Behavioral implications of an embedded tyrannosaurid tooth and associated tooth marks on an articulated skull of Edmontosaurus from the Hell Creek Formation, Montana

Predation or scavenging?

Whether T. rex and other tyrannosaurids were predators or scavengers has been a subject of disagreement in the literature citing various lines of evidence including tyrannosaurid physical attributes and interpretations of tooth marks/pathologies (e.g., Carpenter, 1998; Wegweiser, 2004; Carbone, Turvey & Bielby, 2011; Murphy, Carpenter & Trexler, 2013 vs Horner & Lessem, 1993; Horner, 1994; Therrien et al., 2023 respectively). Given the nature of the fossil record and lack of perfect modern analogues, it is unlikely that this debate will be satisfactorily settled. Further complicating matters, a majority of modern carnivores lie on a predation-scavenger spectrum, rarely relying solely on either strategy to procure food and establishing which end of the spectrum modern carnivores are more inclined is extremely difficult (DeVault, Rhodes Jr & Shivik, 2003; Holtz Jr, 2008). Although some recent literature conceptualizes tyrannosaurids as both predators and scavengers (e.g., Holtz Jr, 2008; Horner, Goodwin & Myhrvold, 2011; Kane et al., 2016), some publications still describe adult tyrannosaurids as “apex predators” (e.g., Therrien et al., 2023).

It may be more pragmatic to conceptualize tyrannosaurids as both predators and scavengers. Further, being able to tease out which behaviour was more commonly practiced may be difficult with the information available for an extinct animal. The best that can probably be attained from the fossil record is how tyrannosaurids apprehended prey and manipulated carcasses. We note that the possibility of bite marks resulting from scavenging cannot be eliminated without direct evidence that the proposed prey animal was alive at the time of the bite (e.g., pathologies resulting from failed predation attempts). Of the pathologies cited as evidence of tyrannosaurid predation, most examples are injuries to the tail of hadrosaurids (Carpenter, 1998; Murphy, Carpenter & Trexler, 2013; DePalma et al., 2013; but see also Bertozzo et al., 2025) with one example of damage to a rib of a lambeosaurine (Wegweiser, 2004) and another possible example that, if the pathology resulted from a bite, preserves evidence of an attack to the head of the ankylosaur Tarchia (Tumanova et al., 2025). Damage to the neural spines of the tail is interpreted as evidence of a bite delivered from behind while the hadrosaurid was trying to escape (Carpenter, 1998; Murphy, Carpenter & Trexler, 2013; DePalma et al., 2013). Although the etiology of the Tarchia skull pathologies is unclear, if they are indeed the result of a predation attempt by a tyrannosaurid, then the bite may have been delivered posterolaterally (Tumanova et al., 2025). However, Molnar (1998) suggested that tyrannosaurids, like modern canids, may have killed their prey by delivering a powerful crushing bite to the snout of a prey animal ultimately cutting off the airway and resulting in suffocation—a scenario virtually identical to the predation scenario proposed here that appears to be most consistent with the emplacement of the tooth in the skull of MOR 1627.

Embedded tooth and group 1 tooth marks

From the information presented above, the embedded tooth likely represents a middle-posteriorly positioned tooth from an adult Tyrannosaurus. The slight curvature of the tooth indicates that the mesial surface of the tooth is closer to the posterior end of the Edmontosaurus skull. Combined, this suggests that the tooth broke off during a bite that came from the front. The presence of tooth marks more anteriorly on the snout of MOR 1627 in the same region as the embedded tooth and on the same surface with a broadly similar orientation as the embedded tooth suggests that multiple teeth made contact during this bite. Under this scenario, the most anteriorly positioned contacting tooth would have penetrated the nasal while more posteriorly positioned teeth contacted the bone. The lack of bone remodelling in this region further indicates that the bite was inflicted at, near or after the time of death. Based on the behaviours of large carnivores, bites to the face, particularly to the snout of a herbivore, are inflicted in an attempt to gain control of a struggling prey animal during predation (mammals, crocodiles), to suffocate the prey animal (mammals), drown the prey animal (crocodiles) and/or to manipulate the carcass after the animal is already dead (Emerson & Radinsky, 1980; Cleuren & De Vree, 2000; Karanth & Sunquist, 2000; Njau, 2006; Njau & Blumenschine, 2006; Van Valkenburgh, 2007; Wiesel, 2010; Westaway et al., 2011; Behrendorff, Belonje & Allen, 2018).

