5. Case Studies

5.1 Happisburgh, Site 3, Norfolk

Zoe Outram and Peter Marshall

The Cromer Forest-bed Formation can be found along the coast of Norfolk and Suffolk. It formed between 2 and 0.5 million years ago and is famous for its rich flora and fauna. Mammoth, rhinoceros and hippopotamus remains have been discovered over the last 250 years. Despite such a long history of investigation, it has only recently yielded evidence for hominin presence.

Excavations at Happisburgh Site 3, Norfolk (see Plate 1), recovered an assemblage of 78 flint artefacts from the fills of a series of stacked, overlapping channels (Parfitt et al. 2010). The deposits contained a remarkable range of remains for a Pleistocene site, including stone tools, and floral and faunal remains (Figure 26). These provided the opportunity for a detailed study of hominin activities and the environment that they occupied.

Stratigraphic evidence indicates that the site was older than MIS 12 (450 ka), when the marine and freshwater deposits associated with Cromer Forest-bed Formation ended and sediments associated with the Anglian glaciation were laid down. Determining more precisely the age of the hominin activities at the site was essential to understand their significance and to place them into a broader context.

Samples for palaeomagnetic dating (see Palaeomagnetism) were obtained from a stratigraphic sequence of deposits below, within and above the artefact-bearing gravels. The sediments all displayed a reversed polarity, placing their deposition in the Matuyama chron (2.52–0.78 Ma).

Biostratigraphic evidence, including the presence of key plant taxa identified from the pollen spectra, suggested that the age of the site was towards the end of the Early Pleistocene. Plant taxa identified include Tsuga (hemlock) and Ostrya-type (hop-hornbeam type), which are unknown in northern Europe after the Early Pleistocene.

Extinct mammals identified included mammoths, equids and voles (genus Mimomys; see The ‘Vole Clock’). Taken together, these biostratigraphic and palaeomagnetic data indicate that the hominin occupation at Happisburgh occurred towards the end of the Matuyama chron, placing the deposits between 990 and 780 ka (Parfitt et al. 2010).

Further evidence of this hominin occupation is provided by the palaoebotanical record, which indicates that it occurred during a phase of climatic cooling in the second half of an interglacial cycle (Figure 27). Using this evidence, Parfitt et al. (2010) suggest that the site was occupied at the end of either MIS 25 (970–936 ka) or MIS 21 (866–814 ka).

Westaway (2011) proposes a younger date of MIS 15c (c. 600 ka), based on reinterpretation of the palaeomagnetic evidence and the suggestion that the pollen and faunal remains on the site are reworked from earlier deposits.

Parfitt et al. (2010, supplement) assign the sediments, based on their composition, to an Early Pleistocene extended Thames. By MIS 15, the northern part of East Anglia, within which Happisburgh 3 is located, was now within the River Bytham catchment, with the Thames positioned further south. Accepting Westaway’s (2011) interpretation would therefore throw into doubt the current understanding of the River Thames and Bytham catchments at this time (White et al. 2018).

If Westaway’s later dating is correct, Happisburgh Site 3 is much in keeping with other indications from the European Continent of the date when the first hominins occupied this part of north-west Europe. If Parfitt et al.’s earlier dating is accepted, then Happisburgh Site 3 has yielded the oldest hominin occupation north of Iberia and the first occupation within the northern boreal zone. This earlier dating has important implications for our understanding of populations, in terms of their migrations and movements, their behaviour, and their ability to adapt and survive different environments, such as the cooler climates recorded towards the end of an interglacial.

These different interpretations of the evidence for the age of Happisburgh 3, which fundamentally depends on the production of an accurate chronology for the site, highlights the importance of scientific dating techniques in the Pleistocene. Nonetheless, the evidence from Happisburgh Site 3 has redefined our understanding of the earliest known occupation of Britain.

5.2 Boxgrove, West Sussex

Danielle Schreve

The site of Boxgrove is located in West Sussex, 12km north of the English Channel coast. Its modern position is significant, as the Pleistocene deposits of archaeological and palaeontological interest lie on top of a wave-cut platform in the Cretaceous Upper Chalk bedrock, indicating that the site once lay at the northern edge of a large marine embayment (Figure 28).

The marine beach associated with the platform reaches a maximum height of 43.5m OD, highlighting considerable tectonic uplift since the deposits were laid down. The site forms the highest (and oldest) of a flight of four marine terraces that represent former sea-level high stands, which extend down to modern sea-level on the West Sussex Coastal Plain.

