Multi-isotopic and morphometric evidence for the migration of farmers leading up to the Inka conquest of the southern Andes

The study investigates whether a significant migration of farming populations occurred into the southern Andes prior to the Inka conquest, and how such movements relate to regional economic, demographic, and social change. The Uspallata Valley (Mendoza, Argentina) marks the southern frontier of Andean agropastoral economies where agriculture decreased southwards between 32°–34° S. Over the last ~2000 years this area shows evidence for mixed subsistence involving maize, quinoa, beans, squash, camelid herding, and foraging. Prior isotopic research suggested variable maize intake linked to climate, with elevated δ13C during AD 1250–1370 associated with warmer summers. The Inka established Uspallata as their southernmost administrative center east of the Andes after ~AD 1400. The valley’s rich mortuary record (AD 800–1500) provides an opportunity to reconstruct life histories, residence, diet, and biological affinities. The study integrates strontium isotope-based mobility, stable isotope paleodiet, craniofacial geometric morphometrics, and radiocarbon-based temporal modeling to test for migrant influx and its timing relative to regional dynamics and the Inka expansion.

Archaeological perspectives on migration have shifted from early twentieth-century diffusionist models to skepticism, and more recently to renewed interest fueled by isotopic and paleogenomic methods. Radiogenic and stable isotope analyses can resolve mobility and dietary sources, while ancient DNA elucidates population history. For the southern Andean periphery, migration has often been underemphasized in explaining economic and demographic change, though exceptions exist. Regional work in central-western Argentina and central Chile documents mixed economies, variable maize consumption, and climatic influences on agriculture, with isotopic signals peaking prior to the Little Ice Age. Research on bioavailable strontium and isoscapes across the Andes demonstrates strong geological controls on 87Sr/86Sr and their utility for provenance and mobility studies. Studies of artificial cranial modification in the Andes interpret vault shapes as potential ethnic markers. Demographic proxies derived from radiocarbon date distributions indicate regional occupational peaks during the Late Intermediate Period, a time of intensified regional identities, economic specialization, and conflict elsewhere along the Andes.

Study area and samples: The Uspallata Valley (northwestern Mendoza, Argentina) lies between the Frontal Cordillera and Precordillera and forms part of a ~400 km biogeographic corridor of intermontane valleys. Human remains (~MNI 196) from seven sites spanning AD 800–1500 were examined. Strontium baseline: 65 modern and archaeological rodent samples with restricted home ranges were analyzed along a ~250 km transect from the Chilean Pacific coast to eastern Andean lowlands to construct a bioavailable 87Sr/86Sr isoscape via Ordinary Kriging. Geological units include Pacific Coast, Coastal Cordillera, Principal Cordillera (east/west), Frontal Cordillera, Precordillera, and foreland. Uspallata is influenced by radiogenic Precordillera Paleozoic formations. Human strontium: 38 human samples (12 teeth, 26 bones) from 30 individuals (15% of MNI) across AD 800–1500 were analyzed at University of Cape Town by MC-ICP-MS. Data referenced to SRM987 = 0.710255; in-house carbonate standard NM95 monitored accuracy and precision. GIS-based focal statistics and reclassification assigned likely residence areas using each group’s mean ±2 SD. Diagenesis checks included Ca/P and U/Ca elemental screening indicating minimal alteration; dry climate context and coherent multi-isotope/morphometric patterns further support signal integrity. Chronology: Radiocarbon dates were calibrated with SHCal20 and modeled in OxCal 4.4. Two overlapping phases (locals vs migrants) were modeled; kernel density estimates (KDE) and a regional estimate for initial Inka presence (~AD 1410) were used for temporal comparisons. Paleodiet: Stable isotopes were measured on bone collagen (δ13Ccoll, δ15N) and bone apatite (δ13Capat). Collagen was prepared with a pseudomorph method; C/N, %C, %N assessed preservation (most samples within 3.0–3.4 C/N). Bone apatite powders (106–179 μm) were reacted with 100% H3PO4 and analyzed via IRMS; δ13C relative to PDB with precision <0.2‰. Dataset combined 22 new individuals (6 teeth, 19 bone) with prior results to total 39 bone samples: 20 from sites with migrants (Potrero Las Colonias, Túmulo III, Usina Sur 2, Monte de Algarrobos) and 19 from sites with only locals (Túmulo I, Túmulo II, Barrio Ramos I). Statistical comparisons used Mann–Whitney tests; isotopic niche width estimated with SIBER (SEA and Bayesian SEAB). Geometric morphometrics: Twenty-one crania were digitized (56 3D landmarks) with Microscribe G2X; intra-observer error r≈0.90. Generalized Procrustes superimposition produced shape variables; sexual dimorphism was controlled by regression of shape on size (residuals used). PCAs were run for whole skull, cranial base, and vault modules; Mahalanobis distances computed; Procrustes ANOVA tested group differences (locals vs non-locals). Contextual archaeology: Mortuary contexts, associated grave goods, and site types (primary burials vs ossuaries) were compiled, with MNI and age/sex structure. Regional KDE of 343 radiocarbon dates from Mendoza provided macro-regional occupational trends.

