Late Pleistocene South American megafaunal extinctions associated with rise of Fishtail points and human population

Late Pleistocene megafaunal extinctions occurred across most continents, with multiple proposed drivers including human hunting, climate change, hyperdisease, habitat modification, and even extraterrestrial impacts. In the Americas, most extinctions followed the Last Glacial Maximum and coincided with the first widespread dispersal of humans, making causal attribution contentious. Recent work indicates the extinction dynamics differed between North and South America: while North America’s losses coincided with the Younger Dryas and the spread of Clovis technology, South America experienced more severe losses (over 40 genera) and a later spread of fluted Fishtail projectile points (FPP). Despite the severity of South American losses, archaeological narratives have often downplayed direct human involvement due to limited clear evidence of megafaunal exploitation and assumptions of earlier ‘pre-Clovis’ human presence. This study asks whether humans were principal drivers of South American extinctions by testing two expectations: (1) an inverse temporal association between human-related signals (FPP, archaeological site density) and megafaunal density, and (2) a positive spatial association (overlap) between FPP distributions and megafaunal ranges. The study integrates temporal density reconstructions from radiocarbon data with species distribution models to assess spatiotemporal relationships among megafauna, FPP, and human populations across South America and within key regions (Pampa, Southern Patagonia, Andes).

Debate over Pleistocene extinctions centers on competing hypotheses: human overkill (Martin), climate-driven changes, hyperdisease, habitat modification, and impact events. In North America, the timing of extinctions around 13–12 ka aligns with the Younger Dryas and the expansion of Clovis culture and its big-game hunting technology, prompting mixed models where both climate and humans played roles. In South America, extinctions were more severe and accompanied by the spread of FPP (slightly later than Clovis) and the Antarctic Cold Reversal, yet direct archaeological evidence of megafaunal hunting is scarcer and limited to a few species. Some archaeologists argue for early ‘pre-Clovis’ colonization and a non-central role of FPP in subsistence; however, paleoecological and paleontological studies increasingly implicate humans as major contributors. The apparent synchrony between the spread of FPP, a surge and subsequent slowdown in human population growth, and the decline of megafauna motivates an integrated temporal-spatial analysis to reassess human involvement beyond strict climate-vs-human dichotomies.

Data compilation: The authors assembled radiocarbon datasets for extinct large mammals (megafauna) and for archaeological sites across South America, combining published compilations and newly reported dates. They also compiled Fishtail projectile point (FPP) occurrences and megafauna occurrences lacking absolute dates but constrained to the terminal Pleistocene, with paleogeographic coordinates from databases such as the Paleobiology Database. Temporal analysis: They applied Summed Calibrated Probability Distributions (SCPD) of radiocarbon dates to estimate temporal changes in density for megafauna, FPP, and archaeological sites across 18–9 ka calibrated BP. They compared regional SCPDs (Pampa, Southern Patagonia, Andes) and used permutation tests against null models to assess significance of temporal peaks and troughs. Spatial analysis: For potential distributions (18–9 ka cal BP for species; 13–10 ka cal BP for FPP), they built Species Distribution Models (SDMs) using MaxEnt, with late Pleistocene/early Holocene bioclimatic layers from PaleoClim. Model performance was evaluated via AUC. They produced stacked suitability maps to estimate local species richness of extinct megafauna. Spatial overlap and niche similarity were quantified using Warren’s I similarity index, and relationships among distributions (FPP, archaeological sites, species) were visualized with non-metric multidimensional scaling (nm-MDS) and a minimum spanning tree of similarity matrices. Analytical focus: Analyses were conducted for South America as a whole and for three key regions with strong evidence of human–megafauna coexistence: Pampa (Argentina, Uruguay, S Brazil), Southern Patagonia, and Andes. The study compared temporal density patterns among datasets, assessed spatial concordance of high-suitability areas, and related these to known climatic phases (e.g., Antarctic Cold Reversal).

