Full Terms & Conditions of access and use can be found athttp://www.tandfonline.com/action/journalInformation?journalCode=uica20

Download by: [University of St Andrews] Date: 24 November 2017, At: 01:58

The Journal of Island and Coastal Archaeology

ISSN: 1556-4894 (Print) 1556-1828 (Online) Journal homepage: http://www.tandfonline.com/loi/uica20

Chronostratigraphy of an Eroding Complex AtlanticRound House, Baile Sear, Scotland

Tim Kinnaird, Tom Dawson, David Sanderson, Derek Hamilton, AlanCresswell & Rebecca Rennell

To cite this article: Tim Kinnaird, Tom Dawson, David Sanderson, Derek Hamilton, AlanCresswell & Rebecca Rennell (2017): Chronostratigraphy of an Eroding Complex AtlanticRound House, Baile Sear, Scotland, The Journal of Island and Coastal Archaeology, DOI:10.1080/15564894.2017.1368744

To link to this article: https://doi.org/10.1080/15564894.2017.1368744

© 2017 The Author(s). Published withlicense by Taylor & Francis Group, LLC© TimKinnaird, Tom Dawson, David Sanderson,Derek Hamilton, Alan Cresswell, andRebecca Rennell

View supplementary material

Published online: 07 Nov 2017.

Submit your article to this journal

Article views: 43

View related articles

View Crossmark data

The Journal of Island & Coastal Archaeology, 0:1–15, 2017Copyright C© 2017 Tim Kinnaird, Tom Dawson, David Sanderson, Derek Hamilton,Alan Cresswell, and Rebecca RennellISSN: 1556-4894 print / 1556-1828 onlineDOI: 10.1080/15564894.2017.1368744

Chronostratigraphy of an ErodingComplex Atlantic Round House,Baile Sear, Scotland

Tim Kinnaird,1,2 Tom Dawson,3 David Sanderson,1

Derek Hamilton ,1 Alan Cresswell,1 and Rebecca Rennell41Scottish Universities Environmental Research Centre, University of Glasgow,

East Kilbride, UK2School of Earth and Environmental Sciences, University of St. Andrews,

St. Andrews, UK3School of Art History, University of Saint Andrews, Saint Andrews, UK4Archaeological Institute, University of the Highlands and Islands – Lews

Castle College, Stornoway, UK

ABSTRACT

A high-resolution chronostratigraphy has been established for aneroding Atlantic round house at Sloc Sàbhaidh (North Uist, Scotland),combining detailed OSL profiling and dating of sediments encom-passing the main bracketing events associated with the monument,radiocarbon AMS dates on bone recovered from excavated featuresand fills within it, and TL dates on pottery and burnt clay. ConcordantOSL and radiocarbon evidence place construction of the wheelhousein the first to second centuries AD, contemporary with dates fromthe primary occupation. Beneath the wheelhouse, clay depositscontaining burnt material, attest to cultural activity in vicinity tothe monument in the preceding second to first centuries BC. At alater date, the southern wall collapsed, was rebuilt, and the interiorspaces to the monument re-structured. The chronology for the laterhorizons identified from the sediment luminescence dates extends tothe second half of the first millennium AD, which goes beyond the

Received 22 February 2017; accepted 26 July 2017.Address correspondence to Tim Charles Kinnaird, School of Earth and Environmental Sciences,University of St. Andrews, St. Andrews KY16 9AL, UK. E-mail: tk17@st-andrews.ac.ukColor versions of one or more of the figures in the article can be found online atwww.tandfonline.com/uica.This is an Open Access article distributed under the terms of the Creative Commons Attribution-NonCommercial-NoDerivatives License (http://creativecommons.org/licenses/by-nc-nd/4.0/), whichpermits non-commercial re-use, distribution, and reproduction in any medium, provided the originalwork is properly cited, and is not altered, transformed, or built upon in any way.

1

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Tim Kinnaird et al.

range of the radiocarbon dates obtained. The data from ceramicsencompass both periods. The juxtaposition of the dating evidenceis discussed relative to short and longer chronologies for this IronAge monument. Corollaries of this research are the implications thatbased on the long chronology, some of the ecofacts (bone) appearto be residual, and that the temporal duration of Hebridean CoarseWare may extend into the second half of the first millennium AD.

Keywords chronology, geoarchaeology, household archaeology

INTRODUCTION

This paper documents the chronostratigra-phy of an eroding complex Atlantic roundhouse at Sloc Sàbhaidh, situated on the westcoast of Baile Sear, a tidal island that lies0.5 km off the south-west coast of NorthUist (Scotland, UK). This monument was ex-posed during a storm in 2005, necessitat-ing the need for archaeological excavationsand intervention. The monument survivedas a stone structure with an internal diame-ter of ca. 12 m and ca. 1.5 m height, with acircular external wall, internally divided bypiers/slabs.

The geomorphology of the west coastof Baile Sear is complex, resulting from thecontinual re-working of sediment stored onthe gently sloping continental shelf that ex-tends into the Atlantic (Gilbertson et al.1996; Hall 1996), coupled with the gradualHolocene rise in sea level since deglacia-tion. Together, these processes have pro-duced the distinctive machair grassland, aplain of fertile calcareous soils that bor-ders the west coast. This highly dynamicenvironment is subject to constant erosionand deposition, highlighting the challengesassociated with dating island and coastalsettlements in this special issue (Krus andThompson).