Tooth marked specimens thought to be produced by tyrannosaurids are often attributed to carcass consumption (Erickson & Olson, 1996; Fowler & Sullivan, 2006; Hone & Watabe, 2010; Hone, Tanke & Brown, 2018; Brown, Tanke & Hone, 2021) rather than predation (Carpenter, 1998; Wegweiser, 2004; Murphy, Carpenter & Trexler, 2013). This is because most examples of tooth marks are very difficult to attribute to tyrannosaurids and subsequently interpret as evidence of predatory behaviour, especially given most tooth marks do not show signs of healing and so were produced at or near the time of death of the animal being fed upon making it hard to determine the sequence of events. Often, tooth mark interpretations involve whether predation or feeding is the most likely mechanism behind the production of those tooth marks but this is by no means definitive. Based on observations of modern carnivores, bites to the snout are typically inflicted as a means of prey capture or carcass manipulation (Emerson & Radinsky, 1980; Cleuren & De Vree, 2000; Karanth & Sunquist, 2000; Njau, 2006; Njau & Blumenschine, 2006; Van Valkenburgh, 2007; Wiesel, 2010; Westaway et al., 2011; Behrendorff, Belonje & Allen, 2018; Somaweera et al., 2018), rather than to consume flesh from the muzzle as relatively little flesh is found in this region; although, removal of flesh from this region is not entirely out of the realm of possibility (Haynes, 1980; Haynes, 1982; Boaz et al., 2000). Consumption of flesh from the snout region of a hadrosaurid is unlikely as very little flesh would have been present and the snout would have been covered with hard-to-digest, heavily keratinized skin and rhampotheca (Fisher, 1981; Haynes, 1982; Novak et al., 2005; Bell, 2014). Thus, two possible interpretations for MOR 1627 exist: emplacement during predation versus emplacement during carcass manipulation, either following predation or while scavenging. Falsification of the scavenging hypothesis would require evidence that the Edmontosaurus was alive when bitten, either through evidence of bone healing around an emplaced tooth (e.g., DePalma et al., 2013) or other means. Although scavenging of MOR 1627 cannot be ruled out, here we argue that in the case of MOR 1627, predation is the more likely mechanism of emplacement for the embedded tooth given the directionality of the bite, the number and relative positioning of tooth marks and the level of articulation of the rest of the skull.