Excavated between 1984 and 1996, Boxgrove is internationally renowned for its spectacular Lower Palaeolithic archaeological record. This includes several hundred ovate bifaces (handaxes) made of flint sourced from the nearby Chalk cliff, its rich and diverse fossil vertebrate assemblage, and the presence of hominin remains (two incisors and a tibia) attributed to Homo heidelbergensis. More than 100 species of vertebrate fauna, together with invertebrates such as molluscs, ostracods and foraminifera were recovered (Roberts and Parfitt 1999).

The deposits at the site were laid down on the wave-cut platform and consist of a sequence of marine sands (the Slindon Sands) laying beneath a series of lagoonal deposits (the Slindon Silts) upon which a stable land surface developed. Palaeoclimatic conditions remained temperate throughout this period.

The Slindon Silts and overlying land surface are the source of the majority of the archaeological and faunal remains (Figure 29). The fine-grained nature of the sediments is such that individual episodes of handaxe knapping can be identified and the flakes refitted to reveal the process of manufacture. Bone and antler hammers used for handaxe manufacture were also recovered, and part of the site contains evidence for the presence of spring-fed pools surrounded by open grassland, which appear to have acted as a focal point for human activity.

This Pleistocene land surface was subsequently covered by silty brickearth and gravels (the Eartham Formation), which were deposited as climate deteriorated and vegetation cover became sparse. This transition to cold climate conditions is supported by the kinds of ostracod and mammal remains found in the upper part of the Slindon Silts and the basal sediments of the overlying Eartham Formation (Roberts and Pope 2009).

Boxgrove has also yielded extensive proof for large mammal butchery, including evidence for the dismemberment of wild horse, red deer, giant deer, bison and three Hundsheim rhinoceroses. The carcasses of these animals are littered with cutmarks from stone tools, indicating a complete process from skinning to the removal of the major muscle blocks and tendons. Where present, carnivore gnaw-marks overlie anthropogenic cut-marks, indicating that humans had first access to the carcasses and to the complete range of body parts.

Pathological evidence for a trauma wound to the shoulder blade of the butchered horse is consistent with impact damage from a large projectile, such as a wooden spear (Roberts and Parfitt 1999). In combination, all the evidence provides a strong indication of hominin hunting capabilities in the Lower Palaeolithic. The large size of the prey tackled and the concomitant meat yield (700kg in the case of a rhinoceros) have implications for palaeo-demography, with groups of up to 50 individuals in the immediate area.

A range of techniques have been used to date Boxgrove since it was first discovered. As with many Palaeolithic and Pleistocene sites, the establishment of a robust chronology has been problematic, particularly in the absence of suitable materials for geochronological techniques at the time of excavation, or indeed methods that extend far enough back in time or provide sufficient resolution (Figure 9).

When the site was first discovered, only three interglacials were formally recognised in the Middle and Late Pleistocene in Britain: the Cromerian, Hoxnian and Ipswichian (Mitchell et al. 1973). However, the unusual character of the mammalian assemblage, containing both post-Cromerian and pre-Hoxnian species, was first detected by Currant (in Roberts 1986), suggesting that Boxgrove might date to a previously unrecognised intermediate episode.

Later palaeomagnetic dating (see Palaeomagnetism) analysis of the Slindon Sands in the 1990s confirmed that the sediments have normal polarity and are therefore younger than 780 ka old but could not provide any further resolution (David and Linford 1999).

The age of the Boxgrove deposits is debated. Different authors suggest pre- and post-Anglian (MIS 12) ages. The evidence for a post-Anglian age can now be questioned, however, based on scientific advances since the original studies were undertaken. For example, Amino-Acid Racemisation put forward as the strongest evidence for later dating (Bowen and Sykes 1999) was not undertaken on the intra-crystalline fraction of the shells.

The mammalian biostratigraphy from the site, in particular the presence of a large number of taxa, strongly implies that the Boxgrove temperate climate sediments must pre-date, rather than post-date, MIS 12. This evidence includes the shrew Sorex (Drepanosorex) savini, the vole Pliomys episcopalis, the cave bear Ursus deningeri, the rhinoceros Stephanorhinus hundsheimensis (Figure 30) and the giant deer Praemegaceros dawkinsi and Praemegaceros cf. verticornis that became extinct in Britain during the Anglian glaciation. This evidence strongly implies that the Boxgrove temperate climate sediments must pre-date, rather than post-date, MIS 12.