BACKGROUND

Settlement on the machair

A number of long-lived prehistoricsettlement mounds have been identified, in-dicating that the fertile machair grasslandshave formed a focus for human occupa-tion since the mid third millennium BC

(Parker Pearson et al. 2004:43). To date,the only identified evidence for prehistoricsettlements on the Baile Sear machair hasbeen found at Sloc Sàbhaidh and CeardachRuadh. The antiquarian, Erskine Beveridge(2001) was the first person to record bothCeardach Ruadh and Sloc Sabhaidh, notingthe presence of midden deposits and arange of artifacts, although he made nomention of structures. Fairhurst and Ritchie(1963) suggested that masonry revealed atCeardach Ruadh indicated the presence of awheelhouse, a complex form of an Atlanticroundhouse, in which a radial arrangementof interior stone piers form the basis forlintel arches to support a corbelled roof.In 1983, Barber (2003), documented froman eroding section at Ceardach Ruadh,the presence of a cultivation horizon thatproduced a radiocarbon date in the LateBronze Age, overlain by extensive middendeposits and domestic structures datingto the Middle to Late Iron Age. Althoughthe excavation did not clarify the nature ofthe settlement, the structural remains mayrepresent one or more wheelhouses. Else-where, in the Hebrides, similar structuresthat have been dated tend to fall within theperiod 25 BC to 380 AD.

Excavation at Sloc Sàbhaidh

The site at Sloc Sàbhaidh remainedunexplored until a storm in 2005 erodedthe coast edge, revealing two drystonestructures. The discovery was reported bylocal heritage group, Access Archaeology,who teamed up with The SCAPE Trust andthe University of St Andrews to record themasonry. Due to the rapid rate of erosion—the coast edge retreated 3 m between

2 VOLUME 0 • ISSUE 0 • 2017

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Chronostratigraphy of a Atlantic round house in Scotland

August and December 2005 (Stentoft et al.2007)—a project to record the two struc-tures commenced in 2006 and continueduntil 2010 (Dawson 2010; MacDonald andMcHardy 2008; Rennell and McHardy 2009;Stentoft et al. 2007).

This manuscript details the excavationand chronology of the better preserved ofthe two structures. The excavation of thewheelhouse at Sloc Sàbhaidh, Baile Sear,was conducted over four seasons between2006 and 2010, and revealed a circular dry-stone wheelhouse with an internal diameterof approximately 10 m. The western part ofthe building (closest to the sea) was suffer-ing active erosion, but much of the struc-ture and associated archaeological depositswere well-preserved, and half of the build-ing was buried within the coastal dune. Ex-cavation focused on elements located onthe beach, partly due to concerns from thecrofters that disturbance of the dune wouldlead to increased erosion in the future. Thismeant that only one half of the structurewas excavated, resulting in a large sectionthrough the deposits being visible to theeast of the excavation.

The excavations revealed a typicalwheelhouse with radial piers and centralhearth. A ‘floor’ (comprising micro layersof burned material mixed with sand), cov-ered a number of pits containing animalbone and pottery, thought to be evidence ofthe primary occupation. At a later date, thesouthern wall collapsed, resulting in struc-tural modification to the building. The re-built wall covered at least one of the wheel-house piers on the southern side, prevent-ing the structure from functioning as a truewheelhouse. Either at the same time orsubsequent to the rebuild, sand was de-posited over the central hearth and south-ern part of the building, placing this halfof the structure out of use. The sand wasnot horizontally bedded, and numerous tipmarks were evident. Occupation continuedin the northern half of the building, and aseries of superimposed hearths and associ-ated deposits were excavated. One of thesehearths, located in the central space, wasdecorated by two intersecting sets of linesin the form of a cross. Evidence of the final

abandonment of the structure is providedby layers of clean aeolian sand.

OSL PROFILING AND DATINGOF SEDIMENT

The site was visited by David Sandersonin September 2010 to collect samples foroptically stimulated luminescence (OSL)profiling and dating. The sediments ex-posed in the section, associated with themonument and its immediate environment,were reviewed and luminescence profilingwas undertaken to evaluate the archaeolog-ical sequences (Figures 1 and 2; see alsoKinnaird et al. 2015, 2017; Sanderson andMurphy 2010), and to identify sampling po-sitions for OSL dating (see supplementarydata S1). At the time of this visit, the floorlayers had been revealed, but excavationhad not yet started, restricting access toearly occupational layers for sampling. Thesequence of pre-existing deposits, construc-tion of the southern and northern walls, thesouthern wall collapse and associated re-build, abandonment of the southern areas,the occupation sequence of the northernareas and its subsequent abandonment,were covered by seven profiles, comprising58 profiling samples and 11 tube samples.