The positioning of the embedded tooth suggests the bite was directed from the front, while the Edmontosaurus was facing the tyrannosaurid. If the tooth was embedded during manipulation of the carcass after the animal was already dead, surely it would have been easier to grab the snout from either side with the upper jaws on one side of the skull and the lower jaws on the other side of the skull. Although rarely described, carrying behaviour in modern carnivores generally appears to leave multiple scores or pits on long bone diaphyses and multiple and/or paired marks on the cranium with punctures and/or furrows being relatively rare (Pokines, 2021). Mammalian carnivores tend to carry carcasses that are smaller compared to their body size, while larger carcasses tend to be consumed on-site or more mobile elements such as limbs (and rarely heads, which appear to be facilitated by grabbing the neck) are disarticulated and carried away to be eaten elsewhere (Windell et al., 2019). The penetrating nature of the tooth of MOR 1627 on the snout indicates high forces were exerted during the bite, something that is not expected during carrying behaviour after the animal has already been incapacitated. In modern carnivores, high bite forces are often delivered to dispatch prey or maintain control of a struggling animal from a position where the prey is unlikely to inflict damage via horns, tusks or teeth. Felids and spotted-tail quolls will apply a single, crushing bite to the head or throat of larger prey to reduce handling time and decrease the chances of injury (Jones & Stoddart, 1998; Mondini & Muñoz, 2008; Meachen-Samuels & Van Valkenburgh, 2009). In quolls the bite is delivered from behind to the braincase while in most felids, the attack begins from behind and the forelimbs are used to maintain a grip on the prey animal while the felid repositions itself to deliver a killing bite to the throat. Legless lizards will apply a similar strategy, biting relatively large prey by the head or neck and waiting for asphyxiation or exhaustion to occur (Wall & Shine, 2007; Mukherjee & Heithaus, 2013). In cooperative species, such as lions or African wild dogs, particularly for more dangerous prey with horns or tusks, bites to the head are used to control and/or drag down prey animals while other members of the group deliver the killing blow (Kruuk & Turner, 1967; Mukherjee & Heithaus, 2013). Other predators such as civets, Komodo monitors, and some canids will employ the opposite strategy, increasing handling time to reduce injury by delivering a series of quick bites to other areas of the body to exhaust prey or induce shock prior to applying the final crushing bite to the head or neck (Eisenberg & Leyhausen, 1972; Jones & Stoddart, 1998; Therrien, Henderson & Ruff, 2005; Meachen-Samuels & Van Valkenburgh, 2009). Crocodiles are ambush predators that will grab the legs or snout of a prey animal at the waters edge before dragging it into the water to be dispatched (Heller, 1918). Such attacks typically occur while the prey animal is drinking, facilitating a front-facing attack by the crocodile. Although a wide variety of hunting strategies can be seen in the animal kingdom, single bites to the head appear to be related to prey control or rapid dispatch of prey. Thus, the single, powerful bite to the snout emplacing the tooth in the nasal of MOR 1627 would appear to be more consistent with behaviours exhibited during predation rather than behaviours observed during feeding.

The high level of articulation of the skull also suggests predation. Most modern mammalian carnivores will consistently consume a carcass starting with higher economy areas, such as the abdominal cavity, and gradually consume lower economy areas, with the head usually being the last region consumed (Blumenschine, 1986; Richardson, 1980; Selva et al., 2003; White & Diedrich, 2012; Haynes & Klimowicz, 2015; Drumheller et al., 2020). Hyaenas are one of the few large mammalian carnivores that will occasionally deviate from this pattern, using multiple crushing bites to open the braincase to consume the brain relatively early on in the carcass consumption sequence (Richardson, 1980; Blumenschine, 1986; Boaz et al., 2000; Wadley, 2020). Comparison of encephalization quotients, which are calculated values quantifying relative brain and body masses, suggest a hyaena-like approach would have been less beneficial for dinosaurian carnivores. In modern megaherbivores such as blue wildebeest (Connochaetes taurinus, REQ = 26.7), impala (Aepyceros melampus, REQ = 27.0), African buffalo (Syncerus caffer, REQ = 23.6) and African elephants (Loxodonta africana, REQ = 73.2) (Hurlburt, 1996), values calculated using the reptilian encephalization quotient formula are much higher than those estimated for hadrosaurs (Iguanodon, REQ = 2.6; “Anatosaurus”, REQ = 2.4; Edmontosaurus, REQ = 2.8) (Hurlburt, 1996; Evans, 2005). Consideration of previous examples attributed to cannibalistic scavenging would be consistent with the interpretation that tyrannosaurids followed a similar consumption sequence to modern, non-hyaenid carnivores wherein higher economy areas such as the abdominal cavity with cranial flesh being utilized later after most of the postcranial flesh has been consumed or decomposed (Longrich et al., 2010; Hone & Tanke, 2015). Under the typical carcass consumption sequence observed in modern carnivores, damage to the cranium is mainly concentrated around areas where the musculature is located (such as the coronoid process), occipital condyles and/or parietal bones and orbital margins (Richardson, 1980). The nasals of ungulates can also be heavily damaged or fragmented during carcass consumption. In some cases, the mandibles will also become disarticulated once the muscles and skin attaching them to the rest of the skull have been consumed (Richardson, 1980). Conversely, under a predation scenario, some damage might be inflicted during prey capture but the skull would remain otherwise unmodified as higher economy areas such as the internal organs in the abdominal cavity would be one of the first areas to receive attention after the animal was killed (Richardson, 1980; Blumenschine, 1986; Selva et al., 2003; White & Diedrich, 2012; Drumheller et al., 2020).

