Gunilla Eriksson

Norm and differenceStone Age dietary practice in the Baltic region

2003

© 2003 Gunilla ErikssonISSN 1400-7835ISBN 91-89338-11-1Jannes Snabbtryck Kuvertproffset HB, Stockholm 2003Cover drawing and design by Sara EricsonLayout by Gunilla Eriksson

Doctoral dissertation 2003Archaeological Research LaboratoryStockholm UniversityS-106 91 Stockholm

AbstractStone Age research on Northern Europe frequently makes gross generalizations about the Meso-lithic and Neolithic, although we still lack much basic knowledge on how the people lived. Thetransition from the Mesolithic to the Neolithic in Europe has been described as a radical shiftfrom an economy dominated by marine resources to one solely dependent on farming. Both theoccurrence and the geographical extent of such a drastic shift can be questioned, however. It istherefore important to start out at a more detailed level of evidence in order to present the overallpicture, and to account for the variability even in such regional or chronological overviews. Fif-teen Stone Age sites were included in this study, ranging chronologically from the Early Meso-lithic to the Middle or Late Neolithic, c. 8300–2500 BC, and stretching geographically from thewesternmost coast of Sweden to the easternmost part of Latvia within the confines of latitudes55–59° N (Fig. 1, Table 1). The most prominent sites in terms of the number of human and faunalsamples analysed are Zvejnieki, Västerbjers and Skateholm I–II. Human and faunal skeletal re-mains were subjected to stable carbon and nitrogen isotope analysis to study diet and ecology atthe sites. Stable isotope analyses of human remains provide quantitative information on the rela-tive importance of various food sources, an important addition to the qualitative data suppliedby certain artefacts and structures or by faunal or botanical remains. A vast number of new ra-diocarbon dates were also presented.

In conclusion, a rich diversity in Stone Age dietary practice in the Baltic Region was demon-strated. Evidence ranging from the Early Mesolithic to the Late Neolithic show that neither chro-nology nor location alone can account for this variety, but that there are inevitably cultural fac-tors as well. Food habits are culturally governed, and therefore we cannot automatically assumethat people at similar sites will have the same diet.

Stable isotope studies are very important here, since they tell what people actually consumed,not only what was available, or what one single meal contained. We should not be deceived toinfer diet from ritually deposited remains, since things that were mentally important was notalways important in daily life. Thus, although a ritual and symbolic norm may emphasize certainfood categories, these may in fact contribute very little to diet. By the progress of analysis of intra-individual variation, new data on life history changes have been produced, revealing mobilitypatterns, breastfeeding behaviour and certain dietary transitions. The inclusion of faunal datahas proven invaluable for understanding the stable isotope ecology of a site, and thereby improvethe precision in interpretations of human stable isotope data. The special case of dogs, though,demonstrates that this animal is not useful for inferring human diet, since dogs due to the numberof roles they possess in human society could, and in several cases has proven to, deviate signifi-cantly from humans in their diet.

When evaluating the radiocarbon data of human and animal remains from the Pitted-Waresite Västerbjers Gotland, the importance of establishing the stable isotope ecology of a site beforemaking deductions on reservoir effects has been further demonstrated.

The main aim of this thesis has been to demonstrate the variation and diversity in human prac-tice, challenging the view of a “monolithic” Stone Age. By looking at individuals and not only onpopulations, the whole range of human behaviour has been accounted for, also revealing thediscrepancy between norm an practice, which frequently is visible both in the archaeologicalrecord and in present-day human behaviour.

Acknowledgements

3

An optional title for this thesis could be Same,same, but different. However, there are a num-ber of people and institutions without whom itwouldn’t have been the same at all, but indeedentirely different.

Two strong women have guided methrough my time as a PhD student, and I amindebted to them both. My supervisor, KerstinLidén, has always been there for me, with feed-back and constructive suggestions, with inspi-ration and support, and with never failing en-thusiasm. She has always pushed me a bit fur-ther, and so often had that slightly differentangle which made me see things in a new light.It has been a true privilege to work with herand I am happy that we will be able to con-tinue our teamwork. My assistant supervisor,Birgit Arrhenius, has contributed not onlywith her wide experience and abundant goodadvice, but also with unexpected views andcomments and unlimited encouragement. Ihave always appreciated her support and feltinspired by her way of opening up new per-spectives.

Ilga Zagorska has opened up her home tome and generously shared both her greatknowledge and her lively approach. I extendwarm thanks to her for her hospitality and forour successful collaboration, which I hope willcontinue for many years. I would also like tothank my other Latvian colleagues, of whom Iwould especially like to mention ValdisBerzins, Anda Berzina and Egita Ziedina.

This work has benefited enormously fromdiscussions and collaboration with LembiLõugas, Janne Storå, Jonathan Lindström,Ingrid Bergenstråhle, Lars Larsson and OlaMagnell, to whom I extend my heartfeltthanks. Thanks go also to Torbjörn Ahlström,Anders Angerbjörn, Erik Bendixen, PerEricson, Tero Härkönen, Christian Lindqvist,Bengt Nordqvist, Eva Olsson, Heike Sieg-mund and Carina Sjögren, who contributed

vital information and data at various points inthe project. I also want to express my appre-ciation of the discussions I have had with fel-low PhD students in the Stone Age business atnumerous workshops and conferences in Swe-den.

Anita Malmius has kept me company dur-ing many long days and nights (although nomornings), and has prevented any laidbackconsensus through her lovely stubbornnessand friendship. My bone chemistry lab-mates,Annica Olsson and Anna Linderholm, havecontributed with good company and providedessential assistance for which I am infinitelygrateful. Present and former lab companions(apart from those already mentioned) haveprovided a stimulating working environment:Malgorzata, Margaretha, Ludde, Charlotte,Sven, Anders, Ann-Marie, Lena, Kjell, Laila,Birgitta, Björn, Emilia, Michael and Liselotte.

Several museums, and especially the fol-lowing people, have been of great help and as-sistance in providing material: Ylva Olssonand Hampus Cinthio at the Lund HistoricalMuseum, Leena Drenzel at the Museum ofNational Antiquities, Mattias Schönbeck andLars Z. Larsson at the Östergötland CountyMuseum and Elisabeth Brynja at the Väster-götland Museum.

The main funding during my period as aPhD student was provided by the Swedish Re-search Council (VR, formerly HSFR) and theFaculty of Humanities, Stockholm University,both of which are gratefully acknowledged.Additional financial support has generouslybeen provided by Berit Wallenbergs stiftelse,Birgit & Gad Rausings stiftelse, Magn. Berg-valls stiftelse, the Swedish Institute, the RoyalAcademy of Sciences, Mårten Stenbergersstipendiefond, Rosa och Viktor Tengborgsstipendiefond, Leonard och Ida Westmansfond, Svenska fornminnesföreningen (Hilde-brands fond) and Greta Arwidssons fond.

4

Sara Ericson has ingeniously designed thecover, and tolerated all my remarks on earlierdrafts. Malcolm Hicks has speedily, accuratelyand with good humour performed the lan-guage revisions of most of the texts, and alsoendured my comments and questions. Despitethe time pressure, Malcolm has managed toreturn the revised manuscripts almost beforehe got them, to paraphrase Monty Python, forwhich I am indebted. None of the remainingerrors are of course anyone’s responsibilitybut mine.

Anette, Ingrid, Liv, and Pei Pei have offeredfriendship and many good laughs, and MVF,

René, Störningsjouren, my old bike and theHon. Twin Soc. have given me energy andstrength when I needed it.

For never-ending love and support I thankMärta, Magdalena, Martin, Johanna, my par-ents, H. S. Amna-xan, Dada Fazila, DrShewbo, Birgitta, Eva, Lisa, Helena and fami-lies. Most of all I thank my darling husband,and Janne, Violetta, Leo and Victor for pro-viding so much love, laughter and deliciousfood. You have been my sunshine and mybackbone and without you this would nothave been possible. Thanx.

Contents

5

1. Introduction ............................................................... 72. Sites studied ................................................................ 93. Method ..................................................................... 124. Results and discussion ............................................... 144.1 Intra-individual variation ..................................... 144.2 Breastfeeding ........................................................ 164.3 Faunal analyses and stable isotope ecology ........... 184.4 Dogs ..................................................................... 194.5 Radiocarbon dating and reservoir effects ............. 224.6 Methodological concerns ...................................... 244.7 Diet and “the Edible” ........................................... 264.8 Mobility ............................................................... 264.9 The non-monolithic Stone Age ............................. 274.10 “Nature” as a norm ............................................ 28

5. Conclusions ............................................................... 306. References ................................................................. 31

List of papers

This thesis is based on the following papers, which will be referred to by their roman numerals.The published papers are reproduced with kind permission from the publishers.

I. Lidén, K., Olsson, A., Eriksson, G. & Angerbjörn, A. ms. Nitrogen isotope analysis ofdeciduous teeth: a tool for tracing weaning patterns. Submitted to Proceedings of theRoyal Society, Series B.

II. Eriksson, G. & Zagorska, I. 2003. Do dogs eat like humans? Marine stable isotopesignals in dog teeth from inland Zvejnieki: Mesolithic on the Move: Papers presented atthe Sixth International Conference on The Mesolithic in Europe, Stockholm 2000, pp.160–168. Oxbow Monograph. Oxford.

III. Eriksson, G., Lõugas, L. & Zagorska, I. 2003. Stone Age hunter–fisher–gatherers atZvejnieki, northern Latvia: radiocarbon, stable isotope and archaeozoology data. Be-fore Farming (www.waspjournals.com) 2003/1 (2), pp. 1–26.

IV. Eriksson, G. ms. Part-time farmers or hard-core sealers? Västerbjers studied by meansof stable isotope analysis. Submitted to Journal of Anthropological Archaeology.

V. Lidén, K., Eriksson, G., Nordqvist, B., Götherström, A. & Bendixen, E. ms. “The wetand the wild followed by the dry and the tame” – or did they occur at the same time?Submitted to Antiquity.

VI. Eriksson, G. & Lidén, K. ms. Skateholm revisited: New stable isotope evidence on hu-mans and fauna. Manuscript.

1. Introduction

7

The simple question “What did Stone Agepeople eat?” covers many dimensions, and thenutritional point of view is only one aspect.First of all, the Stone Age spanned many mil-lennia, during which a number of majorchanges occurred, not least affecting diet, andsecondly, diversity seems to have been a traitcharacterizing even a relatively limited areasuch as the southern Baltic region, which is thefocus of this thesis. Although Stone Age re-search on Northern Europe frequently makesgross generalizations about the Mesolithicand Neolithic (Bonsall et al. 2002; Price 1991;Schulting & Richards 2002b; Zvelebil 1996),we still lack much basic knowledge on how thepeople lived. It is therefore my belief that weneed to start out at a more detailed level ofevidence in order to present the overall pic-ture, and that we need to account for the vari-ability even in such regional or chronologicaloverviews.

The archaeological debate over the Neo-lithic in southern Scandinavia has long beendominated by the notion of archaeologicalcultures. The Swedish Neolithic is tradition-ally held to comprise three “cultures”: thefarming Funnel Beaker Culture, the foragingPitted Ware Culture and the farming BattleAxe Culture (equivalent of the Corded WareCulture). These are defined by certain arte-facts or other features, but are often also con-sidered to represent different economies. Arte-facts and faunal remains rarely give unam-biguous answers, or when they occasionallydo, the accuracy can be questioned. This is onereason why dietary reconstruction by meansof bone chemistry, especially stable isotopeanalysis, has become so important.

