Taxonomic reappraisal of Nihilichnus from taphonomic perspectives of crocodile predatory ecology

Abstract The ichnogenus Nihilichnus has been broadly applied to penetrative feeding traces (bite marks) generated by vertebrates. The type ichnospecies Nihilichnus nihilicus encompasses all pit and puncture traces on vertebrate bone. This definition belies the diversity of pit and puncture morphologies found in the palaeontological and archaeological record and has led to inconsistency in the application and interpretation of these trace fossils. We resolve this ambiguity by splitting previously included morphologies that are too morphologically disparate to include within one ichnospecies. Using a sample of fossil, modern, and experimentally obtained crocodile modified bones, we describe three new ichnospecies under this ichnogenus: rounded pits N. clavus n. isp., bisected pits N. sicarius n. isp., and bisected punctures N. hastarius n. isp.; all of which we contend are morphologically distinct. Using these new ichnospecies, we provide a review, and reclassification when appropriate, of feeding traces previously published under Nihilichnus. We find that N. mortalis, lacking any official documentation, is nomen nudum and thus taxonomically unavailable. This article highlights how the usage of more specific ichnotaxonomic descriptions improves the utility of praedichnia in making palaeoecological inferences.


Introduction
The ichnogenus Nihilichnus describes pit and puncture structures occurring on skeletal material resulting from predation or scavenging (Mikuláš et al., 2006).The trace morphology of this ichnogenus is distinct from other bioerosion traces, presenting singular or multiple depressions or holes in a skeletal substrate, lacking organized boring patterns (Mikuláš et al., 2006).Generally, these traces are inferred to have been produced by vertebrate trace makers, specifically due to contact between teeth and skeletal material (Mazuch et al., 2024;Mikuláš et al., 2006;Rasser et al., 2016;Trifilio et al., 2023).At the time of first description, Nihilichnus was monospecific and the type ichnospecies, Nihilichnus nihilicus, received the same diagnosis as the ichnogenus (Mikuláš et al., 2006, p. 120).The substrate of the type series for N. nihilicus was the cortical bone of cancellous rich bones, as noted in the systematic description (Mikuláš et al., 2006).Bone has been generally accepted as the trace substrate for N. nihilicus (e.g., Jacobsen & Bromley, 2009;Milan et al., 2011;Rasser et al., 2016;Wisshak et al., 2019), if not explicitly stated in the diagnosis (Mikuláš et al., 2006).Mikuláš et al. (2006) mentions but does not describe a second ichnospecies: N. mortalis.Since this time, Nihilichnus has been observed on alternate substrates including ammonite and gastropod shells (Mazuch et al., 2024;Rasser et al., 2016).Three new ichnospecies have been erected: N. covichi (isolated or paired pits and punctures on invertebrate shells), N. quadripertitus (four rows of pits or punctures, occurring in pairs on opposing sides of ammonite shells), and N. sulcatus (punctures to vertebrate bone with branching furrows) (Mazuch et al., 2024;Rasser et al., 2016;Trifilio et al., 2023).
The present taxonomy lumps pit and puncture traces on vertebrate bones, features known to be distinct from one another, into a single ichnospecies: N. nihilicus (we exclude N. sulcatus from this assessment, as it is of a composite morphology, comprising both a puncture and branching furrows) (Binford, 1981;Trifilio et al., 2023).This inhibits the ecological interpretation of vertebrate feeding traces.Pits and punctures were first given a rigid definition by Binford (1981), in his seminal work on zooarchaeology: Bones: Ancient Men and Modern Myths.Binford (1981) clearly separates punctures from pits from a mechanical viewpoint.Therein, Binford (1981) describes punctures as holes passing through the cortical surface following the collapse of the cortical matrix due to surface pressure (p.44).Conversely, pits are generated when the outer bone surface does not collapse, rather the bone crystal resists the penetrating force, forming a simple depression on the surface (Binford, 1981, p. 46).Both these definitions fit within the current definition of N. nihilicus.For this reason, we argue that the trace morphologies currently accepted within N. nihilicus are too disparate to warrant the usage of a single ichnospecies and new ichnospecies should be assigned in order to specify unique morphologies.
