A diagnosis of crocodile feeding traces on larger mammal bone, with fossil examples from the Plio-Pleistocene Olduvai Basin, Tanzania

Abstract

Neotaphonomic studies have determined the patterns of bone damage created by larger mammalian carnivores when consuming mammalian carcasses. Typically, mammalian carnivores gnaw and break bones to various degrees in order to access marrow, grease, and brain tissue. In contrast, crocodiles attempt to swallow whole parts of mammal carcasses, inflicting in the process tooth marks and other feeding traces on some of the bones they are unable to ingest. Although crocodiles are major predators of larger mammals along the margins of protected tropical rivers and lakes, their feeding traces on bone have received little systematic attention in neotaphonomic research.

We present diagnostic characteristics of Crocodylus niloticus damage to uningested mammal bones resulting from a series of controlled observations of captive crocodile feeding. The resulting bone assemblages are composed of primarily complete elements from articulating units, some of which bear an extremely high density of shallow to deep, transversely to obliquely oriented tooth scores over often large areas of the bone, along with shallow to deep pits and punctures. Some of the tooth marks (bisected pits and punctures, hook scores) have a distinctive morphology we have not observed to be produced by mammalian carnivores. The assemblages are also characterized by the retention of both low- and high-density bone portions, an absence of gross gnawing, and minimal fragmentation. Together, the damage characteristics associated with feeding by crocodiles are highly distinctive from those produced by mammalian carnivores.

Modern surface bone assemblages along the Grumeti River in Tanzania's Serengeti National Park contain a mixture of specimens bearing damage characteristic of crocodiles and mammalian carnivores. Comparison of Plio-Pleistocene fossil bones from Olduvai Gorge, Tanzania, to bones damaged by captive and free-ranging Nile crocodiles reveals direct evidence of fossil crocodilian feeding from larger mammal bones associated with Oldowan stone artifacts.

Introduction

Tooth marks and associated damage features inflicted by carnivores on bone are traces of feeding behavior and potentially other aspects of carnivore-carnivore and carnivore-herbivore interactions in modern and fossil settings. Numerous studies have established criteria for recognizing feeding traces of mammalian carnivores, motivated largely by the co-occurrence of carnivore tooth marking and early hominin butchery marking in vertebrate fossil assemblages (Binford, 1981, Bunn, 1981, Potts and Shipman, 1981, Shipman, 1981a, Shipman, 1981b, Haynes, 1983, Shipman and Rose, 1983, Blumenschine and Selvaggio, 1988, Blumenschine and Selvaggio, 1991, Andrews and Fernandez-Jalvo, 1997; and references therein). Tooth marks and accompanying bone damage produced by rodents, predatory small mammals and birds, and non-human primates have also been documented on modern and in most cases fossil bones (e.g., Maguire et al., 1980, Brain, 1981, Farlow et al., 1986, Andrews, 1990, Pickering and Wallis, 1997, Domínguez-Rodrigo et al., 1998, Plummer and Stanford, 2000, Tappen and Wrangham, 2000).

Systematic studies of bone modification by carnivorous reptiles are limited to Fisher's (1981a) observations of bones and teeth partially digested and defecated by alligators and caimans. Fisher found that bones defecated by these crocodilians are usually completely decalcified, a condition that may be unique to these species. The organic matrices of defecated bones are often left intact, but will decompose very quickly. Fisher suggests, however, that these organic matrices might be remineralized and preserved as recognizable fossils if fresh feces are deposited in anaerobic, reducing environments that are subsequently not disturbed. Fisher also found that decalcification during crocodilian digestion will remove all or most enamel of teeth that are defecated, and he describes such fossil enamel-less teeth of Paleocene age attributable to crocodilians (Fisher, 1981a, Fisher, 1981b, Fisher, 1981c).

The only report of modification of bones fed on but not ingested by crocodiles is Davidson and Solomon's (1990) observation of a human killed by a saltwater crocodile (Crocodylus porosus). They described tooth pits (including semi-circular “broken-tooth” marks) and tooth scores (including “incising” tooth marks) as damage features produced by crocodiles. They noted that some of the incised tooth scores resemble stone-tool cut marks, while other scores bear a resemblance to mammalian carnivore tooth marks. They also observed parallel scores, compound and comminuted fractures on the midshaft of limb bones, and the retention of complete limb bone elements.