Further resolution of the likely date of the Boxgrove deposits is provided by The ’Vole Clock‘. The Cromerian Complex interglacials can be divided into an older group, characterised by the presence of the archaic water vole, Mimomys savini, and a younger group characterised by its descendant, Arvicola cantiana terrestris. The presence of the latter at Boxgrove therefore implies a younger age within the Cromerian Complex. Furthermore, the presence of a more derived (i.e. advanced) form of narrow-skulled vole, Lasiopodomys gregalis, suggests a more recent age for Boxgrove within the early Middle Pleistocene Arvicola group.

The preferred position of the Boxgrove Slindon Formation is therefore right at the end of the early Middle Pleistocene, correlated with MIS 13, and with the cold-climate Eartham Formation correlated with MIS 12 (Roberts and Parfitt 1999; Roberts and Pope 2009). The attribution of Boxgrove to MIS 13 also helps to more firmly establish the timing of the earliest Acheulean in Britain, as handaxe sites are currently only known in association with Arvicola (Candy et al. 2015).

These conclusions reinforce the importance of the vertebrate fossil record for chronological determination at Pleistocene sites.

5.3 The Axe Valley at Broom, Devon/Dorset border

Peter Marshall

The sand and gravel exposures at Broom, on the River Axe along the Devon/Dorset border, are of considerable significance in the context of the Lower Palaeolithic and the fluvial terrace stratigraphy of south-west England (Figure 31).

The deposits exposed in three working pits have yielded at least 2,300 Palaeolithic artefacts, an assemblage dominated by handaxes. Like most of England’s river-terrace Palaeolithic archaeology, the contextual information for the assemblage is incomplete. The physical condition of the stone tool assemblage suggests a mixture of locally derived artefacts and pieces that had been transported farther by the river during the Middle Pleistocene.

The archaeology is of both regional importance for the understanding of the Lower Palaeolithic occupation of south-west England and of national significance with respect to the use of chert in the manufacture of the majority of the lithic assemblage (Hosfield and Green 2013).

The traditional model of sediment accumulation at Broom emphasises a tripartite sequence of lower gravels (Holditch Lane Gravel Member); an intervening unit comprising sands, silts and clays (Wadbrook Member); and an upper gravel unit (Fortfield Farm Gravel Member). This sequence resonates with Bridgland’s (1996) model of the typical aggradational terrace. The model's framework associates major fluvial aggradations and incisions with the cyclical shifts from interglacial to glacial recorded in the Marine Oxygen Isotope record (Text Box 1). However, the age of the sediments remained unknown.

The Archaeological Potential of Secondary Contexts project (Hosfield et al. 2007) assessed the interpretative potential of the secondary context archaeological resource for the Lower and Middle Palaeolithic in England. It included excavations at Broom that built on a long history of research to contextualise the artefact collections and date the fluvial sediments. Eighteen OSL samples (see Luminescence dating) were dated (Table 4; Toms 2013; Toms et al. 2005) to provide an absolute chronology for the Middle Pleistocene terrace succession and the artefacts at Broom, and to assess whether the River Axe’s fluvial record matches the classic Bridgland model.

The OSL age estimates from the Wadbrook Member and the Fortfield Farm Gravel Member were combined with relative dating information provided by the stratigraphic relationships between the samples to create a Bayesian chronological model (Figure 32).

Age estimates from the Wadbrook Member come from a single section and were therefore defined sequentially (GL02084<GL03011<GL02083) as their relative stratigraphic position was unambiguous. The Fortfield Farm Gravel Member age estimates derived from several separate sections, and therefore formed part of a Fortfield Farm Gravel phase.

The model provides an estimate for the timing of the transition from the Wadbrook Member to the Fortfield Farm Gravel Member of 311–270 ka (95% probability; Wadbrook/Fortfield Farm Member; Figure 32) and probably 300–280 ka (68% probability).

These results indicate that the Wadbrook Member formed between mid-MIS 9 (interglacial) and early MIS 8 (glacial), and that the Fortfield Farm Gravel Member formed between MIS 8 (glacial) and MIS 7 (interglacial). Combined with the stratigraphic and sedimentary evidence at Broom, these dates show that the Axe valley’s terrace stratigraphy does not fit exactly into existing models of terrace formation. This is a valuable reminder that not all rivers respond in the same way to variations of climate, geology and base level.

