The tectonics of the Mediterranean region are dominated by the convergence between the African and European plates, which is further complicated by the occurrence of various intervening minor plates. This convergence scenario has led to the formation of the Alpide orogenic belts, including the Alps and Apennines, as well as to the development of the many sub-basins that make up the present-day Mediterranean Sea. Among the latter are the ‘neoautochtonous’ basins that began to form on the Tyrrhenian side of the Northern Apennine chain and gradually developed eastward as the orogen migrated toward the foreland (Martini and Sagri 1993). This process resulted in the progressive formation of various sets of tectonic depressions parallel to the rising Apennines and bordered seawards by fringes of topographic highs. These basins have been categorised as either ‘central’ or ‘peripheral’; the former are typically home to Miocene and Pliocene marine deposits, while the latter are mostly filled with Pliocene and Pleistocene continental deposits (Martini and Sagri 1993).

For several centuries, the Neogene successions of the central basins have yielded remarkable finds of marine vertebrate fossils, which have gained international recognition since the latter half of the 19^th century. These fossils include countless specimens of bony and cartilaginous fishes (among which are sharks, rays and chimaeras), cetaceans (both toothed and baleen-bearing whales), sirenians, pinnipeds, seabirds, and sea turtles (e.g., Dominici et al. 2018, and the many references therein). Interestingly, some of the first trace fossils to show evidence of the feeding activity of Osedax come from the Northern Apennine hinterland. They occur as single-entry borings and pits on an isolated cetacean radius from the historic palaeontological locality of Orciano Pisano, and include the holotype of Osspecus tusciae Higgs et al., 2012, which in turn represents the type ichnospecies of Osspecus (Higgs et al. 2012). Unfortunately, the exact geographic and stratigraphic context of the find in question remains uncertain, as is the case for most of the fossils from Orciano Pisano (Landini 1977; Cigala Fulgosi et al. 2009).

The fossil shark teeth dealt with herein originate from clayey badlands at Lucciola Bella (indicative geographic coordinates: 43°02'00"N, 011°45'30"E) and Selvoli (indicative geographic coordinates: 43°03'00"N, 011°40'00"E), two sites in the vicinity of Pienza (Siena Province, Tuscany, central Italy). Both localities (Fig. 1A) belong in the Siena-Radicofani Basin, one of the ‘central’ basins of Tuscany, which is shaped as a NNW-trending tectonic depression that preserves a valuable record of post-collisional deposition in the Northern Apennine hinterland (Martini et al. 2021).

Figure 1. 

Geographic and geological setting. A. Schematic geological map of Tuscany, showing the location of the study sites (Lucciola Bella and Selvoli, indicated by stars). Redrawn and modified from Benvenuti et al. (2014). Abbreviations: Alb – Albegna Basin; BC – Baccinello-Cinigiano Basin; C – Casentino Basin; Cas – Casino Basin; Cec – Cecina Basin; FPP – Firenze-Prato-Pistoia Basin; G – Garfagnana Basin; M – Mugello Basin; Mbam – Montebamboli Basin; MTR – Montespertoli-Tavernelle Ridge; Rad – Radicondoli Basin; Rib – Ribolla Basin; SR – Siena-Radicofani Basin; TF – Tora-Fine Basin; VA – upper Valdarno Basin; VC – Valdichiana Basin; VE – Valdelsa-lower Valdarno Basin; Vel – Velona Basin; Vol – Volterra Basin; B. Landscape view of the marine clays outcropping at Lucciola Bella.