Using the current systematics, we are unable to specifically associate tooth morphologies to correspondingly distinct trace morphologies, an important tool for palaeoecological research.Certain morphological indices in pits and punctures can be correlated with specific actions, behaviours, or tooth morphologies (Njau, 2006;Njau & Blumenschine, 2006;Njau & Gilbert, 2016).Taphonomic studies, both actualistic and palaeontological, have revealed an enormous array of distinct pit and puncture morphologies produced by different vertebrate agents, all of which would be given the same ichnotaxonomic name under the current definition (Njau, 2006;Njau & Blumenschine, 2006, 2012;Njau & Gilbert, 2016).
Pits and punctures, while both penetrative traces, are fundamentally different in their morphology.Punctures pass through the cortical bone (Binford, 1981;Drumheller & Brochu, 2016).The interiors of punctures are often irregular, containing broken cortical bone pieces, cancellous bone, or a hole which passes through the bone surface, opening to either the opposite side of the bone or the medullary cavity (Figure 1).The puncture interior is offset from the cortical surface, and there is a notable separation between the outer margin and the depressed surface (Figure 1).Pits are superficial and are only visible on cortical bone (Binford, 1981;Drumheller & Brochu, 2016).Pits may occasionally depress the underlying cancellous bone, but the interior is contiguous with the cortical bone surface (Figure 1).Pit traces represent a negative mould of the impacting tooth to varying degrees of fidelity according to deformation properties (i.e., brittle versus ductile deformation).Pit interiors have a greater potential to preserve tooth morphology than punctures, preserving the tooth-tobone contact surface.However, both pits and punctures can preserve correlates to distinct tooth morphologies (Drumheller & Brochu, 2014;Njau & Blumenschine, 2006).Njau and Blumenschine (2006) identified morphological indices on pits and punctures that correlate to crocodile tooth morphology.Crocodiles are capable of producing both pits and punctures (Njau & Blumenschine, 2006).Crocodile teeth often possess carinae, a sagittal ridge which gives the tooth a distinct edge and a lenticular outline (Njau & Blumenschine, 2006).Carinae project outwards from the tooth surface (Figure 1).The sides of carinated teeth slope to a distinct ridgeline, even when worn.This morphology is translated onto the bone surface during feeding.Carinated teeth produce traces with a lenticular outline, presenting distinct vertices on opposing sides of the pit or puncture corresponding to the ridge of the carina (Njau & Blumenschine, Figure 1.schematic drawing of crocodile feeding traces using Binford's (1981) classification scheme, based on diagram by Drumheller and Brochu (2016).Drawing not to scale.Bar equals 1 cm, relative to crocodile tooth photographs.Pits (a,b) terminate in the cortical bone, deforming the bone surface in the shape of the tooth.Punctures (c,d) penetrate through the cortical bone, either terminating in the cancellous bone or medullary cavity, or opening to the opposite side of the bone.conical teeth (1) produce rounded pits (a) and rounded punctures (c).carinated teeth (2) produce distinct vertices on opposing sides of the trace outline, generating bisected pits (b) and bisected punctures (d).In the case of bisected pits (b), the edge of the carinated tooth produces a longitudinal vertex within the trace, resulting in a 'V-shaped' cross section.carinae on unworn teeth often penetrate deeper than the remaining tooth surface, producing a distinct groove along this vertex (b).
Here, we re-evaluate the ichnotaxonomy of Nihilichnus using fossils recovered from Olduvai Gorge, Tanzania.This ecosystem presents the faunivorous crocodilian species Crocodylus anthropophagus, previously interpreted to have modified the remains documented here (Brochu et al., 2010;Njau & Blumenschine, 2006).Additional traces presenting the same ichnotaxobases are then documented from bones which have been modified by Nile crocodiles (Crocodylus niloticus) both in the wild and in controlled feeding experiments.We reassign the type species N. nihilicus to an emended diagnosis, constraining its morphology.We split new ichnospecies from the original N. nihilicus definition in order to highlight the morphological diversity we observe.This should help to clarify the behavioural and ecological implications of different morphologies of Nihilichnus and constrain the ichnotaxa applied to this genus to definite, constrained, and translatable ichnospecies.Using these new definitions of Nihilichnus, it should be easier to apply a robust ichnotaxonomic designation to feeding traces found in the fossil record.