These age estimates also provide a chronology for the prolific assemblage of Acheulean (biface) artefacts recovered from the Wadbrook Member. They are notable because they clearly indicate that the Acheulean-dominated assemblage at Broom was deposited at a time when evidence in south-east England suggests the beginning of a shift towards using Levallois prepared-core dominated technologies (e.g. at Purfleet, Essex).

5.4 Marine Aggregate Licence Area 240, North Sea off Great Yarmouth, Norfolk

Peter Marshall

Between December 2007 and February 2008 gravel extraction 11km off the coast of East Anglia in Marine Aggregate Licence Area 240 (Figure 33) produced an important collection of Middle Palaeolithic artefacts (Figure 34) and faunal remains: 88 worked flints, including 33 handaxes, plus woolly mammoth, rhinoceros, bison, reindeer and horse (Tizzard et al. 2014; 2015).

The unweathered nature of many of the handaxes indicates they probably derived from an in situ or a near in situ context before being dredged from the seabed. Although prehistoric material has been recovered since the 1930s from the North Sea through fishing and dredging, the material from Area 240 came from known dredging lanes within it. Thus, unlike many chance finds, the Area 240 material offered the opportunity to establish the geological and geomorphological context of the material and to provide an estimate of the age of the deposits.

Area 240 is in the lower reaches of the Palaeo-Yare river system and for most of the last 1 Ma it has been part of a coastal or inland environment due to lower sea-levels. From 2008 to 2011 the geophysical and geotechnical data were re-examined, and a new geophysical survey undertaken of the area from which the artefacts and faunal remains came. The deposits were also cored to obtain material for OSL dating (see Luminescence dating) and for reconstructing past environmental conditions (Tizzard et al. 2015; Figure 33).

A Bayesian chronological model was constructed using the OSL dates and the stratigraphic sequence (Figure 35). The model suggests that Unit 3b, from which many of the artefacts and faunal material are thought to derive, started to form in 275–192 ka (95% probability; start_unit_3B; Figure 36), probably in 244–204 ka (68% probability), and ended in 223–161 ka (95% probability; end_unit_3B; Figure 36), probably in 210–180 ka (68% probability). The dating suggests that Unit 3b most likely dates to MIS 7 or possibly the beginning of MIS 6.

The material from Area 240 can now be included in the corpus of archaeological sites dated to MIS 7 within the British Palaeolithic record (White et al. 2006). The archaeological record of MIS 7 is important as it represents the final phase of Middle Palaeolithic occupation of Britain before the c. 120 ka hiatus between MIS 6–3, when hominins were absent.

Bayesian chronological modelling of the OSL dates from Area 240 (Table 5) enables us to correlate the key depositional unit thought to have contained the flint artefacts with the sequence of MIS stages (most probably MIS 7); something that could not have been achieved without the dating programme. Together with other investigations in Area 240 (Tizzard et al. 2014; 2015) the results confirmed that submerged landscapes have the potential to preserve in situ Middle Palaeolithic artefacts.

5.5 Pin Hole, Creswell Crags, Derbyshire

Alistair Pike

In ideal circumstances samples for Uranium-Thorium dating would be collected during controlled excavation. Many archaeologically important cave sites were excavated, however, in the 19th or early 20th century using now-outdated excavation methods and recording. This means that caves containing intact and undisturbed Pleistocene deposits are rare in England.

Age constraints for the museum collections derived from these excavations, however, can be obtained if flowstones were left in situ in the excavated cave. These can be sampled and related to the excavated assemblage within the museum collection. Flowstones collected as part of the archaeological assemblage can also be sampled for dating.

For example, during excavations at Pin Hole, Creswell Crags, Derbyshire, in 1925, flowstones were collected (Figure 37). Leslie Armstrong, the excavator, collected stalactites and stalagmites (collectively known as speleothems) believing them to be tools (Armstrong 1932). The three-dimensional position of the bones and artefacts, including the calcite, were recorded, enabling us to reconstruct the stratigraphy (Figure 38). This consists of two units: an upper layer containing Upper Palaeolithic or Mesolithic flint blades and a lower layer containing Mousterian non-flint artefacts along with fauna (including reindeer, spotted hyaena, woolly rhinoceros and horse) and datable speleothems (Jacobi et al. 1998).