Specifically, both Lucciola Bella and Selvoli are located in the Radicofani sub-basin, which corresponds to the southern portion of the Siena-Radicofani Basin, south of the so-called Pienza high (Brogi 2011; Brogi et al. 2014; Martini et al. 2017). The sedimentary fill of the Radicofani sub-basin includes some 1,000 m of continental and marine Pliocene deposits that unconformably overlie continental sediments of Miocene age (Martini et al. 2021, and the many references therein). Two distinct depositional cycles have long been identified within this Pliocene succession. The older cycle is characterised by siliciclastic sediments, including sandstones and mudstones, which transition basinward into offshore mudstones punctuated by turbidites. This cycle spans throughout the Zanclean, corresponding to the biozones MPl1 to MPl4a (as per the zonal scheme of Iaccarino et al. 2007). The younger cycle, which only crops out close to the eastern basin margin, is mostly comprised of Amphistegina d’Orbigny, 1826-rich limestones and nearshore sandstones deposited during the Piacenzian, corresponding to the biozones MPl4b to MPl5a. At Lucciola Bella and Selvoli, the fossiliferous outcrops consist of offshore mudstones that belong in the Zanclean cycle, specifically in the Argille Azzurre Formation (Regione Toscana 2006–2009).

Detailed sedimentological and palaeontological data are available for the Pliocene succession exposed at Lucciola Bella (Fig. 1B), which consists of massive, pervasively bioturbated silty clays with interbeds of normally graded sands (Pascucci et al. 2006; Bianucci et al. 2009). Analyses of the foraminiferal assemblages from the Lucciola Bella section have led to the identification of either the zones MPl 2–3 (Pascucci et al. 2006) or the zones MPl 3–4a (Bianucci et al. 2009), which belong in the Lower Pliocene. Micropalaeontological observations have also resulted in reconstructing an outer shelf/upper slope depositional environment with relatively low oxygen levels and high concentrations of organic matter at the seafloor (Bianucci et al. 2009). Marine vertebrate finds from this site include a skeleton of the extinct delphinid Etruridelphis giulii Bianucci et al., 2009, a rostral spine of a pristid sawfish (Collareta et al. 2017), a dentary beak of an ocean sunfish (Collareta et al. 2025), and a large number of shark teeth whose systematic analysis is still in progress. Macro-invertebrate (mollusc) fossils from the same site were previously noted by Pascucci et al. (2006), and later examined in detail by Spadini and Manganelli (2010) and Spadini (2022).

During the Early Pliocene, the Radicofani sub-basin was shaped as a semi-enclosed embayment sheltered seaward by an emerging structural high (Pascucci et al. 2006). That said, the depositional settings were rather deep, though relatively close to the continent, as evidenced by the local occurrence of nearshore (shoreface to shallow-deltaic), coarse-grained deposits representing the proximal equivalent of the offshore mudstones that are found at the study sites (Martini et al. 2021).

Specimen GAMPS-01970 from Lucciola Bella is an 11.8-mm-high, 11.3-mm-wide lower tooth (Fig. 2A, B). Macroscopically, GAMPS-01970 is remarkably complete but for some wear of the thin enameloid layer at the crown base on the labial tooth surface as well as for some sparse, very minor pitting of the root exterior. A subcircular hole (‘α’ in Fig. 2C), approximately a quarter of a millimetre in maximum diameter, pierces the labial root surface along the distal root lobe, slightly below and beyond the mid-height and mid-length of the latter, respectively. Under the macroscope, the hole in question is readily distinguishable from the minor foramina that pierce the labial root surface as well as from other nearby, smaller, seemingly shallower pits. Micro-CT analyses demonstrate that this hole represents the single external opening of a well-defined, tree-like cavity, which differs from the surrounding dentine pores both in terms of size and shape (Fig. 2D). Below the subcircular aperture, a very short neck expands into a much larger internal chamber, leaving only a thin layer of surface dentine on top (Fig. 2E, F). The chamber is roughly ellipsoidal in shape and sends around a variety of exploratory branches (Fig. 2G, H). These branches range from short to long (sensu Jamison-Todd et al. 2025b), most of them being short and incipient. Some branches are bulbous, but others are lobate and tapering, with the longest narrowing into a point. A short, acutely pointed branch is observed projecting downwards from the trace floor. Measurements of trace α are reported in Table 2.

Figure 2. 

GAMPS-01970, lower tooth of Carcharhinus from the Lower Pliocene of Lucciola Bella, overview and close-ups of the associated Osspecus boring (i.e., trace ‘α’). A, B. Microphotographs of GAMPS-01970 in lingual (A.) and labial (B.) views; C. Close-up of (B.), showing the location of the external opening of the Osspecus boring α; D. Transparent 3D model of GAMPS-01970 and superimposed opaque 3D volume reconstruction of the Osspecus boring α. The whitish areas represent the vascular cavities developing within the tooth dentine (note the large, open pulp cavity); E, F. Two microtomographic cross-sections of the Osspecus boring α; G. Close-up of (D.), taken from a different point of view; H. 3D model of the Osspecus boring α. The external opening of the boring is oriented upward.

Specimen GAMPS-01971 from Lucciola Bella is an 11.0-mm-high, 14.1-mm-wide upper tooth (Fig. 3A, B). Like GAMPS-01970, it is remarkably complete but for some missing enameloid in occurrence of the distal notch. The labial root surface features four subcircular holes. Two large such holes pierce the distal root lobe, one of them being filled with sediment (‘γ’ in Fig. 3C). Micro-CT analyses reveal that these holes, too, represent the external openings of much larger borings. The boring associated with the sediment-free hole (‘β’ in Figs 3C, 4A) is shallow, its deepest point being located less than one millimetre below the external opening, but not much so considering the limited thickness of the distal root lobe at its distal end. The chamber develops subhorizontally. When viewed apically, its outline is roughly trapezoidal, its long axis being aligned with the root lobe itself (Fig. 4D). There are a few exploratory branches, which are short and substraight (Fig. 4E, G). The boring associated with the sediment-filled hole (‘γ’ in Fig. 4B) is strongly eccentric, developing as it does toward the middle of the root (Fig. 4D). The neck is stout and gradually expands into an oblique-oriented chamber (Fig. 4B). The latter sends around some thin, acutely pointed exploratory branches (Fig. 4D, E, H). Each of the two small borings that pierce the mesial root lobe (‘δ’ and ‘ε’ in Fig. 3D) also connects with an internal chamber through a short neck (Fig. 4C). Said chambers are closely interdigitated yet still separated from each other by a thin wall of dentine (Fig. 4B, D, F). Their overall shape is bulbous, with short exploratory branches, but for a single, long, pointed branch associated with trace ε (Fig. 4I, J). Measurements of traces β, γ, δ and ε are reported in Table 2.

Figure 3. 

GAMPS-01971, upper tooth of Carcharhinus from the Lower Pliocene of Lucciola Bella, overview. A, B. Microphotographs of GAMPS-01971 in lingual (A.) and labial (B.) views; C, D. Close-ups of (B.), showing the location of the external openings of the Osspecus borings β and γ (C.), and δ and ε (D.).

Figure 4. 

Close-ups of the four Osspecus borings (i.e., traces ‘β’, ‘γ’, ‘δ’ and ‘ε’) occurring on GAMPS-01971, upper tooth of Carcharhinus from the Lower Pliocene of Lucciola Bella. A–C. Microtomographic cross-sections of the Osspecus borings β (A.), γ (B.), and δ and ε (C.). D. Transparent 3D model of GAMPS-01970 and superimposed opaque 3D volume reconstructions of the Osspecus borings β (yellow), γ (purple), δ (blue) and ε (green). The whitish areas represent the vascular cavities developing within the tooth dentine (note that the pulp cavity is most likely filled with a dense mineral deposit); E, F. Close-ups of (D.), taken from different points of view. The colour legend is the same as in panel (D.). G–J. 3D models of the Osspecus borings β (G.), γ (H.), δ (I.), and ε (J.). The external openings of the borings are oriented upward.

Specimen GAMPS-01972 from Selvoli is a 12.3-mm-high, 12.5-mm-wide lower tooth (Fig. 5A, B). It is well preserved but for some missing dentine and enameloid at its mesial end as well as for a smaller, crater-like damage along the lingual surface of the mesial root lobe (see description below). The basal end of the sediment-filled nutrient groove appears somewhat widened mesially (‘ζ’ in Fig. 5C). Under the Micro-CT, this broadened sector of the groove is shown to connect with a largely irregular chamber through a short neck (Fig. 5D). A number of stout, rounded exploratory branches depart from the trace floor. Because it is continuous with the nutrient groove and intersects multiple vascular cavities within the root dentine, this trace could not be reliably segmented to produce a 3D reconstruction.