Stable isotope analyses of human remainsprovide quantitative information on the rela-tive importance of various food sources, animportant addition to the qualitative datasupplied by certain artefacts and structures orby faunal or botanical remains. The archae-

ological record is not only full of pits, but alsoof pitfalls. Thus, apart from the taphonomicfactors/processes affecting what is left for us asarchaeologists to investigate, intentional de-posits such as burials are imbued with conno-tations which may tell us much about the sym-bolic world of the people concerned but lessabout their everyday lives. This is not to saythat the latter aspect is more important thanthe former, but it is vital to recognize that wemay perceive Stone Age life in a distorted waybecause of the emphasis on certain aspects in aburial or a monument (cf. Parker Pearson1999). One illustration of this is the presenceof numerous tooth pendants deposited ingraves, which do not necessarily tell us whatspecies made the most important contribu-tions to the diet, or the accumulations of pigbones in association with a cemetery, which donot automatically imply the importance ofpork as a meat. Similarly, finds of bone fromdomestic animals such as sheep or cattle arenot inevitably evidence of a farming economy.

The transition from the Mesolithic to theNeolithic in Europe has been described interms of “turning their backs to the sea”(Schulting & Richards 2002c:155), entailing aradical shift from an economy dominated bymarine resources to one solely dependent onfarming. Both the occurrence and the geo-graphical extent of such a drastic shift can bequestioned, however. Here dietary analyses ofindividuals from numerous locations are vitalfor an understanding of how widespread thisallegedly “monolithic” norm was.

Yet another reason for studying dietarypractices during the Stone Age, and by nomeans the least important, is the recurrent useof this period as a key to the notion of “naturalbehaviour” (Audette 1999; Ljungberg 1997),the “biologically normal” or “original life-style” (Lindeberg 1997). In such lines of rea-soning the Stone Age way of life is “natural”,

8

gunilla eriksson

which by definition is good, i.e. the only rightway. The argument is implicitly normative,and the underlying assumption seems to bethat Stone Age people were in some way sav-ages, less influenced by culture than thepresent-day population. Accordingly, we aretold, we should only eat this and that, treat ourbabies this and that way, and appreciate gen-der inequalities – because it is only “natural”.

Apart from being based on biologicalreductionism, such arguments signal an una-wareness of the diversity of human practicesduring the Stone Age. Any attempt to graspthis multiplicity through simple dichotomiessuch as Mesolithic/Neolithic, coastal/inlandor female/male will inevitably fail, as will bedemonstrated in the present study.

9

norm and difference

The fifteen Stone Age sites included in thisstudy range chronologically from the EarlyMesolithic to the Middle or Late Neolithic, c.8300–2500 BC, and stretch geographicallyfrom the westernmost coast of Sweden to theeasternmost part of Latvia within the confinesof latitudes 55–59° N (Fig. 1, Table 1). Themost prominent sites in terms of the number ofhuman and faunal samples analysed areZvejnieki, Västerbjers and Skateholm I–II,although the latter produced very limitedstable isotope data because of poorpreservation. All three sites comprisedextensive cemeteries with associated culturallayers, but while both Skateholm I–II andVästerbjers seem to have been in use forseveral hundred years, Zvejnieki wasevidently used for five millennia (!). They alsodiffer in that Zvejnieki was located inland,whereas Skateholm was situated at a lagoon,and Västerbjers by a narrow bay of the BalticSea. The remaining sites are represented in this

material by just a few individuals from variouscontexts– regular burials as well as isolatedhuman bones scattered in cultural ortransgressed layers. A range of locations in thelandscape are represented, and although thereis an emphasis on hunter–fisher–gatherercontexts for sites with both Mesolithic andNeolithic dates, four sites also includedburials attributed to the farming Corded WareCulture (Sarkani, Selgas, Kastanjegården anda few of the Zvejnieki burials).

The geographical focus is to some extent aconsequence of availability and preservationconditions. There are simply no well-preserved Stone Age human bones availablefrom the northern half of Sweden, nor fromthe mainland of Finland, largely on account ofsoil conditions. Moreover, even for southernSweden, the Mesolithic is not exactlycharacterized by a wealth of well-preservedhuman remains, which is one reason why thepresent work has benefited greatly from the

2. Sites studied

Fig. 1. Locations of the sites included in this thesis. 1 – Uleberg, 2 – Evensås, 3 – Huseby klev, 4– Rolfsåker, 5 – Hanaskede, 6 – Ageröd, 7 – Kastanjegården, 8 – Skateholm, 9 – Alby, 10 – Ire,11 – Västerbjers, 12 – Zvejnieki, 13 – Selgas, 14 – Sarkani.

10

gunilla eriksson

Tab

le 1

. Sit

es in

clud

ed in

the

stud

y, s

orte

d by

dat

e.

site

arch

. dat

eap

prox

.ch

ron

ozon

elo

cati

onco

ntex

tan

alys

ed sa

mpl

espa

per

no.

in14

C d

ate

hum

anfa

unal

Fig.

1(u

ncal

. BP)

(n i

ndiv

idua

ls)

Hus

eby

klev

EM

9000

, 85

00PB

/BO

coas

tal,

oute

r ar

chip

elag

otr

ansg

ress

ed la

yers

5 (5

)9

V3

(dee

p pi

t, t

ent)

Han

aske

deE

M8

800

BO

inla

ndst

ray

find

3 (1

)–

V5

Age

röd

IE

M/M

M79

00–

7400

AT

inla

ndin

term

ingl

ed c

ultu

ral l

ayer

s5

(5)

20

V6

Zve

jnie

kiM

M–L

N82

00–

4200

BO

/AT

/SB

inla

ndce

met

ery,

cul

tura

l lay

ers

43 (

33)

98

II, I

II1

2U

lebe

rgL

M6

600

AT

coas

tal,

oute

r ar

chip

elag

obu

rial

1 (1

)–

V1

Skat

ehol

m I

–II

LM

630

0A

Tco

asta

l, la

goon

cem

eter

y, c

ultu

ral l

ayer

s27

(24

)4

5V

I8

Alb

yL

M/E

N5

300

AT

coas

tal,

lago

onbu

rial

6 (1

)–

V9

Eve

nsås

EN

500

0SB

coas

tal,

oute

r ar

chip

elag

odi

stur

bed

buri

al1

(1)

–V

2Sk

ateh

olm

VI

EN

490

0SB

coas

tal,

lago

onst

ray

find

1 (1

)–

V8

Rol

fsåk

erM

N4

500

SBco

asta

l, in

ner

arch

ipel

ago

buri

al1

(1)

–V

4V

äste

rbje

rsM

N4

300

SBco

asta

lce

met

ery,

cul

tura

l lay

ers

86 (

26)

64

IV1

1Ir

eM

N4

300

SBco

asta

lce

met

ery,

cul

tura

l lay

ers

–1

4I,

IV1

0Sa

rkan

iM

N/L

N4

300

SBin

land

buri

al1

(1)

–II

I1

4Se

lgas

MN

/LN

420

0SB

inla

nddo

uble

bur

ial

3 (2

)–

III

13

Kas

tanj

egår

den

MN

/LN

400

0SB

inla

ndtr

iple

bur

ial

1 (1

)–

III,

V7

Key

: EM

, MM

, LM

= th

e E

arly

, Mid

dle

and

Lat

e M

esol

ithi

c re

spec

tive

ly,

EN

, MN

, LN

= t

he E

arly

, Mid

dle

and

Lat

e N

eolit

hic

resp

ecti

vely

,PB

= P

rebo

real

, BO

= B

orea

l, A

T =

Atl

anti

c, S

B =

Sub

bore

al

11

norm and difference

inclusion of Latvian material. The situationfor Neolithic material is better, although onecould always wish for more. The mainjustification for the choice of material is notaccessibility, however, although thisnecessarily delimits the options available, buta conscious desire to go into detail within alimited area. It was also felt important toinclude sites both east and west of the BalticSea, to avoid the imbalance in much previousStone Age research, which has tended todiscuss either the westernmost parts of

Europe, or only the eastern parts. This has inpart been caused by the geopolitical situationof much of the past century, of course, butwhereas the Iron Curtain has disappeared, thisimbalance has prevailed. This is unfortunate,not least because there is little evidence thatthis division was of relevance to people duringthe Stone Age; for the Baltic in particular, thesea was a uniting feature, not a divisive one(the “Uniting Sea” was also the theme of arecent workshop in Uppsala 2002).

12

gunilla eriksson

There are several elements with naturally oc-curring stable isotopes which can be used inarchaeological studies, two of which are nitro-gen and carbon. While early stable carbon andnitrogen analyses were characterized by sheeroptimism about the possibilities of themethod, the following decades saw a growingawareness of some of the problems and limita-tions (e.g. Katzenberg 1992; Pate 1994;Schoeninger & Moore 1992; Schwarcz &Schoeninger 1991; Sealy 2001). Many of thesehave now been overcome, however, and, afterhaving been successfully used for a quarter ofa century, stable carbon and nitrogen isotopeanalyses are one of the most important tech-niques employed in archaeological researchtoday.

The method has been described in detailelsewhere (e.g. Ambrose 1990; Brown et al.1988; Lidén 1995b), so I will only reiteratehere the basic principles, concepts and limita-tions. The two basic principles underlying sta-ble isotope analysis of bone are (1) “you arewhat you eat”, i.e. body tissue is synthesizedout of components from the diet, and (2) theproportions of the stable isotopes, 13C vs. 12Cand 15N vs. 14N, alter as a consequence of vari-ous biological, physical and geological proc-esses. The first means that a dietary record isincorporated into our skeletons, reflecting thetime of formation or remodelling, and the sec-ond that this record can be traced and thatthere is enough variation to make it a valuablesource of information. For the area of interesthere, the Baltic region in the Stone Age, themost important difference exhibited in carbonisotope values (δ13C values, expressed in permil, ‰, relative to the standard, PDB) is thatbetween the marine input to the diet and theterrestrial or freshwater contribution. The ma-jor information gained from the nitrogen iso-tope value (δ15N, expressed in ‰ relative to thestandard, AIR) concerns levels in the food

web, the value increasing for each additionalstep in the food chain. These stable isotope ra-tios can be measured by mass spectrometry,and simultaneous analyses of the two are valu-able, since variation in a two-dimensionalspace extends the possible combinations rela-tive to information gained from a linear scale.

An important feature facilitating the inter-pretation of stable isotope data from humanremains is the addition of faunal isotope data.Although the approximate stable isotope end-values, i.e. maximum and minimum values formarine vs. terrestrial and herbivorous vs. car-nivorous organisms, have been established forthe region, there can be considerable variationon account of the particular ecology of a site,and faunal remains from the same context asthe human remains should therefore ideally beincluded in the analysis.

The stable isotope analyses are performedon collagen, the predominant protein presentin skeletal tissue, and the values mainly reflectprotein intake (Ambrose & Norr 1993). Ascremation destroys the structure of the colla-gen, only unburned bone will produce reliabledata (DeNiro et al. 1985; DeNiro 1985;Götherström 2001). Bone is constantly beingremodelled during a person’s lifetime, and itsstable isotope signature therefore reflects theaverage diet over a period of several yearsprior to death. Teeth, by contrast, are formedearly in life, and the dentine (the bony sub-stance of teeth, partly covered by the enamel)is not subject to any collagen turnover, whichmeans that the isotopic signal reflects the dietwhich prevailed when the teeth were formed,i.e. in childhood. An analysis that includesboth bone and teeth from adults therefore ineffect expands the population studied to in-clude children who survived into adulthood, agroup otherwise severely underrepresented inarchaeological research.