Fossil sample
We document the partial remains of an adult Pleistocene zebra (Equus oldowayensis) and a juvenile, eland-sized bovid recovered from Olduvai Gorge, Tanzania, previously described in Njau and Blumenschine (2006).The Equus oldowayensis specimen consists of a complete femur and the bovid specimen is a fragmented tibia, lacking a proximal epiphysis.Both specimens present heavy surface modification and were discovered in approximate association in Olduvai Lowermost Bed II (∼1.8 Ma).Excavation was conducted by the Olduvai Landscape Paleoanthropology Project (OLAPP) at the VEK archaeological locality (OLAPP Trench 21, Level 1).

Modern sample
Modern predator modified bones were collected from the Lower Grumeti River in Serengeti National Park, Tanzania, by Njau and Blumenschine (2006).This sample is composed of two tibiae and one humerus of plains zebra (Equus quagga), and two mandibulae of blue wildebeest (Connochaetes taurinus).Bones were collected from the channel bed during the dry season and from the margins of active crocodile pools.Minimal damage to the epiphyseal ends, low frequency of bone surface modification, crocodile-specific trace morphologies, and direct documentation of crocodile feeding in this area all suggest that these traces were generated by Nile crocodiles (Njau & Blumenschine, 2006).

Trace identification and classification
Traces were observed using a 10x magnification hand lens and a 100x magnification Meiji EMZ binocular dissecting microscope.Pit and puncture traces were identified on our sample of predator modified bones.Distinct morphologic characters were identified in comparison to the current diversity of ichnotaxa, in particular those classified under the ichnogenus Nihilichnus.Consistent trends in trace morphology, as observed in our sample and noted by previous authors (e.g., Drumheller & Brochu, 2014, 2016;Njau, 2006;Njau & Blumenschine, 2006, 2012;Njau & Gilbert, 2016;Sahle et al., 2017;Westaway et al., 2011), were used to identify three new morphologies which we believe warrant classification as distinct ichnotaxa.
Criteria for ichnotaxonomy are largely summarized in the seminal work by Bertling et al. (2006Bertling et al. ( , 2022)).
Additional theoretical underpinnings are also discussed by Bertling (2007) and Rindsberg (2018).Using the taxonomic guidelines established by these authors, as well as those described specifically for bioerosion traces by Pirrone et al. (2014) and Wisshak et al. (2019), we describe three new ichnospecies according to the following ichnotaxobases.The primary ichnotaxobase is morphology, more specifically the internal structure, outline geometry, and the depth of penetration relative to bone structure (Bertling et al., 2022).Substrate is also critical to our assessment of these ichnotaxa.Substrate is generally avoided as an ichnotaxobase when describing traces in inorganic materials but is essential to the description of bioerosion (Bertling et al., 2022;Pirrone et al., 2014;Wisshak et al., 2019).The ichnospecies described herein are, in some cases, correlative to trace-maker.However, these ichnotaxa are defined by objective criteria rather than from an interpretation of the producer.

Photography
Following maceration, bones were coated in ammonium chloride to enhance contrast (Njau & Blumenschine, 2006;Njau & Gilbert, 2016).Traces were photographed using a Nikon D7100 Macro Camera.Pictures were taken using variable multisource low angle lighting to ensure illumination of the trace interiors.