Other tooth marks

The spacing between the marks of group 2a on the left jugal are consistent with the spacing of tooth marks previously attributed to tooth marks produced by the premaxillary teeth (Erickson & Olson, 1996; Fowler & Sullivan, 2006; Hone & Watabe, 2010; Rivera-Sylva, 2012; Drumheller et al., 2020; Lei et al., 2020). Hone & Watabe (2010) have suggested that these teeth have been used in “scrape feeding” to remove flesh close to the bone surface. What exactly the tyrannosaurid was consuming in this region is unclear as the adductor musculature—the major source of flesh on the skull—would have been tucked underneath the jugal and postorbital (Rybczynski et al., 2008; Holliday, 2009; Nabavizadeh, 2020a; Nabavizadeh, 2020b) (Fig. 7) and accessing these muscles would have presumably moved these bones out of articulation. Recently, Sharpe et al. (2025) proposed another soft tissue in this region known as the exoparia which is reconstructed as extending from the lateral surface of the jugal flange to the surangular. Removal of this structure would have left the skull largely articulated if most of the adductor musculature underneath was largely intact. However, the tooth marks in group 2a are located near the posterior margin of the postorbital fenestra which is positioned more dorsally than the flange of the jugal where the exoparia is reconstructed (Fig. 7). One possible, albeit less likely explanation for this is the exoparia extends higher on the jugal than originally proposed. A more likely explanation is that the tooth marks in group 2a were produced by teeth of the right premaxilla contacting the bone surface while teeth of the left premaxilla removed the exoparia.

Spacing between tooth marks of groups 2b, 3, 4 and 5 are much greater than those seen for group 2a marks (Fig. 1). Although this could suggest that these tooth marks were produced by more lateral teeth, a scenario where teeth from both premaxillae are contacting the bone can also be envisioned. The stage of tooth eruption, angle at which teeth contact bone and the surface topography of the bone will theoretically impact what teeth contact the bone surface (Hone & Rauhut, 2010; Lei et al., 2020). It is possible that teeth from both premaxillae or at the very least anterior teeth from opposite sides of the skull were involved in the production of group 3–5 tooth marks but only three teeth made actual contact with the bone surface—one tooth from one premaxilla and two teeth from the other or one anterior maxillary tooth on one side and two teeth from the premaxilla on the opposite side. What teeth produced the remaining tooth marks is uncertain given ascertaining tooth spacing is impossible from a single tooth mark. Given the proximity of group 5 tooth marks to m. pterygoideus ventralis, it is possible that these tooth marks were inflicted while feeding on this muscle (Rybczynski et al., 2008; Holliday, 2009; Nabavizadeh, 2020a; Nabavizadeh, 2020b) (Fig. 7). The close proximity of the remaining tooth marks on the right mandible to possible articulation points of various adductor muscles (Rybczynski et al., 2008; Holliday, 2009; Nabavizadeh, 2020a; Nabavizadeh, 2020b) (Fig. 7), would also suggest that these tooth marks were produced during feeding on the adductor musculature. However, it is hard to imagine entire muscle masses being removed without disturbing the overlying jugal in most cases so these tooth marks may either represent removal of only portions of the muscles or consumption of soft tissue other than the adductor musculature. It is also possible that some of the tooth marks concentrated on the coronoid process of the right mandible are associated with the removal of the exoparia which extends from the lateral surface of the jugal flange to the surangular (Sharpe et al., 2025) and could have been removed without disturbing the jugal. Although Sharpe et al. (2025) were unable to determine the exact nature of the exoparia, targeted feeding on the exoparia suggests it had higher nutritive value which would implicate the exoparia was muscular rather than ligamentous in nature. Alternatively, the relatively shallow nature of the exoparia could have made it particularly easy to remove thus making the benefit outweigh the cost of its removal.

