Pre- and postmortem tyrannosaurid bite marks on the remains of Daspletosaurus (Tyrannosaurinae: Theropoda) from Dinosaur Provincial Park, Alberta, Canada

Premortem

The vast majority of the identified marks on the skull are premortem, indicating that numerous injuries or infections occurred at various times in the life of the individual. These must also have occurred prior to death at such a time as to allow for evidence for healing and repair to be visible. Premortem damage is considered as such based on evidence of healing through pathology, or presence of finely pitted reactive bone and also anomalous directions of pits combined with raised ridges (Rothschild & Martin, 1993; Tanke & Rothschild, 2002).

A. A bite on the tip of the snout with bulged and pathological area on the right ascending process of the premaxilla, and apparent fusion of the right and left premaxillary processes (Figs. 1–3). This damage is associated with a small subcircular mark approximately 13 mm in diameter, and 6 mm in depth, on the anteriolateral face of the left premaxilla. A second suboval (16 mm long by 6 mm across, and less than 2 mm deep) mark lies 6 mm posteriorly to this at the juncture of the right premaxilla and right nasal. The long axis of the oval is subparallel to the ramus of the premaxillary process.

B. A lesion which is close in form to a bite-and-drag mark (anterior to dorsally directed) on the anterolateral part of left maxilla (Fig. 2). This is 22 mm long by 8 mm (widest point) and maximum depth of the puncture is 1.5 mm. It is a distinctly different colour from the surrounding bone, being more of a burnt-orange than brown.

C. Comma-shaped damage anteroventral to the promaxillary fenestra on the right maxilla (Fig. 2). This is 22.5 mm long, 6.5 mm wide at the top and 1.5 mm wide at base. It appears to be a bone avulsion that was torn out, and is recessed in the excavation of the maxilla. There is evidence of healing on the maxilla around the edges of this damage, but not apparently on the avulsion. This is close in appearance to some tyrannosaurine bite-and-drag marks (sensu Hone & Watabe, 2010), and here is anteroventrally directed.

D. A series of lesions and injuries on the right maxilla (Fig. 1). Two very small lesions lie on the posteriorly directed ramus of the right maxilla, below the right nasal. The first is of mild osteomyelitis and tracks the upper margin of the right promaxillary fenestra (Fig. 4). This is a small subcircular lesion, 7 mm long and 6 mm tall and 2 mm deep. The second lies in line with the long axis of the left naris, and is a large and subcircular lesion measuring 15 mm in diameter and with a maximum depth of 2 mm. This lies alongside the edge of the maxilla, but does not extend onto the other element.

Just dorsal to the seventh maxillary alveolus is a groove that is posteroventrally directed, it is 16.5 mm long is up to 6.5 mm wide and 2 mm deep (Fig. 5). Similarly, on the ventral margin of maxilla above ninth alveolus is a dorsoventrally directed notch 11 mm tall, 4 mm in width and up to 2.75 mm deep.

A pair of conjoined subcircular depressions are apparent on the anterodorsal edge of the right maxilla, close to the suture with the right premaxilla. A major lesion lies on the anterior part of the right maxilla, represented by a moderately thickened patch (slightly raised by 1 or 2 mm compared to the surrounding bone) of reactive bone that is suggestive of osteomyelitis. This area measures 48.5 mm tall, and is up to 31 mm wide.

E. A large subcircular puncture (9 mm across, approximately 2 mm deep) on the dorsal part of the middle of the left nasal (Fig. 3). The right nasal also has small lesion (10 mm long by 5 mm) just dorsal to the posterodorsal edge of the narial opening. A second lesion occurs on the right side of the right nasal where it contacts the maxilla, lying 30 mm posterior to the posteriormost part of the right naris (Figs. 1 and 2). This lesion is subcircular and 14 mm in diameter, and around 1 mm in depth. It is composed of darker bone colour than the surrounding tissue and is finely pitted both inside the lesion and also around its margin. There are two slightly raised irregularly shaped ‘islands’ of bone that sit within the lesion, these are prominent and have a smoother texture than the other pitted bone.