- Bioavailable strontium isoscape: Rodent 87Sr/86Sr values closely follow geology with minimal overlap among key regions; Uspallata shows radiogenic influence from Paleozoic Precordillera. - Human strontium bimodality: 38 human samples group into two non-overlapping modes (Hartigans’ dip test D=0.10455, p=0.001; Mann–Whitney z=4.7637, p<0.0001): • Locals (n=27; 11 teeth, 16 bones): mean 0.7090 ± 0.0003; range 0.7083–0.7095; overlaps Uspallata/Mendoza Valley baseline (0.7087–0.7096). • Non-locals/migrants (n=11; 1 tooth, 10 bones): mean 0.7073 ± 0.0001; range 0.7072–0.7075; does not overlap local baseline. - Spatial assignment: Local group consistent with residence in/near Uspallata; marine coastal alternative rejected by diet (no marine signal). Migrant values consistent with lowlands SE or highlands N of Uspallata; equifinality leaves broader origins possible. - Temporal pattern: Locals present continuously AD 800–1500. Migrants cluster between ~AD 1280–1420 (modeled medians), largely contemporaneous with the regional occupational peak and just prior to/overlapping initial Inka presence (~AD 1410). - Site patterns: Potrero Las Colonias (ossuary, MNI=119) – all 7 analyzed individuals are non-local; Túmulo III (ossuary, MNI=27) – 1/3 non-local; Túmulo I & II (primary cemeteries) and Barrio Ramos I (Inka-period context) – locals only among sampled individuals. Migrant ossuaries lack grave goods; local burials include items (e.g., lip plug, projectile points, ceramics from Central Chile traditions; Inka-period shell beads and bone projectile points) consistent with local residence. - Craniofacial morphology: PCA and Procrustes ANOVA show significant differences primarily in cranial base shape between locals (Túmulo II) and migrants (Potrero Las Colonias) (F=1.83; p≤0.01), indicating distinct lineages. Whole-skull differences weaker (F=1.18; p=0.06). Cranial vault analysis indicates culturally mediated modifications differ between groups (F=1.90; p≤0.01), with migrants showing a wider range of frontal/occipital flattening patterns. - Paleodiet: Sites with migrants exhibit significantly more enriched δ13C in collagen and apatite (Mann–Whitney p<0.05), indicating high C4 intake; δ15N similar between groups, suggesting comparable trophic level/protein sources (e.g., wild herbivores). Isotopic niche width is smaller for migrants: SEA (corrected) 3.881 vs locals 7.377; Bayesian SEAB migrants 3.224 vs locals 9.913 (p=1 indicating migrants’ ellipse significantly smaller). Migrant δ13Capat values approach highly maize-reliant agricultural populations elsewhere in the Andes. - Macro-regional linkage: The migrant influx coincides with a regional KDE peak of radiocarbon dates in Mendoza, suggesting demographic intensification synchronous with migration.

The integrated isotopic and morphometric evidence demonstrates a distinct influx of non-local individuals into the Uspallata Valley between ~AD 1280–1420, immediately preceding and overlapping the initial Inka presence. Migrants are isotopically homogeneous with low-variance 87Sr/86Sr distinct from the local baseline and present highly enriched δ13C indicative of maize-focused agriculture, in contrast to locals who maintained broader, mixed C3/C4 diets including camelids. Cranial base shape differences support separate ancestral lineages, while vault modification patterns point to differing cultural identities. Mortuary evidence shows contemporaneous burial of locals and migrants, sometimes side by side, though migrant ossuaries lack diagnostic grave goods. The timing aligns with a regional occupational peak and broader Andean Late Intermediate Period trends of settlement reorganization, regional identities, and economic intensification, suggesting push-pull dynamics and potential demographic pressures in source regions. These findings nuance existing models that attribute maize intensification solely to climate-driven in situ adaptation by highlighting a substantial contribution from incoming maize-farming groups. The resulting multiethnic social landscape would have shaped and possibly facilitated Inka strategies of engagement with local populations at the imperial frontier.

This study provides interdisciplinary evidence—strontium isoscapes, human 87Sr/86Sr, stable isotopes, and craniofacial morphometrics—that a migration pulse of maize-farming groups entered the Uspallata Valley ca. AD 1280–1420, just prior to the Inka expansion. Migrants are distinguishable from locals by geographic origin, cranial base morphology, cultural vault modifications, and a narrow, maize-centric dietary niche. The migration coincided with regional demographic intensification and likely contributed to a multiethnic setting leveraged during Inka consolidation. The work underscores the need to integrate migration into models of economic and demographic transitions traditionally framed by climate-driven local adaptation. Future research should expand sample sizes, refine geographic source assignments (e.g., along the intermontane corridor to the north), incorporate paleogenomics to resolve lineages and admixture, and pursue micro-scale life history reconstructions (including trauma and disease) to elucidate causes of death and interaction dynamics. Enhanced modeling integrating climate, demography, and mobility will advance comparative frameworks for transitions to productive economies in the southern Andes.

- Geographic origin ambiguity due to isotopic equifinality: the migrant 87Sr/86Sr range matches multiple regions (e.g., SE lowlands or northern highlands), limiting precise source attribution. - Sample size constraints: Only 38 human Sr samples (30 individuals) and 39 bone isotopic diet samples; morphometric analysis limited to 21 crania, potentially affecting statistical power and representativeness. - Diagenesis risk: Although elemental screening (Ca/P, U/Ca) and contextual factors indicate minimal alteration, bone apatite and bone Sr remain more susceptible than enamel to diagenetic overprint. - Uneven tissue representation: Teeth (childhood signals) are underrepresented relative to bone (adulthood), limiting life-course mobility resolution. - Mortuary context gaps: Migrant ossuaries lack associated grave goods, restricting cultural attribution; absence of clear evidence for violence or disease requires further bioarchaeological investigation. - Demographic proxies: KDE of radiocarbon dates is an indirect indicator and not a direct measure of population size, necessitating cautious interpretation.