Profiles 1 and 2 examined substratesbelow the wheelhouse structure, and anunderlying burnt earth surface which maybe associated with earlier cultural activityin vicinity of the wheelhouse. The datingquestions associated with these materialsrelate to the age of the formation of theburnt surface as well as providing a termi-nus post quem (TPQ) for the constructionof the wheelhouse walls. Profiles 3 and 4examined the stratigraphy of sediments im-mediately outside the southern wall of thewheelhouse, which may be associated withthe original wall construction, a collapse ofthe southern wall, and its subsequent re-building. The dating questions here relateto the timing of the original constructionand/or collapse (evidence provided by a cutfeature seen in section against the southernwall) and rebuilding events relative to theoccupation evidence within the southern

THE JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 3

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Tim Kinnaird et al.

Figure 1. a) Location of Baile Sear (Baleshare), North Uist; b) aerial view of northern part of site,showing upstanding piers [642] and [1018]. Floor deposits not fully excavated; c) planof wheelhouse, showing spatial distribution of OSL samples. Images b) and c) copyrightSCAPE.

chamber. Profile 5 examined the backfilledmaterials in the southern chamber, leadingto the deliberate abandonment of this partof the wheelhouse. The chronology of thisabandonment relative to the wall rebuildis of interest. Profiles 6 and 7 examinedthe infill of the central north and northerninterior areas, including occupational evi-

dence, a later sequence of hearths (whichappeared to be constructed after the back-filling event), and aeolian sands depositedafter the final abandonment.

The results from field profiling were in-formative, with the substrate, archaeolog-ical materials, clean aeolian sands (post-dating abandonment) and modern sands,

4 VOLUME 0 • ISSUE 0 • 2017

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Fig

ure

2.

Ea

stfa

cin

gse

ctio

nil

lust

rati

ng

dep

osi

tsw

ith

inth

ew

hee

lhou

se,w

ith

hea

rth

dep

osi

tsbet

wee

npie

r[1

01

8]

an

dre

vetm

ent

wa

ll[1

04

0]

(Fig

ure

3)

an

ddu

mped

san

dla

yers

toth

eso

uth

(Fig

ure

4).

Pro

file

1is

posi

tion

eda

tO

SL1

on

Figu

re1

.Copy

righ

tSC

AP

E.

all showing promising contrasts in lumi-nescence intensities and depletion ratios(Figure 3; Table S1.1). Moreover, as theseinitial measurements were undertaken onsite and undertaken in near real-time,dating priorities were coupled directly witharchaeological objectives, and relative tem-poral frameworks were constructed forthe sedimentary stratigraphies (Figures 3and 4). The investigated substrate stratigra-phies to the monument, including the sandswhich cover the burnt surface, all representrelatively short temporal sequences. The dy-namic ranges in OSL and IRSL net signal in-tensities for the sands directly beneath thestructure (TPQ for construction) and for the‘floor’ layers within it, are not suggestive ofa substantial temporal break (tabulated inS1). In contrast, the dynamic range in inten-sities between the substrate layers and sedi-ments associated with the re-packing of thesouthern wall after collapse are suggestiveof a temporal break (Figure 3).

With the relative temporal frame-work established, which agreed with therecorded archaeological stratigraphic se-quence, the investigations progressed tomore formal laboratory analysis, first toexploratory luminescence screening andcharacterization, then to more conventionalquantitative quartz OSL dating. Samplepreparation and analysis was undertaken atthe Scottish Universities Environmental Re-search Centre (SUERC). For each profilingposition, paired analyses were undertakenon quartz and polymineral extracts, using asimplified two-step single aliquot regener-ative dose (SAR) OSL protocol (Table S2.1).The data were used to plot depth-doseand depth-sensitivity profiles for eachsection (Figures 3 and 4). These analyseslargely corroborated the hypotheses raisedduring field profiling—strengthening theargument for a prolonged (and potentiallymulti-period) occupation of the wheel-house, but also confirmed that most OSLsamples had failed to capture the primaryoccupational layers.

OSL SAR dating utilizes extractedquartz from the samples to determinethe radiation dose experienced by thesediments since their last zeroing event,

THE JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 5

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Tim Kinnaird et al.

Figure 3. Sequence of deposits south of the wheelhouse; a) east facing section; b) photograph oflower sequence, with profiling positions marked; c) field- and laboratory- luminescenceprofiles, annotated relative to stratigraphy, with main observations noted. Key to symbols(here, and for Figure 4): Red circles, net IRSL signal intensities and depletion indices (inred), Blue circles, net OSL signal intensities and depletion indices (in blue).

assumed to be by exposure to light priorto final deposition. This is combined withdose rate analysis based on field and lab-oratory measurements of environmental

radioactivity. The age is determined asthe ratio of dose divided by dose rate. Inthis work, dose rates for the bulk sedi-ment were quantified using high resolution

6 VOLUME 0 • ISSUE 0 • 2017

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Chronostratigraphy of a Atlantic round house in Scotland

Figure 4. Hearth deposits between pier [1018] and revetment wall [1040]; a) east facing section;b) photograph of sequence, with profiling positions marked; c) field- and laboratory-luminescence profiles, annotated relative to stratigraphy, with main observations noted.

gamma spectrometry (HRGS) and thicksource beta counting (TSBC) in the lab-oratory, coupled with water contentanalysis and in situ gamma dose ratemeasurements (Tables S3.1–3.3). Quartzwas extracted using standard laboratory

procedures, and purity checked with scan-ning electron microscopy (SEM). Equiva-lent doses were determined, initially, on16 aliquots of quartz per sample using thequartz OSL SAR procedure, with additionalinvestigation of further sets of 16 to build

THE JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 7

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Tim Kinnaird et al.

Table 1. Quartz SAR OSL ages for sediments associated with the monument and theimmediate environment.

SUTL Field Dose Rate/ Stored Calendar

No. No. Profile mGy a−1 Dose/Gy Years BP years

2392 OSL1 1 1.08 ± 0.09 2.54 ± 0.13 2350 ± 230 340 ± 230 BC

2393 OSL2 2 1.21 ± 0.09 2.24 ± 0.10 1855 ± 160 AD 160 ± 160

2394 OSL3 2 1.29 ± 0.06 2.45 ± 0.20 1895 ± 180 AD 110 ± 180

2395 OSL4 3 1.24 ± 0.06 1.55 ± 0.02 1245 ± 60 AD 765 ± 60

2397 OSL6 5 1.02 ± 0.06 1.47 ± 0.02 1240 ± 75 AD 770 ± 75

2398 OSL7 5 1.17 ± 0.06 1.51 ± 0.05 1290 ± 75 AD 720 ± 75

2399 OSL8 6 1.10 ± 0.06 1.44 ± 0.07 1313 ± 95 AD 695 ± 95

2400 OSL9 6 1.10 ± 0.06 1.18 ± 0.03 1070 ± 70 AD 940 ± 70

2401 OSL10 6 1.03 ± 0.06 1.12 ± 0.03 1085 ± 70 AD 925 ± 70

2402 OSL11 7 1.10 ± 0.08 2.05 ± 0.22 1875 ± 250 AD 135 ± 250

statistical power for selected pre-heatinggroups (Tables S4.1–4.2). Radial plottingmethods were used to appraise samplehomogeneity, and robust statistics wereused, for aliquots satisfying SAR acceptancecriteria to estimate equivalent doses for agedeterminations.

The individual sediment ages fall intothree groups (Table 1): an earlier set inthe first to second century AD, for thesands beneath the wheelhouse (TPQ forconstruction; SUTL2393-94, 2402); a laterset in the sixth to seventh century AD, forthe sands immediately outside the south-ern wall, leading to the interpretation thatthey are associated with a collapse of thesouthern wall and its subsequent re-build;and the backfilled materials in the southernchamber (dating the re-modelling of theinternal space; SUTL2395, 2397–98); andfinally, a set in the ninth to tenth cen-turies AD, for the last hearth to be fired inthe northern chamber (SUTL2400), and forthe post-abandonment sands (terminus antequem (TAQ) for abandonment; SUTL2401).It was possible to augment (and extend)the sediment chronologies, by returning tothe apparent doses obtained from the ex-ploratory laboratory analyses, and calculat-ing an apparent age using the dose rate esti-mates from the adjacent dating positions.

Moreover, the combination of thequartz OSL SAR ages and profiling dating,

assessed relative to the stratigraphy ofthe wheelhouse, allows for both conven-tional statistical assimilation and Bayesianmethods. Bayesian models have been usedin radiocarbon dating for several years(e.g., Buck et al. 1996) to evaluate datesin sequences that are constrained strati-graphically, interpret these dates in termsof archaeological events, and to reduceage estimate uncertainties (see data S5).The model parameters are presented as amatrix in Figure S5.1. TPQ for constructionis provided by the 4 quartz SAR ages frombeneath the structure (SUTL2392-94, 2401;Figure S5.1). The stratigraphic relationshipbetween the sands sampled beneath thenorthern wall and the burnt surface isnot known. A single sample constrainsthe timing of the cut feature observed insection outside the southern wall, whichhas been interpreted through this datingmodel as relating to the collapse and sub-sequent re-build of this wall (SUTL2395).This event is further constrained by twosamples from within the southern interiorspace, from sediments deliberately back-filled, related to first modification, thenabandonment of this space (SUTL2397-98).The southern wall needs to have been inplace before this space was backfilled, asthe sediments abut this wall. The northerninterior space remained in use after thesouthern chamber was abandoned and two

8 VOLUME 0 • ISSUE 0 • 2017

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Chronostratigraphy of a Atlantic round house in Scotland

samples constrain the period of occupationof the central chamber (SUTL2399-2400).A single sample from clean aeolian sandspost-dates abandonment (SUTL2401) andprovides TAQ for abandonment of thewheelhouse.

Two Bayesian models were investi-gated: model A was populated with thequartz OSL SAR ages (Figure S5.2; Amodel

= 86; Aoverall = 94)); model B was addi-tionally populated with age estimates ob-tained from laboratory profiling (Figure 5 orS5.3; Amodel = 74). Under Model B, the mostprobable duration for accumulation of theburnt surface underneath the wheelhouseconstruction is 240 BC to 80 AD (95% prob-ability) and most probably in 160 BC to5 BC (68% probability; Figure S5.4). Tak-ing account of the model B constraints forthe samples immediately underneath thewheelhouse southern and northern walls,in combination with the constraints im-posed by the underlying burnt surface, weinfer a most probable date for constructionof the wheelhouse between 140 BC to 240AD (95% probability) and most probablybetween 60 BC to 140 AD (68% probabil-ity). The period in which the southern wallcollapsed, was re-built, and by inference thesouthern chamber abandoned, with deliber-ate backfilling, is dated to 330 to 790 (95%probability) and most probably, 475 to 740AD (68% probability). The wheelhouse hadbeen abandoned by 825 to 995 AD (95%probability), and most probably by 870 to950 AD (68% probability).