Descriptive terminology
The terminology used to describe trace fossils is a topic of much debate.Phrases such as 'bite mark' and 'tooth mark' are commonly used by palaeontologists and zooarchaeologists to describe the traces we discuss.Recently, these terms have been criticized by some ichnologists on the basis of etymology and historical usage, suggesting that the word 'mark' generates ambiguity and may result in conflation with non-biogenic structures (Vallon et al., 2015).However, others have argued that terms such as 'bite mark' , 'tooth mark' , and 'gnaw mark' have historic precedence and are unambiguous with the appropriate modifier, just as the word 'trace' is ambiguous when lacking a modifier (Zonneveld, Fiorillo, et al., 2022).The term 'dentalite' was also proposed by Hunt et al. (2018), but as noted by Zonneveld, Fiorillo, et al. (2022), this term is redundant with the more established terminology mentioned above.The terms 'bite mark' and 'tooth mark' are more widely recognizable and used extensively in the taphonomic, zooarchaeological, palaeoanthropological, and palaeontological literature (D' Amore & Blumenschine, 2009, 2012;Drumheller et al., 2020Drumheller et al., , 2023;;Drumheller & Brochu, 2014, 2016;Njau, 2006;Njau & Blumenschine, 2006, 2012;Njau & Gilbert, 2016).We are of the opinion that an embargo against the terms 'bite mark' and 'tooth mark' would unnecessarily separate ichnology from taphonomy, and hinder communication across these subdisciplines of palaeontology.We, therefore, consider the terms 'bite mark' , 'bite trace' , and 'feeding trace' interchangeable, but for consistency hereafter we use the term 'feeding trace' exclusively.When describing feeding trace morphologies, we use the terminology as defined by Binford (1981) and used by subsequent authors (e.g., D' Amore & Blumenschine, 2009Blumenschine, , 2012;;Drumheller & Brochu, 2014, 2016;Njau & Blumenschine, 2006, 2012).These include 'pits' and 'punctures' as previously discussed, as well as 'furrows' and 'scores' , which refer to linear features (Binford, 1981).Furrows, as with punctures, pass through the cortical bone into the cancellous material (Binford, 1981, p. 48).Scores, like pits, are superficial and only present on the cortical bone (Binford, 1981, p. 46).We use the terms 'penetrations' and 'penetrative traces' to describe all traces which resulted from the ingression of teeth into skeletal material, and thus all ichnospecies of Nihilichnus.'Taphonomic agent' , or otherwise 'agent' is used interchangeably with 'trace maker' , both of which refer to the animal which made the feeding trace.For taxonomic diagnoses, we use the terminology established by Mikuláš et al. (2006) to maintain consistency with previously published systematics.Pits, sensu Binford (1981), were referred to as such by Mikuláš et al. (2006).We use the term 'hole' to refer to punctures, a convention set by Mikuláš et al. (2006), and later used by Rasser et al. (2016) and Mazuch et al. (2024).We disagree with the recent assessment by Trifilio et al. (2023), that the term 'hole' is ambiguous, as its original usage in defining N. nihilicus is in context very clearly describing a penetration through the cortical bone wall (Mikuláš et al., 2006).Outlines are described using geometric shapes.Trifilio et al. (2023) noted that the term 'ovoid' is inappropriate as it may imply an egg-shaped outline rather than an ellipse.Both Trifilio et al. (2023) and Rasser et al. (2016) instead described such trace outlines as 'ellipsoidal' .However, we favour the term 'elliptical' , as an 'ellipsoid' refers to a three-dimensional object.

Systematic ichnology
Nihilichnus Mikuláš et al., 2006 Type ichnospecies: Nihilichnus nihilicus Emended diagnosis: Roughly triangular, circular, elliptical, or lenticular holes, or external pits, found on skeletal material, occurring solitarily or in groups, which may show recurring patterns.Outer margin of the concavity is either smooth and continuous with surrounding skeletal material or shows minute irregular jags.Lacks organized boring patterns.
Description: Nihilichnus is referable to all pit and puncture traces, not formed by rasping, boring, etching, or pointed impact (pecking), but rather compression or penetration, found on skeletal material, bone, or otherwise (Rasser et al., 2016).
Remarks: We build upon the original diagnosis by Mikuláš et al. (2006).We add our observed trace morphologies and clarify the range of morphologies included within Nihilichnus.We observe what appears to be the ductile or semi-ductile deformation of cortical bone around certain instances of Nihilichnus, generating a smooth transition from the surrounding bone to the interior of the trace.This feature was also observed by Trifilio et al. (2023), but they describe the transition as soft, which we find to be an ambiguous term.The original diagnosis included that ragged edges were the result of brittle deformation, we have excluded that statement from the diagnosis as it is interpretive (Mikuláš et al., 2006).Mikuláš et al., 2006 Figure 2 Nihilichnus nihilicus Mikuláš et al. (2006, p. 120); Milan et al. (2011, p. 65); Wretman and Kear (2014, p. 3); Jacobsen et al. (2015, p. 458 Trifilio et al. (2023, p. 13).
