(Re)dating Danebury Hillfort and Later Prehistoric Settlements in the Environs: a Bayesian Approach

Investigators: Prof Colin Haselgrove (Leicester); Prof Chris Gosden (Oxford); Prof Gordon Cook (SUERC)

Research Associates: Dr Derek Hamilton (Leicester-SUERC); Dr. Elaine Dunbar (SUERC); Ms Cynthia Poole (Oxford); Ms Lisa Brown (Oxford)

Traditional approaches to dating the Iron Age constructed complex chronologies based on artefact typologies. 14C dating was long neglected, because it was thought to allow less precision than artefact dating and because of the ‘Hallstatt plateau’ between 800–400 BC. The last decade, has however seen major advances in methodology and through specific 14C dating projects (e.g. Garrow et al. 2010; Hamilton 2010; Jay et al. 2012). Two key results are, first, that typological dating produces sequences that are regularly too late, and second, that various phenomena, from chariot burials to settlement shifts, represent brief episodes, rather than being long lived.

This has created gaps in the familiar sequence, with knock-on consequences for the models that govern our perceptions of Iron Age societies (Barrett et al. 2011). An example is the 2nd to 1st century BC void identified by 14C dating of the metalwork that underpins the pottery typologies used to date most settlements. If a re-alignment of insular and continental chronologies is found to be necessary, this will have major implications for our interpretation of the mid to late Iron Age transition in Britain. Similar14C dating programmes have already altered our understanding of other periods of British and world prehistory, as well as individual sites such as Stonehenge. The current project is seen as a step on the way to putting Iron Age chronologies on a firmer footing.

Why Danebury?

Hillforts are the most iconic monuments surviving from Iron Age Britain, dominating academic and public perceptions of the period. This is especially true of Wessex, where many major hillforts have been explored over the years, including Cadbury Castle, Maiden Castle and best known of all, Danebury. Since Barry Cunliffe began excavating there in 1969, this 16 ha hillfort has stood near the centre of archaeological discourse on Iron Age societies in Britain and beyond. Over 20 years, more than half the interior was explored, whilst from 1989–2006, 15 nearby settlements were excavated. Thanks to this programme, no study of Iron Age societies is complete without reference to the Danebury sites and influential (but often conflicting) interpretations of Iron Age societies have emerged based largely on its evidence.

Our ability to choose between or improve on the competing social models remains constrained by the lack of a comprehensive 14C chronology for the Danebury sites. In the 1980s, a limited dating programme was undertaken for Danebury, but Accelerator Mass Spectrometry now allows far smaller samples to be accurately dated. In combination with Bayesian modelling, it is now possible to date many archaeological events to a margin of decades not centuries, and potentially to surmount the problems that previously plagued radiocarbon dating of the Iron Age.

This collaborative project between Leicester, Oxford and the Scottish Universities Environmental Research Centre funded by the Leverhulme Trust (RPG-2013-009) aims to build a new time-line for the Danebury area both by reworking the 1980s results (Fig 1) and by obtaining 300 new 14C dates. As well as clarifying the relationship between the hillfort and other sites in its environs, the results will provide a basis for reassessing the chronology of the southern British Iron Age and for rethinking social and cultural links between this region and the rest of Britain, as well as across the Channel. Another aim is to explore the use of Bayesian approaches for building better chronologies for the earlier Iron Age, hitherto plagued by the 800–400 BC plateau.

Danebury fig. 1

Figure 1: Older, imprecise 14C dates can be built into Bayesian models. As a feasibility study, we modelled 16 14C dates obtained in the 1980s for Danebury Quarry Hollow Sequence A. The results suggest that the ‘Late Hillfort’ began nearer 400 than 300 cal BC and ended a little earlier than previously thought. By adding new dates for this and other sequences, a more detailed chronology will be established.


Barrett, J, Bowden M & McOmish, D 2011 The problem of continuity. Reassessing the shape of the British Iron Age sequence, in T Moore & L Armada (eds) Atlantic Europe in the first millennium BC, 439–48. Oxford: OUP

Garrow, D, Gosden, C, Hill, JD & Bronk Ramsey, C 2009 Dating Celtic art: a major radiocarbon dating programme of Iron Age and early Roman metalwork, Archaeol J 166, 79–123

Hamilton, D 2010The use of Radiocarbon and Bayesian Modelling to (re)write Later Iron Age Settlement Histories in East-Central Britain, Unpublished PhD thesis, Univ Leicester

Jay, M, Haselgrove, C, Hamilton, D, Hill, JD & Dent, J 2012 Chariots and context: new radiocarbon dates and the chronology of Iron Age burials and brooches in East Yorkshire, Oxford J Archaeol 31, 161–89.

Share this page: