Honey-collecting in prehistoric West Africa from 3500 years ago

The study investigates whether and how prehistoric communities in West Africa exploited honeybee products. Honey is an energy-dense, easily digestible sweet resource and, along with bee brood, may have been important in hominin nutrition and tool use. While rock art indicates ancient honey collection and prior lipid analyses have identified beeswax in prehistoric pottery from Europe, the Near East, and North Africa, evidence from West Africa has been lacking. The Nok culture of central Nigeria (mid-2nd millennium BC to 1st millennium BC) presents an opportunity to examine early farming contexts where agriculturalists coexisted with foragers. The research question is whether organic residues in Nok pottery provide direct chemical evidence of bee product exploitation (e.g., beeswax, honey) and what this indicates about subsistence, technology, and the ecological presence of Apis mellifera in Holocene West Africa.

Prior work documents widespread prehistoric depiction of honey harvesting in global and especially African rock art. Lipid residue analyses have shown beeswax in Neolithic and later pottery from Europe, the Near East, and Mediterranean North Africa, indicating long-standing human use of honeybees from at least the 7th millennium BC. Ethnographic and historical sources across Africa highlight honey and bee brood as food, medicine, and ingredients for beverages, including honey-based alcoholic drinks; modern beekeeping and wild harvest are common across the continent. In West Africa, early pottery dates to the 10th millennium cal BC at Ounjougou (Mali), with later, uneven adoption of food production; the Nok culture is notable for terracotta figurines and early iron production, with a diet including millet and cowpea. Chemical profiles of modern beeswax and variation across Apis mellifera subspecies are well characterized; compositional differences among subspecies are minor. This background motivates testing for beeswax biomarkers in Nok pottery to fill a geographic and cultural gap in the archaeological record of honeybee exploitation.

• Sampling: 458 potsherds from 12 Nok culture sites in central Nigeria spanning Early, Middle, and Late Nok periods were analyzed. Vessels were typically everted rim pots with body diameters ~20–30 cm.
• Extraction protocols: Two approaches were used.
  1. Direct acidified methanol extraction of ~2 g cleaned potsherd with internal standard (n-tetratriacontane, 20 µg). Esterification/transesterification with 2–4% H2SO4/MeOH at 70 °C for 1 h; centrifugation; sequential n-hexane extractions; derivatization with BSTFA + 1% TMCS (70 °C, 1 h). Analyses by GC-FID and GC-MS.
  2. Solvent extraction on selected sherds: Ultrasonication with chloroform/methanol (2:1 v/v), 30 min, 2 × 10 mL; silica filtration; trimethylsilylation; analysis by high-temperature GC (HTGC) and HTGC-MS.
• Instrumentation: Agilent 7820A GC-FID with DB1-HT column; helium carrier, temperature program 50–350 °C. HTGC-MS on ThermoScientific Trace 1300 coupled to ISQ MS; EI 70 eV; oven to 380 °C; full scan m/z 50–950 at 2 scans/s. Blanks run with each batch. Identification via mass spectra and retention times, NIST library, and comparison to modern beeswax.
• Biomarker criteria for beeswax: Presence of characteristic series of even-numbered n-alkanoic acids (C20–C32), n-alkanols (C22–C34), and n-alkanes (C23–C35), along with higher molecular weight alkyl palmitate monoesters (C38–C52) and hydroxy palmitate monoesters (C40–C54). Consideration of diagenetic formation of free n-alkanols from hydrolysis of esters and alteration of n-alkane profiles over time.
• Data handling: Lipid concentrations quantified using internal standard; classification of lipid profiles into degraded animal fat, plant-derived, and beeswax-associated profiles; further solvent extraction performed on higher-yield beeswax candidates to confirm intact wax esters and hydroxy wax esters.