Figure 5. 

GAMPS-01972, lower tooth of Carcharhinus from the Lower Pliocene of Selvoli, overview and close-ups of the associated Osspecus boring (i.e., trace ‘ζ’). A, B. Microphotographs of GAMPS-01972 in lingual (A.) and labial (B.) views; C. Close-up of (B.), showing the location of the external opening of the Osspecus boring ζ within the nutrient groove; D. Microtomographic cross-section of the nutrient groove and associated Osspecus boring ζ.

Each of the above described traces — numbering one in GAMPS-01970, four in GAMPS-01971, and one in GAMPS-01972 — may be described as consisting of a single, circular to sub-circular aperture, lacking any rim or platform, extending into the substrate as a short, subvertical canal that leads to a much larger, globular or irregularly shaped chamber provided with a variety of exploratory branches. Owing to these characters, the borings in question are unambiguously assigned to the ichnogenus Osspecus. The next closest match would be Entobia Bronn, 1838, which is commonly attributed to the boring action of endolithic clionid sponges (Wisshak et al. 2019). Entobian borings have been recently discovered on teeth of Jurassic marine crocodylomorphs (Serafini et al. 2024) as well as of Neogene cetaceans (Jamison-Todd et al. 2025a) and sharks (Maisch IV et al. 2020). However, Entobia commonly consists of networks of chambers and galleries (Tapanila 2006), while our examples of Osspecus are typically found isolated from each other.

Based on the observation of short, tapering necks as well as of far-from-spherical chambers, the traces for which volume reconstructions are available (i.e., those occurring on GAMPS-01970 and GAMPS-01971) are further referred to Osspecus tuscia, whose holotype also comes from the Tuscan Pliocene (Higgs et al. 2012). The next closest match would be Osspecus eunicefootia Jamison-Todd et al., 2025, which, however, typically features exploratory branches that are thin and wavy.

Besides the above described Osspecus specimen, GAMPS-01970 also features the occurrence of three similar-sized cavities along the basal margin of the mesial root lobe, just lateral to the nutrient groove (Fig. 6B). Compared to traces α to ε, these cavities exhibit a much larger, irregular external opening, which results in a crater-like morphology (Fig. 6C). These crater-like structures are best interpreted as Osspecus borings in which the thin layer of dentine that originally roofed the traces themselves has been largely lost through collapse. Such structures have been described by several authors, including Higgs et al. (2012), Boessenecker and Fordyce (2015) and Bosio et al. (2024), and termed ‘pockmarks’.

Figure 6. 

Crater-like ‘pockmarks’, indicated by arrowheads, on shark teeth from the Lower Pliocene of Lucciola Bella and Selvoli. A. Isolated pockmark on the lingual surface of the mesial root lobe of GAMPS-01972; B. Three small pockmarks piercing the basal margin of the mesial root lobe of GAMPS-01970; C. Microtomographic cross-section of the same pockmarks as in (B.). D; E. GAMPS-01973, lower tooth of Carcharhinus from the Lower Pliocene of Selvoli, microphotographs in lingual (D.) and labial (E.) views; F. Close-up of (D.), showing an isolated pockmark and a large, crater-like damaged area along the lingual surface of the distal root lobe.

Pockmarks are commonly found on otherwise reasonably well-preserved shark teeth collected at the study site, including GAMPS-01972 (Fig. 6A) as well as GAMPS-01973 (Fig. 6D, E), the latter being a 12.8-mm-high, 11.8-mm-wide, fairly complete lower tooth from Selvoli. While most such craters are similar in size to the above examples of Osspecus, others are significantly larger, extending as they do along the root lobes (Fig. 6E). This is best explained as due to the coalescence of adjacent Osspecus borings followed by the collapse of the delicate roofs thereof (see also Higgs et al. 2014 and Boessenecker and Fordyce 2015).

