Direct dating reveals the early history of opium poppy in western Europe

Opium poppy (Papaver somniferum) can grow in diverse environments and is cultivated worldwide for medicinal, psychoactive, and alimentary uses. Despite its importance, its early history has been under-studied. Archaeobotanical remains (seeds and stigmatic discs, charred or waterlogged) combined with radiocarbon dating can clarify domestication and dispersal during the Neolithic. The wild relative P. somniferum subsp. setigerum occurs naturally in the central and western Mediterranean, but archaeological seeds of wild and cultivated forms cannot yet be distinguished morphologically, complicating interpretation. Previous scenarios about origin and spread relied on indirect dating (associated materials) and on the distribution of early finds across western Europe and the Near East. Near Eastern origins are weakly supported: wild opium poppy does not grow there today, early Near Eastern finds are scarce and possibly intrusive, and there is no evidence along the Neolithic dispersal route before 5300 cal BCE. Conversely, numerous Early Neolithic western European sites (5900–4700 cal BCE) have yielded opium poppy remains, especially in Mediterranean contexts (Impressa/Cardial/Epicardial) and later within the LBK cultural sphere in temperate Europe. However, small seeds are prone to post-depositional movement and many records may be intrusive if not directly dated. Recent advances in compact AMS and optimized pretreatments allow direct dating of minute samples like poppy seeds. This study directly tests: (1) the antiquity of La Marmotta’s capsules (~5500 cal BCE) as an early benchmark; (2) early integration into LBK economies in northwestern temperate Europe (~5300 cal BCE); and (3) timing of arrival in the western Alps (5000–4850 cal BCE), by directly dating opium poppy remains from 11 Neolithic sites dated 5900–3500 cal BCE.

Prior research catalogued early opium poppy finds across the Mediterranean and temperate Europe and discussed possible Near Eastern versus European domestication scenarios. Near Eastern evidence (PPN sites in Israel and Anatolia) is sparse and likely intrusive, and no continuous trail links Near Eastern Neolithic dispersal to western European opium poppy remains. In contrast, multiple Early Neolithic Mediterranean sites (Impressa/Cardial/Epicardial) and many LBK sites in temperate Europe reported seeds, with earliest secure contexts around 5900–5000 cal BCE in the Mediterranean and around 5300–5000 cal BCE in LBK II–V phases. Previous chronologies mostly relied on indirect dating (charcoal, wood, animal bones) rather than direct dating of poppy remains, risking stratigraphic or post-depositional biases. Methodological parallels from broomcorn millet chronologies showed that indirect attributions can be erroneous, underscoring the need for direct AMS dating of the plant remains themselves.

Study design: Eleven Neolithic sites across the central and western Mediterranean, northwestern temperate Europe, and the western Alps (5900–3500 cal BCE) were selected based on context quality and potential early opium poppy records. Cultural attributions included Cardial/Epicardial/Impressa, LBK II–V, BVSG, and Early/Middle Alpine Neolithic. Sampling: Archaeological sediments from pits or ground levels containing daily-life debris were sieved for charred and/or waterlogged opium poppy seeds and capsule fragments (microsamples). When possible, a co-occurring macrobotanical remain from the same sediment sample (macrosample; typically annual crop/weed seeds; one charcoal sample; and one poppy capsule fragment treated as macro) was also dated to test for intrusions. At Los Castillejos, the macrosample came from the same stratigraphic horizon but a different sediment sample. Radiocarbon dating: Due to the very small carbon content of poppy seeds, microsamples underwent gentle acid wash only, then offline combustion on a manual line for very small samples. Resulting CO2 was split when possible and introduced directly into the AMS using a gas interface system (GIS). Macrosamples received standard AAA pretreatment, combustion, and graphitization using an AGE 3 system, with precautions against memory effects by pre-combusting a similar-aged sample. All measurements were performed on the compact AMS ECHOMICADAS. Standards and blanks (Oxalic acid II NIST, phthalic anhydride) were used for normalization, fractionation, and background correction. For small samples, a constant contamination correction was applied (Mc=0.3 µgC, Rc=0.64 F14C). Data reduction for macrosamples used BATS software. Calibration used OxCal 4.4 with the IntCal20 atmospheric curve. Analytical considerations: Microsamples carry larger measurement errors (±50 to ±150 14C years), yielding broad 2σ calibrated ranges (200–400 cal years). Paired micro/macro dating from the same sediment was used to test for possible intrusions, with t-tests assessing agreement. Additionally, two previously published macrosample dates (La Draga, Isolino Virginia) were incorporated.