RADIOCARBON DATING

The archaeological excavations were com-pleted by October 2010. In post-excavation,12 fragments of bone collected from con-texts attributed to primary occupation (foursamples; floor layers, and from pits sealedby these), the hearth sequence in the cen-tral north chamber (two samples), the dis-turbed sediments outside the southern wall(three samples), and the packed materials inthe southern chamber (two samples) weresubmitted to the SUERC radiocarbon facili-ties for AMS dating (Table 2).

The ‘floor’ layers within the wheel-house provide the evidence for the primaryoccupation. These layers, located across theentire internal area (apart from the west-ern edge, which was truncated by erosion),consisted of numerous micro layers of sandmixed with burnt material. It was not pos-sible to follow any one particular ‘floor’, sothe deposits were excavated in 50 mm deepspits by 1 m grid squares. The calibrated ra-diocarbon dates of bone recovered from thefloor layers at 58 to 217 cal AD and 121 to254 cal AD (at 95%, Table 2), and from pitssealed by these horizons, at 129 to 258 calAD and 137 to 345 cal AD (at 95%, Table 2),constrain primary occupation to the secondto third centuries AD.

In the central northern chamber, radio-carbon dates were obtained from bone fromthe layers of peat ash deposited around theseries of superimposed hearths that were inuse after the abandonment of the southernpart of the building. At 241 to 390 cal ADand 242 to 391 cal AD (at 95%, Table 2),these document occupation in the third tofourth centuries AD. The radiocarbon sam-ples are from contexts above and beneaththe OSL samples (SUTL2399 & 2400).

The radiocarbon constraints on the re-structuring of southern part of the wheel-house, are obtained in part from the fill ofthe construction trench/robber cut, and inpart, the backfill of the southern chamber.The radiocarbon dates obtained for bonesenclosed within the fill of the cut featureoutside the southern wall at 753 to 412 calBC, 795 to 541 cal BC and 85 to 239 cal AD(at 95%, Table 2), are not thought to datethe cut, rather to represent residual mate-rials, most probably derived from the mid-den deposits which abut the southern wall.However, they do provide a TPQ for thisevent. In the southern chamber, the bonesfrom within the dumped materials dated at215 to 387 cal AD and 254 to 400 cal AD (at95%, Table 2), would suggest, if these mate-rials were also not residual, that the south-ern interior space was abandoned by thefourth century AD.

Therefore, purely on the radiocarbonevidence, the site history of the Baile Searwheelhouse could be interpreted in termsof a shorter chronology, spanning from

THE JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 9

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Tim Kinnaird et al.

Figure 5. Bayesian age model B.

10 VOLUME 0 • ISSUE 0 • 2017

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Tab

le2

.A

MS

dat

esfr

om

bo

ne

reco

vere

dfr

om

the

Bai

leSe

arw

hee

lho

use

.

δ1

3C

δ1

5N

Cal

ibra

ted

rad

ioca

rbo

nd

ate

rela

tive

tore

lati

veto

C/N

rati

oE

qu

ival

ent

Lab

ref

Co

nte

xt

Age

BP

VP

DB

Air

(mo

lar)

toat

1si

gma

(68

.2%

%)

at2

sigm

a(9

5.4

%)

SUER

C-

4559

610

3917

53±

29−

21.1

4.3

3.4

>P

5/5

&O

SL6

274

to33

4ca

lAD

(48.

9%)

241

to26

5ca

lAD

(19.

3%)

215

to38

7ca

lAD

(95.

4%)

SUER

C-

4559

710

5317

12±

27−

21.5

4.7

3.3

=P5/

4,<

OSL

632

3to

384

calA

D(4

8.7%

)26

0to

284

calA

D(1

9.5%

)25

4to

400

calA

D(9

5.4%

)

SUER

C-

4559

410

3818

44±

29−

21.2

4.5

3.3

=P3/

1313

0to

215

calA

D(6

8.2%

)85

to23

9ca

lAD

(95.

4%)

SUER

C-

4559

510

4924

54±

25−

20.6

4.9

3.3

=P3/

9,>

OSL

474

7to

688

calB

C(2

5.7%

)55

4to

503

calB

C(2

0.5%

)66

5to

645

calB

C(8

.0%

)44

1to

418

calB

C(7

.1%

)46

2to

450

calB

C(3

.4%

)

597

to41

2ca

lBC

(54.

5%)

753

to68

8ca

lB

C(2

7.5%

)66

7to

632

calB

C(1

1.2%

)

SUER

C-

4559

810

5625

26±

29−

21.2

6.1

3.3

P3/

4–6

642

to59

2ca

lBC

(27.

3%)

785

to74

9ca

lBC

(22.

5%)

688

to66

6ca

lBC

(14.

7%)

577

to56

9ca

lBC

(3.7

%)

694

to54

1ca

lBC

(68.

