Healed injury in a nektobenthic trilobite: “Octopus-like” predatory style in Middle Ordovician?

The Lower Paleozoic sediments of the Barrandian area are globally renowned as a classical ex ample of well-preserved skeletal marine fauna, including abundant remains of trilobites. Seve ral tens of morphologically anomalous exoskeletons of trilobites have been collected and docu mented from Cambrian to Devonian clastic sediments and carbonates. One of them, an excep - tionally well preserved, articulated and partly enrolled exoskeleton of the Ordovician nektoben thic trilobite Parabarrandia bohemica (NOVÁK, 1884) exhibits a prominent palaeopathological anomaly in its pygidium. We interpret this anomaly as a healed traumatic injury and attribute this damage to a failed predatory attack. The subsequently healed injury is classified as the ichno genus Oichnus BROMLEY, 1981. The structure on the pygidium is strongly reminiscent of inju - ries caused by octopods and a large cephalopod is proposed as a potential durophagous pred-ator responsible for the herein described trilobite injury. However, an attack from an unknown arthropod while the trilobite was in a soft-shelled stage cannot be excluded.

Here we describe an internal mould of the nileid trilobite Parabarrandia bohemica (NOVÁK, 1884) housed in collections of the National Museum Prague. This partly enrolled specimen shows the below described palaeopathological anomaly on its py gidium.

Previously described cases of anomalous trilobites
From the Barrandian area, twenty anomalous trilobites showing healed traumatic injuries have been reported from late the Mid dle to early Late Ordovician strata; data about these earlier re ports are summarized in Table 1 and Figures 1 and 2.

MATERIAL AND METHODS
The studied specimen of Parabarrandia is preserved as an inter nal mould in a siliceous nodule and is housed in the National Mu seum, Prague (inventory number NM L59869). The external mould is unknown. The specimen was collected by V. Schüs from an unknown locality within the PrahaŠárka area in 1943. Con sidering the lithology, it is likely to have been from the "U trian glu" site, which is within the higher levels of the Šárka Formation (see PERŠÍN & BUDIL, 2009). The specimen was coated with ammonium chloride to enhance contrast and photographed with a digital Canon EOS 70 D camera.

DESCRIPTION OF THE INJURED PARABARRANDIA (FIG. 3)
The dorsoventrally flattened, slightly damaged and partly en rolled exoskeleton is 42 mm wide and 90 mm long. The smooth surface of the internal mould has slender unbranched to irregu larly branched burrows assigned herein, following KRAFT et al. (2020), to the ichnogenera Palaeophycus HALL, 1847, Arachnostega BERTLING, 1992 andPilichnus UCHMAN, 1999. Un branched cylindrical tunnels in the cephalon and the thoracic axis (reaching ~ 2 mm in diameter) are classified as Palaeophycus isp. (Pal in Fig. 3A). Narrow straight to slightly curved tunnels are common at the axial surface of the thorax and are observed in the axial and pleural pygidial surface; several intricate tunnels are noted in thoracic pleurae. Such fine homogeneously distributed tunnels along the internal surface of the exoskeleton are classi fied as Pilichnus isp.
A wide pygidial doublure with fine terrace lines is exposed due to breaks of pleurae and lateral, posteriolateral and posterior pygidial margin. Several minute Pilichnus isp. are also seen in the doublure (Pil in Fig. 3B). Fine, ramified burrows on the sur face of internal moulds with an oval cross-section are classified as Arachnostega isp. (Ar in Fig. 3B). Pilichnus has primarily been described from fine-grained soft substrates (UCHMAN, 1999). In accordance with the limited acceptance of the substrate as an ichnotaxobase (BERTLING et al., 2006), we can also classify as Pilichnus thin branched tunnels made in direct contact with the trilobite shell. Pilichnus built in this way can be transferred to more complex systems corresponding to Arachnostega. We sug gest that transitional forms between Pilichnus isp. (= thin branch tunnels in contact with the shell) and Arachnostega isp. (= open or almost closed networks in contact with the shell) can exist and are observed on the studied specimen.
Two prominent punctures are observed on the left anterior surface of the pygidial pleural field (arrows in Figs. 3B, D). The larger puncture is elliptical with its longer axis oriented parallel with the pleural furrow (a in Fig. 3D). The longer axis of this puncture reaches ~ 1.8 mm. The smaller puncture is rounded and measures 0.9 mm in diameter (b in Fig. 3D). Both punctures rep resent small craters surrounded by an elliptical swelling, which is ~ 9.5 mm wide and 4 mm long (dotted line in Fig. 3D).
Remarks. Morphologically comparable pit developed on up per and lower lamellae of the bilaminar cephalic fringe of the Si lurian trilobite Bohemoharpes ungula was described and figured by ŠNAJDR (1978b, pl. 1, figs. 1-5;1990, p. 62-63), who inter preted this anomaly as resulting from an activity of an endopara sitic organism. Recently, this interpretation was also accepted by DE BAETS et al. (2022, Table 1).

