Malformed individuals of the trilobite Estaingia bilobata from the Cambrian Emu Bay Shale and their palaeobiological implications

Abstract Malformed trilobite specimens present important insight into understanding how this extinct arthropod group recovered from developmental or moulting malfunctions, pathologies, and injuries. Previously documented examples of malformed trilobite specimens are often considered in isolation, with few studies reporting on multiple malformations in the same species. Here we report malformed specimens of the ellipsocephaloid trilobite Estaingia bilobata from the Emu Bay Shale Konservat-Lagerstätte (Cambrian Series 2, Stage 4) on Kangaroo Island, South Australia. Ten malformed specimens exhibiting injuries, pathologies, and a range of teratologies are documented. Furthermore, five examples of mangled exoskeletons are presented, indicative of predation on E. bilobata. Considering the position of malformed and normal specimens of E. bilobata in bivariate space, we demonstrate that the majority of malformed specimens cluster among the larger individuals. Such specimens may exemplify larger forms successfully escaping predation attempts, but could equally represent individuals exhibiting old injuries that were made during earlier (smaller) growth stages that have healed through subsequent moulting events. The available evidence from the Emu Bay Shale suggests that this small, extremely abundant trilobite likely played an important role in the structure of the local ecosystem, occupying a low trophic level and being preyed upon by multiple durophagous arthropods. Furthermore, the scarcity of malformed E. bilobata specimens demonstrates how rarely injuries, developmental malfunctions, and pathological infestations occurred within the species.

Trilobites with malformations are often documented as stand-alone records, with examples ranging from the Cambrian through to the Carboniferous (Owen, 1985;Babcock, 2007;Bicknell et al. 2022a). As such, few publications have presented a thorough record of malformations from the same deposit, especially considering the ratio of malformed individuals in the context of non-malformed specimens of the same species. However, over the last five years, there has been a transition to presenting detailed records of malformed specimens from the same deposit and contextualizing these specimens within a larger population (see Pates et al. 2017;Bicknell et al. , 2022aBicknell & Smith, 2022). To extend this line of enquiry, we considered Estaingia bilobata Pocock, 1964 from the Emu Bay Shale (Cambrian Series 2, Stage 4) of South Australia. This species is exceptionally abundant within the Emu Bay Shale Konservat-Lagerstätte, with the vast majority of specimens preserved as articulated exoskeletons (Paterson et al. 2016;Holmes et al. 2021a, b;Bicknell et al. 2022a). The sheer abundance of specimens increases the likelihood of identifying malformed trilobites. Furthermore, injuries and pathologies are already known in other trilobites from the Emu Bay Shale and adjacent strata (Pocock, 1974;Conway Morris & Jenkins, 1985;Nedin, 1999;Paterson & Edgecombe, 2006;Holmes et al. 2020;Bicknell et al. 2022a). Here we document specimens of E. bilobata that exhibit injuries, pathologies, and teratologies, as well as evidence of possible durophagy (Figs 1-5). Furthermore, given the detailed information available on E. bilobata ontogeny (Holmes et al. 2021a, b), we compare these specimens to a population of non-malformed individuals to illustrate at which size class malformed specimens occur, presenting unique insights into the palaeoecology of this abundant Emu Bay Shale trilobite.