Crocodiles are notorious predators in tropical rivers and lakes, routinely ambushing terrestrial mammals that come down to the water's edge to drink, or scavenging kills of mammalian carnivores (e.g., Attwell, 1959, Cott, 1961, Guggisberg, 1972, Pooley and Gans, 1976). In these settings, crocodile predation is likely to contribute to both the composition and condition of bone assemblages, which may also bear traces of feeding by mammalian carnivores.

Crocodilian tooth marking in fossil assemblages is inferred rarely and anecdotally in the paleontological literature (e.g., Dubois, 1927, Buffetaut, 1983, Pickford, 1996, Frey et al., 2002, Avilla et al., 2004). However, fossil bones of animals consumed partially by crocodiles are likely to be present in many paleontological assemblages beginning with the appearance of these reptiles in the Mesozoic (Schwimmer, 2002). Crocodile predation is focused in aquatic settings conducive to in situ or downstream burial and preservation of bones. This potentially has great relevance for paleoanthropology because many early hominin body fossils and activity traces occur in or near aquatic settings, some of which bear body fossils of crocodilians (e.g., Leakey, 1971, Tchernov, 1976, Tchernov, 1986, Harris, 1978, Feibel et al., 1991; for East African Neogene examples). Recognition of crocodile feeding traces in a bone assemblage is therefore important for identifying the multiple biological agencies of bone assemblage formation and their potential interactions.

Given the habitat-specificity of modern crocodiles (Barker, 1953, Modha, 1967, Neill, 1971, Watson et al., 1971, Graham and Beard, 1973), the recognition of crocodilian trace fossils is also important for assigning a recovered bone assemblage to a specific environmental context. This is relevant to the study of prehistoric hominin land use, in which the habitat-specificity of hominin activity traces is predicted to be strong and highly variable across heterogeneous paleolandscapes (Blumenschine and Peters, 1998).

Building on Davidson and Solomon's (1990) anecdotal observations, we provide a systematic description of feeding damage to uningested bones of larger mammal carcasses inflicted by modern crocodiles, specifically Crocodylus niloticus. Our description is drawn from observations of the morphology of tooth marks on bone surfaces, their anatomical patterning, and the degree of gross damage to and fragmentation of bones. The diagnosis is based on controlled observations of feeding by captive Nile crocodiles at two crocodile farms in Tanzania, and on contrasts in resulting bone damage to that which we have observed on bone collections bearing modifications produced by larger mammalian carnivores including spotted hyenas, lions, leopards, cheetah, and jackals (Blumenschine, 1988, Blumenschine and Selvaggio, 1988, Blumenschine et al., 1996). The distinct bone damage inflicted by mammalian carnivores and crocodiles is apparent in bone assemblages recovered at or near pools in the Lower Grumeti River occupied by free-ranging Nile crocodiles in the Serengeti National Park, Tanzania. The results are relevant to identifying and interpreting the trace fossils of feeding by fossil crocodiles, and by extension, some carnivorous dinosaurs. We provide two examples of fossil bones inferred on the basis of our neotaphonomic study to have been fed on by Plio-Pleistocene crocodiles from Olduvai Gorge. These fossils were recovered in Oldowan paleoanthropological contexts by the Olduvai Landscape Paleoanthropology Project (OLAPP; e.g., Blumenschine et al., 2003). It was the recognition by one of us (RJB) that the damage patterns on these specimens are uncharacteristic of those produced by mammalian carnivores that prompted our study of crocodiles as taphonomic agents.

The feeding behavior and anatomy of crocodiles determines their taphonomic capabilities on bones of larger vertebrates. We use the term crocodile to refer to members of the genus Crocodylus (“true” crocodiles), and the term crocodilian to refer to the family Crocodylidae, which includes the subfamilies Alligatorinae (alligators and caimans), Crocodylinae (“true” crocodiles and African dwarf crocodiles), and Gavialinae (gharials and “false” gharials or Tomistoma; e.g., Grenard, 1991; cf. Brochu, 2003).

Depending on body size and age, living crocodiles include an extremely wide range of food in their diet, from insects to large mammals (Table 1). Crocodiles do not transport prey items long distances as some mammalian carnivores do, and usually consume prey in the water a short distance from the kill site, typically at the water's edge. Crocodiles do not need to eat frequently due to their exothermic regulatory capabilities. Also, the relatively small stomach of adult crocodiles compared to their body size restricts meal size (Parsons and Cameron, 1977, Grenard, 1991), increasing the potential for relatively large carcass parts, including whole bones, to be abandoned. These parts usually sink to the bottom of inhabited pools.

