# Sources of Variation in Carnivore Tooth-Mark Frequencies in a Modern Spotted Hyena (Crocuta crocuta) Den Assemblage, Amboseli Park, Kenya

> How do you tell, from a pile of old bones, whether hyenas or early humans ate first? Studying bones from a spotted hyena den in Kenya’s Amboseli National Park, this study shows that the share of tooth-marked bones depends on how big the pieces are, which bone they come from, and how dense that bone is. The author suggests ways to compare bone collections fairly, starting with the dense middle shafts of limb bones.

Summary by Kersey Lawrence, Original Wisdom. Updated 9 November 2026. Page: https://www.originalwisdom.com/tracking-literature/2007-faith-sources-variation-carnivore-tooth/

## Citation and access

- **Authors:** J. Tyler Faith
- **Year published:** 2007
- **Journal or source:** Journal of Archaeological Science
- **Full citation:** Faith, J. T. 2007. Sources of variation in carnivore tooth-mark frequencies in a modern spotted hyena (Crocuta crocuta) den assemblage, Amboseli Park, Kenya. Journal of Archaeological Science 34(10): 1601-1609.
- **DOI:** 10.1016/j.jas.2006.11.014
- **Access status:** Publisher link only
- **Where to find it:** https://doi.org/10.1016/j.jas.2006.11.014

## Tags

- **Species:** Mammals > Spotted hyena
- **Region:** Africa > Kenya
- **Behaviour and ecology:** Behaviour; Bone accumulation; Denning; Feeding
- **Methods and technology:** Bone damage analysis; Chi-square test; Minimum number of individuals (MNI)
- **Data type:** Existing data; New field data
- **Observer qualification:** Not needed
- **Discipline:** Palaeoanthropology; Taphonomy
- **Year published:** 2007

## Detailed keywords

- **Species:** Studied: spotted hyena (Crocuta crocuta), as the bone collector at the den. Bone remains in the den (Table 1, by number of identified specimens): wildebeest (Connochaetes taurinus, 315), zebra (Equus burchelli, 154), Grant’s gazelle (Gazella granti, 41), Thomson’s gazelle (Gazella thompsoni, 22), domestic cattle (Bos taurus, 20), donkey (Equus asinus, 5), spotted hyena (Crocuta crocuta, 5), jackal (Canis cf. mesomelas, 4), caracal (Felis caracal, 4), springhare (Pedetes sp., 4), bushbuck (Tragelaphus scriptus, 3), leopard (Panthera pardus, 2), warthog (Phacochoerus aethiopicus, 2), sheep or goat (caprovine, 1), lion (Panthera leo, 1), African buffalo (Syncerus caffer, 1), impala (Aepyceros melampus, 1), plus bovids identified only to size class (sizes 1 to 3). The paper gives scientific names only for most of these; the common names here are ours. Comparison assemblages discussed: Hadza foragers’ bone refuse, and experimental assemblages including sheep (Ovis aries) at the Berkeley Hyena Colony.
- **Location and habitat:** Amboseli Airstrip Den, in the open grasslands of Amboseli National Park, southern Kenya. The den is a 20 by 4 m open trench in calcrete (a hard limestone crust), with small covered passages and burrows in its low (30 to 40 cm) walls. The bones are stored at the Palaeontology Department of the National Museums of Kenya.
- **Topics:** carnivore tooth marks on long bones; zooarchaeology; taphonomy (how bone assemblages form); whether early humans (hominins) or carnivores reached carcasses first; hunting versus scavenging by early humans; spotted hyena den bone accumulations; bone fragment size; prey body size classes; differences between skeletal elements and portions (ends, near-end shafts, middle shafts); bone density and density-mediated destruction; bone weathering stages; experimental versus archaeological assemblages.
- **Statistics or models:** Spearman’s rank-order correlation; chi-square tests with adjusted residuals; t-tests; number of identified specimens (NISP) and minimum number of individuals (MNI). Bone density values are published wildebeest values (Lam et al., 1999). Two small inconsistencies: the text gives the increase in tooth-marked middle shafts from size 2 to size 3 mammals as 11.8%, but Table 3 (63.4% and 75.3%) gives 11.9%; and Table 4 gives femur proximal ends as 2 of 3 tooth-marked as 66.6%, where 66.7% would be expected. Neither changes the findings.
- **Data type:** New field data: the bone collection was studied in 2006. The bones themselves were mapped and collected from the den surface by Andrew Hill in 1975 (over 1600 specimens); the den had been used by hyenas over several decades before 1975. Bone density values come from an existing published dataset.