Description: Referred to as 'punctures' by Binford (1981).Cortical bone has been pierced by a taphonomic agent, either destroying it or pushing it into the internal cavity.Interior of depression may be composed of: (a) separated and vertically offset cortical bone fragments surrounded by either cancellous bone or the medullary cavity; (b) exposed cancellous bone; (c) an opening to the medullary cavity of the bone; or (d) a hole passing through the entire bone.Trace outlines vary greatly.Some morphologies reflect cross-sectional shape of the penetrating agent.Occasionally the outer surface of the tooth is preserved in the depression, imprinted its form into the cancellous bone, or the depressed cortical bone fragments.
Remarks: Substrate is necessary, as an ichnotaxobase, to distinguish N. nihilicus from other forms of Nihilichnus.We specify in the diagnosis that N. nihilicus occurs in bone and that as a puncture the trace passes through the cortical bone.Multiple outline morphologies of N. nihilicus are observable.N. nihilicus often lacks the morphological characters necessary to ascribe a taphonomic agent to the trace.Rounded punctures, N. nihilicus with circular or elliptical outlines, are likely produced by animals with rounded, conical teeth (Njau & Blumenschine, 2006).Crocodiles produce rounded punctures, but because conical teeth are fairly ubiquitous among vertebrate faunivores, punctures with circular to elliptical outlines are usually not diagnostic to any particular agent (Njau & Blumenschine, 2006).Punctures with polygonal outlines, such as the triangular-shaped punctures described by Mikuláš et al. (2006), are likewise observed associated with crocodile feeding (Njau & Blumenschine, 2006, 2012).As modern crocodilian teeth are rounded to lenticular in cross section, such traces are not diagnostic trace makers with triangular teeth (Njau & Blumenschine, 2006;Njau & Gilbert, 2016).
Nihilichnus sulcatus Trifilio et al., 2023 Holotype: MCC 197-V Emended diagnosis: Elliptical holes with branching furrows.Both holes and furrows penetrate through the cortical bone into the cancellous bone, but neither contains filling or bioglyphs.The holes are perpendicular to the bone surface and have sharp edges, whereas the furrows have a smooth bottom, 'U-shaped' cross-sectional morphology and smooth edges.
Description: N. sulcatus is composed of punctures to vertebrate bone which are contiguous with furrows.This morphology includes only traces which penetrate through the cortical bones, thus more superficial trace morphologies are excluded.
Remarks: We do not observe N. sulcatus in our crocodile feeding trace sample.N. sulcatus was originally attributed to tooth slippage following tooth impression (Trifilio et al., 2023).However, the deep incision of the surrounding grooves into the cancellous bone fit more closely to Binford's (1981) definition of furrows.Furrows are not typically generated by crocodiles (Njau & Blumenschine, 2006).Likewise, furrows are not typically associated with bone slippage, a common phenomenon during crocodile feeding (Drumheller & Brochu, 2014;Njau & Blumenschine, 2006).Instead, they are most commonly attributed to gnawing; a behaviour distinct from slippage (Binford, 1981).Gnawing is not performed by crocodiles and is instead characteristic of mammalian carnivores (Njau & Blumenschine, 2006).Binford (1981) notes that as a bone is continuously gnawed, the initial puncture is obliterated as a furrow is developed.N. sulcatus thus appears to represent an intermediate between a puncture and a fully developed furrow.For this reason, perhaps N. sulcatus should be included within the gnawing trace ichnogenus, Machichnus, or recombined into a new ichnogenus (Mikuláš et al., 2006).
Nihilichnus hastarius n. isp.Nihilichnus Sadlok (2022Sadlok ( , p. 1196) ) Holotype: IUPC 101303 Etymology: Latin, hastarius: of or belonging to the spear; in reference to similarity to wound produced by a double-edged blade Diagnosis: Lenticular holes in vertebrate bone, passing through the cortical bone.Outline presents two vertices on opposing sides of trace.
Description: Referred to as 'bisected punctures' by Njau (2006) and Njau and Blumenschine (2006).The trace is a puncture with a lenticular (lens-shaped) outline, possessing two distinct vertices.As with N. nihilicus, the interior may be preserved as either separated cortical bone fragments or cancellous bone, or the puncture may pass into the medullary cavity, or through the bone entirely.If the interior is preserved, the sides may slope to a central, longitudinal vertex.