Nile crocodiles are the largest extant crocodilians, capable of taking animals larger than themselves, including adult African buffalo (e.g., Pienaar, 1969). They first appeared in East Africa during the Pliocene, in some places co-existing with their direct ancestor, Crocodylus lloidi, the predominant East African crocodilian from the base of the Neogene to the mid-Quaternary (Tchernov, 1986). C. niloticus is a less robust species with a slightly narrower and longer snout than its brevirostrine (short- and broad-snouted) ancestor, but the two are very similar postcranially. Two other, often sympatric crocodilians that lived in East Africa during the Plio-Pleistocene are Crocodylus cataphractus and Euthecodon brumpti. While C. lloidi and E. brumpti went extinct by the mid-Quaternary, C. cataphractus survives in Lake Tanganyika today, and C. niloticus spread progressively throughout East Africa, predominating in inland waters.

The Nile crocodile snout has elongated slightly since the Pliocene, without significant morphological changes to the posterior region of the cranial skeleton, or to the postcranial skeleton (Tchernov, 1986). Indeed, all modern (eusuchus) crocodilians are morphologically conservative in their postcranial region, varying interspecifically only in skull morphology (e.g., Densmore and Owen, 1989, Schwimmer, 2002) in response to dietary specializations (Iordansky, 1973, Langston, 1973) or ontogeny (Steel, 1973). Brevirostrine forms rely on larger terrestrial mammal prey, as Tchernov (1986) inferred for C. lloidi, while longer-snouted forms such as C. niloticus include more fish in the diet. Exclusive piscivory characterizes the most longirostrine (long- and narrow-snouted) species, such as C. cataphractus and the modern gharial (Gavialis gangeticus) (Iordansky, 1973), and as inferred for the ancient E. brumpti (Tchernov, 1976, Tchernov, 1986). The contrast can also be seen among age groups within a single species. For example, very young Nile crocodiles have a relatively elongated skull suited to catching insects and small fish, while the snout of very large and old individuals is more robust, broad, and short, essentially restricting them to mammalian prey (Table 1).

The craniodental morphology common to crocodilians is most suitable for grasping and stabbing rather than for chewing or cutting (e.g., Edmund, 1969, Bellairs, 1970, Langston, 1973). A massive skull and large gape allow crocodiles to swallow large food items without intensive oral processing. Unlike mammalian carnivores, crocodiles do not possess precise tooth-to-tooth occlusion, and their lateral teeth angle inward and interlock mesio-distally. The anterior dentary teeth lie labial to the premaxillary tooth row. The absence of occlusion restricts side-to-side jaw movement and helps to prevent premature dislodging of a functional tooth from the socket (e.g., Pooley, 1989). In contrast to mammalian carnivores, crocodiles are not adapted to acquiring within-bone nutrients through gnawing and bone fragmentation.

All crocodilians retain the primitive thecodont dentition of the archosaurians, in which teeth are set in deep alveoli (Romer and Parsons, 1978, Lubkin, 1997) and replaced repeatedly during life through resorption of the root and shedding of the crown (Owen, 1840-1845, Mummery, 1924, Poole, 1961, Edmund, 1962, Erickson, 1996a, Erickson, 1996b). Each successional tooth grows to be larger than its predecessor, with tooth replacement slowing in older individuals (Poole, 1961). Interlocking of upper and lower teeth is preserved during succession, although great variation in tooth size is evident in single individuals due to variation in eruption age of each tooth.

Crocodile teeth are robust, pointed, and conical or cylindro-conical, flattened slightly bucco-lingually. Juvenile crocodiles have slender and sharp teeth (effective, for example, for piercing insect exoskeletons) compared to adults, whose more robust and blunt teeth are suited for crushing and tearing large vertebrate carcass parts. The tooth crowns are bicarinated, each displaying a carina, or keel-shaped ridge, on the anterior and posterior faces (Fig. 1). When newly erupted, the carinae display a continuous series of very fine denticles, enhancing their sharpness. Carinae with worn-away denticles still form a pronounced ridge that becomes rounded and blunt with further wear. The carinae of the anterior dentition are sharper and more pronounced (e.g., Poole, 1961). With the exception of the teeth located distally in the dentary and maxilla, rounded linear ridges (folds) run vertically from the full circumference of the neck, meeting at the tip of the crown. In addition to their variable eruptive morphology and wear, common chipping of the apex of the crown expectedly leads to diverse tooth mark morphologies, even on a single bone fed on by one crocodile.