DISCUSSION
The Ordovician record of injured skeletal invertebrates in the Barrandian area includes gastropods, cephalopods, and trilobites. Up to now, only two injured brachiopods are known (BUDIL & FATKA, unpublished observation).

Injured trilobites
In the Barrandian area, injured trilobites have been classified to nine genera (Table 1). The oldest recorded are the few trilobites from the Šárka Formation; similarly, there are rare specimens showing healed injuries in the overlying Dobrotivá Formation (Table 1, Fig. 2). The most abundant injured trilobite specimens are observed in the late Sandbian Letná Formation, and the youngest malformed specimen was described from the late Sand bian early Katian Vinice Formation (Table 1, Fig. 2).
In following sections 5.3.1 -5.3.3, the current knowledge on the supposed lifestyle of trilobite specimens is summarised. This summary reviews the potential predators. Some trilobites were able to eliminate the predation pressure by cryptic behaviour (see FATKA & BUDIL, 2014), while other heavily skeletonised species or good swimmers effectively used passive defensive strategies.

Benthic trilobites
In the Barrandian area, most Ordovician trilobites with healed traumatic injury after failed predatory attacks have been classi fied as benthic and nektobenthic forms.

Dalmanitina
Both disarticulated parts and articulated exoskeletons and Dalmanitina are very common in the Letná Formation (FATKA et al., 2021). PŘIBYL & VANĚK (1976, p. 9) classified Dalmanitina as a good swimmer occasionally burrowing in the top layer of a shallow water bottom.
Asaphellus BUDIL et al. (2007, p. 68) andMERGL et al. (2008, p. 277) as signed this genus to large benthic predators. GIBB et al. (2010), and more recently also NETO DE CARVALHO & BAUCON (2016) documented cooccurrence of the trace fossil genera Rusophycus and Cruziana and articulated exoskeletons of the asaphid trilobite Asaphellus. Such close association of the puta tive tracemarker and its trace documents a benthic life of these large and heavily skeletonised trilobites.

Nektobenthic trilobites
Most pelagic trilobites were poorly streamlined (see FORTEY 1985), and it is supposed that they swam quite slowly. Some larger trilobites like Parabarrandia show a hydrofoil shape, with the head end prolonged into an elongate "nose," comparable to extant sharks (FORTEY, 1985) and are hypothesized to have swum much faster.

Potential predators
In the Ordovician of the Barrandian area, injured gastropods are ascribed to cephalopods, echinoderms and arthropods (HORNÝ, 1996(HORNÝ, , 1997a. Injuries to cephalopods were likely interpreted to be made by other cephalopods (AUBRECHTOVÁ, 2015; AU BRECHTOVÁ & TUREK, 2018). The malformed Parabarrandia bohemica described and considered here also requires an expla nation. ALPERT & MOORE (1975), WHITTINGTON & BRIGGS (1985), ŠNAJDR (1980ŠNAJDR ( , 1981 and RÁBANO & AR BIZU (1999) proposed that sea anemones, anomalocarids and cephalopods caused the injuries of Cambrian and Ordovician tri lobites. BICKNELL et al. ( and 2022 suppose that tri lobites could damage other trilobites. BRETT & WALKER (2002, p. 94) suggested that priapulids, nautiloid cephalopods, phyllo carid crustaceans and other arthropods (e.g., eurypterids) were likely Ordovician durophagous predators. The recently described specimen of Dalmanitina with a malformed and regenerated eye is interpreted as an unsuccessful attack by a cephalopod or a large arthropod (FATKA et al., 2021). From the morphology of the mal formed Parabarrandia, combined with the large size of the taxon, we exclude predators including sea anemones, anomalocarids, echinoderms, and priapulids as the injury makers. Consequently, cephalopods and arthropods are the potential culprits.

Arthropods
The length of the carapace of planktic phyllocarids does not ex ceed 50 mm in the Ordovician (RACHEBOEUF & CRASQUIN, 2010). Consequently, phyllocarids are excluded as a potential cul prit of the herein studied trilobite. Presuming a benthonic mode of life for Paleozoic marine chelicerates (for eurypterids see BRADY, 2001), sublethal predator−prey interactions between chelicerates and nektobenthic trilobites like Parabarrandia might be possible.

Feeding post-mortem
After KRAFT et al. (2020) and other authors, producers of Palaeophycus apparently preferred an easily accessible and nourish ing food that was easily consumed, e.g., their trace makers selec tively oriented on decaying soft tissues. The Arachnostega and Pilichnus traces are oriented in a manner suggesting systematic feeding. These trace makers spent more time in a carcass. The occurrence of Palaeophycus, Arachnostega and Pilichnus in the internal mould of the Parabarrandia attests to a post-mortem feeding activity on the trilobite carcass. Also, the perfect articu lation of the trilobite exoskeleton suggests a carcass, not an exu vium (see VALLON et al., 2015).

