Direct evidence of trophic interaction between a crocodyliform and a large terror bird in the Middle Miocene of La Venta, Colombia

Abstract

Direct evidence of predation and other trophic relationships provide valuable information about trophic interactions between species in palaeo-communities. Data on ecological interactions amongst extant apex predators open a unique opportunity to better understand how sympatric apex predators coexisted or interacted with each other in the past. Here, we describe direct evidence of a predation or scavenging event in which we hypothesize that a medium-sized caiman (possibly Purussaurus neivensis) consumed (either through scavenging or through direct predation) a large terror bird. The distal part of a left tibiotarsus from a phorusrhacid had four pits inflicted on the cortical bone, and no signs of healing, suggesting it did not survive this trophic event. This record contributes to our current understanding of prey consumed by P. neivensis in the wetlands of the Pebas System of South America and indicates that large phorusrhacids might have had higher predation risk than previously expected. This study provides evidence of a trophic relation between apex predators and the complexity of trophic interactions in the diverse vertebrate palaeo-community of La Venta in the Middle Miocene of northern South America.

1. Introduction

Being the largest vertebrates with specialized traits for carnivory, apex predators have played key roles in shaping the structure of extant and extinct communities through their direct effects on prey abundance and behaviour [1,2]. They usually forage on even larger preys, in general primary or secondary consumers, but this simplified view on the role of apex predators might underestimate the real complexity of trophic relationships [3]. Although two or more sympatric apex predators may coexist through niche partitioning within the same ecosystem [4], agonistic encounters between them are in fact common on extant biological communities either by predation (e.g. jaguar preying on caimans and crocodiles in the Neotropics [5,6]), by fighting over a corpse [7,8] or even by the scavenging of one apex predator on another apex predator’s corpse.

Fossil tooth marks are direct evidence of these trophic interactions and provide crucial data for testing hypotheses on the feeding ecology of extinct taxa and the trophic structure of ancient communities [9,10]. The relatively scant information on the nature and frequency of inter-specific relationships in palaeo-communities (e.g. competition or predation) has challenged our current understanding of consumer–resource interactions in the past. Indirect evidence of consumer–resource interactions or the reconstruction of general palaeo-diets relies on stable isotope [11,12] or microwear analyses (amongst other methods) [13,14], but trophic events carved in the bone of prey are glimpses of past community structure dynamics [15]. Evidence of trophic interactions between species has also been used to better understand community structure of past vertebrate assemblages [15]. For example, Pujos & Salas-Gismondi [16] reported taphonomic evidence of the predation of a large ground sloth by the largest crocodylian from the Middle Miocene Pebas Formation, Drumheller et al. [9] provided evidence of phytosaurs preying on paracrocodylomorphs during the Late Triassic, and Botfalvai et al. [17] documented tooth marks of a potential crocodylian on a bothremydid turtle and a hylaeochampsid crocodylian. A recent study has also examined the remains of a Miocene sea cow with evidence of tooth marks inflicted by both a crocodylian and a shark [18].

Intra-specific and inter-specific evidence of predation in the fossil record has also been recorded on many large vertebrate groups including crocodylomorphs, theropods, synapsids and placental mammals [19–22]. Nonetheless, direct evidence of scavenging proved more difficult to study in the past and often relies on the examination of the ichnological record [23,24], or on evidence of bioerosive activity [24]. Given that, in many cases, prey processing and subsequent tooth marks inflicted in the prey’s bones are similar in a predatory or scavenging event, they seldom provide evidence to confidently sort out these two predator–prey interactions. Thus, many records of tooth marks in the fossil record need to be interpreted as potentially being inflicted in a predatory or a scavenging interaction, without the possibility of excluding any of these hypotheses. Nonetheless, some studies with high frequency of tooth marks have potentially inferred scavenging interactions, rather than direct predation [23].

In tropical ecosystems, semi-aquatic crocodyliforms forage on both aquatic and terrestrial prey, as water sources become obligatory resources for terrestrial species. This kind of trophic dynamic might have ruled proto-Amazonian ecosystems since semi-aquatic crocodyliforms were abundant and enormous wetlands endured during most of the Miocene [25]. In the fossil record, there is also evidence of large caimans preying on large terrestrial sloths [16], and evidence of trophic interactions in palaeo-communities further contributing to our understanding of the ecological relationships between sympatric taxa, including those between apex predators [9].