- Temporal dynamics (South America): Megafaunal SCPD is very low at ~18 ka BP, increases from ~17.5 ka, and grows rapidly between ~15.3 and 12.9 ka BP, followed by a dramatic decline from ~12.9 to 11.6 ka BP. After ~11.6 ka BP, only a few megafaunal genera persist, and many early Holocene dates have been questioned. - FPP timing: FPP appear around ~13 ka BP, rise rapidly, and peak between ~12.4–12.2 ka BP, then decline, with the technology virtually disappearing by ~10.9 ka BP. - Human archaeological signal: Archaeological site SCPD is low from ~15 to 13 ka, then increases with a peak around ~12.5 ka BP, followed by a slight decline to ~11.6 ka BP and a subsequent rise. - Temporal associations: The megafaunal decline begins just after FPP appear (~13 ka), with a tight fit between explosive FPP growth and a steep megafaunal decline over the next ~600 years. Regional SCPDs (Pampa, Patagonia, Andes) show broadly similar megafaunal density peaks (ca. 13.5–12.5 ka) and nearly identical FPP density patterns. - Spatial patterns: SDMs show high potential distributions for many megafaunal species and FPP in open environments of the Pampas and Southern Patagonia; the Andes appear marginal for both. Areas of highest potential FPP distribution coincide with highest local megafaunal species richness and higher densities of human occupations. - Model performance: SDMs for megafauna, FPP, and archaeological sites performed well (AUC ~0.91–0.99). - Niche overlap: FPP distributions show high similarity with megafauna adapted to Pampas/Patagonia (e.g., Megatherium americanum, Glossotherium robustum, Equus neogeus), and lower similarity with gomphotheres in northern South America (e.g., Cuvieronius hyodon, Notiomastodon waringi), suggesting the latter were not principal prey of FPP hunters. - Demography-technology linkage: Rapid expansion of FPP parallels a rapid increase in human population density; population growth slows sharply around ~12.5 ka, shortly after the megafaunal decline begins and just before FPP frequency peaks and decreases.

The temporal coupling of the rise of FPP and human population growth with the abrupt decline of megafauna suggests a strong human predation signal, consistent with a big-game hunting strategy. Spatially, the significant overlap of FPP potential distributions with high megafaunal species richness in open steppe environments (Pampas and Patagonia) reinforces the interpretation that FPP were designed for hunting large mammals. The weaker overlap with northern gomphotheres indicates prey selectivity rather than indiscriminate hunting. The archaeological signal’s early growth (post-13 ka) and synchronization with FPP proliferation supports either rapid adoption of a highly effective hunting technology or that FPP users were among the earliest widespread colonizers of South America. Climatic and environmental changes likely contributed to vulnerability (e.g., low reproductive rates of large-bodied species), but the timing of megafaunal decline does not align as well with major climatic shifts as it does with the expansion of FPP and human demography. Even moderate targeted predation could have triggered trophic cascades and network destabilization, contributing to widespread extinctions, while leaving limited direct archaeological evidence due to taphonomic and sampling biases and carcass transport behaviors. The persistence of guanaco may relate to broader geographic distribution and habitat availability, mitigating extinction risk despite human hunting. Overall, the findings support humans as principal drivers, with climate as a facilitating secondary factor.

The study integrates radiocarbon-based temporal reconstructions with climate-informed species distribution models to demonstrate a strong spatiotemporal association among Fishtail projectile points, human demographic expansion, and the decline of South American megafauna. Results indicate that direct human predation was the main driver of megafaunal decline, with late Pleistocene environmental changes and indirect anthropogenic effects creating conditions that enabled a community-wide collapse. Future research should resolve species-specific trajectories and interactions, refine regional chronologies, quantify the relative contributions of direct hunting versus indirect ecosystem impacts, and improve the integration of archaeological, paleontological, and paleoclimatic datasets.

- Direct archaeological evidence of megafaunal hunting in South America is sparse and uneven, likely due to taphonomic loss, sampling biases, carcass processing and transport behaviors, and low archaeological visibility of kill sites. - SCPD-based temporal proxies can be influenced by preservation, sampling intensity, and dating practices; early Holocene megafaunal dates remain debated. - Spatial overlap and niche similarity analyses provide indirect evidence of trophic interaction; they cannot by themselves confirm prey selection or hunting intensity. - Some taxonomic and occurrence records (e.g., gomphothere species distinctions) are contested, potentially affecting distribution models. - Regional heterogeneity (e.g., the Andes as a marginal zone) suggests varying human–megafauna dynamics that may not be fully captured by continent-wide models. - Assumptions about the timing and nature of earliest human presence (pre-Clovis vs later arrivals) introduce uncertainty in interpreting initial human–megafauna interactions.