Attack on living specimen of Parabarrandia
Two prominent punctures penetrate the trilobite exoskeleton and are surrounded by swelling. The morphology in NM L59868 il lustrates, that this exoskeletal anomaly occurred in life of the Parabarrandia, likely during the "papershelled" stage of HEN NINGSMOEN (1975) or "soft-shelled" stage of SPEYER & BRETT (1985). The other possibility is in vivo attack by a culprit capable of boring.
Confirmed living marine perpetrators drilling their prey are mainly gastropods and octopod cephalopods (VERMEIJ, 2002, p. 385). Most drill holes are interpreted to be caused by predatory gastropods such as naticids and muricids. BROMLEY (1981) pro posed that the ichnofossils that are made by naticid drilling were Oichnus paraboloides BROMLEY, 1981 and muricids made Oichnus simplex BROMLEY, 1981.
The holes in the pygidium of Parabarrandia are morpho logically comparable to drill holes found in modern molluscs and crustaceans (e.g., ARNOLD & ARNOLD, 1969;BOYLE & KNOBLOCH, 1981;NIXON & MACONNACHIE, 1988;HARPER, 2002). The elliptical outline and dimensions of our drill holes are in accordance with the morphology of the ichno species Oichnus ovalis BROMLEY, 1993, an ichnofossil inter preted to be the result of boring by octopod cephalopods (see BROMLEY, 1993;WISSHAK et al., 2015). Further, the drilling of two or even three holes in one shell is a strategy known to be deployed in some species of recent octopods (NIXON & MA CONNACHIE, 1988). NIXON (1979NIXON ( , 1980 reported that in recent Octopus vulgaris, the drilling activities are carried out by a salivary papilla lying just below the radula. The role of saliva produced by sali vary glands was later shown to be important for a successful at tack, as it contains a wide spectrum of paralysing and proteolytic substances (NIXON, 1988, p. 709). Some of them are responsible for the breakdown of the musculoskeletal attachment mecha nism in crabs within 20 min of capture (NIXON, 1984). Similar breakdown of the musculoskeletal attachment mechanism would probably mean the same for trilobites.
The key ichnogenus Oichnus BROMLEY, 1981 and other morphologically similar ichnotaxa have been recently revised (WISSHAK et al., 2015). For the creation of our trace, drilling behaviour seems to be the most plausible because of the absence of sharp edges typical for biting, combined with the diminutive, protected space.
In terms of systematic ichnology, the herein described struc tures from Parabarrandia are attributable to the ichnospecies Oichnus ovalis BROMLEY, 1993. Ancient, fossilised structures were interpreted as octopus borings, based on observations from studies of the recent octopods (BROMLEY, 1993;NIXON, 1979NIXON, , 1980NIXON & MACONNAICHE 1988). Based on these obser vations, we interpret the structures observed in the specimen of Parabarrandia studied here as resulting from an "Octopuslike" predatory attack.
Origination of the swelling on the internal mould. We sup pose that the trilobite was attacked during the "softshelled" stage. The thin exoskeleton was probably drilled (= "Octopus like" predatory style). Consequently, in the injured area, the soft tissue under the unbiomineralised exoskeleton overdeveloped. This swelling would have been recorded during exoskeletal hard ening with swelling expressed both on the external and internal surfaces of the exoskeleton.

CONCLUSION
(1) The exoskeletal anomaly seen at the left pygidial side of Parabarrandia represents a partly healed injury after a failed preda tory attack during life.
(2) Two scenarios explain this anomaly: a -The healed injury classified as the ichnospecies Oichnus ovalis BROMLEY, 1993 can be interpreted as an exoskeletal anomaly which originated after a failed "octopuslike" strategy of the predatory attack. This preferred interpretation reflects the nektobenthic lifestyle of Parabarrandia and the nektonic life style of the suspected predator.
b -The morphology and the extent of the swelling surround ing both punctures combined with the noticeable absence of any crack of the surrounding exoskeleton indicates the high flexibility of the cuticle during the attack. The attack resulted in two re stricted perforations (punctures) followed by plastic deformation of the exoskeleton copying the swelling. In such cases, the injury could result from attack of an unknown predatory arthropod.
FINANCIAL SUPPORT: This research was supported by the Czech Science Foundation (GACR) project no. 1814575S and by Cooperatio GEOL (OF).
COMPETING INTERESTS: The authors declare that they have no conflict of interest.

ACKNOWLEDGEMENT
We acknowledge both reviewers Lothar H. VALLON (Østsjael lands Museum, Faxe, Denmark) and Russell D.C. BICKNELL (University of New England, Armidale, New South Wales, Aus tralia) for their helpful review and the linguistic improvements that made on our text. M. VALENT (National Museum, Prague) is acknowledged for his help with photographing and tracing the origin of the specimen. This is a contribution to the IGCP 653 "Filling the gap between Cambrian Explosion and the GOBE".