Geological and palaeoenvironmental context
The Emu Bay Shale forms part of the Kangaroo Island Group in the Stansbury Basin, which crops out on the north coast of Kangaroo Island (South Australia), and is interpreted as a nearshore delta complex (Gehling et al. 2011;Jago et al. 2021). The Emu Bay Shale is considered to be c. 512 Ma (Pararaia janeae Zone; Cambrian Series 2, Stage 4) in age (Bengtson et al. 1990;Paterson & Jago, 2006;Jago et al. 2020), based primarily on the co-occurrence of the emuellid trilobite Balcoracania dailyi Pocock, 1970 and an associated radiometric date of 511.87 ± 0.14 Ma in the Billy Creek Formation of the Arrowie Basin (Paterson & Edgecombe, 2006;Betts et al. 2018).
In the Big Gully area of Kangaroo Island, the Emu Bay Shale is 78 m thick at the shoreline and c. 61 m thick near Buck and Daily quarries, where the material considered here was sourced (c. 400 m inland; see García-Bellido et al. 2009 for a detailed locality map). Here, the formation commences with a thin (up to 2 m) interval of polymict conglomerate, the base of which is interpreted as a sequence boundary separating it from the underlying Marsden Sandstone. Above this, the conglomerate transitions sharply into a package of dark grey, laminated mudstone with siltstone and fine sandstone event beds that become thicker and more common towards the contact with the sandstones of the overlying Boxing Bay Formation (Gehling et al. 2011). The mudstone interval of the lower Emu Bay Shale is interpreted as representing a relatively deep-water prodelta environment, and hosts the Konservat-Lagerstätte that contains a diverse biota (50þ species, mainly within the basal c. 15 m of the formation), with many taxa considered to   (Holmes et al. 2020). Intact moult configurations of these species suggest that they lived in or close to the low-oxygen environment of the prodelta setting, although were likely subject to occasional 'mass kill' events resulting from fluctuations of the oxycline (Paterson et al. 2016). The prevalence of dorsum-down trilobite specimens in the deposit (up to 90% for E. bilobata; Drage et al. 2018) may indicate a physiological response to oxygen stress.

Methods
All Estaingia bilobata specimens housed in the South Australian Museum (Adelaide, South Australia) palaeontological collection (SAMA P) were examined for evidence of malformations.
Identified specimens were photographed under normal (fibre optic) lighting conditions using a Canon EOS 5D digital camera with a Canon MP-E 65 mm 1-5× macro lens. The linear measurement dataset published by Holmes et al. (2021a) was used to determine where malformed Estaingia bilobata specimens were located in bivariate space compared to non-malformed specimens. The cephalic and trunk (thorax þ pygidium) lengths in the dataset were used (Fig. 6a). Where possible, measurements from malformed specimens were added (Supplemental Information 1). Not all specimens documented here could be included due to partial preservation. Data were natural-log (ln) normalized and plotted. Data points were coded for malformation presence or absence.

4.b. Injury
Exoskeletal breakage through accidental injury, attack or moulting complications (sensu Bicknell et al. 2022a). Injuries are usually L-, U-, V-, or W-shaped indentations across the exoskeleton (Babcock, 1993;Bicknell et al. 2022a). They can also be expressed as the reduction and rounding of exoskeletal sections, or as a 'single segment injury' (Bicknell et al. 2022a, d). These features generally show cicatrization and/ or segment repair and regeneration. Occasionally, injured exoskeletal areas recover abnormally, resulting in the fusion of exoskeletal sections and a lack of segment expression (Owen, 1985;Bicknell et al. 2022a). Injuries do not extend across the entire specimensuch breaks likely reflect post-mortem, taphonomic processes (Leighton, 2011).

Results
Examination of Estaingia bilobata specimens uncovered several examples of injuries, pathologies, and teratologies. We also found evidence for possible durophagous predation or post-mortem scavenging in the form of mangled exoskeletons.

5.a. Injuries
Six injured Estaingia bilobata are identified. Most injuries are unilateral, show L-, U-, W-shaped morphologies (Figs. 1, 2; Table 1) and are located on the trunk, showing evidence for pleural spine regeneration and rounding (Fig. 1c, d) and cicatrization (Figs 1e,f,2a,b). However, there is a single specimen with a highly deformed posterior trunk showing fused thoracic segments (Fig. 2e, f). There is only one specimen that displays a cephalic injury, represented by a broken genal spine that has partially regenerated ( Fig. 2c, d).

5.b. Pathologies
Only one specimen shows evidence for pathological growths on the pygidium in the form of two ovate neoplasms (Fig. 3). Pathological growth on the right side of the pygidium is twice the size of the left neoplasm.