## Synopsis

### 1. What they did

Archaeologists use carnivore tooth marks on broken limb bones to work out whether early humans or carnivores got to a carcass first. If carnivores ate first, the middle shafts of the limb bones tend to be heavily tooth-marked; if humans stripped the meat and marrow first, carnivores leave fewer marks there. But the percentages of tooth-marked bones vary widely even among experiments where the order of feeding is known, which has led some researchers to question the method. The author set out to find the sources of that variation.

The author studied the bones collected in 1975 from the surface of the Amboseli Airstrip Den, a natal and communal spotted hyena den used for many generations. At the time, a group of 15 spotted hyenas lived there, so digging was not allowed. The collection is dominated by wildebeest and zebra, followed by Grant’s and Thomson’s gazelles.

- The sample. All mammal limb bone pieces with little weathering (732 specimens), so that tooth marks were not hidden by cracking or flaking.

- What was recorded. For each piece: species, body size class, which bone, which part of the bone (end, shaft near the end, or middle shaft), maximum length, and whether and where tooth marks were present. Each piece was checked under strong light with a hand lens, following published guidelines for identifying tooth marks.

- What was compared. Tooth-mark frequencies against fragment length, prey body size, the different bones and their parts, and published bone density values.

### 2. What they found

Fragment size

- Longer pieces were more likely to be tooth-marked. The share of marked pieces rose from 14.3% for pieces up to 20 mm long to about 95% for pieces 80 to 100 mm long. Overall, 71.0% of pieces were marked.

Prey body size

- Middle shafts of larger (size 3) mammals were marked more often than those of smaller (size 2) mammals: 75.3% versus 63.4%, a significant difference.

- But the larger animals’ bone pieces were also longer (64.1 mm on average, against 45.1 mm). When only pieces of similar length (20 to 60 mm) were compared, the difference almost disappeared (59.0% versus 57.7%, not significant).

Different bones and parts of bones

- Tooth marks were not spread evenly across the different bones. There were more marked tibia shafts near the knee, humerus shafts and ends near the elbow, and radius middle shafts than expected, and fewer marked humerus middle shafts.

- Bone ends varied hugely, from 15.4% to 100% marked depending on the bone. Shafts near the ends varied from 45.8% to 75.0%.

- Middle shafts were much more consistent, from 71.6% to 82.4% across all bones.

Bone density

- For bone ends, the less dense the bone, the more often the surviving pieces were tooth-marked (a significant negative relationship).

- Shafts near the ends showed a weaker negative relationship that was not significant.

- Middle shafts showed no significant relationship with bone density.

### 3. What they concluded

- Tooth-mark frequencies depend on fragment size, on the mix of bones and bone parts in an assemblage, and, for ends and near-end shafts, on bone density. These come from carnivore behaviour and the physical properties of bone, so they are likely to apply to bone collections made by other bone-crunching carnivores, and to archaeological sites.

- Anything that breaks bones into smaller pieces (trampling, sediment pressure, weathering), not only carnivores, can lower the share of tooth-marked pieces. This makes it risky to compare simple experiments with archaeological sites that have been through many destructive processes.

- Because carnivores eat the least dense parts first, the pieces that survive from those parts carry more tooth marks. Processes that destroy low-density bone at a site can remove exactly the most heavily marked pieces.

- The author suggests three corrections: compare pieces of similar length; analyse each bone and bone part separately; and, given density effects, rely mainly on dense middle shafts, which show no link with density and are consistent across bones.

- These variables are not the full story, and more research is needed to find other sources of variation.

## Cite this summary

Lawrence, K. 2026. Summary of: Sources of Variation in Carnivore Tooth-Mark Frequencies in a Modern Spotted Hyena (Crocuta crocuta) Den Assemblage, Amboseli Park, Kenya (Faith, 2007). Original Wisdom, Tracking Literature. https://www.originalwisdom.com/tracking-literature/2007-faith-sources-variation-carnivore-tooth/