Banjo Enclosure

A banjo enclosure is a distinctive class of Middle-to-Late Iron Age (c. 400 BCE–50 CE) earthwork settlement and pastoral monument across southern and central Britain, characterized in plan view by a small sub-circular, oval, or sub-rectangular banked-and-ditched enclosure (typically 0.2 to 0.8 hectares in internal area) approached through a single long, narrow, parallel-ditched entrance avenue (the "neck") that flares outward at its distal end into funnel-like antennae ditches—imparting a striking visual resemblance to a musical banjo when viewed from the air. First recognized as a coherent monument class in 1966 by Royal Commission archaeologist B. T. Perry from vertical and oblique aerial cropmark photographs over the Hampshire and Wiltshire chalklands, more than 200 banjo enclosures have now been mapped through aerial archaeology, geophysical magnetometry, and National Mapping Programme surveys across Wessex (Hampshire, Wiltshire, and Dorset), the Cotswolds, the Upper Thames Valley, Sussex, and Northamptonshire, with recent airborne laser scanning and salvage excavations revealing continental La Tène analogues in northern Gaul.

Unlike monumental multivallate hillforts such as Danebury or Maiden Castle—which enclose 5 to 20 hectares atop prominent defensive hill summits—a banjo enclosure is a compact, non-military earthwork typically situated on a gentle chalk valley slope, false crest, or plateau margin between 80 and 180 meters above sea level:

1. The Sub-Circular Central Compound ("The Body"): The main enclosure measures between 45 and 95 meters in internal diameter (enclosing 0.15 to 0.80 hectares, averaging ~0.4 hectares). It is bounded by a V-profiled rock-cut ditch 1.5 to 3.0 meters deep and 2.5 to 4.5 meters wide, with the excavated chalk or limestone rubble piled into an internal bank (or occasionally an external bank, indicating livestock containment rather than military defense).
2. The Parallel-Ditched Approach Avenue ("The Neck"): Projecting outward from a single entrance gap in the enclosure circuit is a straight, narrow sunken corridor flanked on both sides by parallel ditches and banks. This entrance avenue measures 30 to 90 meters in length (occasionally exceeding 120 meters, as at Bramdean in Hampshire) but only 3 to 8 meters in internal width—creating an architectural bottleneck that restricted visitors and livestock to a single controlled file.
3. Flaring Antennae Ditches and Outer Compounds ("The Peghead"): At its outer end, the parallel avenue ditches do not simply terminate; instead, they splay outward at 45° to 90° angles like funnel wings (antennae ditches) or merge directly into linear boundary ditches and coaxial Celtic field trackways. At complex sites such as Gussage Cow Down and Micheldever Wood, two or more banjo enclosures sit nested inside a larger multi-hectare outer compound or dyke system.

Since B. T. Perry's initial synthesis, archaeological explanations of why Iron Age communities expended substantial labor digging 80-meter-long ditched avenues have evolved from a purely utilitarian livestock model to a nuanced high-status pastoral and ceremonial residency model:

• The Livestock Funnel and Sorting Hypothesis: In early landscape analyses, the flaring antennae ditches and narrow parallel-walled avenue were interpreted primarily as a mechanical stock-handling funnel: herds of cattle (Bos taurus), sheep (Ovis aries), and semi-feral horses grazing on open upland downland pastures could be driven between the wide outer antennae wings, funneled single-file down the narrow 4-meter avenue for autumn counting, culling, branding, or veterinary inspection, and penned inside the central ditch-and-bank enclosure.
• High-Status Residential Homesteads and Feasting Arenas: Open-area excavations conducted from the late 1970s onward—at Micheldever Wood (by Peter Fasham), Nettlebank Copse (by Barry Cunliffe), Owlesbury, and Winterborne Kingston (by Miles Russell)—demonstrated that banjo enclosures were densely occupied domestic and ceremonial homesteads. Excavated interiors routinely contain one to three circular timber roundhouses (10 to 14 meters in diameter, marked by eaves-drip drainage gullies and porch postholes), elevated four-post granaries, and dozens of subterranean beehive-shaped chalk storage pits identical to those inside hillforts. Compared to ordinary unenclosed Iron Age farmsteads or simple hillslope enclosures, banjo enclosures yield rich assemblages of high-status metalwork and imported feasting goods, including cast-bronze terret rings and bridle bits, iron currency bars, ceramic salt-drying briquetage, Late Iron Age gold and silver coinage, and pre-conquest Roman Dressel 1 wine amphorae imported from Republican Italy via Hengistbury Head.

Zooarchaeology and bioarchaeology of storage pits inside banjo enclosures reveal structured ritual depositions closely paralleling those at Danebury:

1. Feasting Bone Deposits: Pit fills at Micheldever Wood and Nettlebank Copse contain articulated joints of prime beef, neonatal lamb bones (indicating spring birthing right outside the enclosure), and high proportions of pig (Sus scrofa) and horse (Equus caballus) remains associated with seasonal communal feasting.
2. Structured Human Pit Burials: At Winterborne Kingston in Dorset and Micheldever Wood in Hampshire, decommissioned grain storage pits along the inner perimeter ditch and entrance avenue contained crouched human inhumations, isolated human crania, and concentrated neonate/infant burials placed alongside complete ceramic vessels and sacrificed corvids or dogs.

Radiocarbon dating and ceramic stratigraphy across Wessex and the Cotswolds show that banjo enclosures emerged during the Middle Iron Age (c. 400–200 cal BCE)—characterized by regionally distinctive Saucepan Pottery styles—and reached their greatest architectural elaboration during the Late Iron Age (c. 100 BCE–43 CE) alongside wheel-thrown Belgic terra nigra ceramics. During the early Romano-British period (mid-1st to 2nd centuries CE) following the Claudian invasion of 43 CE, most banjo enclosures had their avenue ditches intentionally backfilled or remodeled into rectilinear compound ditches—frequently evolving directly into the site of a masonry Roman villa (as at Bramdean and Rockbourne), whereas in non-villa western regions analogous enclosed homesteads persisted as Cornish rounds.

1. Perry, B. T. (1966). "Some recent discoveries in Hampshire." In Rural Settlement in Roman Britain (CBA Research Report 7, pp. 39–42). Council for British Archaeology. https://doi.org/10.5284/1081702
2. Fasham, P. J. (1987). A Banjo Enclosure in Micheldever Wood, Hampshire. Hampshire Field Club and Archaeological Society Monograph 5. https://doi.org/10.1017/S0003598X00074500
3. Cunliffe, B., & Poole, C. (2000). The Danebury Environs Programme: The Prehistory of a Wessex Landscape, Vol. 2, Part 5: Nettlebank Copse, Wherwell, Hants, 1993. Oxford University Committee for Archaeology Monograph 49. https://doi.org/10.1017/S0079497X0000150X
4. Lang, A. T. O. (2016). "Defining banjo enclosures: Investigations, interpretations, and wider contexts." Proceedings of the Prehistoric Society, 82, 341–361. https://doi.org/10.1017/ppr.2016.10