Pillow Mound

A pillow mound (known in medieval and early modern estate documents as a burga, clapper, berry, or coney-earth) is a low, rectangular, oblong, or cigar-shaped artificial mound of earth and stone bordered by a shallow perimeter quarry ditch, constructed across Britain and Ireland between the 12th and 19th centuries CE as an engineered breeding habitat within a managed rabbit warren (coneygarth). Measuring typically 10 to 40 meters in length, 4 to 10 meters in width, and 0.4 to 1.2 meters in height, pillow mounds were built to provide dry, well-drained, easily dug soil and internal stone-lined gallery networks for the husbandry of the introduced European rabbit (Oryctolagus cuniculus). Because pillow mounds were frequently sited on open upland moorlands, chalk downlands, and abandoned prehistoric earthworks—and because their elongated, ditch-flanked profiles closely resemble Neolithic long barrows, oval barrows, or Bronze Age tumuli—they were long misclassified in 19th-century archaeology as prehistoric burial mounds until historical, stratigraphic, and zooarchaeology investigations established their function in medieval manorial economies.

Understanding the architecture of pillow mounds requires examining the biogeographical history of the European rabbit (Oryctolagus cuniculus). Confined following the Last Glacial Maximum to its native refugium in the Iberian Peninsula and southern France, O. cuniculus was not native to post-glacial Britain or Ireland (and is absent from British Mesolithic, Neolithic, Bronze Age, and Iron Age faunal assemblages, except as modern intrusive burrowing contaminants verified by radiocarbon dating).

1. Norman Introduction and Elite Status (12th–14th Centuries CE): Although isolated imported rabbits appeared at Roman sites such as Fishbourne Palace and Lynford in the 1st–2nd centuries CE, self-sustaining breeding populations were introduced to Britain by Anglo-Norman lords in the mid-to-late 12th century CE—initially on offshore islands (the Isles of Scilly by 1176, Lundy, Skomer, and Caldey) and rapidly thereafter into royal and baronial mainland parks (coningera or coneygarths).
2. Physiological Vulnerability and the Need for Artificial Mounds: Adapted to the warm, dry, sandy Mediterranean soils of Iberia, early medieval rabbits in Britain were vulnerable to damp, cold, heavy clay soils and winter waterlogging, which flooded natural burrows and caused high kit mortality from hypothermia and coccidiosis. At the same time, rabbit meat (especially tender suckling kits, or laurices) and dense winter rabbit pelts commanded luxury prices under royal Free Warren charters, equivalent per pound to skilled artisan wages.
3. Engineered Drainage and Burrowing Substrates: To protect these prized exotic animals from winter flooding and predators (foxes, polecats, and stoats) while concentrating them for harvesting with nets and ferrets (Mustela furo), warreners constructed elevated earthen pillow mounds whose encircling ditches drained rainwater away from the friable central soil bank.

Stratigraphic excavations of pillow mounds across Wales (by Douglas Hague, Christopher Spurgeon, and David Austin), Dartmoor (by R. Hansford Worth and Stephen Greeves), and the Cotswolds have revealed that pillow mounds were carefully built subterranean stone-and-earth structures:

1. Sub-Mound Stone Tunnels (Clappers or Flues): Before heaping up the earthen mound, warreners laid out a geometric grid of drystone channels directly on the leveled ground surface—constructed of two parallel lines of upright stone slabs or bricks roofed with flat capstones, measuring 15 to 25 centimeters wide and 15 to 30 centimeters high (matching the body diameter of an adult rabbit).
2. Spine-and-Rib (Herringbone) Layouts: A central longitudinal stone spine tunnel typically ran the full length of the mound, intersected at intervals of 1.5 to 3.0 meters by perpendicular side branches (ribs or eyes) that opened out into the flanking quarry ditch, where stone sills prevented rainwater from running into the burrows.
3. Nesting Chambers and Ferreting Access: At the inner junctions of the stone tunnels, warreners built expanded square or circular stone boxes (breeding stops) roofed with large removable flat slabs sitting just beneath a thin covering of turf. During the winter harvesting season (November to February, when pelts were thickest), warreners laid purse-nets over the outer side-openings along the perimeter ditch and sent a muzzled ferret down the central spine tunnel—or lifted the central nesting capstones—to bolt the rabbits into the nets without digging up the mound.

From the 18th century until the 1960s, antiquarians such as Richard Colt Hoare and Ordnance Survey mapmakers labeled pillow mounds as "Long Barrows" because medieval and Tudor warreners established warrens on unplowed downlands and modified existing prehistoric round barrows, bank barrows, cairns, and hillfort ramparts (as at Hambledon Hill, Maiden Castle, Dolebury Warren, and Minchinhampton), incorporating residual prehistoric flints and potsherds from the topsoil.

In modern field survey (codified by O. G. S. Crawford and Tom Williamson), pillow mounds are distinguished from Neolithic long barrows and Bronze Age clearance cairns by four diagnostic criteria:

• Scale and Aspect Ratio: Whereas Neolithic long barrows measure 30 to 100 meters long and 1.5 to 3.5 meters high with deep flanking quarry ditches, pillow mounds are lower (0.4 to 1.0 meter high), flat-topped, uniform in width (4 to 8 meters wide), and bordered by a narrow, shallow ditch (1.0 to 1.5 m wide and 0.3 m deep) that often wraps around all four sides.
• Geometric Clustering and Cross-Contour Alignment: Pillow mounds occur in regimented groups of 5 to 120 mounds, often arranged in parallel rows running down the hillslope across the contours to maximize drainage inside the flanking ditches, accompanied by stone vermin traps for stoats and weasels.
• Internal Stone Flues: Ground-penetrating radar (GPR) and excavation expose narrow 20-cm-wide herringbone drystone rabbit channels rather than megalithic chambers or cists.
• Stratigraphic Superposition Over Medieval Ridge-and-Furrow: At deserted medieval villages (DMVs) depopulated after the 14th-century Black Death, pillow mounds sit stratigraphically on top of medieval ridge-and-furrow and strip lynchets—proving a Late Medieval or Post-Medieval (15th–18th century CE) date.

Optically stimulated luminescence (OSL) dating of buried ditch silts combined with Tudor and Stuart estate surveys, tithe maps, and place-name evidence (Warren Farm, Coneygarth, Conyger, Cwninger in Welsh) anchors the peak construction of British pillow mounds between 1350 and 1750 CE, before 19th-century agricultural enclosure and the natural adaptation of feral rabbit populations rendered artificial warren mounds obsolete.

1. Williamson, T. (2006). "The archaeology of rabbit warrens." Journal of Archaeological Method and Theory, 13(2), 135–159. https://doi.org/10.1017/S0003598X0009493X
2. Sykes, N., & Curl, J. (2010). "The rabbit." In Extinctions and Invasions: A Social History of British Fauna (pp. 116–126). Windgather Press / Oxbow Books. https://doi.org/10.2307/j.ctv13gvh3h
3. Austin, D. (1988). "Excavations and survey at Bryn Cysegfan, Llanfihangel-y-Creuddyn, Dyfed." Medieval Archaeology, 32(1), 130–165. https://doi.org/10.1080/00766097.1988.11735505
4. Gould, D. (2017). "The distribution and chronology of pillow mounds on Dartmoor." Proceedings of the Devon Archaeological Society, 75, 195–222. https://doi.org/10.1080/00794236.2017.1390421