- Dataset: 22 total samples considered from 11 Neolithic sites; 12 direct AMS dates on opium poppy remains (seeds or capsules) and 10 macrosamples (including 8 newly processed and 2 previously published). - Chronology: Calibrated 2σ range for directly dated opium poppy spans 5622–4050 cal BCE. - Mediterranean origin and early use: La Marmotta (central Italy) capsules directly dated to 5617–5480/5622–5478 cal BCE, confirming mid-6th millennium presence within the species’ putative wild range. Le Taï (Epicardial, S France) and La Draga (Late Cardial, NE Spain) fall broadly between 5311–4717 cal BCE. - Northwestern temperate Europe (LBK sphere): Remicourt–Fond de Momalle (Belgium) microsample 5305–4850 cal BCE; paired charcoal macro 5329–5211 cal BCE (possible old-wood effect). These support presence from the earliest LBK occupation regionally (ca. 5300–5200 cal BCE) west of the Rhine. - Western Alps: La Gillière 2 (Switzerland) microsample 4999–4726 cal BCE, indicating arrival by 5000–4850 cal BCE; Isolino Virginia (Italy) microsample 4796–4057 cal BCE with published macro 4932–4606 cal BCE; Le Chenet des Pierres (France) microsample 4338–4050 cal BCE shows continued presence into the 5th millennium. - Intrusions and inconsistencies: - Buchères (BVSG, N France): Microsample 4837–4546 cal BCE slightly younger than macro 4988–4797 cal BCE; t-value just above 5% suggests potential minor intrusion or slightly later pit chronology, but still Neolithic. - Těšetice (Czech Republic): Microsample 5010–4549 cal BCE disagrees with macro 5318–5084 cal BCE (t-value >5%), suggesting intrusion at this multiperiod site; still within Neolithic range. - Los Castillejos (Spain): Macro (barley) fits Early Neolithic (5209–5005 cal BCE), but opium poppy microsample 3329–2704 cal BCE is much younger, indicating intrusion from later layers. - Smólsk 4 (Poland): Poppy seed dated modern (intrusive); macro weed seed 5306–5067 cal BCE fits Neolithic context. - Agreement cases (no significant difference between micro and macro): La Marmotta, Le Taï, La Draga, Remicourt–Fond de Momalle, Isolino Virginia, La Gillière 2. - Dispersal pattern: Evidence supports early cultivation/use in the Mediterranean by mid-6th millennium BCE, rapid dispersal west of the Rhine by 5300–5200 cal BCE, and later introduction into the western Alps (~5000–4800 cal BCE), becoming widespread from the later 5th millennium.

The study directly addresses the long-standing question of when and from where opium poppy was first used and how it dispersed across western Europe. Direct AMS dates on poppy remains validate a mid-6th millennium BCE presence within the Mediterranean, consistent with a European origin and early cultivation within the species’ natural range. The alignment of micro- and macrosamples at several key sites reduces the likelihood that these remains are intrusive. In northwestern temperate Europe, opium poppy appears integrated into LBK economies from their earliest regional phases (ca. 5300–5200 cal BCE) west of the Rhine, supporting dispersal via contacts between Cardial and LBK groups. Ecological plasticity of poppy likely facilitated rapid adoption across diverse environments, including loess soils. In the western Alps, the later appearance (5000–4800 cal BCE) suggests different regional rhythms linked to delayed Neolithization in mountainous areas and possibly secondary acquisition via Rhine corridor networks. The detection of intrusive seeds at certain sites (Los Castillejos, Těšetice, Smólsk, and possibly Buchères) underscores the necessity of direct dating for accurate cultural-chronological placement. Overall, the findings refine the chrono-geographical framework for poppy’s early history and illuminate inter-cultural contacts and exchange networks at the dawn of European farming.

This study establishes the first robust chrono-geographical framework for the early history of opium poppy in western Europe using direct AMS dating of minute botanical remains. Key contributions include: (1) confirmation of mid-6th millennium BCE use in central Italy (La Marmotta); (2) early dispersal into northwestern temperate Europe, with presence west of the Rhine by 5300–5200 cal BCE; and (3) later introduction to the western Alps around 5000–4800 cal BCE, with widespread occurrence from the later 5th millennium. The work demonstrates the critical value of directly dating small plant remains to avoid stratigraphic and post-depositional biases, revealing several intrusive instances. Future research should expand direct dating across additional regions (including adjacent European areas and the Near East) and integrate morphometric analyses of seeds to investigate domestication processes and to distinguish wild versus cultivated forms during the Early Neolithic.

- Morphological indistinguishability between wild and cultivated Papaver somniferum subspecies prevents definitive classification of archaeological seeds. - Small seed size yields relatively large radiocarbon errors (±50–150 14C years), producing wide calibrated ranges (often 200–400 years at 2σ). - Potential intrusions in stratified, multiperiod sites can mislead context-based chronologies; even paired dating may not fully resolve complex site formation processes. - One macrosample was charcoal (long-lived taxon) and may suffer from an old-wood effect. - Limited number of early, well-preserved waterlogged Mediterranean sites prior to 5200 cal BCE creates temporal gaps. - Geographic coverage, while broad, remains incomplete for eastern LBK regions and the Near East.