8%)

795

to72

7(2

9.9%

)

SUER

C-

4560

512

2617

29±

29−

21.5

4.4

3.3

=P6/

3,<

OSL

925

5to

306

calA

D(4

0.5%

)31

2to

346

calA

D(2

5.6%

)24

1to

390

calA

D(9

5.4%

)

SUER

C-

4560

612

2817

28±

29−

21.4

5.0

3.3

=P6/

2&

OSL

825

5to

305

calA

D(3

8.6%

)31

2to

327

calA

D(2

5.4%

)37

0to

377

calA

D(4

.2%

)

242

to39

1ca

lAD

(95.

4%)

SUER

C-

4559

911

1918

89±

29−

21.1

5.2

3.3

PO

68to

135

calA

D(6

8.2%

)58

to21

7ca

lAD

(95.

8%)

SUER

C-

4560

311

2318

27±

29−

21.1

5.1

3.3

PO

137

to22

5ca

lAD

(68.

2%)

121

to25

4ca

lAD

(92.

7%)

87to

104

cal

AD

(2.3

%)

SUER

C-

4559

312

1018

08±

27−

21.1

5.0

3.2

PO

208

to24

2ca

lAD

(31.

0%)

166

to19

6ca

lAD

(23.

1%)

139

to15

9ca

lAD

(14.

1)

129

to25

8ca

lAD

(90.

8%)

299

to31

9ca

l(4

.6%

)

SUER

C-

4558

910

1717

70±

29−

21.2

4.8

3.2

P7/

1−/3

228

to26

3ca

lAD

(29.

5%)

278

to32

9ca

lAD

(38.

7%)

137

to34

5ca

lAD

(95.

4%)

Abb

revi

atio

ns

asfo

llow

s:>

abo

ve,<

bel

ow

;PO

=P

rim

ary

occ

up

atio

n.

THE JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 11

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Tim Kinnaird et al.

the first to second centuries AD to thefourth century AD, with re-structuring ofthe southern interior space by the secondto third century AD, its abandonment, andthen restricted occupation in the centralnorth and northern areas into the fourthcentury AD.

TL DATING OF HEATED MATERIALS

The sediment samples and chronology donot directly register the primary occupa-tion, which appears as a temporal gapin the Bayesian sequences of both theOSL dating samples and the generalizedchronology from the profiling samples(these archaeological layers were unexca-vated at the time of OSL sampling). Theevidence for primary occupation includesradiocarbon dating samples from primarycontexts and Hebridean Coarse Ware, tradi-tionally attributed to the first half of the firstmillennium AD. Some of this material how-ever appeared to come from parts of thesite which post-date the step in sedimentchronologies (Figure 5; although these lay-ers also produced radiocarbon dates indica-tive of a shorter temporal sequence).

Here, luminescence methods are ap-plied, to determine the period of lastheating of fragments of a ceramic hearthassociated with the lower part of the‘hearth sequence’ in the central chamber,and a series of pottery sherds taken fromadjacent units. Initially, 8 sub-samples offragmentary material from the clay hearth,and 14 pottery sherds were examined (ofwhich 11 were taken through dating proce-

dures). In this study, the sherd size limitedthe yields of quartz available from interiorportions, and also the quantity of materialavailable for dose rate determinations toless than the 20 g de-minimus sample sizewhich is normally used for beta and gammaanalysis at SUERC. Accordingly, procedureswere adapted to permit 10 g radionuclideanalyses where needed. Quartz SAR inves-tigations from the sherds were limited toa single pre-heating temperature, basedon the preheating data obtained from thesediments from this site. The quartz OSLresults from both the hearth material andthe sherds produced unsatisfactory results,with apparent ages from the last 500 to1,000 years and failures of dose recoverytests in the laboratory under the pre-heatingregime used for the ceramics. Accordingly,the feldspar extracts were also investi-gated using a series of different stimulationand quantification approaches. Multi-stimulation single aliquot regenerative-additive (SARA) methods were used withprolonged 16 hr preheating at 120 °C onthe natural cycle to stabilize fading (seesupplementary data S6). Within the avail-able precision, the majority of these datasupport the long chronology of the sedi-ment sequence and imply that coarse warecomes from the later first millennium AD.

Subsequently, the investigations wereextended, to include five sherds recov-ered from pits in the wheelhouse floor inareas adjacent to the lower sedimentarysequences in the central and northernchambers. Further studies were also un-dertaken using quartz extracts from the ce-ramic hearth to assess the potential of using

Table 3. Feldspar TL SARA results: Total dose rates, stored dose and age estimates.

SUTL from Pottery Dose rate/ Stored Years Calendar

No. context ID mGy a−1 dose/Gy BP years

2787 1017 P685 2.08 ± 0.1 2.89 ± 0.49 1.4 ± 0.3 AD 620 ± 250

2788 P713 2.11 ± 0.1 2.75 ± 0.17 1.3 ± 0.1 AD 720 ± 100

2789 P716 2.12 ± 0.1 2.76 ± 0.32 1.3 ± 0.2 AD 710 ± 160

2790 1123 P1636 2.46 ± 0.1 3.14 ± 0.31 1.3 ± 0.1 AD 740 ± 140

2791 1167 P1218 2.49 ± 0.1 3.59 ± 0.11 1.4 ± 0.1 AD 580 ± 70

12 VOLUME 0 • ISSUE 0 • 2017

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Chronostratigraphy of a Atlantic round house in Scotland

Table 4. Quartz OSL SAR results: Total dose rates, stored dose and age estimates.