The cracking observed on the left and right jugals and the right mandible that parallel the bone fiber is consistent with subaerial exposure prior to burial (Behrensmeyer, 1978; Pokines et al., 2018). The level of cracking on the left jugal is far less pronounced than that on the right jugal and mandible, suggesting that the skull was positioned left-side-down, leaving it exposed for a shorter period of time. The presence of this cracking on both sides of the skull indicates that skull of MOR 1627 would have been sitting exposed to the surface for some period of time, likely with the right side of the skull exposed for longer than the left (Behrensmeyer, 1978). Whether this exposure was before or after the majority of the tooth marks on the skull were inflicted cannot be determined as there is little difference in the physical appearance of tooth marks inflicted prior to weathering and those inflicted on the surface of weathered bone (Vachirawongsakorn, 2019).

Carcass abandonment?

The high level of articulation of the skull also suggests something about the nature of the skull at the time it was buried. As previously discussed, the level of articulation suggests that some, if not most of the adductor musculature had not been consumed by hungry carnivores. This adductor musculature could have also prevented disturbance of the elements during burial in a higher energy environment, as indicated by MOR 1627 being recovered from a sandstone. It therefore appears likely that at least some of the adductor musculature was left intact at the time of burial. Although it is possible the skull was buried before carnivores could have consumed this region, it seems odd that the exoparia and possibly some of the other adductor muscles were consumed while the majority of the adductor musculature remained untouched. Presumably, removal of the jugal and quadratojugal to access the underlying musculature would have been relatively easy for a tyrannosaurid to have accomplished if it were sufficiently motivated given the deeper adductor musculature (e.g., mAMEP and mAMEM) would have accounted for a majority of the flesh on the skull. Although it is possible for the skull to have been buried before the skull could have been disarticulated to access this musculature, the weathering present on the jugal suggests the skull sat subaerially exposed for some period of time prior to burial. Considering the behavior of modern carnivores, we propose that MOR 1627 is an example where the carcass of a predated animal had been mostly, but not completely, utilized before it was abandoned. Given tyrannosaurids appear to have followed the typical carcass consumption sequence of modern non-hyaenid carnivores (e.g., Hone & Tanke, 2015), more superficial tissue from the skull (e.g., the exoparia) was likely removed after most of the postcranial tissue had been utilized, producing the remaining tooth marks on MOR 1627. During the process of carcass consumption, modern carnivores will disarticulate, fragment and/or carry off parts of the postcranial skeleton away from the main carcass, especially when competition between carnivores, either of the same or different species, is higher (Capaldo & Peters, 1995; Haynes & Klimowicz, 2015; White & Diedrich, 2012; Wadley, 2020). Of particular significance are observations regarding the consumption of a large kudu carcass reported by Wadley (2020). After a few weeks of being fed upon by carnivores, all that remained were a few fragments of the postcranial skeleton (about 5.5 g worth) and the skull. Although the skull had been dragged away from its original position, the skull was virtually untouched and the brains had not been fed upon despite the presence of hyaenas at the site. If MOR 1627 underwent a similar process before it was abandoned, only a few postcranial fragments would have been present and could have easily washed away during burial. There is also precedence in the modern record for the same carnivores to repeatedly visit a kill site over a period of time before abandoning the skull (sometimes with the neck also articulated) with relatively little tooth marking or flesh removal (Richardson, 1980; Haynes, 1982; Capaldo & Peters, 1995; Haglund, 1997; Haynes & Klimowicz, 2015; Wadley, 2020). Given the spacing of the tooth marks on the left jugal and right mandibles are consistent with that of a Tyrannosaurus, and the embedded tooth is also that of a Tyrannosaurus, it is plausible for MOR 1627 to represent a scenario where the same animal, or group of animals, repeatedly visited the carcass over a period of time before abandoning the skull, with most of the adductor musculature intact, in favor of a fresher carcass.