The Osspecus specimens described above are most parsimoniously regarded as indicative of feeding by zombie worms, as no other organisms are known to produce borings with such distinctive morphology in hard organogenic phosphatic substrates (Jamison-Todd et al. 2025b). That all these traces excavate the root dentine, leaving the enameloid-coated crowns seemingly unaltered, is consistent with the behaviour of extant Osedax spp. feeding on artificially deployed shark teeth, as is the observation of a single boring opening within the nutrient groove of GAMPS-01972 (Rouse and Goffredi 2023). It is worth noting here that shark tooth roots are comprised of osteodentine – a rather porous, organic-rich, and overall bone-like variety of dentine (Berkovitz and Shellis 2017).

Extant Osedax spp. are believed to exploit both lipids and collagen for nutrition. Although small quantities of lipids may occur in the shark tooth dentine (López-Álvarez et al. 2016), the main nutrition source of past and present shark tooth-inhabiting zombie worms is most likely collagen, which abounds in the organic-rich shark tooth roots (LeGeros 1981; Enax et al. 2012).

To the best of our knowledge, the putative occurrence of Osedax-related bioerosional features on fossil shark materials has been evoked by as few as three previous works. Specifically, Amalfitano et al. (2019) interpreted some circular borings on a vertebral centrum from a partially articulated skeleton of Cretoxyrhina mantelli (Agassiz, 1835) from the Upper Cretaceous (Turonian) of northeastern Italy as possibly due to the feeding activity of zombie worms. However, this inference is contentious, as the 3D morphology of the traces in question remain unknown, and there is no evidence at present for extant Osedax spp. colonising the endoskeletal elements of sharks from observations in the wild, nor from deployment experiments (Rouse and Goffredi 2023). Feichtinger et al. (2020) tentatively referred some microborings on teeth of Cretacladoides noricum Feichtinger et al., 2018 from the Lower Cretaceous (Berriasian–Valanginian) of Upper Austria as possibly related to Osedax or some Osedax-like organism. Here again, nothing is known on the 3D morphology of the traces in question, which are characterised by very small aperture diameters (< 0.1 mm) and significantly predate the earliest unambiguous occurrences of Osspecus. Lastly, a Megalodon tooth collected at an as-yet unnamed deep-sea locality close to Johnston Atoll (northern Pacific Ocean) has recently been hypothesised to lack the root due to bioerosion by Osedax (Pollerspöck et al. 2025). No ichnological features that may either support or falsify this fascinating, if speculative, hypothesis appear to be preserved on the tooth in question. All things considered, our bioeroded Carcharhinus teeth from the Italian Pliocene provide the first fossil evidence for Osedax exploiting shark tooth dentine – and, more broadly, any shark-derived material.

Osedax is a major taphonomic agent in the marine environment owing to its remarkable capability of eroding whale bones, which in turn is due to the acid secretion that accompanies the absorption of bone-trapped nutrients (Higgs et al. 2011; Tresguerres et al. 2013). Where Osedax worms are abundant, bioerosion by these bone-eating siboglinids may even induce the rapid destruction of entire whale skeletons. Thus, it is widely believed that the feeding habits of Osedax have had a prominent role in shaping the cetacean fossil record over most of the Cenozoic. Now, however, considering the well-established occurrence of Osedax-related borings on the skeletal remains of other marine vertebrates (including Cretaceous marine reptiles such as plesiosaurs and mosasaurs), the ‘Osedax effect’ — whatever its actual extent — is unlikely to have solely impacted the cetacean fossil record. This interpretation is further strengthened by the fact that multiple deployment experiments have shown various extant members of Osedax to be able to be feed on the bones of other marine and terrestrial mammals, birds, and reptiles, and lately also on shark teeth, exploiting the latter’s dentine. By documenting the occurrence of Osspecus on fossil teeth of Carcharhinus, the present paper serves as a powerful demonstration that feeding on shark tooth dentine can and does occur in at least some wild populations of Osedax. Thus, we are led to ponder whether an ‘Osedax effect’ may exist with respect to the selachian fossil record.