The upper jaw of an adult Nile crocodile has an average of 36 erupted teeth, while the lower jaw has an average of 30 teeth (cf. Brazaitis, 1973, Iordansky, 1973). The premaxillary fourth tooth is distinctly enlarged, bearing pronounced carinae suited to gripping struggling prey. The enlarged upper tenth tooth and lower eleventh tooth are robust and less pointed, and are important for crushing weaker bones. A notch occurs between the eighth and ninth dentary teeth to allow room for the protruding upper tenth tooth. Similarly, a notch occurs on the upper jaw between both the fifth and sixth teeth, and the twelfth and thirteenth teeth. These notches accommodate in the closed jaws the enlarged lower fourth and eleventh teeth, respectively.

Review of the literature indicates that the sequence by which crocodiles seize, kill, and consume large mammalian prey can be divided into six stages (cf. Bramble and Wake, 1985, Hiiemae and Crompton, 1985, Cleuren and De Vree, 1992, Schwenk, 2000):

• Prey capture: Crocodiles capture prey by lunging or leaping, usually in ambush from water (e.g., Grenard, 1991). Most often the anterior portions of the jaws are involved in apprehending the prey (Schwenk, 2000, p. 50), although the cheek teeth may also be used (e.g., Cleuren and De Vree, 2000, p. 348). The muzzle, skull, lower limbs, ankle and knee joints, or other convenient parts of the prey may suffer injuries inflicted by the teeth of the reptile, amplified for large crocodiles by the blunt force trauma inflicted by their massive head. Tooth marking on bone expectedly accompanies these injuries.

• Killing: After acquisition, the struggling prey is dragged and subdued underwater until it dies (e.g., Pooley and Gans, 1976). The seized-upon prey is positioned further back in the mouth before a forceful bite is applied by pressing the jaws together in order to crush and compress the carcass into the mouth. The reptile may spend a considerable amount of time at this stage with its head pointing upwards to initiate swallowing (Attwell, 1959). The enlarged premaxillary fourth teeth grip and hold the prey while the tenth maxillary teeth and the eleventh mandibular teeth may be involved in crushing relatively weak bones. These and other teeth occasionally penetrate flat bone, and even the thicker cortical bone of long bone midshafts, which in the process can be fractured obliquely or longitudinally along the shaft (Davidson and Solomon, 1990).

• Reduction: This stage includes gross dismemberment of the carcass into pieces that can be swallowed. It is accomplished through vigorous shaking. Large crocodiles will also employ both forceful battering of bones on rocks, and the “death-roll,” in which it spins its whole body along the longitudinal axis while anchored by the jaw to a carcass part (e.g., Schmidt, 1944, Attwell, 1959, Green, 1988). The death-roll is often performed on the same carcass by two or more crocodiles working at opposite ends and spinning in opposite directions. This action expectedly produces abundant and deep tooth-marking on bones at grasping sites (e.g., distal limb elements). We have also observed a single crocodile performing the death-roll on a relatively large carcass.

• Defleshing: This behavior facilitates skeletal disarticulation. Usually, a number of crocodiles will surround a carcass, each ripping off a mouthful before retreating to the periphery for consumption. For fully fleshed parts, the animal will grab a muscle mass with the side of the jaw or the anterior teeth before tearing and pulling in a death-roll action. Crocodiles typically do not deflesh bone as cleanly as do mammalian carnivores or vultures. Rather, large scraps of near-bone flesh that could not be removed by the anterior dentition remain on abandoned parts. Defleshing may lead to relatively shallow tooth-marking on any meaty bone.

• Swallowing: A reduced or whole carcass part is maneuvered in the mouth accompanied by a rapid series of light bites before being swallowed. If a food item is still too large to swallow, the animal will deliver several fast-closing bites or work the food item back and forth in its mouth, head held up and biting in many locations. Depending on the size of the food item, a forceful bite may be delivered during this phase of carcass manipulation in order to crush larger bones or skeletal units such as the rib cage. However, jaw adduction may not be complete if the bone is too robust. Many tooth marks, including punctures, are likely inflicted at this feeding stage.

• Carcass abandonment: Some carcass parts escape consumption if the prey is large relative to the size or number of crocodiles present. This is particularly true of large parts, which are discarded either complete or broken and allowed to sink to the bottom of the pool. The discarded fragments are referred to by Davidson and Solomon (1990) as “flown-off” pieces. It is the feeding traces on these abandoned whole and fragmentary bones that we aim to document in this report.