SUTL From Pottery Dose rate/ Stored/ Years Calendar

No. context ID mGy a−1 Gy BP years

2787 1017 P685 2.01 ± 0.10 2.79 ± 0.07 1.4 ± 0.1 AD 630 ± 80

2788 P713 2.05 ± 0.10 2.70 ± 0.10 1.3 ± 0.1 AD 700 ± 80

2789 P716 2.06 ± 0.09 2.60 ± 0.05 1.3 ± 0.1 AD 750 ± 60

2790 1123 P1636 2.40 ± 0.11 2.82 ± 0.04 1.2 ± 0.1 AD 840 ± 60

2791 1167 P1218 2.43 ± 0.09 3.26 ± 0.05 1.3 ± 0.1 AD 680 ± 50

2651 clay hearth — 2.12 ± 0.12 3.53 ± 0.04 1.7 ± 0.1 AD 350 ± 90

2653 — 2.12 ± 0.12 3.16 ± 0.05 1.5 ± 0.1 AD 520 ± 90

2654 — 2.04 ± 0.15 3.64 ± 0.05 1.8 ± 0.1 AD 230 ± 140

more aggressive pre-heating to resolve theunsatisfactory quartz behavior observed in2014. With higher pre-heating temperaturesof 250–280 °C satisfactory dose recoverybehavior was observed in this material, andalso in the later sherds. For these materials,both quartz OSL SAR and feldspar SARAanalyses were undertaken, and yieldedluminescence dates (Tables 3 and 4). Forthe hearth, the individual ages (after refine-ment of the dating protocol), fall withinthe period AD 350 ± 100 (SUTL2651), AD520 ± 90 (SUTL2653) and AD 230 ± 140(SUTL2654), with statistical combinationssuggesting a date of AD 410 ± 60. It mustbe noted that these dates are from undec-orated fragments of the hearth; as such, wecan only indirectly relate these to the deco-rated central portion of the hearth, with theassumption that they were well-fired andthe luminescence signals reset. The datesfall within the period between constructionand early occupation, and later occupationassociated with the modification of theinternal space of the wheelhouse. Broadlyconcordant quartz and K feldspar lumines-cence ages were obtained for the potterysherds, with individual quartz OSL SAR agesin the range sixth to seventh centuries AD,and K feldspar post-IR TL SARA ages in therange fifth to seventh centuries AD.

DISCUSSIONS AND CONCLUSIONS

All chronological evidence has been as-sessed relative to the stratigraphy of the

wheelhouse and its immediate environ-ment using conventional assimilation andBayesian statistical methods, in combina-tion with our archaeological observations,and the field- and laboratory-luminescenceprofiling data.

For the early site history, the OSL andradiocarbon chronologies are concordant,with construction of the wheelhouse inthe first to second century AD, contem-poraneous with the primary occupation.Moreover, both provide evidence of ear-lier cultural activity in vicinity to themonument. Beneath the wheelhouse, claydeposits containing burnt material, attestto cultural activity nearby in the second tofirst century BC. Midden deposits whichabut the exterior walls of the wheelhousecontain materials dated from the earlyIron Age onwards, and it is known thatother structures, possibly of earlier date, lieadjacent to the wheelhouse.

Thereafter, the chronological evidencefor the later site history is more ambigu-ous, with the radiocarbon evidence suggest-ing a shorter chronology, and the OSL sedi-ment chronologies and TL dates on pottery,a longer chronology.

The short chronology. The site historyof the wheelhouse spans from its construc-tion in the first to second centuries ADto abandonment after the fourth centuryAD. In this temporal framework, the col-lapse of the southern wall, its rebuild, andthen subsequent abandonment of the south-ern interior space all occurred in the sec-ond to third century AD, with occupation

THE JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 13

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Tim Kinnaird et al.

limited to the central north and northern ar-eas thereafter.

The long chronology. In this temporalframework, the collapse of the southernwall, its re-build, and the subsequent mod-ification of the southern interior space,including its deliberate backfilling andabandonment, all occurred in the later partof the first millennium AD, significantly af-ter the wheelhouse construction and earlierphases of occupation. After abandonmentof the southern areas, the central north andnorthern areas remained in use, with oc-cupation through the seventh to the ninthcentury AD. The wheelhouse must havebeen abandoned shortly after, as registeredby the first occurrence of clean, aeoliansands starting from the ninth century AD.Later evidence of aeolian activity throughthe Little Ice Age (up to sixteenth centuryAD) is also recorded in post-abandonmentsands.

Apparent ages from OSL profiling aug-ment this temporal framework, suggestingthat the northern chamber was in useduring all period of occupation, from theIron Age construction of the wheelhouse(ca. AD 150) through to its abandonment.Evidence for the earliest occupation phasein the southern chamber, is only regis-tered by a single profiling sample from the‘floor’ layers (first to second century AD;profile 5).