F. A lesion on the right jugal that penetrates the bone fully (Fig. 1). A semicircular area which is estimated at 11 mm tall and approximately 7 mm across (based on the lack of a posterior part) is missing from the element. Mildly reactive bone is present around the edge of this lesion, increasing the overall length of the damage to 24 mm by 14.75 mm, with an anteriorly a swollen area showing evidence of osteomyelitis. A small premortem lesion, ventral to the above described one, is also present and measures 11 mm long by 8.5 mm tall.

G. A pit on the left lacrimal is 6.5 mm across and 3 mm deep. See also below.

H. A posteriorly directed lesion, that includes a pit in the anterior part, that lies on the left posterior edge of the postorbital descending process, just above the tip of the ascending process of the left jugal (Fig. 6). The lesion is 2 mm deep, 13 mm long and 5 mm wide and shows signs of healing, with the medial face shows swelling, and is invaginated into the cavity. Medially, there is a slight overlap of the fragments that make up the lesion showing that after the break occurred and before or during healing one part slipped over the other, prior to being fused into their current positions.

I. A deep and elongate lesion with a rimmed margin on the right postorbital (Fig. 6). It lies halfway down the anterior edge of descending process. The lesion is 18.5 mm in length and 9 mm wide (including the raised edges) and with a maximum depth of 2.5 mm. The surface of the lesion shows mildly reactive bone.

J. There is a light patch of osteomyelitic bone, diagnosed by texture changes, present on the left quadratojugal, in the middle of the element.

K. Damage to both sides of the nuchal crest. When seen in posterior view, the nuchal crest is an unusual shape and apparently the two sides have both been damaged though in different ways (Fig. 7). Although partly broken, a natural, finished bone surface is present on much of the dorsal margins. The right side has a semicircular section missing, lined by normal cortical bone, and suggestive of a hole in the flange. This is similar in form to an injury seen in a specimen of Troodon that may have been a cyst (Currie, 1985) or a tooth puncture wound (Tanke & Rothschild, 2002). On the left side of the crest, the dorsal edge curves very rapidly in a ventral direction and, although now broken, would not have met the rising ventrolateral margin. Again however, the edges are natural, suggesting the shape is mostly genuine and not the result of breakage, though this is considered to be likely be pathological or trauma induced. There is only limited distortion elsewhere in the element, suggesting again the feature is probably genuine.

L. Damage to the lateral margins of the dentary (Fig. 8). These are present mostly on the ventral side and this element, and these consist of subcircular puncture wounds, or pairs of elongated ridges that are indicative of former scores. One point on the extreme ventral surface is barely visible in lateral or medial view, and is represented by a slight bump associated with a slight score. An additional puncture lies on the lateral side of the right dentary, ventral to the meeting point between the sixth and seventh alveoli. This mark is 4 mm tall by 5 mm long and 1 mm in depth. There is a small rectangular piece of bone that lies on the anterior edge of the mark that appears to have been lifted up (an avulsion) and settled back into place and subsequently fused back with the bone.

M. Two holes, and a large patch of ostomyelitis are on the right surangular. This ostomyelitic bone is stronger anteriorly, and then fades towards the mandibular fenestra. Additional osteomyelitic bone lies above and posterior to the fenestra and there is still more along the dorsal edge of this element, including small ‘hump’ on the posterior part of the surangular (Fig. 10). Two damage points from preparation and/or excavation also lie on the right surangular

N. The posterior carina of fifth right maxillary tooth has suffered serious ablation (Fig. 5). This damage is here attributed to occlusion wear with the dentary tooth row (Schubert & Ungar, 2005). The first and third teeth of the right maxilla both have their tips broken and with subsequent wear leading to smoothing of the affected areas. The eleventh tooth of the right dentary has also suffered damage which is interpreted here as being a result of premortem malocclusion.

A number of other very fine marks and cracks are present in places across the cranium. These may represent possible osteopathic marks and traces, but could be taphonomic artefacts based on the partial separation of material and the dorsoventral collapse of the nasals and other parts. As they are minor it is difficult to determine their origins, and these are not discussed further. In addition to the cranial damage, one dorsal rib shows evidence of having been fractured, but this was well healed at the time of death.