Remarks: N. hastarius is similar in formation mechanics to N. nihilicus but possesses a unique outline.As with N. nihilicus, the cortical bone has been punctured by the taphonomic agent, either destroying it or pushing it into the internal cavity.The lenticular outline matches the cross section of carinated crocodile teeth.N. hastarius may appear similar to bisected pits (N.sicarius), especially if the carinated tooth left behind a negative imprint in the cancellous bone.However, N. hastarius are differentiable from N. sicarius as they pass through the cortical surface.Bisected punctures are not observed associated with mammalian carnivores, rather they are diagnostic of crocodile feeding (Njau & Blumenschine, 2006).
Interior of pit evenly slopes to singular point with no obvious facets.Concavity may either be conical or hemispherical.
Remarks: Rounded pits do not pass through the cortical bone.The interior surface is not separated or offset, rather the pit interior slopes to the trace margins and is contiguous with surrounding cortical bone.N. clavus lacks facets and internal features, differentiating these traces from N. sicarius.Rounded pits are not typically diagnostic of any one taphonomic agent, particularly when the interior is poorly preserved.Smooth conical teeth are ubiquitous to a wide variety of predators, with possible causal agents ranging from mammalian predators to crocodiles (Njau & Blumenschine, 2006;Njau & Gilbert, 2016).
Nihilichnus Sadlok (2022Sadlok ( , p. 1196) ) Holotype: IUPC 101303 Etymology: Latin, sicarius: assassin; in reference to similarity to wound produced by a double-edged blade, but more superficial damage than N. hastarius Diagnosis: Lenticular pits on the surface of vertebrate bone, not passing through the cortical bone.Width to length ratio is greater than 0.25, and sides slope to an internal vertex dividing pit, resulting in a 'V-shaped' cross section.
Description: Referred to as 'bisected pits' by Njau (2006), Njau and Blumenschine (2006), and Njau and Gilbert (2016).Interior of pit consists of deformed cortical bone, with the sides sloping to a central bisecting plane.Interior of pit contains distinctive vertex or groove.Outline is lenticular, with distinct vertices aligned along the same axis as the pit vertex.
Remarks: Bisected pits are commonly associated with crocodilians.The teeth of crocodilians, particularly those found anterior in the tooth row, possess distinct carinae, and are lens-shaped in cross section.These lenticular teeth produce pits with sloping sides and a lenticular outlines.Often carinae contact the bone deeper than surrounding tooth, resulting in a deep groove at the centre of the trace.Even when crocodile teeth are significantly worn, the carinated teeth are still lenticular in cross section and possess sides which slope to a ridgeline.Thus, while the internal vertex may vary in depth, carinated teeth consistently produce bisected pits.As with N. clavus, the trace does not pass through the cortical bone, rather the cortical bone is depressed and remains contiguous with surrounding bone surface.Carinated teeth are also thought to generate bisected punctures (N.hastarius), which unlike N. sicarius pass through the cortical bone (Njau & Blumenschine, 2006).
Nihilichnus mortalis Mikuláš et al., 2006 Holotype: None Diagnosis: None Description: None Remarks: The ichnospecies Nihilichnus mortalis was mentioned in the abstract of Mikuláš et al. (2006) but was not described therein.Therefore, this ichnospecies is nomen nudum.Rasser et al., 2016 Holotype: SMNS 101527 Emended diagnosis: Solitary to rarely paired holes or external pits, with mainly elliptical to sometimes circular outline, in invertebrate shell.The hole margin is straight and perpendicular to the inner and outer surfaces.Very few small scratches may be present on the shell surface in close vicinity to the holes.
Description: A puncture or pit to the whorl of an invertebrate shell.Traces have only been observed on one side of the shell.
Remarks: The taxonomic definition of N. covichi stands as originally described.We emend the wording of the diagnosis to be explicit in the fact that these traces include pits and punctures and that the substrate is an invertebrate shell.Given the diversity of shelly fauna in the fossil record, this definition may need to be further restricted according to substrate taxon.These traces are not observed in our crocodile feeding trace sample, as our analysis was restricted to vertebrate fossils.

Holotype: CHMHZ-Mch0001
Diagnosis: Roughly circular openings occurring on a fourfold of nearly parallel, slightly divergent rows.The rows form an angle of 20° at most, while the openings maintain a regular distance and are situated more or less opposite each other along the length axis; the distance between the holes is approximately an order of magnitude greater than their diameter.
