2000 Year-old Bogong moth (*Agrotis infusa*) Aboriginal food remains, Australia

The study investigates whether Bogong moths (Agrotis infusa), historically documented as a seasonal Aboriginal food in the Australian Alps, can be identified archaeologically, thereby illuminating the deep-time history of insect harvesting in Australia. Ethnographic and historical accounts describe mass harvesting and communal consumption of insects worldwide, including Aboriginal Australian use of various insects such as cossid moth larvae and, notably, Bogong moths in the Eastern Uplands. Despite rich ethnohistoric descriptions of Bogong moth feasts, conclusive archaeological evidence of their processing or consumption has been lacking. The research aims to bridge this gap by applying use-wear and biochemical residue analyses to a grindstone from Cloggs Cave to detect microscopic remains indicative of Bogong moth processing and associated cooking activities, and to securely date these practices within the site's stratigraphy.

Prior research and ethnographic reports document widespread entomophagy across diverse cultures, including mass-harvesting strategies (e.g., Shoshone and Paiute practices) and the nutritional importance of insects. In Australia, Aboriginal groups are known to have consumed various insects, notably 'witchetty grubs' (cossid moth larvae). Historical accounts from the 1830s–1850s describe large Aboriginal gatherings to harvest and consume Bogong moths in the Australian Alps, including cooking methods, stirring that detached wings and legs, and pounding moths into cakes that were smoked for preservation. Archaeologically, insect foods have been understudied, and previous putative indicators (e.g., smooth pebbles fluorescing under UV) are equivocal due to non-specific fluorescence of many organic and mineral substances. Methodological advances in residue analysis emphasize that morphological diagnostics can be altered by grinding, heat, and environmental factors, underscoring the need for biochemical staining (e.g., Picrosirius Red) combined with high-powered microscopy and use-wear studies to differentiate plant and animal residues and identify collagen and lignin.

Site and stratigraphy: Cloggs Cave, in GunaiKurnai Country (southern foothills of the Australian Alps), was re-excavated in 2019–2020 to refine stratigraphy and occupation chronology. Squares P34 and P35 (50×50 cm) were excavated adjacent to the 1971–1972 pit to 2.28 m depth. The sequence includes SU3A–3G (subsidence crater infill) overlain by 73 finely stratified ash-rich layers (SU2A–SU2BU), and capped by SU1A–SU1C. The grindstone derived from SU2 microstratified ash layers. Dating: Chronology integrated AMS radiocarbon, OSL, and U-series dating. The SU2 hearth layers containing the grindstone are dated to 2002–2117 cal BP at their base (uncal. 2091 ± 16 BP; median 2062 cal BP) and 1567–1696 cal BP at their top (uncal. 1724 ± 16 BP; median 1632 cal BP). Additional site-wide ages include rapid SU3 accumulation c. 6090 ± 1140 years ago and subsequent slower sedimentation until c. 4400 cal BP, followed by SU2 ash deposition until c. 1567–1696 cal BP, and a later single fire 6–281 cal BP. OSL used single-grain SAR protocols with minimum age models to derive burial doses; dose rates assessed by in situ gamma spectrometry, beta counting, cosmic ray contributions, and water content. Grindstone description and use-wear: The artefact is a tabular sandstone grindstone (10.5×8.3×2.2 cm; 304 g) with two flat parallel worked surfaces (A and B), exhibiting fine unidirectional striations, lowered but not levelled topography, and plucked/ripped quartz grains. Surface texture and lack of developed polish indicate it was not used to process siliceous plants or shape ground stone axes. Stereozoom and metallographic microscopy assessed use-wear features. Residue sampling and analysis: Nine residue 'lifts' were extracted from worked and control areas across both surfaces using ultrapure water, lifted to slides, dried, and stained. Biochemical staining employed Picrosirius Red (PSR) to detect collagen (animal protein) via colorimetric change and birefringence under cross-polarised light. Acidified Phloroglucinol (Wiesner reagent) tested for lignin. High-powered microscopy (Leitz Dialux 22, 200×–400×) with plane/part/cross-polarised light documented residues. Control samples from non-worked areas assessed background contamination. Comparative reference slides were prepared from dried Bogong moth wings and hind legs, pounded on glass, stained with PSR, and examined to establish diagnostic morphology and birefringence for wing scales/segments. Data interpretation: Residue distributions, associations with use-wear, and comparisons with controls informed use-relatedness. Identification criteria included collagen fibre colour/birefringence patterns, presence/absence of starch and phytoliths, lignin response, morphology and dimensions of Lepidoptera wing segments, and evidence of carbonisation indicating exposure to heat.