Postmortem

Postmortem damage on the material is assessed based on the lack of any indications of healing (swelling, reactive bone etc.) with tooth-marks being identified as being sub-parallel traces damaging the bone cortex. One clear set of bite-marks appears on the medial face of the posterior part of the right dentary. A number of other elements show possible evidence of biting with breaks aligning between elements or with localised impacts on bone.

i. Four tooth marks on posteromedial part of dentary are considered to be the result of biting (Figs. 11 and 12). These are well spaced (around 15–20 mm between each) and light—they barely graze the surface of the bone. There is no evidence of healing of these implying they were inflicted either postmortem, or very shortly before death. However, the position of these marks on the medial face of the extreme posterior part of the dentary suggests they were unlikely to be premortem.

ii. A dislodged bone fragment lies between the two posteriomost alveoli preserved in the right dentary. This fragment of bone has apparently been forced down into the space between the alveoli (Figs. 11 and 12). Given that the original position of the dentary as recovered was inverted, this suggests that a strong, but localised, impact broke and drove this fragment into this position. It is therefore indicative of a bite, indeed probably the same one that delivered the sub-parallel damage described above in section (i). We consider this damage unlikely to be the result of trampling given that the element is otherwise intact, and also not likely due to fluvial action given the relative positions of the dentary and cranium.

iii. Breaks across the posterior parts of the right dentary. As preserved these are considered to be likely the result of a bite. The edges of the breaks are rough and indicative of a break of green bone, and are roughly in line with the series of four scores on the dentary described above (i). The lack of fluvial damage to the rest of the specimen suggests this is an inflicted injury. Similarly, the extreme dorsal anterior part of the surangular is damaged and may be linked to the bite at the posterior part of the dentary. Part of the dentary posterior to this point has separated from the rest of the element and is now lost; however, there are photographs of this part prior to this loss showing the original condition of the material (Fig. 12).

iv. Broken process on the right splenial. Approximately 59 mm of the anterodorsal process of the right splenial was broken off when the element was found (Fig. 13). There is damage to the dorsal margin of the broken part and the dorsal margin of the splenial as a whole. Restoring this broken process and then placing the whole element in position on the jaw (using Brochu, 2003, Fig. 40 as a guide) this break aligns very well with the indented bone fragment in the dentary (ii) and the associated bite marks (i). This provides evidence that the right mandible was originally complete and one bite (or more, but very close to each other in position and delivery) damaged both the right dentary, and into and through, the posterior part of the right splenial (Fig. 14).

Indeterminate

There are a number of pieces of the skull that present evidence of damage or breakage, but in which it cannot easily be determined if this was damage from trampling, bites, or general disarticulation/transport, etc. Therefore, although these may represent additional evidence for postmortem damage from another animal, we take a conservative approach here and do not directly assign them as such.

1. There is some form of scrape-like mark down the dorsal part of the left maxilla, close to the suture with the left nasal.

2. The left maxilla shows numerous breaks and damage. The base of the left maxilla (4.5 cm posterior to the anterior edge of the antorbital fenestra) is broken as is the posterioventral part of the maxilla, and some pieces of this element have separated from the main body (Fig. 2). These breaks can be aligned with the damage seen to the base of the left lacrimal, which is missing the anterior process and most of the ventral process, as well as the broken palatal shelf. All three areas (maxilla, lacrimal and palatal shelf) show evidence of breaks of green bone, and not just disarticulation or recent erosion. It is at least possible that this area was collectively damaged and/or removed as a result of a single or multiple bites to the area. This is plausible given the clear bite marks to the left dentary that shows postmortem bites to the skull, but the lack of distinct tooth marks on this part of the skull make it hard to confirm. This may simply be the result of damage through trampling, but then it would be difficult for an animal to step on a skull and break only one side, or leave the palate largely intact while the maxilla was shattered.

3. The anterior part of left pterygoid is broken and missing. The break occurs in a position approximately in line with the midpoint of the maxilla-jugal suture (based on comparisons with other tyrannosaur skulls) and may be linked to the above described missing portion of the right maxilla and palate (point 2).