The Middle Miocene fossiliferous deposits of La Venta (see detailed geological description in [26]) have yielded one of the most diverse vertebrate fossil assemblages of South America, with over 150 vertebrate taxa described including almost 100 mammals [27]. While very well known for the unique primate diversity, the fossil assemblage from La Venta badlands was also home to a complex community of crocodylomorphs, turtles, fishes and many mammals including South American native ungulates (SANUs), xenarthrans, rodents, marsupials, dolphins, manatees and bats [27]. Ecological modelling of the ancient vertebrate community at La Venta has suggested an atypically low diversity and abundance of terrestrial predators [28]. Recent studies have provided evidence that apart from massive aquatic crocodyliforms, such as Purussaurus neivensis, terrestrial predators such as large sebecids [29], phorusrhacid birds [30] and sparassodont metatherians [31] might have filled the apex predator niche of La Venta’s palaeo-community.

We reveal here taphonomic evidence of a unique trophic interaction between two apex predators from proto-Amazonia by describing the tooth marks perpetrated by a large aquatic caiman on a terrestrial terror bird in the Middle Miocene of La Venta, Colombia, and discussing its ecological implications.

2. Methods

We described tooth marks found on the fossil of a large phorusrhacid bird (MT-0200), recently described for the Middle Miocene assemblage of La Venta [30]. The fossil comes from the middle portion of Chunchullo Beds in La Victoria Formation (figure 1).

Figure 1.

(a) Map of equatorial South America showing the geographic location of La Venta (Tatacoa Desert). (b) Caudal view of distal tibiotarsus of the phorusrhacid MT-0200, showing the position of tooth marks. (c) Detail of the first pit. (d) Detail of the second pit. (e) Detail of the third pit. (f) Detail of the fourth pit. Scale bar in (b) is 50 mm. Scale bar in (c) to (f) is 5 mm.

Morphological terminology for the description of tooth marks follows [32–36]. Ichnotaxonomic definitions follow Mikuláš et al. [37] and LaBarge & Njau [38]. A handheld surface scanner (Artec Spider) was used to generate a mesh surface model saved as a Polygon File Format (.ply) of MT-0200. The model was transformed into .stl format using MeshLab v. 2023.12. The model was cropped to extract a strip of the surface containing the tooth marks. This strip was manipulated to generate a digital ‘negative’ of the tooth marks. This ‘negative’ was in turn compared with tooth series specimens of crocodyliform predators from La Venta, in order to identify the potential identity of the trace maker. Specimens used for comparisons are deposited at UCMP (University of California Museum of Paleontology), MT (Museo La Tormenta), IGM/SGC (Servicio Geológico Colombiano, Colombia) and MHNLT (Museo de Historia Natural La Tatacoa).

3. Results

(a). Prey

MT-0200 is represented by a partial tibiotarsus, comprising the distalmost part of the diaphysis and the entire distal epiphysis of a Phorusrhacidae, and particularly to a large phorusrhacinae bird as described by [30], an assemblage of large, cursorial terror birds with unresolved phylogenetic affinities [38].

There are four tooth marks on the posterior surface of the lateral condyle and distal diaphysis of the tibiotarsus (figure 1). The tooth marks are aligned, forming a row running obliquely in a disto-lateral to proximo-medial direction. Sets of two or more toothmarks that have a relatively parallel disposition are considered ‘serial’ tooth marks, implying they were inflicted by several teeth in a single jaw movement [34,36]. The greatest diameter of the tooth marks is roughly aligned with the general direction of the row. All tooth marks represent pits, rather than punctures, as circular depressions where the cortical bone is vertically offset, but not displaced to the level of the cancellous bone or the medullary cavity [32,33,36,38]. All pits are shallow and well separated, indicating that the cortical bone was penetrated only by the apical most part of the teeth. The bottom of the pits is hemispherical, rather than conical, and becomes better defined towards the proximo-medial end of the row. There are no signs of healing associated to the pits, and they can be assigned to the ichnogenus Nihilichnus [37].