5.c. Teratologies
Three examples of teratologies are observed. The first shows abnormally developed posterior segments on the right side of the thorax, including the apparent fusion of certain axial rings (e.g. T5 and T6) and pleurae, some of which bifurcate distally (Fig. 4a, b). The second specimen displays very minor, partial fusion of distal sections of two anterior pleurae on the left side of the thorax (Fig. 4c, d). The third specimen has an additional 'half-segment' immediately behind T3 on the right side of the thorax. This additional segment begins at the sagittal line and rapidly expands to form the right half of an axial ring and a normal pleura (Fig. 4e, f). This has resulted in a highly asymmetrical trunk morphology, with the additional 'half-segment' also causing the posteriorly adjacent segment to bend.

5.e. Bivariate space
The distribution of malformed Estaingia bilobata in bivariate space illustrates a clear overall size pattern (Fig. 6). The majority of malformed individuals cluster among the largest holaspid stages (Fig. 6b). The exceptions are two smaller specimens that are separate from other malformed individuals. There is no marked distinction between where injured, pathological, and teratological specimens are located in shape space.

Discussion
The specimens of Estaingia bilobata documented here show evidence for the three major groups of trilobite malformations (Owen, 1985). This is rare among Cambrian trilobites, as the vast majority of malformed specimens record injuries, with limited evidence for teratologies and pathologies (Bergström & Levi-Setti, 1978;Bicknell & Paterson, 2018;De Baets et al. 2022). As such, E. bilobata presents a unique opportunity to study the palaeobiological significance of malformation types in one trilobite species from a single deposit. The majority of observed Estaingia bilobata injuries (Figs. 1, 2a (Rudkin, 1979;Babcock, 1993;Bicknell & Pates, 2020;Zong, 2021a, b). Indeed, select E. bilobata injuries  are very similar to those observed on the co-occurring Redlichia takooensis and R. rex that have been attributed to failed predation (Conway Morris & Jenkins, 1985;Bicknell et al. 2022a). It is highly likely that these injuries in E. bilobata also reflect failed predation. Evidence for cicatrization (Figs 1a, b, e, f, 2a, b) and regeneration (Fig. 1c, d) at injury sites indicates that these individuals survived and continued to moult after attacks. Other evidence for predation on, or post-mortem mastication of, E. bilobata is represented by the mangled exoskeletons (Fig. 5). Similar examples of R. takooensis have been interpreted as rare evidence of possible durophagy (e.g. Bicknell et al. 2022a, fig. 1C-E). Such evidence of durophagy, combined with examples of failed predation, further bolsters the idea that E. bilobata was a key prey item within the Emu Bay Shale biota. This smaller trilobite species likely occupied a relatively low trophic level within the ecosystem , especially given its extreme abundance (>80% of individuals across all known species), and yet injured specimens of E. bilobata are exceedingly rare. Further, most injured specimens cluster among the larger individuals in bivariate space. This may suggest that only larger individuals were capable of escaping and recovering from attacks, preserving records of failed predation. Alternatively, these specimens may preserve evidence of older injuries that recovered through subsequent moulting events. In either case, this pattern indicates a possible record of survivorship bias within the trilobite fossil record. One injured specimen of Estaingia bilobata shows a substantially disrupted posterior thorax (Fig. 2e, f). This malformation is comparable to rare specimens of other Cambrian trilobite species with highly disrupted and wrinkled pleurae adjacent to the injury (Conway Morris & Jenkins, 1985;Bicknell et al. 2022a). This condition is more commonly observed in the trunk of younger iso-and macropygous trilobites (Ormiston et al. 1967;Šnajdr, 1981;Owen, 1985;Rudkin, 1985). Disruption of pleurae indicates that the individual may have been attacked during the soft-shelled stage (shortly after moulting), resulting in a highly deformed exoskeleton that became fixed in subsequent moults (Conway Morris & Jenkins, 1985;Rudkin, 1985).