4. The anterior ramus of the right palatine is broken. Approximately 20 mm of bone missing based on comparison to the complete left element and the remainder of the anterior ramus is 67.5 mm long. This was found loose in the quarry, and was later restored to the main body of the right palatine. The posterior parts of both posterior rami of the right palatine are also broken and lost.

5. Left exoccipital has been broken away from the braincase. The braincase is relatively complete, but the left exoccipital suffered damage at its base and three loose pieces of this were found in the quarry. Two of these pieces can be fitted to the main body with the respective broken edges lining up well. There are no traces of bite marks or punctures suggesting that this was a natural break or caused through trampling.

6. A number of marks on the dentary teeth and isolated crowns may be pre- or post-mortem. These are not considered further as they may be as a result of combat, postmortem damage or malocclusion (Fig. 15). Two disarticulated teeth (one already broken) show possible marks that are similar to tooth marks on tyrannosaurine teeth held in the Tyrrell collections (e.g., TMP 1988.050.0145, TMP 1988.050.0146) but it is not possible to determine when these marks occurred.

Taphonomy/burial and evidence for scavenging

Collectively the specimen most closely resembles Taphonomic Mode B of Eberth & Currie (2005) for the Dinosaur Park Formation. That is, although the specimen is largely incomplete the vast majority of the recovered material relates to a single individual skeleton. These are primarily recovered in palaeochannel facies (Eberth & Currie, 2005) and some common themes in such theropod specimens (their Table 24.2) suggest the possibility of local reworking, and that the animal may have died through drowning or disease/injury. Easily transported microvertebrate fossils were found scattered in the matrix that surrounded the tyrannosaur suggesting some fluvial action. The preserved microvertebrate specimens do not reveal anything unusual with respect to the usual Dinosaur Provincial Park fauna but the presence of champsosaur, crocodylian, and trionychid material among others does also indicate potential fluvial conditions. However, the presence and preservation of the small bone fragments that could be identified and restored to various parts of the main skull and dentary especially show that the water movement was slow.

The specimen as recovered was somewhat dissociated and it is difficult to separate out which of the disarticulation and damage to the skeletal elements may be the result of feeding, and which from simply collapse of a decomposing carcass or possible transport. Certainly the specimen suffered some damage from another theropod, and the damage to the femur suggests damage through trampling. However, the lack of much of the postcranium and the presence of elements from other species in the quarry points to a degree of transport in the local area that moved elements around at times.

For example, the position of the left jugal and lack of damage compared to the left lacrimal may be as a result of the dissociation of the former, followed by a bite or trampling damaging to the latter, but could also have resulted from a bite damaging the lacrimal with the jugal being separated as a result and falling away (Fig. 2). Thus, both the positions of marks and breaks on the individual elements (and lost parts or broken rami found elsewhere in the quarry) must be assessed against their likely original positions. Furthermore, a heavy bite on thin bones (as with a cranium) would not leave deep tooth impressions (cf. Erickson et al., 1996; Hone & Watabe, 2010) but instead might lead to shattering, or at least breaking, of elements. Thus, tooth marks might not be left or easy to discern, despite a strong bite. Our inferences below are therefore tentative given that some major breaks appear in patterns that imply breakage of multiple aligned elements, but without the definitive evidence that a tooth mark would provide.

There is a lack of bite marks on the remaining parts of postcranial skeleton, including areas where meat would be plentiful on a fresh carcass, or even mid-stage carcass consumption (following Blumenshine, 1987) e.g., the ilium, femur, ribs. However, tyrannosaurids seem to leave far more bites on material that do other carnivorous theropods, and even may leave multiple traces with relatively ‘careful’ feeding (Hone & Watabe, 2010). However, given the amount of missing material it is likely that the specimen was exposed for some time prior to burial. The lack of bite marks may therefore be the result of the loss of much material which could have borne them, although and the lack of shed teeth (both of tyrannosaurs and other small carnivores e.g., see Hone et al., 2010) also collectively suggest that the material was not fed upon extensively.