The pit on the disto-lateral end of the row lies in a short distance medial to the articular surface of the tibiotarsus. This pit is the smallest of the series, with a roughly elliptical outline. Its major and minor diameters are 5.0 and 4.1 mm, respectively, its maximum depth is 2.4 mm, and the distance from the next pit in the series is 5.8 mm. This pit can be referred to the ichnospecies N. clavus, which includes circular to elliptical pits on the surface of bone [38].

The second pit is the deepest in the series (approximately 3 mm), but its margins are poorly defined on its disto-lateral half. Its major and minor diameters of 10.4 and 8.9 mm, respectively, and the distance from the following pit in the series is 17.3 mm. It is irregularly elliptical in outline, with a triangular indentation (hereafter ‘bisected pits’ sensu [34]) on the proximo-medial margin of the pit. From this indentation, an irregular groove runs along the proximo-medial half of the pit bottom. Two scores (defined as elongated tooth marks that do not reach the cancellous bone) run from the pit, extending distally towards the articular surface but not reaching it. These scores may represent drag snags formed during the initial contact with the bone surface [35]. The third pit of the series is the shallowest in the series (approximately 1.7 mm), with an elliptical outline. Its major and minor diameters are 9.8 and 8.6 mm, respectively, and the distance to the next pit is 14.3 mm. On the disto-lateral end, there is a shallow bisected pit.

The fourth pit, located on the proximo-medial end of the row, is approximately 3.4 mm deep, with major and minor diameters of 10.1 and 7.7 mm, respectively. This pit is more defined than the others and seems to represent a more detailed mould of the corresponding tooth. There are well-developed bisected pits on both ends of this pit so that its outline is lenticular, rather than elliptical. On the bottom of the pit, there is a shallow, irregular groove running between both indentations. The second, third and fourth pits all share characteristics of ‘bisected pits’ [34]. Bisected pits lie either on one or both ends of the major diameter, and a distinct groove runs along the bottom of both the second and fourth pits. These features refer these pits to the ichnospecies Nihilichnus sicarius [38].

4. Discussion

(a). Identity of the trace maker

Bone assemblages produced in tropical terrestrial environments feature a mixture of elements modified by carnivorous mammals and crocodyliforms [34]. This pattern could potentially be found in fossils from the La Venta biota, where, in addition to phorusrhacid birds [29], medium- to large-sized vertebrate predators include eusuchian and sebecid crocodyliforms as well as sparassodont metatherians [29,31,39,40].

Living mammalian carnivores and crocodylians display different modes of prey processing. Carnivorous mammals tend to gnaw and break the epiphyses of long bones in order to consume cancellous tissue and marrow [41,42], whereas crocodylians separate portions of the carcass and swallow them whole [34]. Long bones modified by mammals show some degree of gnawing at their extremes, often bearing toothed or crenulated edges [41], while bones consumed by crocodylians may show tooth-marking patterns on diaphyses, but epiphyses are left ungnawed [34].

Tooth marks reported on the distal tibiotarsus of a psilopterine phorusrhacid from the Middle Eocene of Argentina are largely concentrated on the distal articular surface [43]. This damage pattern, consistent with gnawing, indicates consumption by a carnivorous mammal. In contrast, MT-0200 features a single row of pits or serial tooth marks, mostly restricted to the diaphysis, and leaving the articular surface intact, as expected from consumption by a crocodylian. In addition, the presence of ‘bisected pits’, referred to N. sicarius, indicates that the traces were made by teeth with distinct carinae on their mesial and distal edges, more typically found in crocodyliforms than in carnivorous mammals [33,34].

At least eight crocodylian taxa have been recognized from the fossil assemblage at La Venta (see appendix 1 in [27]), including the terrestrial sebecid, Langstonia huilensis, the gryposuchine gavialoid, Gryposuchus colombianus, the putative tomistomine, Charactosuchus fieldsi and five caimanine alligatorid species [29,40,44,45]. Caimanines include small, generalist forms referred to Caiman sp. and Eocaiman sp., together with specialized forms such as the blunt-toothed Kuttanacaiman iquitosensis Salas-Gismondi et al. [46], the ‘duck faced’ Mourasuchus atopus, and the giant P. neivensis [29,40,47].