The collagen turnover rate in bone varies

3. Method

13

norm and difference

with age and the skeletal element concerned,but there are also differences within a bone ele-ment, with a slower turnover rate in compactbone (see Lidén & Angerbjörn 1999 for a re-view on factors affecting collagen turnover).Since the preservation of bone collagen is as arule better in compact, cortical bone, skullbones or the diaphyses of long bones are gen-erally preferred over the spongy, trabecularbone of the epiphyses and various flat and ir-regular elements of the skeleton. This does notimply that the latter bone elements cannot beused for analysis, of course. On the contrary,the sampling strategy, extraction protocol andquality criteria applied ensure that only datafrom intact collagen will be considered.

The main aspect separating bone chemistryfrom conventional archaeological data, espe-cially in connection with graves, is not the bio-logical character or that the data were scien-tifically produced, but the fact that the datawere not intentionally deposited at a burial orin any other ritual. Accordingly, there was nocommunicative or normative intent which re-sulted in a certain stable isotope value at the

time of deposition, making it relatively unbi-ased as compared with grave goods, for in-stance. On the other hand, the diet, andthereby indirectly the isotopic signature, isculturally governed, of course, and the body isloaded with meaning (Johannisson 1997;Liukko 1996). Furthermore, bone chemistry iscontextually independent, meaning data willbe produced regardless of the presence of con-textual information. Contextual data will addto the value of the findings, but cannot be re-garded as a prerequisite for their employment.Consequently, individuals can be studied whoare habitually excluded from burial analysisbecause they lack grave goods, are subadult,fragmented, found in multiple burials or dis-persed in cultural layers, or have otherwisebeen treated differently from what we perceiveas regular burials. Inclusion of those who devi-ate from our preconception of a “normal”burial improves the representativeness of thematerial and enhances our possibilities for un-derstanding the full range of human behav-iour in prehistory.

14

gunilla eriksson

4.1 Intra-individual variation

An important advance brought about by theapplication of δ13C and δ15N analyses toprehistoric skeletal remains is the possibilityfor tracing dietary variation on the individuallevel, i.e. to obtain dietary life histories forindividuals. Stable carbon and nitrogenisotope analyses of human bone and/ordentine have only rarely been employed earlierto study intra-individual variation (but seeSealy et al. 1995; Wright & Schwarcz 1999),since most previous analyses have beenconcerned either with animal tissue (e.g.Balasse et al. 1999; Hobson 1998; Koch et al.1995; Wiedemann et al. 1999), or with dentalenamel, often focussing on other stableisotopes such as strontium and oxygen (e.g.Balasse et al. 2002; Balasse et al. 2001; Wright& Schwarcz 1998).

In the case of archaeological subjects forwhom whole crania were available, both boneand teeth were subjected to analysis (see inparticular Paper III). The strategy involvedsampling the first, second and thirdpermanent molar teeth for each individual,where present. The samples were taken with adentist’s drill directly below the crown of eachtooth. The reasons for sampling several teethfrom each individual instead of takingappositional samples from one tooth arerelated to the morphology of human teeth.The main tissue making up the teeth is dentine,with a composition similar to that of bone,inside which are the cavities formed by thepulp chamber and root canal(s). The dentineof teeth is laid down in angled layers startingfrom the crown and proceeding down to theroot (Hillson 1996), and in order to obtain asample representing as limited a time offormation as possible, drilling shouldaccordingly take place perpendicular to thelongitudinal axis of the tooth (Fig. 2). Since

the crown is coated with enamel, a very hardmineral substance, drilling of the crownwould make separation of the dentine fromthe enamel difficult, and could also cause thetooth to fracture, so the samples had to bepositioned below the crown. Molar teeth as arule have several roots, however, and one rootis too thin to produce large enough samples,especially since the surface layer must alwaysbe discarded to avoid contamination, so theonly possible position for sampling was justbelow the cervix. This position was also idealsince it caused very little damage to the tooth –if removed from the jaw prior to sampling, thetooth would hardly show any visible signs ofdrilling once replaced in position (Fig. 2c) – animportant consideration for archaeologicalspecimens.

The approach of using the three types ofmolar was employed in order to obtainsamples representing three age spans. Thespecific sections sampled on each tooth typecould be estimated to correspond to three agecategories based on the timing of formation ofeach tooth section (Fig. 3) (Hillson 1996; Reidet al. 1998). Although there is considerableindividual variability in the timing of toothformation and the ages assigned to eachcategory should accordingly be regarded onlyas approximations, the ages of formation ofthese teeth do not overlap, but follow in asequence. The second molars, if unavailable,were occasionally replaced by premolars,however, since the timing of crown formationin the latter overlaps to a great extent withformation of the second molars. Although thesampling strategy necessarily involved theinclusion of several consecutively formedlayers of dentine, it excluded those portions ofthe dentine formed at the very beginning andend of tooth formation, thereby narrowingdown the age span represented by eachsample.

4. Results and discussion

15

norm and difference

Even bone can be used for tracing intra-individual change in certain cases, since thecollagen turnover rate varies in different boneelements. This is likely to be most pronouncedin growing individuals, as was demonstratedby the analysis of both skull bone and ahumerus from a newborn, or possiblystillborn, infant at Zvejnieki (Paper III), whichexhibited large differences caused by the

mother changing her diet during pregnancy(bearing in mind that the bones were formedin utero). The fact that short-term changes indiet are recorded in bone from children couldfurthermore be utilized to trace seasonalmobility at the population level, since whereasthe isotopic signature of adult bone is levelledout due to the slow collagen turnover rate, theisotopic record of children would show a

Fig. 2. (a) Simplified cross-section of an incisor showing the dentine layering, which starts atthe crown (adapted and redrawn from Steele & Bramblett 1988). (b) All the permanent toothsamples were drilled from the same section of the tooth, just below the cervix, illustrated herewith a second molar. (c) Sampling caused very little damage to the material; it was hardlyvisible once the tooth was back in position in the jaw (here a mandible). Note the sample drilledfrom the first molar, indicated by an arrow.

16

gunilla eriksson

lesser degree of averaging and thus highervariability (Fig. 4). It should also be possibleto demonstrate this high variability in teeth aswell as bone, provided that the timerepresented by each sample is short enough torecord only one season.

4.2 Breastfeeding

Breastfeeding and weaning practices haveimportant health implications for both themother and the child, and could affectpopulation growth and mobility, but they arealso ultimately a matter of control overreproduction, making this issue an importantfield of study of relevance to both modern lifeand prehistory. Maybe this is also why normsfor breastfeeding practice are often sodogmatic, although they range from theintroduction of the first solids right after birthto the end of weaning at the age of 5–8 years(Coates 1993; DeLoache & Gottlieb 2000;Dettwyler 1995; Eriksson et al. 2000; Fildes1995; Gartner & Stone 1994; Jelliffe & Jelliffe

1978; Quandt 1995; Riordan 1993; Short1984; Stuart-Macadam 1995).

The term “weaning age” is somewhatmisleading, as it implies that weaning is adistinct event, whereas in most cases it shouldbe regarded as a process, beginning with thefirst introduction of other foods and endingwith the last breastfeeding (Riordan &Auerbach 1993)(the exception, of course,would be abrupt weaning, which turns thisprocess into a distinct event). Regarded thisway, the weaning process is a part of thebreastfeeding practice, and will in many casesconstitute the main part of it.

One important contribution to theinvestigation of intra-individual changes isthus the possibility to trace breastfeedingpatterns in individuals by means of stablenitrogen isotope analysis of deciduous teeth(Paper I). Previous studies of prehistoricbreastfeeding and weaning have analysedbone from whole populations, includingchildren, typically plotting the δ15N valuesagainst age at death (Fig. 5), and inferringweaning age from the position of the peak and

Fig. 3. The collagen of the skull and the various teeth sampled was formed at a specific age ineach case, and samples from these represent different ages in the life of the individual. Skullfrom Västerbjers burial 93, photo by the author (published with permission from the Museumof National Antiquities).

17

norm and difference

subsequent drop in the curve (e.g. Fogel et al.1989; Katzenberg et al. 1996; Richards et al.2002; Schurr 1997). There are severalproblems with this approach, however: (1) theactual weaning age will in fact be earlier thanthat deduced from the curve produced fromthe bone isotope values, since there is a time-lag between the intake of a certain food andthe incorporation of the stable isotope valueinto the skeletal tissue due to collagenturnover (cf. Lidén & Angerbjörn 1999), (2)the bone isotope values represent childrenwho died during childhood, potentiallybecause of early weaning, which causesproblems of representativeness, (3) theevidence of weaning only shows an averagefor the population and does not takeindividual variation into account. The firstproblem can be overcome by correcting for thetime-lag, although there is a need for moredetailed data on the effect of collagen turnoverand growth on isotope values (cf. Lidén &Angerbjörn 1999). The second problem isrelated to the specific demography of deadpopulations, which differs considerably from

living ones; the fact that we are faced withthose who never survived into adulthood mustnever be forgotten, and will inevitably cause abias in the data. This fact underlines theimportance of the progress made in tracingchildhood diet by analyses of the permanentteeth of adults (Papers III, IV). The problem ofusing population averages instead ofindividual data is also one ofrepresentativeness, since it does not accountfor the variability in breastfeeding practiceswithin one population (cf. Fig. 6). Eventhough norms for how and when weaningshould take place may be rigid, practices maydiffer significantly from the norm, as in somany other instances. Furthermore, eachinfant–mother pair is unique, and the weaningprocess may therefore differ between siblings,even though the mother is the same (cf. PaperI).

All stable isotope studies of the weaningprocess in prehistoric populations should alsobear in mind that particular food taboosduring pregnancy and lactation could causestable isotope signatures to deviate from the

Fig. 4. Hypothetical bone stable nitrogen isotope values plotted against age at death for apopulation using isotopically different food sources over the course of a year. The variationrecorded for growing individuals is levelled out for adults because of the slower collagenturnover rate.

0

3

6

9

12

15

18

0 20 40 60 80

age (years)

δδ15

N (

‰)

18

gunilla eriksson

expected. Food avoidances associated withpregnancy or lactation have been attested invarious parts of the world (Eichinger Ferro-Luzzi 1980a; Eichinger Ferro-Luzzi 1980b;Fieldhouse 1996; Wilson 1980). Moreover,since the deciduous teeth are eventually shedand replaced by permanent teeth, the earliestrecord of dietary practice will be lost whenchildren grow older, although there will still beevidence of their early diet in the first molars,for example. It could even be speculated thatthe drop in δ15N values in the first molars(corresponding to an age of three years) in theVästerbjers population (Paper IV:table 7) to alevel significantly lower than for the thirdmolars (early adolescence) or for adult bonefrom the same individuals could be interpretedas representing the final phase of the weaningprocess. Such a drop was frequently seen atthis stage in the modern individuals studied(Paper I:Fig. 3), and a corresponding drop,though at an earlier age, was also recorded forthe Pitted-Ware child at Ire (Paper I:Fig. 4).