Description: Four rows of circular to elliptical pits or punctures in ammonite shell.Two rows occur on each side of the ammonite shell.Rows occur in pairs, converging at an 18-to 20-degree angle.Remarks: Unlike other forms of Nihilichnus, N. quadripertitus represents Nihilichnus situated within defined patterns.This has allowed for the diagnosis of the likely trace maker to a mosasauroid (Mazuch et al., 2024).These traces are not observed in crocodile feeding trace sample, as our analysis was restricted to Plio-Pleistocene vertebrate fossils.

Discussion
Ichnotaxonomy, as a branch of zoological nomenclature, allows for the systematic classification, observation, and interpretation of animal behaviour in the fossil record (Rindsberg, 2018;Vallon et al., 2016).Based entirely on morphology, ichnologists cannot rely on other forms of evidence to refine the classification of taxa (Rindsberg, 2018).We address a major issue of trace ambiguity with regard to praedichnia, traces left in bones as a result of predatory behaviour (Ekdale, 1985;Hunt & Lucas, 2021;Vallon et al., 2016).Feeding traces are often ascribed according to very broad definitions, defeating the utility of their classification due to an inability to systematically identify trace maker and the process of carcase utilization (Hunt & Lucas, 2021).
With the exception of N. quadripertitus, the prior diversity of Nihilichnus did not allow for an accurate interpretation or trace maker nor how animals in the past consumed prey.We split the ichnospecies N. nihilicus into four ichnospecies.These ichnotaxa will provide a better record of animal activity in the fossil record.We define these ichnotaxa based on morphological induces that are both unique and allow for interpretation of trace maker to varying degrees of accuracy.
We emend the diagnosis for N. nihilicus to exclude pits and restrict this ichnospecies to the morphology observed in the holotype: a non-lenticular puncture on vertebrate bone (Mikuláš et al., 2006).The remaining trace morphologies we observe in our sample of crocodile-modified bones are classified here into three new ichnospecies, separated according to the specificity of the biological information they preserve and their morphological distinctiveness.
'Bisected punctures' , as observed by Njau (2006) and Njau and Blumenschine (2006), fall outside the holotype morphology of N. nihilicus, and can be strongly ascribed to a specific tooth morphology.These are classified into a new puncture ichnospecies, N. hastarius: a puncture on vertebrate bone with a lenticular, or lens-shaped, outline.
Pits record trace maker tooth morphology with much greater accuracy, thus we define pits differently from punctures and diagnose them according to their internal morphology.Rounded pits with hemispherical or conical cross sections are named N. clavus, and lens-shaped pits with a longitudinal vertex and a 'V-shaped' cross section are named N. sicarius.N. clavus and N. sicarius are discrete states, representing rounded pits and bisected pits, respectively.Pits outside these two defined morphologies are currently not applicable to any taxonomic designation beyond Nihilichnus.Some states such as triangular pits and other discrete morphologies may be assigned to additional Nihilichnus ichnospecies with further study.
Given the information we observe within individual traces we favour an approach to the ichnotaxonomy of Nihilichnus based on the morphology of individual tooth impressions.Mazuch et al. (2024) notably favour the alternative approach, describing an ichnospecies of Nihilichnus based on a pattern of tooth impressions, preserving a trace of the entire jaw.As our sample of feeding traces generally lacks discernable patterns, the former approach is more applicable and in this context conveys more relevant ecological information.
Other proposed ichnotaxa for pit and puncture patterning include Mandaodonites (Cruickshank, 1986), representing elliptical tooth impressions along a sigmoidal curve, and Heterodontichnites (Rinehart et al., 2006), composed of both circular and elongated tooth impressions in a straight or gently curved series, which is attributed to heterodonty in the trace maker (Wisshak et al., 2019).Wisshak et al. (2019) observed that Heterodontichnites is likely a junior synonym of Mandaodonites, but Zonneveld (2022) later found Mandaodonites to be nomen erratum, which would restore priority to Heterodontichnites.Heterodontichnites may conflict with Nihilichnus, as Heterodontichnites would be composed of traces which, in isolation, could be construed as Nihilichnus.However, Wisshak et al. (2019) questioned whether the outline differences between the punctures reported by Rinehart et al. (2006) can be substantiated given the poor preservation of the referred material.The validity of Heterodontichnites, and whether this ichnotaxon may preclude the assignment of some pit and puncture traces to Nihilichnus, will require further assessment.Most Nihilichnus groupings, including N. quadripertitus, lack the outline heterogeneity stipulated in the generic definition of Heterodontichnites, and the traces reported herein for Nihilichnus do not conflict with Heterodontichnites or Mandaodonites (Cruickshank, 1986;Rinehart et al., 2006).