Differing tooth morphologies in crocodyliforms at La Venta allow us to assess potential candidates for the trace maker (see figure 2). The teeth of L. huilensis, like those of other sebecids, are labio-lingually compressed, curved in distal direction and have serrated edges with denticles [29,40,44]. This morphology, known as ziphodont [32], differs from that of extant crocodylians and carnivorous mammals but is present in the Komodo dragon Varanus komodoensis. Controlled feeding trials on V. komodoensis reveal that this species seldom leaves pits on bone surface, whereas the prevailing type of tooth marks are scores, some of them striated by serrated edge [48]. Similar tooth marks have been attributed to extinct ziphodont reptiles, such as theropod dinosaurs [49], but are significantly different from those observed on MT-0200, indicating that the tooth marks studied herein are unlikely to have been produced by sebecid crocodyliforms. Stenorostrine eusuchian crocodylians from La Venta are represented by G. colombianus and Ch. fieldsi [29,40]. These narrow-snouted forms have very distally curved, slender conical crowns with pointed apices. Such tooth morphologies would leave pointed conical pits different from the hemispherical bottom surfaces observed on MT-0200.

Figure 2.

Comparisons between tooth marks on MT-0200 and eusuchian crocodyliform dentitions from the Honda Group. (a) Cropped surface model of MT-0200 including tooth marks, in external (above) and lateral (below) views. (b) Gryposuchus colombianus, VPPLT 1744, posterior left maxillary and dentary, showing maxillary teeth 15 and 17, and dentary teeth 19–21 (reversed image). (c) Mourasuchus atopus, left dentary showing teeth 4–7 and 9–10 (reversed image). (d) Kuttanacaiman sp. VPPLT 294, posterior right dentary showing teeth 9–10 and 13–15. (e) Caiman sp., VPPLT 1660, right maxillary showing teeth 4–12. (f) Purussaurus neivensis, MT−1096, left maxillary of juvenile specimen showing maxillary teeth 8–10 (reversed image). Scale bars are 10 mm.

Among eusuchian crocodyliform teeth from La Venta, the morphology of tooth marks on MT-0200 is more consistent with that of the posterior blunt crowns found in most caimanines. The only exception is M. atopus, in which all teeth are conical and pointed, although less curved than those of gavialoids. The depth of pits on MT-0200, as well as the distance between them, indicates that these tooth marks were made by the crown tips of large, relatively blunt teeth. K. iquitosensis has a very specialized dentition, forming a flat crushing ‘battery’ in which the posterior most teeth are flat and lie at the same level [49]. This configuration does not match the tooth marks on MT-0200. The posterior teeth of Caiman sp. and Eocaiman sp. are far more consistent with the shape expected from the tooth marks. However, teeth in known specimens of these caimanines are not large enough to produce the type of tooth marks observed on MT-0200. The deepest parts of the three larger pits are separated by approximately 27 mm, which is roughly equivalent to the distances between crown tips. Comparisons with specimens of black caiman, Melanosuchus niger, suggest that the traces were likely inflicted by a large caimanine, between 4.6 and 4.8 m long (figure 3). In the current fossil assemblage of La Venta, the best match for a large caiman in this size range would be a juvenile or subadult specimen of the giant caimanine P. neivensis, the largest crocodyliform in the La Venta Fauna.

Figure 3.

(a) Superimposed 3D models of MT−0200 and a skull of extant black caiman (Melanosuchus niger UF-Herp-53600). Both models at the same scale, the M. niger model is archived on MorphoSource (https://doi.org/10.17602/M2/M359353). Total length estimated for the specimen of M. niger is 4.84 m and is used here as an analogy to a medium-sized Purussaurus (skull or mandible not available) [50]. Scale bar is 50 mm. (b) Detail of the superimposed model, showing the close match between the teeth of UF-Herp-53600 and the tooth marks on MT-0200. Scale bar is 50 mm. (c) Artistic representation of a caimanine alligatorid (Purussaurus neivensis) preying on a large phorusrhacid bird. (d) Artistic representation of a caimanine alligatorid (Purussaurus neivensis) scavenging on a large phorusrhacid bird. Illustrations by Julian Bayona Becerra.