Breastfeeding is a complex process whichinteracts with a number of biological,ecological, economic, social, cultural andindividual factors, and no single factor willaccount for all the variation in practices, sothat breastfeeding in prehistory can beexpected to vary widely. The quest for an“original” or “natural” manner ofbreastfeeding (Knutsson 1995; Ljungberg1995) is a vain undertaking.

4.3 Faunal analysesand stable isotope ecology

The extensive analyses of faunal remains pre-sented here contributed substantially to the in-terpretation of the human data (Papers II, III,IV, VI). An understanding of the stable isotopeecology of a site is crucial to its comprehen-sion, and will in many cases provide detailedinformation which it would not otherwise bepossible to glean. At Västerbjers (Paper IV),the inclusion of faunal analysis demonstratedthe importance of seals in the human diet on avery detailed level, and also indicated that theproportion of fish ingested may previouslyhave been overestimated (Fig. 7). Further-

more, it was shown that the importance ofpork in the diet could be ignored. Without thefaunal baseline it would only have been possi-ble to conclude that marine protein consti-tuted a significant portion of the diet, but notto go into detail regarding how and to whatextent marine resources were exploited.

For Skateholm I and II (Paper VI) it is evenmore obvious that without the faunal data wecould not have achieved the same level of con-fidence in our interpretations. Although boththe human and animal bones were in poorcondition and very few produced collagen forstable isotope analyses, the few that did gener-ated enormously important data, indicatingthat certain people travelled between the westcoast and the Skateholm lagoon during theLate Mesolithic. Similarly, the Ageröd faunalisotope data (Paper V) lead to the interpreta-tion that the human bones recovered therewere from individuals who had contact withpeople from the west coast, as indicated by thepresence of dog bones, but who themselvesgained their subsistence from the east coast.The mobility patterns suggested by the com-bined faunal and human analyses at bothAgeröd and Skateholm could not have beeninferred from the human data alone.

The Zvejnieki Stone Age complex (PapersII, III) offers another example of the signific-

Fig. 5. Bone stable nitrogen isotope valuesplotted against age at death. Weaning wasinferred to occur at age one year in this case(reproduced from Katzenberg & Pfeiffer1995).

19

norm and difference

ance of faunal data. It was possible by refer-ence to the various faunal species included todemonstrate not only the predominance offreshwater fish in the diet, especially duringthe oldest periods for which the cemetery wasused, but also to suggest that the whole popu-lation had at least three or four major con-stituents in their diet.

The complex natural history of the BalticSea makes it especially vital to establish thestable isotope ecology when studying sites inthis region. Salinity and sea level have variedconsiderably both spatially and temporallywithin the region, emphasizang theimportance of analysing faunal remains fromthe same site and of the same date as thehuman material, since the stable isotopeecology could be specific to the site andperiod. This is also exemplified by the “fossilfuel effect”, which has caused δ13C values todecrease globally by some 1.5‰ comparedwith the pre-industrial era (Marino &McElroy 1991).

The analysis of animal bones could alsoassist in challenging widely held views about

certain animals and their relations to humanpopulations, e.g. the issue of pigs in theMiddle Neolithic on Gotland. The stableisotope data provide no support for thehypothesis that these were domestic, or“freeland pigs”, but suggest instead that theywere wild boar or possibly feral pigs.Nevertheless, these animals were obviously ofgreat symbolic importance to the Pitted WareGotlanders, as displayed by the numerousfinds of boar tusks, mandibles and pig bonesin Pitted Ware cemeteries (Paper IV), althoughthey seem to have offered the meat to theirdogs rather than consuming it themselves. Asfor the dogs and their relation to the humanpopulation, they deserve a section of theirown.

4.4 Dogs

Four of the sites studied yielded both humanand dog bones, namely Ageröd (Paper V),Skateholm (Paper VI, Eriksson & Lidén2002), Zvejnieki (Papers II, III) and

Fig. 6. Stable nitrogen isotope values for two children from Pitted-Ware sites on Gotland,plotted against crown formation age and estimated age based on bone collagen remodelling.Weaning was initiated at six months for both children, but completed at 1.5 years for the Irechild and not until 4 years of age for the Västerbjers child. (Note: the data points for theVästerbjers child were from ITM measurements corrected by a factor calculated for theparticular run. Three of the samples were replicated at GEO and conformed entirely to thecorrected values.

14

15

16

17

18

19

20

0 1 2 3 4 5 6 7 8

age (years)

15N

(‰)

Ire 7B

Västerbjers 67:2b

20

gunilla eriksson

Fig. 7. Individual isotopic values for Stone Age faunal remains at Västerbjers and Ire, groupedby category (top). Humans and dogs plotted against isotopic expectancy values for individualsliving entirely off any of four groups of potential foods (bottom). For ease of interpretation, thehuman and dog values affected by lactation (i.e. measurements from dog teeth and childskeletons) were excluded from the plot.

21

norm and difference

Västerbjers (Paper IV). It has previously beenclaimed that dogs could be used as proxies forhumans in dietary studies, whereupon a stableisotope analysis of dog remains could beregarded as providing a good approximationfor the human diet (Clutton-Brock & Noe-Nygaard 1990; Noe-Nygaard 1988; Persson1998; Schulting & Richards 2000; Schulting& Richards 2002a). However, as is evidentfrom the analyses at these four sites, dogs areof limited or no value as substitutes for peoplein dietary reconstruction (Fig. 8).

The two dogs analysed at Ageröd hadabout 3‰ more positive δ13C values than thetwo human subjects, indicating a much highermarine protein input. The δ13C value for thegrey seal remains examined at this site,–19.2‰, indicates in turn that this seal hadbred in the Baltic, which was still at theAncylus Lake stage, i.e. it had a much lowersalinity than in later times, and suggests thatthe people could have consumed protein fromthe Baltic, whereas the dogs evidentlyobtained their marine protein from the westcoast of Sweden. The interpretation isimpeded by the mixed layers at this site,spanning a period of some 800 years. Thehuman remains analysed at Skateholm showan exceptionally high range of carbon stableisotope values, but the one dog analysed

nevertheless gave a result lying far beyond thisrange, as was also true of the nitrogen isotopevalue. At Zvejnieki the dogs exhibited anextraordinary variability in stable isotopevalues, with evidence of a freshwater fish dietfor one group, whereas others evidently fed onseals and a third group seems to havecomprised terrestrial scavengers. The stableisotope analysis of human remains, on theother hand, indicated a diet of predominantlyfreshwater fish. Finally, the people atVästerbjers relied almost entirely on seals,while the dogs were in most cases fed on fish,and possibly occasionally pork. The detailedfaunal analyses of the sites discussed herefacilitated the interpretation and helpeddistinguish between potential food sources fordogs and humans.

It is clear from the above that the dogs atboth inland and coastal sites ranging from theseventh to the third millennium BC deviatedsubstantially in their diet from the humanpopulation. This is furthermore apparentfrom various other stable isotope studies ofdog and human remains (e.g. Katzenberg1989; White et al. 2001; White & Schwarcz1989). Although it is still possible that thedogs at some sites could have had diets equalto those of the people, it would be totallyerroneous to deduce the human diet from

Fig. 8. Stable isotope plot for dogs (open symbols) and humans (closed symbols, mean±s.d.,number of individuals analysed in brackets) at four sites. Data from Paper IV (Västerbjers),Paper V (Ageröd), Paper VI (Skateholm I–II), Paper III and unpublished material (Zvejnieki).

0

1

23

4

5

67

89

10

1112

1314

15

16

1718

-26 -25 -24 -23 -22 -21 -20 -19 -18 -17 -16 -15 -14 -13

δ13C (‰)

δ15N

(‰)

Ageröd (2)

Skateholm I–II (3)

Västerbjers (26)

Zvejnieki (27)

22

gunilla eriksson

analyses of dog remains only.Having established this, it must be said that

analyses of dogs could nonetheless be ofimportance in their own right. Dogs in allprobability had a number of differentfunctions and roles in prehistory, both ritualand practical, which were not necessarilymutually exclusive (Olsen 2000; Serpell1995). Practical use of their skills includedroles as hunting partners, draught animals,guard dogs, herding animals, and scavengers,while their wool (Schulting 1994), fur (Noe-Nygaard 1995), meat (Serpell 1995) and teeth(Paper II) are also known to have been used.The company and affection of dogs and thestatus conferred by their possession were otherimportant traits, as was their use in rituals(Crockford 2000). Dogs are easy to moveabout and are likely to have accompaniedpeople who went from the interior to the coastor vice versa, as suggested by the differencesexhibited between the human and dog isotopesignatures reported here (Fig. 8, Papers II–VI).Moreover, because of their value, they arelikely to have been traded and exchangedbetween various groups of people.

Although there are no general criteria thatapply cross-culturally to distinguish betweenritual and non-ritual dog deposits (Olsen2000), the numerous finds and variousmanners of dog deposition at Stone Age sitesin the Baltic region bear witness to the

importance of these animals and their ritualsignificance (Benecke 1987; Larsson 1990;Lepiksaar 1984; Lõugas et al. 1996; Paaver1965). The treatment displayed in burial ritescould differ greatly from their daily treatment,however, and the dogs buried with fullceremony in graves of their own could stillhave been regarded as “polluted” creatures intheir lifetime (e.g. Serpell 1995) – anambivalence which again illustrates thediscrepancy between norm and practice.

4.5 Radiocarbon dating andreservoir effects

Radiocarbon dating is of great importance forthe interpretation of the archaeologicalrecord, and thereby for the application ofdietary analysis. Correctly applied andcombined with archaeological data,radiocarbon dating may help in sorting outthe stratigraphy at a site, narrowing down thelikely time of use, or establishing an absolutedate for an archaeological event. Conspicuousexamples are the human bones recovered atHuseby klev (Fig. 1), the earliest humanremains hitherto found in Sweden, whereradiocarbon dating of these and various othermaterials and structures aided in establishingthe early dates, and also in demonstrating that

Fig. 9. An “original” and “natural” family? Mobility and sexual division of labour in thechildren’s book Tomtebobarnen by Elsa Beskow, originally published in 1910.

23

norm and difference

the stratigraphically separate zones did notoverlap chronologically (Nordqvist 2000).Radiocarbon results have been crucial for thepresent study in several instances (Papers III,IV, V and VI), and therefore some issues ofrelevance need to be brought up here.

The importance of making carefulassessments of the archaeological problem tobe solved by radiocarbon dating has beenpointed out by Nelson (1998), for example,who used the terms “archaeological event”and “radiocarbon event” to distinguishbetween the event of archaeological interest,such as a burial, and the time represented bythe dated sample. If dating teeth from a buriedadult, the discrepancy between theradiocarbon event (in this case the childhoodof the interred person) and the archaeologicalevent of interest (the burial) could be as muchas 50 years, depending on the age of thesubject and the teeth used for dating. Thisneed not be a problem, however, if it is takeninto account when calibrating the date. In thestudy of Västerbjers (Paper IV), where thiswould be applicable, no age offset correctionswere applied to the radiocarbon dates, inorder not to confuse the discussion on thereservoir effect, since the former correctionmust be applied after calibration, whereas thereservoir age correction should always beapplied prior to calibration. This may well bedone in a future study going into the details ofindividual graves, however.