• Lipid recovery: 66 of 458 sherds (14.4%) yielded interpretable lipid profiles, with concentrations from 7 to 1815.6 µg g−1 (one non-beeswax animal fat outlier at 13.2 mg g−1).
• Profile categories among lipid-yielding sherds: 14 showed degraded animal fat (dominant C16:0 and C18:0), 27 had complex plant-type distributions, and 25 exhibited beeswax-associated profiles.
• Beeswax evidence: 25 sherds displayed characteristic even-numbered n-alkanoic acids (C20–C32), n-alkanols (C22–C34), and n-alkanes (C23–C35). Of these, 19 were selected for solvent extraction; five (NOK003, NOK081, NOK119, NOK120, NOK167) unambiguously contained wax esters (C40–C52, maximizing at C46) and hydroxy wax esters (notably C48, C50), confirming beeswax residues. One additional vessel (NOK376) showed wax esters in very low abundance. Remaining beeswax-like profiles likely had concentrations too low to preserve intact esters.
• Concentrations: Among beeswax-associated sherds, 15 had <100 µg g−1 and 10 had >100 µg g−1 lipid concentrations. The highest beeswax-bearing concentration was NOK119 at 1815.6 µg g−1 (~1.8 mg g−1). Other high concentrations included NOK167 (~0.9 mg g−1) and NOK376 (~0.6 mg g−1).
• Mixing with animal products: Some beeswax-confirmed vessels (e.g., NOK119, NOK167) also contained abundant C16:0 and C18:0 fatty acids, indicating possible co-processing of bee products with animal fats.
• Temporal distribution: Beeswax biomarkers occur in Early Nok, Early Middle Nok, and Later Middle Nok periods; none detected in Common Era sherds (only 8% of CE sherds yielded any lipids). Overall, beeswax was present in 38% of lipid-yielding sherds; it was most frequent in Early Nok pottery (55% of lipid-yielding sherds), and 38% in Early Middle and Later Middle Nok.
• Ecological implication: Findings provide biomolecular evidence for exploitation of Apis mellifera in Holocene West Africa, consistent with A. mellifera adansonii distribution, though subspecies cannot be distinguished chemically.

The detection of beeswax lipids in over one-third of lipid-yielding Nok vessels provides direct chemical evidence for bee product exploitation, most plausibly linked to honey collecting and processing, within early West African farming contexts ca. 1500–500 BC. The characteristic beeswax biomarker suite, including intact wax and hydroxy wax esters in several sherds, indicates melting or gentle heating of combs and/or processing or storage of honey. High lipid concentrations and co-occurrence of animal fat markers in some vessels suggest mixing bee products with animal fats, potentially for culinary use or preservation of meat in honey (paralleling ethnographic practices). The absence of beeswax in Common Era sherds may reflect different vessel uses and/or poorer lipid preservation. The results align with ethnographic accounts across Africa of intensive honey use (food, brood consumption, medicinal applications, and beverages) and with archaeological evidence elsewhere for technological uses of beeswax. While pots may sometimes be used as hives in Africa, the small size of Nok vessels argues against hive use in this case. The study extends the geographic scope of archaeochemical evidence for honeybee exploitation and supports the significance of honey and bee products in Nok subsistence and technology, while also contributing to understanding the palaeoecology of A. mellifera in West Africa.

This study provides the first direct chemical evidence for honeybee product exploitation in prehistoric West Africa by identifying beeswax residues in Nok culture pottery dating to around 3500 years ago. Beeswax biomarkers, including intact wax and hydroxy wax esters, were present in numerous vessels, indicating processing and/or storage of honey and related hive products and, in some cases, mixing with animal fats. The findings underscore the importance of honey collecting within early farming communities and expand the known temporal and spatial record of human-honeybee interactions. Future research should apply lipid residue analysis to earlier West African pottery, including contexts as early as the 10th millennium cal BC, to test for even older bee exploitation and to explore the nature of potential honey-based beverages, acknowledging the analytical challenges of detecting ancient fermentation.

• Low lipid recovery (14.4% of sherds) limits sample representativeness.
• Many beeswax-like residues were at low concentrations, hindering detection of intact wax esters; diagenesis may alter biomarker distributions.
• Direct identification of honey (sugars) is not possible with these methods; beeswax is used as a proxy for honey processing.
• Potential contamination issues (e.g., plasticizers in some extracts) constrained further analyses for specific samples.
• No beeswax found in Common Era sherds, which may reflect use differences or poorer preservation rather than absence of bee exploitation.
• Chemical methods cannot distinguish Apis mellifera subspecies.
• Identification of ancient fermentation (e.g., honey-based alcoholic beverages) remains analytically difficult and was not demonstrated conclusively.