Implications of the longer chronology,are that some of the artifacts appear tohave been subject to post-depositionalmovements. This is true of both the boneand pottery, dated here; although, the bonemust be residual. Notwithstanding this,the temporal duration of Hebridean CoarseWare must extend well into the secondhalf of the first millennium AD. The ob-servations in respect of ceramic traditions,coupled with the diversity of mineralsobserved by electron microscopy, suggeststhat further work is needed to examine theduration of ceramics on other sites in theHebrides.

Whilst a detailed chronostratigraphyhas been established for the constructionand early occupational sequences of thewheelhouse at Sloc Sàbhaidh, questions still

remain as to the age and duration of thelater occupational sequences in the cen-tral north and northern areas, and the re-structuring of the interior space. The datingproject has created two sequences, raising anumber of questions that are still being dis-cussed and showing that the dating of whatwas thought to be a well-understood monu-ment type can be problematic in sand duneenvironments.

ACKNOWLEDGEMENTS

The excavation team would like to thankHistoric Scotland (now Historic Environ-ment Scotland) and the University ofSt Andrews for providing funding. Thisproject would not have been possible with-out the enthusiasm and support of the Ac-cess Archaeology group and a number ofother residents of North Uist. Thanks aredue to North Uist Estates, and especially tothe MacDonalds of Crois Moraig, on whosecroft the excavation took place.

ORCID

Derek Hamilton http://orcid.org/0000-0003-4019-3823

REFERENCES

Barber, J. 2003. Bronze Age Farm and Iron AgeFarm Mounds of the Outer Hebrides. Edin-burgh: The Society of Antiquaries of Scotland.

Beveridge, E. 2001 [1911]. North Uist: Its Ar-chaeology and Topography with Notes uponthe Early History of the Outer Hebrides. Edin-burgh: Birlinn.

Buck, C. E., W. G. Cavanagh, and C. D. Litton.1996. Bayesian Approach to Interpreting Ar-chaeological Data. Chichester: John Wiley &Sons, Ltd.

Dawson, T. 2010. Baile Sear Community Ex-cavation. Sloc Sàbhaidh, Baile Sear, NorthUist, Season 4. Data Structure Report. Unpub-lished report. St. Andrews: The SCAPE Trustand the University of St. Andrews.

14 VOLUME 0 • ISSUE 0 • 2017

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017

Chronostratigraphy of a Atlantic round house in Scotland

Fairhurst, H., and W. Ritchie. 1963. Baleshare Is-land, North Uist. Discovery and Excavationin Scotland: 31.

Gilbertson, D., J. Grattan, and J.-L. Schwenninger.1996. A stratigraphic survey of the Holocenecoastal dune and machair sequences. In TheOuter Hebrides: The Last 14,000 Years(D. Gilbertson, M. Kent, and J. Grattan,eds.):72–101. Sheffield: Sheffield AcademicPress.

Hall, A. 1996. Quaternary geomorphology of theOuter Hebrides. In The Outer Hebrides: TheLast 14,000 Years (D. Gilbertson, M. Kent, andJ. Grattan, eds.):5–12. Sheffield: Sheffield Aca-demic Press.

Kinnaird, T. C., J. Bolós, S. Turner, and A.Turner. 2017. Optically-stimulated lumi-nescence profiling and dating of historicagricultural terraces in Catalonia (Spain).Journal of Archaeological Science 78:66–77.

Kinnaird, T. C., D. C. W. Sanderson, andG. Bigelow. 2015. Feldspar SARA IRSLdating of very low dose rate aeolian sed-iments from Sandwick South, Unst, Shet-land. Quaternary Geochronology 30(B):168–174.

MacDonald, K., and I. McHardy. 2008. BaileSear Community Archaeology Project. SlocSàbhaidh, Baile Sear, North Uist. Season 2(2007) Data Structure Report. Unpublishedreport. St. Andrews: The SCAPE Trust and theUniversity of St. Andrews.

Parker Pearson, M., N. Sharples, J. Symonds, J.Mulville, J. Raven, H. Smith, and A. Woolf.2004. South Uist: Archaeology and History ofa Hebridean Island. Stroud: Tempus.

Rennell, R., and I. McHardy. 2009. Baile SearCommunity Archaeology Project. Sloc Sàb-haidh, Baile Sear, North Uist. Season 3(2008) Data Structure Report. Unpublishedreport. St. Andrews: The SCAPE Trust and theUniversity of St. Andrews.

Sanderson, D. C. W., and S. Murphy. 2010. Us-ing simple portable OSL measurements andlaboratory characterisation to help under-stand complex and heterogeneous sedimentsequences for luminescence dating. Quater-nary Geochronology 5:299–305.

Stentoft, K., T. Dawson, and L. McKenna. 2007.Baile Sear Community Archaeology Project.Sloc Sàbhaidh, Baile Sear, North Uist. Unpub-lished report. St. Andrews: The SCAPE Trustand the University of St. Andrews.

THE JOURNAL OF ISLAND & COASTAL ARCHAEOLOGY 15

Dow

nloa

ded

by [

Uni

vers

ity o

f St

And

rew

s] a

t 01:

58 2

4 N

ovem

ber

2017