With the software currently available, suchas OxCal or Calib (both freely accessible onthe Internet), calibration of radiocarbon datesproduced from collagen is a relativelystraightforward matter for terrestrialherbivorous samples, but a marine correctionshould be applied before calibration forsamples with a considerable marine influence,such as seals or humans living off seals(Arneborg et al. 1999; Stuiver & Braziunas1993; Taylor 1992). This reservoir effect isprincipally caused by upwelling of water fromlower depths in large basins (oceans, seas orlarger lakes) and its mixing with surface water.As the water from deeper levels has not had thesame carbon exchange with the atmosphere, itcontains lower amounts of 14C than thesurface water and thus exhibits an apparent

radiocarbon age (Taylor et al. 1996). It is wellknown that samples from the Baltic Sea mayincorporate carbon with such marinereservoir ages (e.g. Olsson 1986), and becauseof the complicated natural history of theBaltic, the extent of this effect has fluctuatedwith time, so that the discrepancy must beestablished separately for any given period.Moreover, it can be expected to have variedwidely within the basin, both vertically andhorizontally, due to the circulation systemimposed by the topography, salt-water inflowand freshwater runoff (Bonsdorff &Blomqvist 1993; Ojaveer & Elken 1997). Aless well-known phenomenon, perhaps, is thefreshwater reservoir effect, which was firstbrought to light by Lanting & van der Plicht(1996; Lanting & van der Plicht 1998), andhas recently been observed in the DanubeGorges, referred to as the Iron Gates, wherehumans consuming large amounts offreshwater fish produced radiocarbon datesseveral hundred years older than terrestrialherbivores from the same contexts (Cook et al.2002; Cook et al. 2001).

To estimate the extent of the reservoir effectfor a given archaeological setting, thecustomary approach is to date material fromthe same closed context, such as a burial,which can be assumed not to be affected byany reservoir effect (e.g. bone or antler from aterrestrial herbivore) along with the humanbone potentially affected by it. Although thereis always a risk that the dated material willhave been in circulation for some time beforedeposition, i.e. that the radiocarbon eventmay be much earlier than the archaeologicalevent, this has to be balanced against theimportance of dating a closed context, i.e. tohave a reliable association between theinterred human and the animal.

This methodology was applied atVästerbjers (Paper IV), demonstrating aconsiderably smaller marine reservoir effectfor seal hunters on Middle Neolithic Gotlandthan previously suggested, 70±40radiocarbon years. The fact that the Baltic isbrackish and of limited size could probablyaccount for this effect being much smaller thanin large oceans, where the marine reservoireffect is generally estimated to between 400

24

gunilla eriksson

and 500 radiocarbon years, with someregional differences (e.g. Arneborg et al.1999). Considering the fact that the Zvejniekipopulation (Paper III) had a highconsumption of freshwater fish, one has totake into account the possibility of afreshwater reservoir effect for this site. Theextent of the effect for an Iron Gates subjectwith a 100% intake of freshwater fish wasestimated by Cook et al. (2002) at roughly500 radiocarbon years. However, twoindividuals from a multiple burial atZvejnieki, who must be considered coeval onarchaeological grounds, exhibitedradiocarbon dates which differed by around300 radiocarbon years, but in the “wrongdirection” relative to their diets, the individualwith the higher consumption of freshwaterfish as indicated by stable isotope analysishaving the later date, contrary toexpectations. It may be that the reservoir effectfor Zvejnieki (if present at all) is considerablysmaller than that found at the Iron Gates. Anattempt to estimate the extent of anyfreshwater reservoir effect has been initiated,and the findings will be discussed in a futurepaper (Lidén & Eriksson forthcoming). Thechronological trends seen at Zvejniekinevertheless seem to be valid regardless of anyreservoir effect (see Paper III for a detaileddiscussion).

4.6 Methodological concerns

There are some pitfalls involved in the analysisand subsequent interpretation of stableisotopes in collagen which I would like tobring to focus here in addition to thediscussion of intra-individual change, faunaldata and representativeness. One concernscontamination, i.e. the inclusion of substancesother than that intended for analysis, whichcould be caused by both prehistoric burialpractices and modern techniques of handlingskeletal remains at an excavation or in thelaboratory. Another is diagenesis, chemicalalteration caused by degradation of theskeletal remains. Both processes will distortthe stable isotope signature. Fortunately, thereare a number of precautions and measures

that can be employed in order to avoidcontamination and ensure the integrity of thecollagen.

The sampling strategy employed here wasto discard the outermost layer when drillingsamples from a tooth or bone, preceded ifnecessary by immersing the element in distilledwater and subjecting it to ultrasonication forless than one minute to removecontaminations. The collagen extractionprotocol used (Brown et al. 1988) includes anultrafiltration step, which selects for high-molecular weight remnants, thereby isolatingthe collagen from degraded protein remnantsor from any contaminating substances whichare smaller than the >30 kD fraction. Inaddition to the precautions againstcontamination or degraded collagen duringextraction, there are a number of qualitycriteria which must be fulfilled for the sampleto produce reliable stable isotope data. Theseinclude an atomic C/N ratio within the range2.9–3.6 for unaltered collagen (DeNiro 1985),carbon and nitrogen concentrations withinthe limits for collagen from modern bone(Ambrose 1990), a given extraction yield (cf.van Klinken 1999), and visual appearance(Ambrose 1990).

The only source of nitrogen in thediet is the amino acids in protein, andexperiments have shown that carbon fromingested protein is routed to collagen(Ambrose & Norr 1993; Gearing 1991). Theδ13C and δ15N values for collagen thereforemainly reflect protein intake. The collagenδ13C value is enriched by some 5‰ comparedwith ingested protein δ13C (DeNiro 1985),although this enrichment factor variesbetween tissues and even between compoundsin the same tissue. Consequently it is veryimportant that any lipids, carbonates,contaminating substances or degradedcollagen should be removed from the collagenprior to analysis, in order not to distort theδ13C values. This fact should be emphasized,since different tissues and compoundstypically have different stable isotopesignatures even if they derive from the sameindividual. There is a commonmisunderstanding that the δ13C range forcollagen is applicable to other compounds as

25

norm and difference

well, a possibility most commonly discussed inconnection with the radiocarbon dating oforganic residues on pottery (e.g. Edenmo et al.1997, but see Isaksson et al. ms. for adiscussion; Persson 1997). A δ13C value of–22‰ should thus only be interpreted asterrestrial if derived from collagen.

Stable isotope analysis gives a measure ofhow uniform the diet is throughout apopulation, in that the standard deviation ofδ13C for a population with a homogeneousdiet has been estimated to be around 0.3‰(Lovell et al. 1986), a higher standarddeviation thus indicating differential foodintake. This was reported for a populationmainly living off terrestrial herbivores,however, and may be less valid for sealhunters, for example. The approach used atVästerbjers (Paper IV), to estimate ranges forexpectancy values based on faunal data, hasproved successful and should be appliedwhere possible (see also Schwarcz 1991). Thestandard deviation is nevertheless animportant measure, although the absolutevalue of 0.3‰ may not be applicable to allecological settings. The use of differentskeletal elements for analysis could alsoinfluence the variability and should thereforebe taken into account.

Ideally, the samples analysed should beconsistently from the same skeletal elements,and from individuals representative of thewhole population with regard to age, sex andmanner of deposition. However, thearchaeological and antiquarian reality is suchthat it has not always been possible tooptimize the selection of individuals or ofskeletal elements for analysis, because thematerial was not available, sampling notpermitted, or the elements of interest were notsufficiently well preserved (cf. Papers III, IV).Taken together, these deficiencies of courseimpede the assessment of possible genderdifferences, for instance. The apparentunderrepresentation of children andsubadults, on the other hand, is alleviated bythe fact that teeth from adults actuallyrepresent children that survived childhood – agroup that is normally invisible in thearchaeological record.

The present study was initiated in 1997,

and the many hundreds of samples includedwere processed continuously over the years,the mass spectrometry and elemental analysesbeing run in several batches at twolaboratories. Unfortunately, the precision andconsistency of the data from one of theselaboratories, that of the Institute of AppliedEnvironmental Research (ITM) at StockholmUniversity, proved insufficient for the presentpurposes, replicate measurements of some 150samples carried out at the Dept. of Geologyand Geochemistry, Stockholm University(hereafter GEO), having shown pooragreement. The ITM experienced problemswith uneven voltage, which may account forsome of the irregularities. Furthermore, therewas a consistent difference in the observed (byGEO) vs. the assumed values of the standardsused by ITM, showing a 0.6‰ offset in δ13Cand 0.4‰ in δ15N (i.e. the values recorded atGEO were higher in both cases). In addition,the reference gas used at ITM was isotopicallynot within the same range as the measuredsamples, which reduced the accuracy of themeasurements, and the ITM performedrecalibrations for every run, which resulted indifferences that varied for each batch ofsamples. Thus the mean inter-lab differencefor each ITM run varied from –0.2‰ to+0.9‰ for δ13C, and from –1.0‰ to +0.8‰ forδ15N, the standard deviations being as a rule ofthe order of 0.3‰ or less for both δ13C andδ15N (with some exceptions).

Not all the samples run at ITM could besuccessfully re-measured – either becausethere were insufficient amounts of collagenleft from the previous measurement, orbecause the remaining collagen had been usedfor radiocarbon dating. This was unfortunatefor several reasons, of course, not onlybecause the extractions were time-consumingand the analyses costly, but above all becausethe archaeological remains are not infinite. Inworking with prehistoric remains, measuresare always taken to avoid inflictingunnecessary damage on the material. Thisincludes drilling as small samples as possible,avoiding destruction of any morphologicalcharacters, and, not least, always making anassessment of the amount of informationgained in relation to the number of samples

26

gunilla eriksson

and the damage caused to the archaeologicalcollections. It is frustrating, of course, to havecaused damage to the material withoutachieving any data, but the same is alsoexperienced with material of insufficientquality and is “part of the deal”, so to speak.

The strategy in dealing with these problemswas (1) to reproduce as many measurementsas possible at GEO, (2) to calculate the offsetsin both δ13C and δ15N for each run at ITM, (3)to discard data derived from those samples forwhich the analysis could not be replicated atGEO, or (4) if the analysis could not bereplicated but the sample was of keyimportance and offsets could be calculatedfrom the same run, to correct the values. Thelatter correction was applied to only tensamples (Papers V, VI), where the crucialimportance of the sites made an estimation ofthe values justifiable. Nevertheless, it must bekept in mind that the precision of the estimatesis poorer than for the other samples.

4.7 Diet and “the Edible”

Our attitudes towards food are often strongand less rooted in biology than one would atfirst imagine. Although we all rely by natureon certain nutrients for our survival andhealth, diet is one of the most forcefulexamples of culturally dependent phenomena(Fieldhouse 1996). Cultural and social aspectsnot only govern what is considered edible in agiven society, but they also regulate whencertain foodstuffs or dishes should be eaten, orsometimes even by whom they should beeaten. Social identity and position could thusaffect both the amount and frequency ofaccess to certain foodstuffs, and the overallvariation in diet (Hastorf 1999). We consumeonly a limited proportion of the potential foodresources available, a fact that is equally validfor the present day and for Stone Age people,and thus what is considered edible in a givensociety is not predictable from the accessibleresources alone, neither is it self-evident thatfoodstuffs which were important on asymbolic or ritual level were actuallyconsumed to any great extent (Arrhenius1987; Fieldhouse 1996).

Since food habits are so closely bound tocultural and social identity, they are likely tobe as strong indicators of group identity as anyartefacts customarily used for attribution toan archaeological “culture”. The concept ofwhat is edible is accordingly central to theunderstanding of cultural identity formation.Feelings of disgust towards certain foodcategories are largely culturally conditioned,and will function as a divider between “us”and “them”. Similarly, the attraction ofcertain exotic edibles will sometimes serve as adriving force for conflicts or for inducingchange.