Sharks have inhabited the Earth’s oceans for some 250 million years (Maisey 2012), leaving behind a substantial palaeontological record (Cappetta 2012), most of which consists of dental and dermal remains due to the fact that teeth and scales are heavily mineralised in sharks, while the endoskeleton is largely cartilaginous, and as such, poorly suited for fossil preservation. The fossil record of selachians is further enhanced by their unique dental replacement pattern, whereby teeth are shed and replaced continuously throughout their lives (Kemp 1999; Gillis and Donoghue 2007; Tucker and Fraser 2014; Berkovitz and Shellis 2017), so much so that some modern sharks can lose thousands or even tens of thousands of teeth during their lifetime (Brisswalter 2009), which in turn results in the accumulation of abundant dental remains on the seafloor (Litvinov 2007).

Considering such an abundant fossil record, and taking into account that zombie worms and sharks appear to have coexisted for at least 100 million years (Danise and Higgs 2015; Jamison-Todd et al. 2025b), how could the occurrence of Osedax-related modifications on shark teeth go unnoticed for such a long time? That Osedax’s capability to bore into and feed on shark tooth dentine has been disclosed only recently does certainly contribute to explain this widespread invisibility, in that Osspecus-like traces may have simply not been searched for, at least not routinely, on fossil shark teeth. However, search bias alone is unlikely to account for the pattern, as the discovery of Osspecus on fossil bones of marine turtles and birds (Kiel et al. 2011; Danise and Higgs 2015) has predated the identification of Osedax on their modern counterparts (Rouse et al. 2018).

The point we would like to make here is that Osspecus borings represent somewhat self-destroying evidence for ancient Osedax populations. Indeed, Osspecus borings are unlikely to be preserved on bones (or teeth for that matter) that are not buried halfway through biostratinomy, as Osedax itself progressively consumes its own substrate through feeding (Jamison-Todd et al. 2024). In this regard, it is noteworthy that the Osspecus borings dealt with herein are remarkably small-sized compared to those made by extant Osedax spp. on bone, with total volumes falling well below 1 mm³ (Table 2), versus maximum boring volumes of up to some tens of cubic millimetres in many living zombie worm species (Higgs et al. 2014: tab. 2). Also remarkable is the fact that the volume of the largest individual Osedax-related boring recorded by Higgs et al. (2014: tab. 2) significantly exceeds that of the root of e.g. GAMPS-01970 (99 mm² vs 76 mm²), and even the total volume of the same tooth (98 mm²). Therefore, our Italian Pliocene specimens of Osspecus were either excavated by an idiosyncratic, likely miniaturised Osedax species that produced very small-sized borings, or they represent ichnogenetically incomplete (sensu Belaústegui et al. 2016) examples of traces that could have grown much larger. Since there is no indication that the exploitation of shark tooth dentine is exclusive to highly specialised members of Osedax (Rouse and Goffredi 2023; but see also Smith et al. 2015), the latter hypothesis is the most promising. Furthermore, considering that some large-sized species of Osedax have roots that can penetrate the substrate by several centimetres (Smith et al. 2015), destruction by zombie worms is likely very efficient, and often completely successful, in the case of inconspicuous food items such as shark tooth roots.

The fossil record, both offshore and inland, may preserve some evidence of the root-destroying effectiveness of Osedax. As mentioned, an otodontid tooth from the midst of the Pacific Ocean has been recently hypothesised to lack the root due to the feeding activity of Osedax (Pollerspöck et al. 2025). Rootless specimens are also known to comprise approximately one half of the shark teeth trawled from the seafloor at the Nazca and Sala-y-Gomez seamounts, in the southern Pacific Ocean (Litvinov 2007). Many other shark tooth assemblages from deep seafloors are known to include a large fraction of rootless specimens (e.g., Belyaev and Glickman 1970). Moving inland, some time-averaged marine vertebrate accumulations have been reported to be home to rootless shark teeth, with some specimens even missing most of the dentine ‘core’ of the tooth crowns while invariably featuring essentially well-preserved enameloid cusps (e.g., Boessenecker et al. 2014). That this peculiar preservation style is exclusive to certain shark tooth-rich beds is suggestive of the local action of a specific taphonomic agent (Boessenecker et al. 2014). Although widespread phenomena of root destruction are often ascribed to fungal and/or bacterial bioerosion (e.g., Underwood et al. 1999; Bosio et al. 2022), it is possible that a prominent role is also played by zombie worms.