If the animal was relatively well-fleshed the when the carnivore fed upon it, then it would be possible to consume large amounts of material without breaking bones or leaving tooth marks. However, the lack of damage and separation of gastralia and other fine elements like the supradentary suggests this was not part of a normal carcass consumption pattern. The cranium had clearly disintegrated at least in part and elements had separated along suture lines prior to burial, and not primarily as a result of being bitten.

The cranium at least must have been undergoing some decay when the bite occurred. This is based on the extrusion of the dentary teeth, which must have been held in by their ligaments in order not to have fallen out entirely (Fig. 8), but still loosely enough attached that they were all partly extruded from their sockets. Similarly, the loose teeth that had fallen into the palate must have decayed to the point that the ligaments were no longer sufficient to hold them in place. We cannot easily estimate how long it may have taken for tyrannosaur tooth ligaments to decay and separate, but presumably this would have been minimally a number of days rather than hours. Therefore the action that lifted the dentary and repositioned it, presumably coincident with the marks delivered to the dentary, occurred some days after death.

Based on the orientation of the dentary teeth and the preserved position of the dentary, this must have been moved anteriorly (Fig. 17). Given that it was originally in a natural articulation (based on the dentary teeth lying in the palate), it must have been lifted to come into its final resting place. It could not simply have been moved forwards by fluvial action as the teeth of the dentary and maxilla would have interlocked. The delivery of the bite that broke this part of the skull is also very unlikely to have been during intraspecific combat or a predation event given the position of the bite marks up inside the jaws and in such a posterior location. The Daspletosaurus would need to have opened the mouth to an extraordinary degree to allow the jaws of the other animal to have reached this position. A decaying carcass however would likely present no such difficulties.

As described above, two other areas of damage are hard to confirm as bites as opposed to damage through trampling. In addition to the damage to, and movement of, the left dentary, there is a possible bite into the right jaw, damaging the right dentary, right splenial and perhaps the right pterygoid. A possible second bite would have been into the left side of the face, damaging the left lacrimal and left maxilla. The left jugal was displaced and was recovered close to the main part of the skull (see Fig. 16). If this element has dissociated through decomposition prior to the putative bite it would explain why this is undamaged, despite the major trauma to that side of the face.

There may also have been a bite to the missing left dentary. Based on Daspletosaurus having four teeth in each premaxilla, 17 for each maxilla (Currie, 2003, though he notes this specimen has an unusually low maxillary count and it may be as little as 13) and 18 for each dentary, the total number of teeth in the skull would have been up to 74. A total of 31 erupted teeth are present in situ in the skull and dentary (not including the incipient replacement teeth), and 19 loose teeth were collected from the quarry, a total of 50 teeth. These teeth match those in the cranium for size, shape and colour (and retain roots) and are not considered shed teeth by another animal. Therefore, we conclude that originally the left dentary must have been present in the immediate area and shed nearly all of its teeth at some point before it was lost.

Finally, the skull is the tallest element of the specimen, and even allowing for stratification of some pieces of the postcranium, may have been the uppermost material in the quarry. Thus, assuming there was a low level of sediment/water in the environment, the skull may have been the only exposed piece, or the most exposed piece, when the scavenging took place. This is somewhat speculative, but fits the overall pattern of marks on the skull not seen on any other material and the apparent trampling of a buried femur. This may therefore explain why one, or even both, dentaries were bitten—they were simply the only exposed part of the skeleton when the encounter took place and thus did not follow the expected carcass consumption patterns.

As with the premortem injuries, it is not possible to distinguish easily between cannibalism and feeding by another tyrannosaurid. Currie (2005) notes that Gorgosaurus is generally more common in the Dinosaur Park Formation, and thus based simply on numbers would appear to be a more likely candidate to have fed on this animal than a second Daspletosaurus. Cannibalism is known in North American tyrannosaurines (Longrich et al., 2010) and thus should not be ruled out in Daspletosaurus. Certainly it seems most likely that a tyrannosaurid, if not necessarily a conspecific, was responsible for the postmortem feeding traces left here.