Palaeopathological evidence for intraspecific combat in ankylosaurid dinosaurs

Abstract

Ankylosaurid dinosaurs were heavily armoured herbivores with tails modified into club-like weapons. These tail clubs have widely been considered defensive adaptations wielded against predatory theropod dinosaurs. Here we argue instead that ankylosaurid tail clubs were sexually selected structures used primarily for intraspecific combat. We found pathological osteoderms (armour plates) in the holotype specimen of Zuul crurivastator, which are localized to the flanks in the hip region rather than distributed randomly across the body, consistent with injuries inflicted by lateral tail-swinging and ritualized combat. We failed to find convincing evidence for predation as a key selective pressure in the evolution of the tail club. High variation in tail club size through time, and delayed ontogenetic growth of the tail club further support the sexual selection hypothesis. There is little doubt that the tail club could have been used in defence when needed, but our results suggest that sexual selection drove the evolution of this impressive weapon. This changes the prevailing view of ankylosaurs, suggesting they were behaviorally complex animals that likely engaged in ritualized combat for social dominance as in other ornithischian dinosaurs and mammals.

1. Introduction

Ankylosaurid dinosaur tail clubs are a rare form of weaponry without close analogues among living taxa [1]. Modified distal caudal vertebrae interlock tightly to form a stiff handle, and enlarged terminal osteoderms envelop the tail tip [2,3]. Tail clubs were capable of delivering and withstanding forceful blows and would have been highly effective weapons [4–6]. Many authors note that the tail club could have been used for reactive defence against predators (e.g. [2,4,5,7–12]), and popular science depictions of ankylosaurids overwhelmingly show them using their tail clubs for defence against theropods, especially tyrannosaurs (electronic supplementary material, S1). However, in many extant animals with elaborate weaponry, these weapons are used primarily for intraspecific combat and have evolved under sexual selection [13], which raises the alternative hypothesis that the ankylosaurid tail club evolved under sexual selection for intraspecific display and combat. Few palaeontologists have considered this alternative hypothesis; Coombs [14] considered intraspecific combat using the tail club implausible. Arbour [4] proposed that antipredator defence may not have been the primary function of the tail club, and instead, tail clubs may have been used largely for intraspecific combat, with clubs swung into the flanks of an opponent as in flank-butting in bison [15] or necking in giraffes [16].

Novel evidence for the hypothesis that ankylosaurid tail clubs primarily functioned as weapons of intraspecific combat is present in the exceptionally preserved holotype skeleton of the ankylosaurid Zuul crurivastator (ROM 75860; [17], figure 1), which has nearly complete and exquisitely preserved skin overlying in situ osteoderms across the entire dorsum. Most significantly, ROM 75860 includes multiple pathological osteoderms that are present in localized, bilaterally symmetric regions on the flanks (figure 1). This specimen sheds considerable new light on the function of ankylosaurid armour, as evidence for ritualized intraspecific combat rather than failed predation. Here we use multiple lines of evidence to test the competing hypotheses that (1) intraspecific combat or (2) predation drove the evolution of the ankylosaurid tail club.

Figure 1.

Pathological osteoderms in the holotype of Zuul crurivastator. Above, composite photograph of the skull, first cervical half ring, body block and tail block. Fossil material is brown and surrounding rock matrix is grey. Below, interpretive illustration showing non-pathological osteoderms in white and pathological osteoderms in red. Individual osteoderms are marked with letter–number combinations (e.g. A1) for discussion within the text. CD, distal osteoderm of the cervical half ring; CL, lateral osteoderm of the cervical half ring; CM, medial osteoderm of the cervical half ring; knob, tail club knob.