There are other issues that may arise when defining Nihilichnus patterns while comparable pit and puncture traces can be found in isolation.N. covichi, as currently diagnosed, does not conflict with N. quadripertitus, as N. covichi excludes feeding traces which occur in patterns of more than two (Rasser et al., 2016).Under the present definition, a singular pit or puncture on an ammonite shell would be diagnosed as N. covichi, rather than N. quadripertitus.The actual ecological relationship represented, however, would align more closely with those represented by N. quadripertitus.As we stipulate here in the case of N. nihilicus, it may become necessary to further delineate these ichnotaxa based on the substrate of preservation.We advocate for the description of Nihilichnus based on individual traces rather than groupings, in order to prevent these potential conflicts and to maintain consistency in ichnotaxonomic description.Likewise, assigning ichnotaxa based solely on individual pits and punctures would help prevent Nihilichnus from becoming too morphologically diverse (thus becoming a 'wastebasket taxon').Given the diversity of pit and puncture morphologies within osseous substrates, and the major ecological differences between predators of vertebrates and predators of shelly fauna, it may also be necessary to separate pit and puncture traces on calcareous substrates into separate ichnogenera from those found in bones.A future revision to exclude non-osseous substrates from Nihilichnus would prevent the ichnogenus from becoming too broad to convey meaningful ecological and behavioural information.

Application of Nihilichnus to vertebrate palaeoecology
Generally, trace fossils are named on the basis of observable behavioural differences translated into a substrate (Bertling et al., 2022;Vallon et al., 2016).However, in the case of praedichnia, individual traces, regardless of morphology, often present similar behavioural signals (Hunt & Lucas, 2021).Pit and puncture morphologies are ethologically similar, all resulting from direct perpendicular force applied to a skeletal substrate.The utility of splitting these ichnotaxa lies in our ability to record and interpret ecological and behavioural information in the fossil record, not directly through individual traces, but by measuring their patterns of occurrence (i.e., Njau & Blumenschine, 2006).
One can draw more specific behavioural conclusions from the patterning of feeding traces.However, both mammalian carnivores and crocodiles lack uniformity in their trace patterns, and while patterns of modification do exist, they are not expressed in a way that is consistent enough to describe ichnotaxa.Therefore, we propose to treat these praedichnia as countable units of measurement for describing predator bone modification and interpreting behaviour and ecology.One can identify the tooth morphology of the trace maker using the ichnospecies we describe.Then, using either taxonomically diagnostic ichnotaxa, such as N. sicarius and N. hastarius, or recorded patterns of gross bone surface modification, it becomes possible to identify the likely trace maker.The frequency of occurrence of these traces, or lack thereof, as well as their placement, density, and association with other trace morphologies can then provide important behavioural and ecological information regarding the animal that made the praedichnia (Binford, 1981).
Based on the taphonomic descriptions of Njau andBlumenschine (2006, 2012), the following patterns would emerge when observing Nihilichnus within the context of Plio-Pleistocene East Africa.Crocodiles generate N. nihilicus, N. clavus, N. sicarius, and N. hastarius, while mammalian carnivores typically only generate N. clavus and N. nihilicus.Mammalian carnivores produce many examples of N. clavus on individual bones, particularly on the epiphyseal ends of bones, reflecting gnawing behaviour (Binford, 1981).Crocodiles lack the ability to chew and generate N. clavus and N. sicarius together at lower frequencies on individual bones and more evenly distributed across the bone surface (Njau, 2006;Njau & Blumenschine, 2006).The total sample of Nihilichnus on crocodile modified bones is, proportionally, higher in punctures (N.nihilicus and N. hastarius) than mammalian carnivores, due to the immense bite force of crocodiles (Njau & Blumenschine, 2006).