior to overcome environmental stresses (e.g., waves and tidal action) that can cause severe damages to hard substrates. For example, species of the family Echinometridae are considered the most destructive borers of tropical coasts (Asgaard and Bromley, 2008; Glynn and Manzello, 2015). Through the action of the teeth of Aristotle’s lantern, echinoids produce cup-shaped pits or, more frequently, elongated grooves in which they settle more or less permanently. As said, echinoids are vagile animals, so they are not imprisoned in their burrows but are free to leave them, if necessary, by lowering their spines to pass the narrow entrances (Asgaard and Bromley, 2008).

Until now, in the field of UCH, the boring action of echinoids has been observed only on the architectural structures of the roman “Villa of the dolia’’ in Palaia Epidaurus as superficial sub-circular traces produced by the species Paracentrotus lividus Lamarck, 1816 (Figure 9C). In this site, echinoids reduced the presence of the algal turf, common on horizontal surfaces in shallow waters, due to their grazing activity. On the contrary, echinoids were often observed in association with encrusting red algae (Corallinales), because they are not able to feed on them, due to the calcified thallus.

Furthermore, it has been hypothesized that the sub-circular or ovoid furrows (0.5- 3 cm in diameter, up to 1 cm in depth) observed on some of the tusks recovered from the shipwreck of Bajo de la Campana (Figure 9D) could be attributed to echinoid action (Antonelli et al., 2019). Unfortunately, in this case no body’s elements (e.g., teeth, spines) were present inside the furrows, so the attribution of these bioerosion traces to their activity can be only hypothesized.