Teratological trilobite specimens with an additional 'halfsegment' are almost completely unknown. To our knowledge, the only other example of an additional 'half-segment' within the thoracic region is a malformed Emuella polymera Pocock, 1970 that has multiple 'half-segments' (Pocock, 1974;Owen, 1985;Paterson & Edgecombe, 2006). One further record of this 'halfsegment' condition is known from a Toxochasmops McNamara, 1979 pygidium from the Upper Ordovician (Katian) of Norway (Nielsen & Nielsen, 2017). These rare teratologies likely record damage to a generative zone and the resultant propagation of additional, unaligned segments (Pocock, 1974;Owen, 1985). It is noteworthy that two of the three examples represent early Cambrian forms, one of which (E. polymera) belongs to the family Emuellidae, which is characterized by unusually large numbers of trunk segments (Paterson & Edgecombe, 2006). Emuellids in particular may have been more prone to teratologies as they added new segments through euanamorphosis (Paterson & Edgecombe, 2006). These trilobites certainly had unusual patterns of trunk segment development compared with other forms (Holmes et al. 2021a). This may have increased the likelihood of developmental malformations, indicating possible evidence for increased developmental plasticity within Cambrian trilobites compared to younger forms (Webster, 2007). However, the rarity of these exceptional specimens suggests that these malfunctions seldom occurred.
Fused exoskeletal sections are rarely observed in thoracic pleurae (see Öpik, 1975, pl. 17), but more commonly recognized in pygidial regions (Strusz, 1980;Owen, 1985;Babcock, 1993;Nielsen & Nielsen, 2017;Bicknell & Smith, 2021). Such teratologies have been attributed to genetic malfunctions (Nielsen & Nielsen, 2017) and/or moulting complications producing a malformation that propagated in subsequent moults (Bicknell & Smith, 2021). As the teratological thoracic segments (Fig. 4c, d) lack any evidence of an injury, this rare malformation was likely caused by a developmental abnormality that impacted a limited exoskeletal region. This may record segments failing to separate or articulate correctly when they budded off from the pygidium.
The morphologies of exceptionally preserved appendages in two Emu Bay Shale arthropod species have been linked to durophagy, potentially on Estaingia bilobata. Firstly, the short, stout gnathobasic spines on the biramous appendages of the very large trilobite Redlichia rex (Fig. 7e, f) have been shown to be comparable with other known durophages, such as the  (Botton, 1984;Bicknell et al. 2018aBicknell et al. , b, 2021Bicknell et al. , 2022aHolmes et al. 2020). This has permitted the linking of injuries identified on R. takooensis specimens with failed predation by the larger R. rex, and injuries to R. rex as possible instances of attempted cannibalism (Bicknell et al. 2022a). Individuals of R. rex (and potentially R. takooensis, although morphological details of the appendages are unknown) may have also consumed E. bilobata, as supported by centimetre-scale coprolites containing E. bilobata remains (Daley et al. 2013, fig. 7E-H) that have been attributed to R. rex (Daley et al. 2013;Bicknell et al. 2022a). The gnathobasic spines of the Emu Bay Shale chelicerate Wisangocaris barbarahardyae Jago et al., 2016 (Fig. 7c, d) show similar morphologies to those of R. rex and other known durophages, and were likely also adapted for crushing biomineralized material . This is supported by the presence of shelly cololites in certain specimens of W. barbarahardyae, some containing fragments of E. bilobata (Fig. 7a, b; Jago et al. 2016, fig. 5c-g). Given the maximum reported size of W. barbarahardyae is c. 60 mm, this arthropod may have targeted and consumed E. bilobata meraspides. Juveniles of E. bilobata are abundant in the Lagerstätte interval and range from c. 1 to 5 mm in length (Holmes et al. 2021a, b). This is consistent with the size of E. bilobata pleurae preserved within a W. barbarahardyae cololite (Fig. 7b). The size range and sheer abundance of E. bilobata within the Emu Bay Shale suggest that this small trilobite occupied an important position within the lower trophic levels of the local ecosystem, and was an important prey item for multiple species of predator.