Revisiting the Jersey “Mammoth Drive”: What Really Happened at La Cotte de St Brelade?
For decades, a dramatic scene has been used to explain two famous heaps of mammoth bones found at La Cotte de St Brelade on Jersey (Channel Islands): Neanderthal hunters supposedly stampeded herds up rugged ground and over a cliff to their deaths, then butchered the fallen animals below. It’s a gripping story—but recent research shows it’s almost certainly wrong.
Below I revisit the evidence, explain why the classic “cliff-drive” doesn’t fit what we know about mammoths or the site’s Ice Age landscape, and summarize what specialists now think happened instead.
The Place: A Ravine on a Changing Ice Age Coast
La Cotte is not a flat “kill-field” above sheer precipices. Reanalysis of archives, sediments, and geomorphology shows the site was a deeply incised headland ravine system opening onto a broad, now-submerged plain—the Channel River valley. The topography would have funneled wind, sediment, carcasses, and people into confined spaces, preserving dense archaeological build-ups. In short: a complex ravine and rock-shelter landscape, not a convenient cliff-edge corral.
The Finds: Two “Bone Heaps,” Not Two Mass Deaths
The famous deposits comprise two stratified concentrations (“bone heaps”) dominated by mammoth, with woolly rhino also present. Detailed taphonomic study shows:
Cut marks and green-bone breakage (bones fractured while fresh) are common.
Carnivore damage is scarce, arguing against hyena or lion dens as the main accumulator.
Cranial parts are unusually abundant (over half the fragments), consistent with on-site processing for brains, sinuses, and rich facial tissues.
Stone tools and debitage cluster with the bones, indicating people worked around these piles.
Together, these patterns fit repeated human carcass processing and dumping at a favored spot, not animals falling en masse from above.
Why the “Cliff-Drive” Fails the Behavior & Ecology Test
Woolly mammoths were steppe-tundra grazers that thrived on expansive, relatively open grassland mosaics (“mammoth steppe”), not jagged cliff lines. Driving large proboscideans up steep, rocky ground toward a ravine brink is ecologically and biomechanically implausible.
Proboscideans don’t spook like gazelles
Modern elephants (our best behavioral analogs) are notoriously difficult to panic-herd. Adult females defending young often stand ground, form defensive circles, or charge, and mixed herds can actively repel threats. Experimental and observational work shows nuanced risk assessment rather than blind flight, even to predator cues. While any wild animal can stampede under extraordinary stress (e.g., wildfire), orchestrating a sustained drive of multi-ton animals toward a cliff is not a realistic, repeatable hunting method.
A Better Fit: A Persistent Place for Big-Game Processing
Archaeologists now interpret La Cotte as a “persistent place” repeatedly used by Neanderthals over long timescales. In such places, hunters likely:
Opportunistically acquired mammoths (and rhino) on the surrounding plain—through ambushes at bottlenecks, intercepts near water, or reworking natural deaths and winter-kills.
Field-butchered and transported select parts to the ravine’s sheltered working floors.
Broke bones for marrow and fat, cached or consumed resources, and dumped refuse in consistent spots—creating the distinct “heaps.”
This model explains the bone composition, cut marks, spatial patterning, and the repeated nature of the deposits without requiring impossible cliff drives.
The Landscape Story Matters
Earlier cliff-drive ideas pictured La Cotte perched above a “featureless coastal plain” where herds could be funneled at will. Updated reconstructions show a dynamic Ice Age coastline with ravines, ledges, and loess-covered slopes. The “bone heaps” look less like the bottom of a kill-fall and more like refuse nodes within a complex ravine workspace used again and again as sea levels and climates shifted.
Where Research Is Heading
La Cotte remains a globally important Neanderthal mega-site. Recent fieldwork and conservation efforts are stabilizing the cliff face, refining stratigraphy, and applying modern analyses (e.g., microstratigraphy, new dating, residue/isotope work). There’s also a new push to catalog and analyze older, pre-1960s collections with today’s methods—likely to sharpen the picture further.
Bottom Line
The mass “mammoth drive over a cliff” at La Cotte is not supported by current geoarchaeology and taphonomy.
The deposits make far more sense as the accumulated by-products of repeated Neanderthal butchery and carcass processing at a favored, sheltered ravine locale overlooking a productive grassland.
This reinterpretation aligns with mammoth ecology and proboscidean behavior, and it offers a more realistic window into Neanderthal subsistence: flexible, strategic, and intimately tied to persistent places on a changing Ice Age coast.
Sources and Further Reading
Scott, B., et al. (2014). A new view from La Cotte de St Brelade, Jersey. Antiquity. (Reassesses the site’s landscape and undermines the cliff-drive model.) Cambridge University Press & AssessmentResearch Explorerarchaeologydataservice.ac.uk
Shaw, A., et al. (2016). The archaeology of persistent places: the Palaeolithic case of La Cotte de St Brelade, Jersey. Antiquity. (Presents the “persistent place” framework; details cut marks, bone breakage, tool–bone spatial links.) Cambridge University Press & Assessment
Smith, G.M. (2015). A contextual reassessment of La Cotte de St Brelade. (Reports high cranial fragment proportions and other taphonomic signatures of intensive processing.) hesp.irmacs.sfu.ca
Natural History Museum / Jersey Heritage overviews of La Cotte (site background, MIS context, artifacts, and climate). Jersey Heritage
UCL Archaeology South-East news (2023–2025) on current stabilization, excavation seasons, and renewed analyses of legacy collections. University College London+1
Guardian and Smithsonian news summaries of the 2014 reinterpretation (useful for lay overviews). The GuardianSmithsonian Magazine
Mammoth ecology: Zimov/Zimova et al. on the mammoth steppe (open, high-productivity grasslands); Guthrie on diet and habitat. media.longnow.orgPleistocene & Permafrost Stiftungrhinoresourcecenter.com
Proboscidean behavior: studies on elephant risk assessment/antipredator responses (playback experiments; responses to lion calls; megaherbivore effects on ungulates). These support the idea that herds don’t reliably “flee to cliffs.” WKU TopScholarPMCBlumstein Lab
Learn more about Dreamer Books: An Ice Age Saga at dreamerliteraryproductions.com