• Secure context and age: The grindstone derives from finely stratified ash layers dated to 2002–2117 cal BP (base; median ~2062 cal BP) to 1567–1696 cal BP (top; median ~1632 cal BP), indicating use 1600–2100 years ago. • Use-wear: Fine unidirectional striations, lowered/undulating topography, and plucked quartz grains indicate grinding activity inconsistent with axe-sharpening or seed processing. • Animal residues: Collagenous residues (amorphous collagen, fibres including reticular fibres, woven collagen) occur at mid-range densities across central worked areas (Samples 1, 4 on Surface B; Sample 5 on Surface A), at approximately three times the frequency of controls. Woven collagen exhibiting birefringence suggests processing of fauna; the association with carbonised residues (Sample 2) indicates cooking. • Plant/fire residues: Moderate densities of carbonised plant residues, including wood-like structures with bordered pits and lignin, were identified (notably Sample 2; partial carbonisation in Samples 4 and 5), consistent with use of twigs/bark as fuel and the grindstone’s proximity to or inclusion within hearths. • Absence of plant food processing markers: Despite high densities of amorphous cellulose (Samples 1, 2, 4, 5), no starch grains or phytoliths were observed in any extraction, indicating a true absence of seed/plant food processing rather than taphonomic loss. • Insect remains: High density of variably carbonised insect wing segments in Sample 6 (Surface A), with additional fragments in Samples 2 and 4. Under PSR staining and cross-polarised light, 26 wing segments matched metrical and morphological criteria for Bogong moth (Agrotis infusa) based on reference material, including patterns of damage (ripped/torn, rectangular fragments) consistent with historical cooking practices where stirring detached wings/legs and some moths were pounded into paste. • Additional insect elements: A probable moth hind leg was identified (Sample 6). Proteinaceous material was observed associated with wing tangles, supporting an animal origin. Overall, the grindstone preserves microscopic remains of ground and cooked Bogong moths, representing the first conclusive archaeological evidence of insect foods in Australia and the first documented residues of insect processing on stone artefacts globally.

The findings directly address the longstanding absence of archaeological evidence for Bogong moth exploitation despite detailed ethnohistoric accounts. The combination of use-wear and PSR-based collagen detection, along with diagnostic Lepidoptera wing morphology and controlled comparisons, demonstrates that the Cloggs Cave grindstone was used to process and cook Bogong moths between ~1600 and 2000 years ago. This situates documented nineteenth-century Aboriginal practices—stunning with smoke, cooking on heated earth, stirring that detaches wings and legs, and pounding into preservable cakes—within a much older temporal framework, extending such seasonal feasting and associated social scheduling at least 65 generations into the past. The study also clarifies that earlier archaeological inferences based on UV fluorescence of pebbles are equivocal due to non-specific fluorescence, whereas biochemical staining offers specific detection of animal collagen. The presence of carbonised wood residues and amorphous cellulose without starch/phytoliths corroborates a scenario of insect processing in or near hearths rather than plant food grinding. Methodologically, the study showcases an effective protocol to detect culturally important but archaeologically elusive insect foods, opening avenues to reassess other sites in the Australian Alps and beyond.

Microscopic residues on a small sandstone grindstone from Cloggs Cave provide the first unequivocal archaeological evidence for the processing and cooking of Bogong moths (Agrotis infusa) in Australia, securely dated to 1600–2100 years ago. This discovery integrates ethnographic narratives of seasonal Bogong moth harvesting into deep-time Aboriginal histories and demonstrates a robust residue analysis approach—combining use-wear, PSR collagen staining, and comparative reference material—for detecting insect foods on stone artefacts. Future research should apply these methods systematically across alpine and foothill sites to map the spatial-temporal extent of Bogong moth use, explore variability in processing tools and techniques, and investigate the broader role of insect resources in past subsistence, mobility, and social networks.

• Evidence derives from a single grindstone at one site, limiting generalizability without further finds. • Residue preservation is subject to taphonomic processes; mechanical grinding, heat, and environmental factors can alter morphology, necessitating biochemical confirmation. • No starch/phytoliths were observed; while interpreted as true absence of plant processing, this relies on preservation expectations. • Minor inconsistencies between paired radiocarbon measurements (blind test) were noted (within three sigma), though not affecting interpretations. • Earlier UV fluorescence-based identifications are acknowledged as unreliable; this underscores the need for cautious interpretation of non-specific proxies.