That said, we must emphasise that although Osedax is no longer regarded as a whale bone specialist (Rouse et al. 2011), large natural populations of this siboglinid have hitherto been reported only from whale bones and skeletons (Smith et al. 2015). Furthermore, there is no indication at present that Osedax may live and feed on the largely cartilaginous elements that comprise the selachian endoskeleton (Rouse and Goffredi 2023), such that the ‘Osedax effect’ is not expected to affect the preservation of these materials. Also, shark teeth are often small and lightweight, so much so that those lying on the seafloor in high-energy depositional settings may be easily mobilised by bottom currents, thereby hindering the successful colonisation by zombie worms. On a different but related note, due to their inconspicuous volume and organic content, very small teeth such as those of many deep-sea sharks are unlikely to be frequently colonised by Osedax spp. On the other hand, we may expect Osedax to be able to exploit large, modified dermal denticles such as the fin spines of some shark species, which share the same general architecture as teeth, featuring a dentine base and ‘trunk’ coated by an enameloid-bearing ‘cap’ (Tovar-Ávila et al. 2008). Further deployment experiments may shed more light on this and other aspects of Osedax’s exploitation of shark-derived dental and dermal tissues.

All things considered, an ‘Osedax effect’ likely exists in relation to sharks in the form of a progressive destruction of the dentine portions of the teeth, including the root and the ‘core’ of the crown. However underinvestigated this selachian-related ‘Osedax effect’ may be, it is likely less generalised than hypothesised for whales and other marine tetrapods, as it appears to leave the largely cartilaginous shark endoskeleton essentially unaffected.

The recent observation of Osedax colonising and feeding on modern shark tooth roots under experimental conditions has raised the question whether the shark tooth dentine represents the primordial nutrition source of this peculiar siboglinid genus (Rouse and Goffredi 2023). By demonstrating that such a behaviour can and does occur in the wild (i.e., outside experimental conditions), our discovery of Osspecus on Pliocene shark teeth gives this hypothesis some credibility, especially in relation to an early Mesozoic or Palaeozoic scenario for the origin of Osedax (Rouse and Goffredi 2023). Such a scenario is at least consistent with the fossil history of sharks, which have wandered the Earth’s seas for some 250 million years, with the palaeontological record of chondrichthyan-like teeth being even deeper, extending as it does back some 425 million years (e.g., Andreev et al. 2022). Although the rates of tooth replacement of the Palaeozoic ‘sharks’ were much lower than those characterising the present-day selachians (Botella et al. 2009), which in turn would imply a more limited availability of teeth on the seafloor, the remarkable antiquity of the cartilaginous fishes could still provide a reasonable context for a deep-time scenario for the emergence of the zombie worms, which nonetheless may also have taken place on bony fish skeletons (Rouse and Goffreddi 2023). In principle, a systematic search for Osedax-related borings on fossil marine vertebrate remains (including shark teeth) from lower Mesozoic and Paleozoic deposits may thus reveal that the zombie worm lineage is much older than thought (Rouse and Goffreddi 2023). However, this tantalising hypothesis is not without some major problems.

First and foremost, we are to remember that extensive natural populations of Osedax are only known to occur on the skeletal remains of cetaceans, and that the same appears to hold true for the Cenozoic record of Osspecus. However limited the Cretaceous evidence for Osedax may be, it seems to conform to a similar pattern, where Osspecus concentrates on marine reptile bones. Considering all the published occurrences of both Osspecus and Osedax, there is no evidence at present for a widespread diffusion of the latter on substrates other than the bony remains of marine tetrapods. This pattern is not particularly suggestive of Osedax having originated on shark teeth, exploiting the latter’s dentine as its original nutrition source.