2. Methods

(a) . Hypothesis 1: intraspecific combat underlies the evolution of the ankylosaurid tail club

Animals with sexually selected weaponry often have highly ritualized intraspecific combat behaviours with repeated patterns of body alignment [13,18,19], and injuries are often found in concentrated areas associated with weapon deployment [18,20,21]. We predict that intraspecific combat trauma in ankylosaurids would be localized to the regions of the body within range of the tail. Ankylosaurids had limited dorsoventral movement in the tail [4], precluding strikes to the dorsal surface of an opponent's back and pelvis. Instead, tail strikes would have landed on the head, shoulders, flanks, hips and/or lateral tail. ROM 75860, the holotype of Z. crurivastator, has several pathological osteoderms within the in situ integument (electronic supplementary material, S1). We identified osteoderms as pathological, rather than taphonomically damaged, if they had the following features: (1) dramatic shape differences from adjacent osteoderms, such as missing apices or aberrant keel shapes, (2) keratin sheath covering missing apices/aberrant keel shapes, and/or (3) presence of reactive bone texture. We tested the hypothesis that osteoderm pathologies were not randomly distributed across the body by dividing osteoderms into regions and using a chi-squared test in PAST v. 4.09 [22] (electronic supplementary material, S1).

Sexually selected traits are often positively allometric [23,24], and allometry in the size of the tail club knob could support the hypothesis that tail clubs evolved as sexually selected weapons of intraspecific combat. We built a dataset to explore the relationship between knob width and femur length, and knob width and caudal vertebral width (electronic supplementary material, S1 and S2).

(b) . Hypothesis 2: predation underlies the evolution of the ankylosaurid tail club

If tail clubs arose as antipredator defensive structures, there could be a correlation between the presence or absence of a tail club and predator morphology, or between tail club morphology and predator morphology. Tail clubs have several variable morphological traits, the most easily quantifiable of which is knob width. Larger tail club knobs can deliver more forceful impacts during a tail strike [4], so larger tail club knobs may have been more effective defensive weapons against larger predators. We collected body mass estimates from previously published literature for the largest known predator in the same formation as an ankylosaurid species, and included the largest known tail club knob from that formation (electronic supplementary material, S1 and S2).

3. Results

ROM 75860 includes a nearly complete suite of osteoderms and integument (figure 1). Along the trunk and anterior pelvis, multiple osteoderms are missing their apices and have unusual shapes (figure 2). Right osteoderms D3 and F3 are each missing about one-quarter to one-third of their apical length, leaving a circular cross-section across the osteoderm. The broken edges have smooth textures, suggesting some bone remodelling has occurred, but they lack an enlarged fracture callus. Left D3 is similarly broken, but the keratinous sheath of the osteoderm has grown over the broken area, covering it completely. Left osteoderms C3 and E3 are missing large portions of the apex and trailing edge of the spike, resulting in concavities. In both osteoderms, the bone along the broken edges is smooth, and the keratinous sheath completely covers the damaged areas. This healing pattern is consistent with those observed in crocodilian osteoderms [25].

Figure 2.

Details of pathological and non-pathological osteoderm morphology in ROM 75860 (Zuul crurivastator). B2R and E3R are non-pathological flank osteoderms. F3R and D3R are pathological flank osteoderms missing the tips of the apex, and the keratinous sheath has not grown over the tip. D3L is a pathological flank osteoderm missing a large portion of the apex, and the keratinous sheath has overgrown the damaged region. C3L and E3L are pathological flank osteoderms with highly modified morphologies, missing large portions of the trailing posterior edge and with the keratinous sheath covering the damaged region.

Pathologies are not randomly distributed across the body of ROM 75860. A chi-squared test of normal versus pathological osteoderms on the dorsum versus the flanks yielded a statistically significant p-value of 0.0039. A runs test of the distribution of transverse rows with or without pathological osteoderms yielded a statistically significant p-value of 0.0016. Of particular interest regarding these damaged and pathological osteoderms is that (1) they appear to represent different stages of healing, which could be interpreted as multiple distinct trauma events or different rates of wound healing in response to infection, and (2) they are relatively symmetrically arranged in a narrow region of the body over the posterior trunk and pelvis.