The Uranium-Thorium ages are scattered (Table 6), reflecting the variable ages of the speleothems before they became incorporated in the archaeological layer. However, the youngest age (64 ka) provides a maximum age (terminus post quem) for the fauna and Middle Palaeolithic artefacts with which they are associated, and also a maximum age for the archaeological assemblages in the level immediately above.

This study is important because the maximum age of 64 ka supported the idea that hominins were absent in Britain during the preceding interglacial (Ipswichian; MIS 5e) but returned at the end of MIS 4. Additionally, the distinctive fauna at Pin Hole, chronologically constrained by these Uranium-Thorium dates and additional Electron Spin Resonance (see Text Box 4) and Radiocarbon dating, is critical in defining a stage in the formal mammalian biostratigraphy for the Late Pleistocene of Britain (Currant and Jacobi 2001).

5.6 Lynford Quarry, Mundford, Norfolk

Peter Marshall and Zoe Outram

In 2002 a palaeochannel was observed during archaeological monitoring at Lynford Quarry, Mundford, Norfolk. The palaeochannel had a dark organic fill containing mammoth remains and associated Mousterian stone tools in situ, as well as debitage buried under 2–3m of bedded sands and gravels (Figure 39). Well-preserved in situ Middle Palaeolithic open-air sites are unusual in Europe and exceedingly rare in England, so this site is of international importance.

The palaeochannel and associated deposits with archaeological remains were subsequently excavated and recorded by the Norfolk Archaeological Unit (Boismier et al. 2012). This provides a range of spatial, palaeoenvironmental and taphonomic information about the deposits and their formation. It also provides information on the associated hominin activity, including for investigating questions of diet, land use and habitat.

The lithic assemblage was dominated by Mousterian tools — handaxes and bifacial scrapers characteristic of the Late Middle Palaeolithic, c. 59–38 ka. The handaxes are of a form frequently associated with Neanderthals. Reliable dating evidence for this activity is important to better understand when England was re-occupied by hominins after the cold stage of MIS 4.

Biostratigraphic evidence from the faunal remains suggested that the site was older than 30 ka and probably older than c. 41 ka, based on the known presence of woolly mammoths (Mammuthus primigenius), woolly rhinoceros (Coelodonta antiquitatis) and spotted hyena (Crocuta crocuta).

Seventeen OSL dates (Table 7; see Luminescence dating) and eight radiocarbon dates (Table 8; see Radiocarbon dating) from the site are incorporated into a Bayesian chronological model (Figure 40). Amino Acid Racemisation analysis failed because of poor preservation of shells. Prior information about the relationship between samples is derived from direct stratigraphic relationships and from the sedimentological model for the formation of the site.

The model establishes a chronological framework for fluvial activity with the infilling of the channel (Association B) estimated to have started in 76–60 ka (95% probability; First association_B; Figure 40) and probably in 72–63 ka (68% probability). Fine-grained organic sediments continued to be deposited in the channel until 65–52 ka (95% probability; OxL-1340; Figure 40) and probably 62–54 ka (68% probability) when beds of laminated sands began to accumulate.

Radiocarbon measurements from two mammoth bones recovered from the Association B channel organic sediments are close to the reliable limits of the technique (see Radiocarbon dating) and suggest that the true age of the faunal material is probably in excess of 50 ka. The model suggests this dating and highlights one of the challenges faced when investigating Middle Palaeolithic sites: that the earlier part of the period is beyond the range of radiocarbon dating.

The hominin activity recorded at Lynford can therefore be dated to late MIS 4 and/or MIS 3 as Neanderthals re-occupied England after a long hiatus during the cold stage of MIS 4.

5.7 Gransmoor, East Yorkshire

Peter Marshall

A working sand and gravel quarry about 1km west of the village of Gransmoor, East Yorkshire, exposed Lateglacial sediments that had accumulated in a kettle hole within fluvio-glacial deposits laid down at the end of the Late Devensian.

The most complete and comprehensively studied sequence came from more than 2 metres of aquatic and semi-terrestrial deposits that overlay several metres of glacial sands in a working face on the north side of the quarry (Figure 41). Palynological, coleopteran, molluscan and geochemical studies of samples from this sequence enables a detailed reconstruction of environmental and climatic change during the Lateglacial period (Walker et al. 1993; Lowe et al. 1995).