4.8 Mobility

One important aspect of the various Stone Agegroups studied is their degree of mobility. Ashas been pointed out by several researchers,there is no inevitable relationship betweenagriculture and sedentism, or betweenforaging and mobility (Orme 1981; Kent1989; Kelly 1992). On the whole, any clear-cut distinction between mobility andsedentism is difficult to make – it is rather amatter of degree. Large cemeteries such asZvejnieki, Skateholm and Västerbjers are ofcourse indicative of territoriality as such, butthis in effect says little about mobilitypatterns. The stable isotope evidence givessome more hints, but these are not applicablein all instances. At Skateholm, the wide rangeof human carbon isotope values and theirrelation to the faunal data reveal that some ofthe interred individuals must have spent sometime on the coast beyond the Baltic Sea proper,i.e. beside the Kattegatt, Skagerrak or theNorth Sea. Similarly, one Zvejnieki individualshowed signs of having a marine input in hisdiet, although it could not be determinedwhether this implied a change of residence orcommuting between the interior and the coast.An additional analysis of a tooth or anotherbone element from this individual could givefurther support for one or other of thesealternatives. The Västerbjers stable isotopedata show that all the individuals studied weredistinctly coastal, even though this couldentail both sedentism and mobility. For the

27

norm and difference

inland site of Ageröd, the faunal data indicatethat the values obtained from human remainswere from people who derived some of theirprotein from the Baltic, implying some kind ofmobility, while the Alby individual providessome indications as to differential input ofmarine protein between childhood andadulthood, but the differences are not hugeand the marine predominance isoverwhelming at all age for which data areavailable. The Hanaskede individual, on theother hand, does seem to have moved about abit, with a more pronounced marinecontribution during his childhood andpossible traces of seasonal mobility. Whilemarine indications for individuals depositedinland are evidence of some kind of mobility,the opposite, i.e. terrestrial/freshwaterindications for coastal individuals, could beinterpreted as evidence either of mobility or ofa different form of subsistence, whether basedon foraging or farming. This is the case for theindividuals from Rolfsåker and Skateholm VI,who were both recovered from coastal sitesbut were markedly terrestrial in their dietaryfocus. For the remaining individuals analysed,nothing can be said about mobility, since thereare only individual values which do notcontrast with the location.

4.9 The non-monolithic Stone Age

What is apparent when considering all thesites included in the present study in the lightof the stable isotope results is the greatcomplexity and diversity of the Stone Age inthis region. Two sites stand out as particularlyimportant for understanding this period in theprehistory of the Baltic, Zvejnieki andVästerbjers (Papers II, III, IV). Both comprisecemeteries with large numbers of inhumationsassociated with extensive cultural layers, andare characterized by hunter–fisher–gatherergroups. To fit a traditional perspective, whichsees the Mesolithic/Neolithic transition as oneencompassing a shift from foraging tofarming, both sites would thus have to beMesolithic, which is obviously not the case.The inland site, Zvejnieki, includes bothMesolithic and Neolithic burials and cultural

layers, while Västerbjers has a much narrowertime span, the Middle Neolithic (disregardingthe intrusive Bronze Age deposits).

The transition to farming has often beenportrayed as something inevitable andirreversible, i.e. once foraging groups cameinto contact with farming, they were bound toadopt it – it was supposedly a point of noreturn ( but cf. Bailey & Milner 2003; Rowley-Conwy 2001; e.g. Zvelebil 1996). Thearchaeological evidence from Västerbjers andZvejnieki indicate that this is anoversimplified view, however. There arenumerous indications of contact with coevalfarming groups (i.e. the Corded-Ware orBattle-Axe Culture) at Västerbjers and otherPitted-Ware sites on Gotland, but no evidenceof any agricultural activity, neitherpastoralism nor horticulturalism (Paper IV).The occurrence of battle-axes, antler stabbingweapons and sheep bone in Pitted-Wareburials show that they were well aware of thepossibility of practizang agriculture, but theyactively chose not to do so. The hypothesisthat the Corded-Ware and Pitted-WareCultures on Gotland represented the samepeople, who commuted between inland areasand the coast, can be refuted once and for all(cf. Andersson 1998, p. 73; Carlsson 1991;Carlsson 1998, p. 59ff; Persson 1986). In myopinion the Pitted Ware populations onGotland should be regarded as a separategroup with their own cultural identity, closelyconnected with seals (see Paper IV for adetailed argument).

Zvejnieki, on the other hand, does seem toinvolve a shift in diet, although not fromforaging to farming, but from a dietcompletely dominated by freshwater fish, toone also including noteworthy contributionsof birds and mammals, although still with aconsiderable intake of fish, and showing amuch higher variability than earlier. A parallelto the Västerbjers contacts with people whopractised farming could be observed in thecoastal contacts evident in the Zvejniekipopulation. Although there are obvious signsof frequent contact with people living by thecoast, there is as yet only one example of anindividual actually moving between the coastand the interior (Paper III).

28

gunilla eriksson

As for the other sites studied, there is nounambiguous evidence suggesting that onesingle factor, whether location, chronology orcultural attribution, accounted for all thevariability. There is simply no general pictureof either the Mesolithic or the Neolithic whichapplies to all the sites, despite the fact that theyare located within such a limited area. Theflexibility of the human mind defies allattempts to assign people to discrete andunchangeable entities which apply to largeregions.

The concept of variation is central to thepresent study; referring here to (dietary)difference or lack of difference within orbetween individuals or populations. Variationis also often taken to mean deviation from animplicit norm or standard, as is commonlyseen in the approach to the archaeologicalrecord, where the attempt to look for thetypical often results in neglect of the atypical,i.e. the variation. This tendency is obvious indiscussions of Neolithic “cultures”, since afixed and static notion of culture is unable todeal with variation, whether you call itculture, economy, tradition, ideology, lifestyleor social organization. The same goes for theone-sided arguments on the characters of theMesolithic and Neolithic, in which peripheralparts of Western Europe are held up asstandards of neolithization in Europe, basedon selective data from the British Isles which itis claimed are “consistent with what is knownfrom elsewhere in western Europe outsidenorthern and eastern Scandinavia” (Schulting& Richards 2002b:1015, emphasis added).Apart from being a very complicated way ofsaying Denmark, I resent the way data fromSweden, Åland, Estonia and Latvia areignored and marginalized. Moreover, theevidence from Denmark is much lessunequivocal than is suggested by this way ofdescribing the alleged transition. Stablecarbon isotope values reported for the Danishcoastal megalith Klokkehøj, for example, areintermediate with regard to the terrestrial–marine continuum, ranging from –19.9 to–18.1‰ (Lidén 1995a), i.e. they involve up toone third marine protein (employing the samemanner of calculation as Schulting &Richards 2002b).

The concepts of normality and thenormative have been challenged by Strassburg(2000), who attempted to solve this problemin his study of the South ScandinavianMesolithic by making the deviant, or queer,into the norm, claiming that cemeteries such asSkateholm represent queer and dangerousindividuals who had to be controlled and“normalized”. He didn’t really solve anythingby merely reversing the roles between the“normal” and the “queer”, however, since hestill failed to account for the variation.Moreover, although imaginative, hisinterpretations seem to have left the actualarchaeological evidence behind. A more viableperspective on the non-normative wouldperhaps be not to think of in terms of theconstant battle and resistance portrayed byStrassburg, but instead to concentrate on adiscrepancy between norm and practice,which need not involve a conflict. This gapbetween how things should be, i.e. the norm,and how things are, i.e. practice, could bebridged if one sees the norm as flexible,perceiving deviations from it as provisionalexceptions, and allowing for both individualvariation and human agency which may alterthe norm in the long run. Such an ability tofind excuses for all kinds of irregularbehaviour is a human trait which is quite easyto relate to, and which would account for thedifferences between norm and practice foundat Zvejnieki (Paper III) and Västerbjers (PaperIV), for instance. That people tend to giverather idealized accounts of their behaviour,which deviates significantly from thatobserved, is also a fact that is described intextbooks covering any kind of fieldwork inliving populations (e.g. Renfrew & Bahn1991; Rössner 1988).

4.10 “Nature” as a norm

Fieldhouse (1996:155), in his account of 19thcentury food reform, states: “Food reformershave always had a pre-occupation with a ‘nat-ural diet’. To call a food natural is to call itgood.”. This is equally true of today’s advo-cates of a “Neander diet” or “biologically nor-mal” food (Audette 1999; Ljungberg 1997).

29

norm and difference

The romantic idea of the happy Stone Age sav-age representing something original andpurely natural, not affected by culture and un-touched by the evil forces of present-day civili-zation, is of course nonsense, and is based onseveral misconceptions. First of all, we have todefine who these original Stone Age peoplewere – early hominids, Neanderthals,Palaeolithic Homo sapiens, any pre-Neolithicpeople, or just any hunter–gatherers at all?Having defined this, we then have to ask whatthey actually consumed. For pre-Neanderthalhominids, the evidence is very scanty, withconcurrent hypotheses pointing to scaveng-ing, predating or gathering as the means of ob-taining food (Lee-Thorpe et al. 1994; Sillen etal. 1995), whereas stable isotope studies onNeanderthals have yielded few results(Bocherens et al. 1991; Pettitt et al. 2000;Richards et al. 2001).. Even so, it must be explained why their diet isof relevance to modern humans – after all,they all became extinct. The key traits of ourspecies, Homo sapiens sapiens, appear to beadaptive and innovative abilities, which ledthese individuals to populate new ecologicalniches throughout prehistory, inevitably lead-ing to changes in diet, and to invent new tech-nology, including new ways of preparing food.This resulted in a great diversity in food strate-gies for hunter–gatherers, and in such a richrange of edibles that it is pointless to describethe variability by reference to “nature”. Toclaim that evolution shows that we should liveaccording to some allegedly “original” normis to ignore not only the diversity of humanpractice but also the actual facts of evolution.

The lactose tolerance of adults in cattle herd-ing populations in Africa and the Middle East,for example, as well as in the populations ofnorthern Europe, is a genetic adaptationwhich seems to have first occurred during theNeolithic (McCracken 1971). Thus those ofus who have this ability to break down the lac-tose in cow’s milk are no longer “original”.

A normative reference to supposedly “nat-ural” traits is also frequent in arguments thattry to justify present-day gender inequalities,in that biological essentialists tell us that suchdifferences are normal because men andwomen are intrinsically different (Thornhill1994) (Fig. 9). Not only is it claimed (withoutsupporting evidence) that the division of la-bour in traditional societies and in prehistoryis always a strict one, without overlap betweenmales and females, but it is also maintainedthat there will never be any gender equality aslong as women give birth – “her role as amother was given to her by evolution”(Burenhult 2002, my transl.). The prescriptiveand predictive value of prehistory in the handsof a biological essentialist is seemingly endless,and, needless to say, completely lacking incredibility. It seems to be based on a nature–nurture dichotomy, entailing a core of biology(nature) upon which some cosmetic casing inthe form of culture (nurture) is applied. Biol-ogy and culture cannot be separated that way,however, and since they continuously interact,one cannot extract one from the other. Conse-quently, to use an alleged situation in prehis-tory as a norm is to abuse archaeology and todisregard the archaeological evidence of diver-sity.