Secondly, and perhaps more importantly, although the timing of the main evolutionary events within the siboglinid clade remains largely unexplored by means of molecular clocks, the available estimates suggest that Osedax did not originate prior to the Cretaceous (Vrijenhoek et al. 2009; Rouse et al. 2011; Taboada et al. 2015). These estimates are broadly consistent with the chronostratigraphic distribution of Osspecus, which has its first appearance in the Albian (uppermost Lower Cretaceous) (Danise and Higgs 2015; Jamison-Todd et al. 2024) (Fig. 7). That said, the recent discovery of multiple Osspecus ichnospecies on bones from Cenomanian (lowermost Upper Cretaceous) deposits has been taken as evidencing a similarly high number of co-occurring Osedax species, which in turn would be indicative either of a rapid diversification of the zombie worms after their emergence in Albian times or of an origin older than previously thought (Jamison-Todd et al. 2024). Although it is somewhat likely that Osedax originated earlier than the Albian first occurrence of Osspecus, that the latter has never been reported to occur on marine vertebrate remains from the reasonably well-prospected Jurassic (Jamison-Todd et al. 2024) is quite suggestive of a genuine absence before the beginning of the Cretaceous. Ultimately, that the molecular clocks and the fossil record converge in suggesting a Cretaceous age for the emergence of Osedax (Fig. 7) is hard to dismiss. If a Cretaceous origin is assumed for Osedax, then hypothesising that such an origin occurred on a substrate other than the bony tissues of (large) marine vertebrates may not be the most parsimonious solution. All things considered, shark tooth roots are unlikely to have been the ancestral substrate and nutrition source of Osedax.

Figure 7. 

Stratigraphic range of marine vertebrate groups associated with trace fossil evidence for ancient Osedax populations. The molecular age estimate for the origin of Osedax is after Vrijenhoek et al. (2009). Data on modern and fossil occurrences of Osedax are after Danise and Higgs (2015), Rouse et al. (2018), Jamison-Todd et al. (2023, 2024, 2025a, b), Ibrahim et al. (2025) and the present work. Note that only reports substantiated by microtomographic analyses are included here, and that Osspecus occurrences originating from roughly coeval deposits were subsumed into a single record if relative to the same marine vertebrate group.

At this point, it can be speculated that the colonisation of inconspicuous substrates such as loose shark teeth would generally begin in the vicinity of large whale falls, as an offshoot (as it were) of the whale fall community, and that the same could have happened in Mesozoic times starting from large marine reptile skeletons. By becoming home to small, peripheral Osedax colonies, the colonised teeth would act as critical benthic islands as well as ecological stepping stones between remote whale falls, which in turn have long been regarded as serving as dispersal and evolutionary stepping stones for a variety of chemotrophic deep-sea invertebrates (Smith et al. 2015, and the many references therein). Ultimately, that many ocean floors are known to be covered with a ‘carpet’ of shark teeth (e.g., Litvinov 2007) is at least consistent with this hypothesis.

A major stratigraphic gap still exists between the youngest Mesozoic occurrence of Osspecus, which dates back to the latest Cretaceous (Jamison-Todd et al. 2024), and its oldest Cenozoic occurrence, which dates back to the late middle Eocene (Ibrahim et al. 2025) (Fig. 7). Whether the near-ubiquitous availability of shark teeth at the seafloor helped Osedax to survive through the end-Cretaceous (K–Pg) mass extinction is an open question. On the one hand, considering that sea turtles are the only extant group of marine vertebrates whose Mesozoic fossil record has hitherto yielded evidence of colonisation by Osedax (Danise and Higgs 2015), such survival may have been possible first and foremost on chelonioid skeletons and bones. On the other hand, considering that the K–Pg event has been recently shown to have had little impact on at least some elasmobranch communities (Feichtinger et al. 2025), deep-sea accumulations of shark teeth may well have played a role in permitting the post-Cretaceous survival of zombie worms.