Age-depth models (Parnell et al. 2011) are crucial for establishing the temporal framework of paleoenvironmental archives such as those from Gransmoor. Twenty-five radiocarbon measurements (see Radiocarbon dating) are available from the site (Table 9).

The samples were processed according to methods outlined in Mook and Waterbolk (1985). The acid-insoluble, alkali-soluble (‘humic acid’) and the alkali- and acid-insoluble (‘humin’) fractions were dated.

Radiocarbon measurements from six samples (SRR-) were determined by liquid scintillation spectrometry (Harkness and Wilson 1972), although 19 samples produced insufficient carbon dioxide for conventional dating, and so sub-samples were sent for graphitisation and dating by accelerator mass spectrometry (AA-), as described by Slota et al. (1987) and Linick et al. (1986).

The two measurements from 1.70m are statistically inconsistent at the 5% level (T’=44.5, T'(5%)=3.8, ν=1; Ward and Wilson 1978) and we have preferred the terrestrial macrofossil sample (AA-12004) over the humic fraction of the bulk sediment sample (SRR-3875) for the age of this horizon, as the macrofossil date shows better agreement (cf. Blockley et al. 2004). The four basal samples have elevated δ¹³C values consistent with a hard-water reservoir effect (Bayliss and Marshall 2022, section 1.6), which would make their dates anomalously old. They have therefore been excluded from the age-depth model shown in Figure 42.

Age-depth modelling was implemented with rBacon version 3.2.0 (Blaauw and Christen, 2011) in R (R Core Team, 2021) using IntCal20 (Reimer et al. 2020). The accumulation rate prior was set at 10 yr/cm as a gamma distribution (Figure 42 — upper middle). A second parameter (acc.shape), which controls how much influence the accumulation rate has on the model, was set at the default 1.5 recommended by Blaauw and Christen, (2011).

Radiocarbon age distributions in rBacon are derived from the Student’s-t distribution, which produces calibrated distributions with longer tails than the Normal model (Christen and Pérez 2009). The longer tails on radiocarbon dates, and a prior assumption of unidirectional sediment accumulation, mean in most cases excluding outliers is not necessary when using rBacon. Thus, unlike previous attempts to produce age-depth models for Gransmoor using OxCal (e.g. Blockley et al. 2004; Elias and Matthews 2013), no radiocarbon dates, apart from SRR-3875 and those with a hard water offset, were excluded from the rBacon model.

Figure 42 shows the resulting age-depth model. The sequence from Gransmoor is estimated to span the period from 14,170–13,630 cal BP (95% probability; 2.35m; Figure 40) to 11,910–11,380 cal BP (95% probability; 0.23m; Figure 42), with deposition of sediments occurring in the Windermere interstadial (GI-1) and Loch Lomond stadial (GS-1; see Figure 8). Age-depth modelling allows ‘events’ in a sedimentary sequence that have not been directly dated to be plotted against time, as opposed to depth, with quantified estimates of their chronological uncertainties (Blaauw et al. 2007).

To illustrate this, a feature of the pollen record from Gransmoor is the decline and subsequent recovery of Betula values that follow an initial abrupt rise for the genus (Sheldrick et al. 1997, Figure 3). The decline from >60% to below 20% total land pollen in less than 100mm can be estimated from the age-depth model (Figure 43) to have taken place between 13,690–13,350 cal BP (95% probability; Figure 44a) and 13,540–13,270 cal BP (95% probability; Figure 44b) over an interval of 30–240 years (95% probability; Figure 44c).

The age-depth model can also be used to provide a chronology for the reconstructed temperature changes derived from the fossil insect assemblage (Atkinson et al. 1987; Elias and Matthews 2013). Figure 45 illustrates that Late Pleistocene temperatures at Gransmoor oscillated rapidly on a large scale.

Finally, an Upper Palaeolithic antler barbed point embedded in a piece of wood was recovered at a depth of 1.69m in the sediment sequence (Sheldrick et al. 1997; Figure 46). This artefact was not directly dated as it was considered too valuable to be sub-sampled. The age-depth model suggests that this artefact was deposited in 13,360–13,110 cal BP (95% probability; 1.69m).

The sequence provided by the stratigraphy at Gransmoor assures strong prior beliefs for the construction of the age-depth model and highlights the importance of sequence in the construction of Bayesian models.

Back to contents | Scientific dating methods | Practicalities