Our macroevolutionary analyses were ambiguous because of missing data from a patchy fossil record (electronic supplementary material, S1 and S2). We did not find any significant correlations between ankylosaurid distal tail width and apex predator body mass, nor body mass of tyrannosauroids. Large-bodied theropods appear in the fossil record well before the evolution of either a stiff tail or tail club knob in ankylosaurids, and knob widths, although variable both within and between species, do not increase in size through geological time. There were insufficient data to draw conclusions regarding allometry in knob size.

4. Discussion

Using multiple lines of evidence, we find strong support that ankylosaurid tail clubs were used for intraspecific combat and likely evolved under sexual selection, and fail to find compelling evidence that predation influenced tail club evolution. ROM 75860 has multiple healed osteoderm pathologies that are best interpreted as resulting from intraspecific combat rather than failed predation attempts or disease. We argue this based on (1) the localization of injuries to the flanks, rather than the dorsum where a club strike could not easily land, (2) the localization of injuries to the posterior thoracic and anterior pelvic region, which would minimize the chance of fatal injuries to the combatants, and (3) the presence of fully healed and partly healed osteoderms, potentially suggestive of multiple injury events. Animals with sexually selected weaponry often have highly ritualized intraspecific combat behaviours with repeated patterns of body alignment [13,18,19], and injuries are often found in concentrated areas associated with weapon deployment [18,20,21]. Although direct comparisons are difficult because of incomplete preservation, it is interesting that no pathologies on the flank osteoderms have been noted in exceptional specimens of nodosaurid ankylosaurs (e.g. [26,27]), which lack tail clubs.

Large, bipedal tyrannosaurs were the most likely predators for ankylosaurids. Vetted direct evidence of healed predation attempts by tyrannosaurs on ornithischians is limited to a single record—a partially healed tail wound on a hadrosaurid linked to an embedded tyrannosaur tooth [28]. However, hadrosaurids lacked both body armour and tail weaponry, whereas a rear attack on an ankylosaur would be a high-risk predation strategy (the largest, most sharply pointed osteoderms are found on the flanks of Zuul). Moreover, although there is evidence that large-bodied mammalian predators commonly target flanks of prey [29], the body–size relationship in these interactions is opposite that of a tyrannosaur/ankylosaur interaction. Ultimately, we would expect tyrannosaur-inflicted injuries to be either randomly distributed across the body, more common on the dorsum given that subadult and adult tyrannosaurs were significantly taller than ankylosaurids, or in the vulnerable neck region, and more common on subadult prey. The limited range of movement for ankylosaurid tails also argues against its origin as an antipredator weapon. Although fractured and healed fibulae in tyrannosaurs have been proposed as possible injuries from ankylosaurid tail club strikes [30], these injuries could have been caused by many factors, such as other prey or conspecific interactions. Ankylosaurids were rare components of North American dinosaur faunas [31–33] and thus were probably not a primary prey item for North American tyrannosaurs.

Unfortunately, missing data because of the patchy nature of the fossil record mean our macroevolutionary analyses yield inconclusive results. We found no significant correlation between knob width and the body mass of the largest theropod or tyrannosauroid in a given ecosystem, which could be interpreted in several ways. Predators may not have had a selective influence on tail club knob size in general. Alternately, ankylosaurid tail clubs do not increase in size with increasing predator size because they reached a mechanical optimum early in their evolution that could not be surpassed. Weapons tend to evolve towards larger sizes, but larger weapons can also exert significant mechanical and physiological costs [34–36]. Additionally, ankylosaurs coexisted alongside large apex predators for millions of years before tail clubs appear in this clade, and the other lineage of ankylosaurs, the nodosaurids, never evolved tail clubs despite sharing their ecosystem with the same predators as conspecific ankylosaurids.

Allometric growth is a hallmark of many sexually selected structures, including weapons (e.g. [23,24], but is difficult to test in ankylosaurids because of the small sample sizes and missing data both for individual species and in Ankylosauridae overall. Data from the comparatively large sample size of Pinacosaurus grangeri indicate that osteoderm growth was delayed relative to the rest of the skeleton [37]. Juvenile Pinacosaurus lacked large osteoderms except in the cervical region; the tail club knob is absent in juveniles with a snout–posterior ilium length of about 68 cm, but present when individuals have a snout–posterior ilium length of about 170 cm (V. M. Arbour 2009, personal observation). Tail club knob size for individual ankylosaurid taxa varies considerably [38], which may reflect, at least in part, ontogenetic growth. Regrettably, the sex of any existing ankylosaurid specimen is currently unknown (and likely to remain unknown given the challenges associated with sexing extinct dinosaurs [39]), and so the presence or absence of sexual variation in tail clubs is also unknown. Delayed growth of the tail club knob supports intraspecific combat or display, rather than a primarily antipredator function for this structure, as has been argued for delayed osteoderm growth in crocodilians [40,41] and squamates [42,43].

Extant vertebrates use a variety of morphological adaptations and behavioural responses to avoid predation, but direct combat with an attacking predator is overall rare. Most terrestrial vertebrates avoid predation by evading detection, by evading capture, or, if captured, by making consumption difficult using defensive armour [44]. To the best of our knowledge, there are no studies demonstrating that predation was a major selective pressure driving the evolution of cranial weaponry in extant vertebrates, but there are many examples of sexual or social selection for weaponry in prey animals. Although sexually selected weapons can also be used for defence against predation (e.g. the horns of bovids, especially female bovids, [45,46]), or as visual deterrents to predators (e.g. elk [46]), there are also examples where weapons used for intraspecific fighting are not used in combat against predators (e.g. moose kick rather than use their antlers [47]). Additionally, there is a growing body of evidence suggesting that, at least in squamates, osteoderms may not function primarily as antipredator defences, but that sexual selection and intraspecific combat may play at least a partial role in their evolution [42,43,48–50].

In summary, we do not find strong evidence for the hypothesis that tail clubs evolved specifically as defensive structures, but assert that sexual selection, rather than predation, appears to have been the primary mechanism that drove the evolution of these iconic structures. This changes the prevailing scientific and cultural view of ankylosaurids, suggesting that they were behaviourally complex animals with sophisticated visual displays and weaponry that likely engaged in ritualized combat for social dominance, as in other ornithischian dinosaurs and mammals.

Data accessibility

All data are included in the electronic supplementary material [51]. All fossils discussed in the paper are reposited in publicly accessible institutions.

Authors' contributions

V.M.A.: conceptualization, data curation, formal analysis, funding acquisition, investigation, methodology, resources, visualization, writing—original draft, writing—review and editing; L.E.Z.: conceptualization, investigation, methodology, supervision, writing—original draft, writing—review and editing; D.C.E.: conceptualization, data curation, funding acquisition, investigation, methodology, project administration, resources, supervision, writing—original draft, writing—review and editing.

All authors gave final approval for publication and agreed to be held accountable for the work performed herein.

Conflict of interest declaration

We declare we have no competing interests.

Funding

This work was supported by the Natural Sciences and Engineering Research Council of Canada (NSERC Discovery Grant RGPIN-2020-04012, Postdoctoral Fellowship, Canada Graduate Scholarship and Michael Smith Foreign Study Supplement to V.M.A., and NSERC Discovery Grant RGPIN-2018-06788 to D.C.E.), an Izaak Walton Killam Memorial Scholarship (V.M.A.), a L'Oréal-UNESCO for Women in Science Supplement (V.M.A.), an Alberta Ingenuity Award (V.M.A.), grants from the Dinosaur Research Institute (V.M.A.) and funding from the Louise Hawley Stone Charitable Trust, Royal Ontario Museum (D.C.E.).