Reconstruction
La Ferrassie, in the Dordogne of southwestern France, preserves repeated occupations in a partly collapsed karstic cave complex. Excavations beginning in the early twentieth century uncovered remarkably complete adult Neanderthal skeletons and the remains of several children, making the site central to research on Neanderthal anatomy and treatment of the dead. Later excavations demonstrated that La Ferrassie is not one simple shelter deposit: its western, northern, and eastern sectors accumulated differently, and artifacts or bones sometimes moved downslope or were reworked.
Food evidence comes principally from animal bones rather than preserved meals. Middle Paleolithic layers in the recently excavated western sector are dominated by large bovids—bison, aurochs, or animals that cannot be distinguished more closely from fragmentary remains. Red deer are also abundant, while roe deer and wild boar occur in a layer associated with relatively wooded conditions. Reindeer appear in smaller numbers in these Middle Paleolithic deposits and become dominant in later, colder Upper Paleolithic layers. These species lists describe faunal assemblages, not an exact menu: carnivores and geological processes also contributed to or altered some accumulations.
Human food processing is nevertheless directly supported. Animal bones from several layers bear cut and scraping marks, impact damage associated with marrow extraction, and anvil marks. Some bones were also used as stone-tool retouchers. These traces show that Neanderthals butchered carcasses, removed flesh, and broke fresh bones to obtain calorie-rich marrow. They worked with Levallois flakes and scrapers, but the surviving evidence does not reveal portion sizes, meal combinations, seasoning, or whether particular species were preferentially roasted. Concentrations of burned and calcined bone demonstrate fire-related activity, yet intact hearth sequences are poorly preserved, and burning cannot automatically be equated with cooking. No securely identified plant meal has been reported from La Ferrassie, so plants should remain an acknowledged gap rather than be reconstructed as specific foods.
The skeletons add a different dimension to the occupation record. A multidisciplinary study concluded that intentional burial is the most parsimonious explanation for the approximately two-year-old LF8 child, although later geoarchaeological work disputed particular observations about sediment inclination and showed that some supposed burial mounds were natural freeze-related features. New radiocarbon work also places LF1 and LF8 much later than older estimates, near the Châtelperronian period.
La Ferrassie's significance therefore lies in evidence for recurring shelter use, carcass processing, marrow extraction, technological activity, and possibly deliberate burial—not in a recoverable cuisine. There is no demonstrated culinary lineage from these meals to a modern dish. Their defensible modern connection is methodological: microscopic damage, proteins, ancient DNA, isotopes, and radiocarbon dating now let archaeologists reconstruct parts of food behavior that early excavators could not see.
Historical context
Europe was passing through the climatically unstable Late Pleistocene, with relatively mild intervals alternating with severe cold phases over centuries or millennia. Neanderthals lived in small, mobile hunter-gatherer populations without agriculture, domesticated livestock, pottery, or permanent towns. La Ferrassie lay within a dense archaeological landscape of limestone caves and shelters around the Vézère and Dordogne valleys. During the later occupations, approximately 45,000–40,000 calibrated years before present, Neanderthals, Châtelperronian industries, and the arrival and spread of Homo sapiens formed part of a rapidly changing western European population history.
Evidence
Archaeological
StrongExcavations have documented stratified stone-tool assemblages, animal remains, combustion-related material, and multiple Neanderthal skeletons. Modern work shows that preservation and depositional history vary substantially among site sectors.
Zooarchaeological
StrongFaunal analysis identifies large bovids, red deer, roe deer, wild boar, reindeer, and other animals. Cut marks, scraping, fresh-bone breakage, anvil damage, and marrow-extraction impacts directly support human processing, although not every animal bone represents consumed food.
Residue chemical
ModerateMicro-FTIR analysis confirms thermally altered and calcined bone, while palaeoproteomics has strengthened taxonomic identification of bones associated with LF1. These analyses do not identify a recipe or prove that all burning resulted from cooking.
Food identification
ProbableLarge bovids and cervids were demonstrably processed by humans and are probable foods. Species presence alone is not dietary proof, and direct botanical evidence for meals is presently insufficient.
Dating
ModerateOSL and radiocarbon programs provide a coherent broad sequence, but early excavation, separated stratigraphic sectors, reworking, contamination, and intrusive deposition complicate individual skeleton dates. Associated fauna place LF1 at approximately 42.6–39.8 ka cal BP, while a fragment attributed to LF8 dates to approximately 41.7–40.8 ka cal BP.
Preparation method
ModerateButchery, flesh removal, and marrow extraction are directly supported by bone modifications. Cooking remains uncertain because poorly preserved combustion evidence cannot securely be connected to particular foods.
Geographic attribution
StrongThe site is securely located at La Ferrassie near Savignac-de-Miremont in Dordogne, southwestern France.
Historical interpretation
InterpretiveIntentional burial is supported as the most parsimonious explanation for LF8 by one multidisciplinary analysis, but details remain debated. Geoarchaeological research identifies some previously claimed burial mounds as natural periglacial features and disputes aspects of the published sediment reconstruction.
Sources
- 1.Antoine Balzeau, Dorothea Mylopotamitaki, Frido Welker, et al. (2026). Contextualising the La Ferrassie 1 Homo neanderthalensis skeleton through palaeoproteomic analysis and radiocarbon dating. Journal of Archaeological Science 188, 106515. doi:10.1016/j.jas.2026.106515Scientific literature
- 2.Vera Aldeias, Dennis Sandgathe, Shannon J. P. McPherron, et al. (2023). Site Formation Histories and Context of Human Occupations at the Paleolithic Site of La Ferrassie (Dordogne, France). Journal of Paleolithic Archaeology 6, article 30. doi:10.1007/s41982-023-00159-7Scientific literature
- 3.Sarah C. Pederzani, Vera Aldeias, Harold L. Dibble, et al. (2021). Reconstructing Late Pleistocene paleoclimate at the scale of human behavior: an example from the Neandertal occupation of La Ferrassie (France). Scientific Reports 11, article 1419. doi:10.1038/s41598-020-80777-1Scientific literature
- 4.Antoine Balzeau et al. (2020). Pluridisciplinary evidence for burial for the La Ferrassie 8 Neandertal child. Scientific Reports 10, article 21230. doi:10.1038/s41598-020-77611-zScientific literature
- 5.Sahra Talamo, Vera Aldeias, Paul Goldberg, et al. (2020). The new 14C chronology for the Palaeolithic site of La Ferrassie, France: the disappearance of Neanderthals and the arrival of Homo sapiens in France. Journal of Quaternary Science 35(7), 961–973. doi:10.1002/jqs.3236Scientific literature
Limitations
- majordate start/date end
The supplied range of approximately 70,000–45,000 BCE reflects older estimates and excludes recent dates for two prominent individuals. Fauna closely associated with LF1 model to approximately 42.6–39.8 ka cal BP, and a Neanderthal fragment attributed to LF8 dates to approximately 41.7–40.8 ka cal BP. These are roughly 40,650–37,850 BCE and 39,750–38,850 BCE, respectively, so the card's 45,000 BCE end date is materially too early for important skeletons represented by its plural title.