30

gunilla eriksson

In conclusion, a rich diversity in Stone Agedietary practice in the Baltic Region has beendemonstrated. Evidence ranging from theEarly Mesolithic to the Late Neolithic showthat neither chronology nor location alonecan account for this variety, but that there areinevitably cultural factors as well. Food habitsare culturally governed, and therefore wecannot automatically assume that people atsimilar sites will have the same diet.

Stable isotope studies are very importanthere, since they tell what people actuallyconsumed, not only what was available, orwhat one single meal contained. We shouldnot be deceived to infer diet from rituallydeposited remains, since things that werementally important was not always importantin daily life. Thus, although a ritual andsymbolic norm may emphasize certain foodcategories, these may in fact contribute verylittle to diet. By the progress of analysis ofintra-individual variation, new data on lifehistory changes have been produced, revealingmobility patterns, breastfeeding behaviourand certain dietary transitions. The inclusionof faunal data has proven invaluable for

5. Conclusions

understanding the stable isotope ecology of asite, and thereby improve the precision ininterpretations of human stable isotope data.The special case of dogs, though,demonstrates that this animal is not useful forinferring human diet, since dogs due to thenumber of roles they possess in human societycould, and in several cases has proven to,deviate significantly from humans in theirdiet.

When evaluating the radiocarbon data ofhuman and animal remains from the Pitted-Ware site Västerbjers Gotland, the importanceof establishing the stable isotope ecology of asite before making deductions on reservoireffects has been further demonstrated.

The main aim of this thesis has been todemonstrate the variation and diversity inhuman practice, challenging the view of a“monolithic” Stone Age. By looking atindividuals and not only on populations, thewhole range of human behaviour has beenaccounted for, also revealing the discrepancybetween norm an practice, which frequently isvisible both in the archaeological record andin present-day human behaviour.

31

norm and difference

Ambrose, S. H. 1990 Preparation and characteri-zation of bone and tooth collagen for isotopicanalysis. Journal of Archaeological Science 17,431—451.

Ambrose, S. H. & Norr, L. 1993 Experimentalevidence for the relationship of carbon isotoperatios of whole diet and dietary protein to thoseof bone collagen and carbonate. In PrehistoricHuman Bone: Archaeology at the MolecularLevel (ed. J. B. Lambert & G. Grupe), pp. 1—37. Berlin: Springer-Verlag.

Andersson, C. 1998 Kontinuitet i de mellan-neolitiska samhällsstrukturerna. In AktuellArkeologi VI, vol. 35 (ed. C. Fredengren & I.Tarsala), pp. 65—75. Stockholm: Dept. ofArchaeology, Stockholm University.

Arneborg, J., Heinemeier, J., Lynnerup, N.,Nielsen, H. L., Rud, N. & Sveinbjörnsdóttir, Á.E. 1999 Change of diet of the Greenlandvikings determined from stable carbon isotopeanalysis and 14C dating of their bones.Radiocarbon 41, 157—168.

Arrhenius, B. 1987 Prehistoric diet and archae-ology in the laboratory. In Swedish Archae-ology 1981—1985 (ed. A. Carlsson, D.Damell, P. Hellström, Å. Hyenstrand & A.Åkerlund), pp. 113—122. Stockholm: SvenskaArkeologiska Samfundet.

Audette, R. 1999 Neanderthin: Eat like a Cave-man to achieve a Lean, Strong, Healthy Body.New York: Saint Martins Press.

Bailey, G. & Milner, N. 2003 Coastal hunter—gatherers and social evolution: Marginal orcentral? Before Farming 2002/3_4 (1), 1—22.

Balasse, M., Ambrose, S. H., Smith, A. B. & Price,T. D. 2002 The seasonal mobility model forprehistoric herders in the South-western Capeof South Africa assessed by isotopic analysis ofsheep tooth enamel. Journal of ArchaeologicalScience 29, 917—932.

Balasse, M., Bocherens, H. & Mariotti, A. 1999Intra-bone variability of collagen and apatiteisotopic composition can be used as evidence ofa change of diet. Journal of ArchaeologicalScience submitted.

Balasse, M., Bocherens, H., Mariotti, A. &Ambrose, S. H. 2001 Detection of dietarychanges by intra-tooth carbon and nitrogenisotopic analysis: An experimental study ofdentine collagen of cattle (Bos taurus). Journalof Archaeological Science 28, 235—245.

Benecke, N. 1987 Studies of early dog remainsfrom northern Europe. Journal of Archaeolo-

gical Science 14, 31—49.Bocherens, H., Fizet, M., Mariotti, A., Lange-

Badre, B., Vandermeersch, B., Borel, J. P. &Bellon, G. 1991 Isotopic biogeochemistry (13C,15N) of fossil vertebrate collagen: Applicationto the study of a past food web includingNeandertal man. Journal of Human Evolution20, 481—-492.

Bonsall, C., Macklin, M. G., Andersen, D. E. &Payton, R. W. 2002 Climate change and theadoption of agriculture in north-west Europe.European Journal of Archaeology 5, 9—23.

Bonsdorff, E. & Blomqvist, M. 1993 Bioticcouplings on shallow water soft bottoms -examples from the northern Baltic Sea.Oceanography and Marine Biology: AnAnnual Review 31, 153—176.

Brown, T. A., Nelson, D. E., Vogel, J. S. &Southon, J. R. 1988 Improved collagenextraction by modified Longin method.Radiocarbon 30, 171—177.

Burenhult, G. 2002 Det ofullkomliga djuret:Kropp, själ och livsstil i ett evolutionärtperspektiv. Stockholm: Natur och Kultur.

Carlsson, A. 1991 Kust/inlands-problematikenunder mellanneolitisk tid i Södermanland,Östergötland och på Gotland - en fråga omperiferi och centrum? In Sentrum—periferisentra og sentrumdannelser gjennom før-historisk og historisk tid, vol. 64 (ed. B. Wik),pp. 117—123. Trondheim.

Carlsson, A. 1998 Tolkande arkeologi och svenskforntidshistoria: Stenåldern. Stockholm Studiesin Archaeology. Stockholm: Department ofArchaeology.

Clutton-Brock, J. & Noe-Nygaard, N. 1990 Newosteological and C-isotope evidence onMesolithic dogs: Companions to hunters andfishers at Star Carr, Seamer Carr and Konge-mose. Journal of Archaeological Science 17,643—653.

Coates, M.-M. 1993 Tides in breastfeedingpractice. In Breastfeeding and HumanLactation (ed. J. Riordan & K. G. Auerbach),pp. 3—26. Boston: Jones and Bartlett Publish-ers.

Cook, G. T., Bonsall, C., Hedges, R. E. M.,McSweeney, K., Boroneant, V., Bartosiewicz, L.& Pettitt, P. B. 2002 Problems of dating humanbones from the Iron Gates. Antiquity 76, 77—85.

Cook, G. T., Bonsall, C., Hedges, R. E. M.,McSweeney, K., Boroneant, V. & Pettitt, P. B.

6. References

32

gunilla eriksson

2001 A freshwater diet-derived 14C reservoireffect at the Stone Age sites in the Iron GatesGorge. Radiocarbon 43, 453—460.

Crockford, S. J. 2000 A commentary on dogevolution: Regional variation, breed develop-ment and hybridisation with wolves. In DogsThrough Time: An Archaeological Perspective,vol. S889 (ed. S. J. Crockford), pp. 295—312.Oxford: Archaeopress.

DeLoache, J. & Gottlieb, A. (ed.) 2000 A World ofBabies: Imagined Childcare Guides for SevenSocieties. Cambridge: Cambridge UniversityPress.

DeNiro, M., Schoeninger, M. J. & Hastorf, C. A.1985 Effect of heating on the stable carbon andnitrogen isotope ratios of bone collagen.Journal of Archaeological Science 12, 1—7.

DeNiro, M. J. 1985 Postmortem preservation andalteration of in vivo bone collagen isotoperatios in relation to palaeodietary reconstruc-tion. Nature 317, 806—809.

Dettwyler, K. A. 1995 A time to wean: Thehominid blueprint for the natural age ofweaning in modern human populations. InBreastfeeding: Biocultural Perspectives (ed. P.Stuart-Macadam & K. A. Dettwyler), pp. 39—73. New York: Aldine de Gruyter.

Edenmo, R., Larsson, M., Nordqvist, B. & Olsson,E. 1997 Gropkeramikerna - fanns de? Materiellkultur och ideologisk förändring. In Regionaltoch interregionalt: Stenåldersundersökningar iSyd- och Mellansverige, vol. 23 (ed. M.Larsson & E. Olsson), pp. 135—213. Stock-holm: Riksantikvarieämbetet.

Eichinger Ferro-Luzzi, G. 1980a Food avoidancesduring the puerperium and lactation inTamilnad. In Food, Ecology and Culture:Readings in the Anthropology of DietaryPractices, vol. 1 (ed. J. R. K. Robson), pp.109—117. New York: Gordon and BreachScience Publishers.

Eichinger Ferro-Luzzi, G. 1980b Food avoidancesof pregnant women in Tamilnad. In Food,Ecology and Culture: Readings in the Anthro-pology of Dietary Practices, vol. 1 (ed. J. R. K.Robson), pp. 101—108. New York: Gordonand Breach Science Publishers.

Eriksson, G. & Lidén, K. 2002 Mammalian stableisotope ecology in a Mesolithic lagoon atSkateholm. Journal of Nordic ArchaeologicalScience 13, 5—10.

Eriksson, G., Sjögren, C. & Lidén, K. 2000 Ingenamning utan avvänjning: Vem eller vadbestämmer? Amningsnytt 4/00, 6–7.

Fieldhouse, P. 1996 Food and Nutrition: Customsand Culture. Cheltenham: Stanley Thornes.

Fildes, V. 1995 The culture and biology ofbreastfeeding: An historical review of westernEurope. In Breastfeeding: Biocultural Perspec-tives (ed. P. Stuart-Macadam & K. A. Dett-wyler), pp. 101—126. New York: Aldine deGruyter.

Fogel, M. L., Tuross, N. & Owsley, D. W. 1989Nitrogen isotope tracers of human lactation inmodern and archaeological populations. InAnnual Report from the Director of theGeophysical Laboratory at the CarnegieInstitution of Washington 1988-1989, pp.111—117.

Gartner, L. M. & Stone, C. 1994 Two thousandyears of medical advice on breastfeeding:Comparison of Chinese and Western texts.Seminars in Perinatology 18, 532—536.

Gearing, J. N. 1991 The study of diet and trophicrelationships through natural abundance 13C.In Carbon Isotope Techniques (ed. D. C.Coleman & B. Fry), pp. 201—218. San Diego,California: Academic Press Inc.

Götherström, A. 2001 Acquired or InheritedPrestige? Molecular Studies of Family Struc-tures and Local Horses in Central Svealandduring the Early Medieval Period. Theses andPapers in Scientific Archaeology. Stockholm:Archaeological Research Laboratory, Stock-holm University.

Hastorf, C. A. 1999 Cultural implications of cropintroductions in Andean prehistory. In ThePrehistory of Food: Appetites for change, vol.32 (ed. C. Gosden & J. Hather), pp. 35—58.London: Routledge.

Hillson, S. 1996 Dental Anthropology. Cam-bridge: Cambridge University Press.

Hobson, K. A. 1998 Stable isotope analyses oftooth annuli reveal temporal dietary records:An example using Steller sea lions. MarineMammal Science 14, 116—129.