This study provides evidence for a large aquatic crocodyliform consuming a large predatory bird either by predation or scavenging. Although inter-specific competitive killing can be common within vertebrate communities [51], evidence of direct trophic interactions between apex predators remains as a topic that has been historically understudied. Prey is most often represented by herbivores and other animals that are not on the top of the trophic web (secondary or tertiary consumers) [2]. Thus, this anecdotal account of an aquatic apex predator feeding on a terrestrial apex predator further contributes to elucidating the complexity of trophic webs in vertebrate extant and palaeo-communities. In the fossil record, anecdotal events such as the tooth marks inflicted by an aquatic phytosaur on a terrestrial paracrocodylomorph support the existence of antagonistic interactions between apex predators in the past [9].

During the Cenozoic Era of northern South America, and particularly in the areas influenced by the large Pebas system, large crocodylomorphs comprised an important predatory guild [16]. Evidence of predation by caimans has been evidenced through tooth marks inflicted in the carapace of a turtle [52], and on the hindlimb of a large terrestrial sloth (Pseudoprepotherium sp.) [16]. Apex predators may often meet at the aquatic and terrestrial interphase [9]. Terrestrial apex predators may preferentially wander close to water sources—especially in dry seasons—as these areas may set an ideal scenario for terrestrial predators to feed on aquatic taxa [5,6], or to attempt to prey on other taxa obliged to drink water from water sources. Thus, aquatic predators can prey on terrestrial vertebrates—including apex predators—as they approach water sources, similar to what is known from aquatic water holes in Africa [53].

The evidence of predation on a relatively small Psilopterinae in southern South America [35] opens the possibility for larger sparassodonts such as Lycopsis longirostrus Marshall, 1977, and Duckecynus magnus Goin, 1997, to have also posed predation pressure on some of the largest cursorial terror birds (i.e. MT-0200) at La Venta. The taphonomic evidence described above largely suggests the pits in the distal section of the tibiotarsus of this large phorursrhacid were inflicted by a medium-sized caimanine, possibly Purussaurus. If the interaction between both apex predators was when both animals were alive (not a scavenging behaviour), it must have taken place when the bird of terror approached a water source (rivers) and resulted in the predation of the victim that did not survive based on the lack of regenerative evidence on the pits inflicted by the large caiman. Contrarily, if the interaction took place as a scavenging event, the large terror bird must have visited areas near the aquatic systems and could have died from intra-specific competition, predation by another terrestrial predator or simply died close to a water body where a Purussaurus was able to retrieve it and feed from its corpse. Nevertheless, at present there is no sufficient evidence to discard or corroborate that the Purussaurus predator was not feeding on a dead terror bird (scavenging). This hypothesis might be further supported by the fact that the stratigraphic layer where the terror bird was found has been proposed to have had water channels running through it during the time in the Middle Miocene when this interaction between apex predators took place [30,54].

In conclusion, this study provides evidence of a unique trophic interaction between some of the most emblematic apex predators in the Miocene of South America providing evidence of the complex inter-specific trophic networks at La Venta and other diverse palaeo-ecological communities.

Ethics

This work did not require ethical approval from a human subject or animal welfare committee.

Data accessibility

This article has no additional data.

Declaration of AI use

We have not used AI-assisted technologies in creating this article.

Authors’ contributions

A.L.: conceptualization, funding acquisition, investigation, writing—original draft; J.W.M.-B.: formal analysis, investigation, visualization, writing—original draft; F.J.D.: investigation, writing—original draft; S.B.C.: investigation, writing—original draft; L.G.O.-P.: conceptualization, investigation, writing—original draft; C.A.P.-R.: investigation, writing—original draft; R.S.-G.: conceptualization, investigation, writing—original draft.

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

Conflict of interest declaration

We declare we have no competing interest.

Funding

We thank Fondos de Investigación at Facultad de Ciencias at Universidad de Los Andes (INV-2023-162-2829) for financial support.