Jelliffe, D. B. & Jelliffe, E. F. P. 1978 Human Milkin the Modern World. Oxford: OxfordUniversity Press.

Johannisson, K. 1997 Kroppens tunna skal: Sexessäer om kropp, historia och kultur. Stock-holm: Norstedts.

Katzenberg, M. A. 1989 Stable isotope analysis ofarchaeological faunal remains from SouthernOntario. Journal of Archaeological Science 16,319—329.

Katzenberg, M. A. 1992 Advances in stableisotope analysis of prehistoric bone. In SkeletalBiology of Past Peoples: Research Methods (ed.S. R. Saunders & M. A. Katzenberg), pp. 105—119. New York: Wiley-Liss.

Katzenberg, M. A., Herring, D. A. & Saunders, S.R. 1996 Weaning and infant mortality:Evalutating the skeletal evidence. Yearbook ofPhysical Anthropology 39, 177—199.

Katzenberg, M. A. & Pfeiffer, S. 1995 Nitrogenisotope evidence for weaning age in a nine-teenth century Canadian skeletal sample. InBodies of Evidence: Reconstructing historythrough skeletal analysis (ed. A. L. Grauer), pp.221—235. New York: Wiley-Liss.

Kelly, R. L. 1992. Mobility/sedentism: Concepts,archaeological measures, and effects. AnnualRevue of Anthropology 21:43–66

33

norm and difference

Kent, S. 1989. Cross cultureal perceptions offarmers as hunters and the value of meat. InFarmers as hunters: The implications ofsedentism. Edited by S. Kent. New Directionsin Archaeology. Cambridge University Press.Cambridge. pp. 1–17.

Knutsson, H. 1995 Slutvandrat? Aspekter påövergången från rörlig till bofast tillvaro. Aun.Uppsala: Societas Archaeologica Upsaliensis.

Koch, P., Heisinger, J., Moss, C., Carlson, R. W.,Fogel, M. L. & Behrensmeyer, A. K. 1995Isotopic tracking of change in diet and habitatuse in African elephants. Science 267, 1340—1343.

Lanting, J. N. & van der Plicht, J. 1996 Wathebben Floris V, skelet Swifterbant S2 envisotters gemeen? (What do Count Florence V,skeleton Swifterbant S2 and otters have incommon?). Palaeohistoria 37/38, 491—519.

Lanting, J. N. & van der Plicht, J. 1998 Reservoireffects and apparent 14C-ages. The Journal ofIrish Archaeology 9, 151—165.

Larsson, L. 1990 Dogs in fraction - symbols inaction. In Contributions to the Mesolithic inEurope: Papers presented at the FourthInternational Symposium “The Mesolithic inEurope”, Leuven 1990 (ed. P. M. Vermeersch& P. V. Peer), pp. 153—160. Leuven: LeuvenUniversity Press.

Lepiksaar, J. 1984 Die frühesten Haustiere derskandinavischen Halbinsel, insbesondere inSchweden. In Der Beginn der Haustierhaltungin der “Alten Welt”, vol. 9 (ed. G. Nobis), pp.221—266. Köln: Böhlau Verlag.

Lidén, K. 1995a Megaliths, agriculture, and socialcomplexity: A diet study of two Swedishmegalith populations. Journal of Anthropologi-cal Archaeology 14, 404—417.

Lidén, K. 1995b Prehistoric Diet Transitions: AnArchaeological Perspective. Theses and Papersin Scientific Archaeology 1. Stockholm:Archaeological Research Laboratory, Stock-holm University.

Lidén, K. & Angerbjörn, A. 1999 Dietary changeand stable isotopes: A model of growth anddormancy in cave bears. Proceedings of theRoyal Society of London: Series B 266, 1779—1783.

Lindeberg, S. 1997 Hälsa och ohälsa underforntiden - om evolutionsmedicinens möj-ligheter. In Ajvide och den moderna arkeologin(ed. G. Burenhult), pp. 105—130. Stockholm:Natur och Kultur.

Liukko, A. 1996 Mat, kropp och social identitet.Stockholm: Pedagogiska institutionen,Stockholms universitet.

Ljungberg, T. 1995 Vad är naturligt för mitt barn:en introduktion till hur man som förälder kanta hand om sitt barn på ett mer naturligt ochbiologiskt ursprungligt sätt. Nyköping: Exiris.

Ljungberg, T. 1997 Den biologiska människan -humanetologi och stenålderns livsmönster. In

Ajvide och den moderna arkeologin (ed. G.Burenhult), pp. 105—130. Stockholm: Naturoch Kultur.

Lõugas, L., Lidén, K. & Nelson, D. E. 1996Resource utilization along the Estonian coastduring the Stone Age. In Coastal Estonia:Recent Advances in Environmental andCultural History, vol. 51 (ed. T. Hackens, S.Hicks, V. Lang, U. Miller & L. Saarse), pp.399—420. Rixensart: Council of Europe.

Lovell, N. C., Nelson, D. E. & Schwarcz, H. P.1986 Carbon isotope ratios in palaeodiet: Lackof age or sex effect. Archaeometry 28, 51—55.

Marino, B. D. & McElroy, M. B. 1991 Isotopiccomposition of atmospheric CO

2 inferred from

carbon in C4 plant cellulose. Nature 349,127—131.

McCracken, R. D. 1971 Lactase Deficiency: AnExample of Dietary Evolution. CurrentAnthropology 12, 479—517.

Nelson, D. E. 1998 Radiocarbon dating definitionsand their consequences, pp. 1—9. Vancouver:Archaeology Dept, Simon Fraser University.

Noe-Nygaard, N. 1988 δ13C-values of dog bonesreveal the nature of changes in Man’s foodresources at the Mesolithic—Neolithictransition, Denmark. Chemical Geology(Isotope Geoscience Section) 73, 87—96.

Noe-Nygaard, N. 1995 Ecological, sedimentary,and geochemical evolution of the late-glacial topostglacial Åmose lacustrine basin, Denmark.Fossils and Strata. Oslo: ScandinavianUniversity Press.

Nordqvist, B. 2000 Coastal Adaptations in theMesolithic: A study of coastal sites withorganic remains from the Boreal and Atlanticperiods in Western Sweden. GOTARC Series B.Gothenburg Archaeological Theses. Göteborg:Dept. of Archaeology, Göteborg University.

Ojaveer, E. & Elken, J. 1997 On regional subunitsin the ecosystem of the Baltic Sea. In Proceed-ings of the 14th Baltic Marine BiologistsSymposium: Biodiversity in the Baltic ecosys-tems, Littoral processes, Anaerobic processes(ed. E. Ojaveer), pp. 156—169. Tallinn:Estonian Academy Publishers.

Olsen, S. L. 2000 The secular and sacred roles ofdogs at Botai, North Kazakhstan. In DogsThrough Time: An Archaeological Perspective,vol. S889 (ed. S. J. Crockford), pp. 71—92.Oxford: Archaeopress.

Olsson, I. U. 1986 13C-variationer - deras orsak,storlek och användbarhet. Fornvännen 81,114—120.

Orme, B. 1981. Anthropology for Archaeologists:An Introduction. London: Duckworth

Paaver, K. 1965 Formirovanie teriofauny iizmencivost mlekopitajuscich Pribaltiki vgolocene. Tartu.

Parker Pearson, M. 1999 The Archaeology ofDeath and Burial. College Station: TexasA&M University Press.

34

gunilla eriksson

Pate, F. D. 1994 Bone chemistry and paleodiet.Journal of Archaeological Method and Theory1, 161—209.

Persson, P. 1986 Några kommentarer till en listaöver STY boplatsfynd från den svenskavästkusten. In Stridsøksetid i Sydskandinavien:Beretning fra et symposium 28.—30.X1985 iVejle (ed. C. Adamsen & K. Ebbesen), pp.266—275. København: Forhistorisk Arkæo-logisk Insititut, Københavns Universitet.

Pettitt, P.B., Trinkaus, E., Smith, F.H., Karavanic,I., & Paunovic, M. 2000. Neanderthal Diet atVindija and Neanderthal Predation: TheEvidence from Stable Isotopes. The Proceed-ings of the National Academy of Sciences, USA97:7663-7666.

Persson, P. 1997 Kontinuitet mellan senmesolitiskoch mellanneolitisk fångstkultur i Syd-skandinavien. In Till Gunborg: Arkeologiskasamtal, vol. 33 (ed. A. Åkerlund, S. Bergh, J.Nordbladh & J. Taffinder), pp. 379—388.Stockholm: Stockholms universitet.

Persson, P. 1998 DNA från ben från megalitgravari centrala Västergötland. In 3. NordiskeSeminar om Biologisk Antropologi: ClaraLachmann Symposium 1998. AntropologiskLaboratorium, Retsmedicinsk Institut,Københavns Universitet.

Price, T. D. 1991 The Mesolithic of NorthernEurope. Annual Review of Anthropology 20,211-233.

Quandt, S. A. 1995 Sociocultural aspects of thelactation process. In Breastfeeding: BioculturalPerspectives (ed. P. Stuart-Macadam & K. A.Dettwyler), pp. 127—143. New York: Aldinede Gruyter.

Reid, D. J., Beynon, A. D. & Ramirez Rozzi, F. V.1998 Histological reconstruction of dentaldevelopment in four individuals from amedieval site in Picardie, France. Journal ofHuman Evolution 35, 463—477.

Renfrew, C. & Bahn, P. 1991 Archaeology:Theories, Methods, and Practice. London:Thames and Hudson.

Richards, M. P., Pearson, J. A., Molleson, T. I.,Russell, N. & Martin, L. 2002 Stable isotopeevidence of diet at Neolithic Catalhöyük,Turkey. Journal of Archaeological Science 29.

Richards, M. P., Pettitt, P. B., Stiner, M. C. &Trinkhaus, E. 2001 Stable isotope evidence forincreasing dietary breadth in the Europeanmid-Upper Paleolithic. Proceedings of theNational Academy of Sciences 98, 6528—6532.

Riordan, J. 1993 The cultural context of breast-feeding. In Breastfeeding and Human Lacta-tion (ed. J. Riordan & K. G. Auerbach), pp.27—48. Boston: Jones and Bartlett Publishers.

Riordan, J. & Auerbach, K. G. 1993 Breastfeedingand Human Lactation. Boston: Jones andBartlett Publishers.

Rowley-Conwy, P. 2001 Time, change and the

archaeology of hunter—gatherers: Howoriginal is the ‘Original Affluent Society’. InHunter—Gatherers: An InterdisciplinaryPerspective, vol. 13 (ed. C. Panter-Brick, R. H.Layton & P. Rowley-Conwy), pp. 39—72.Cambridge: Cambridge University Press.

Rössner, S. 1988 Boken om fetma. Stockholm:Natur och kultur.

Schoeninger, M. J. & Moore, K. 1992 Bone stableisotope studies in archaeology. Journal ofWorld Prehistory 6, 247—296.

Schulting, R. J. 1994 The hair of the dog: Theidentification of a Salish dog-hair blanket fromYale, British Columbia. Canadian Journal ofArchaeology 18, 57—76.

Schulting, R. J. & Richards, M. P. 2000 The use ofstable isotopes in studies of subsistence andseasonality in the British Mesolithic. InMesolithic Lifeways: Current Research fromBritain and Ireland, vol. 7 (ed. R. Young), pp.55—65. Leicester: University of Leicester.

Schulting, R. J. & Richards, M. P. 2002a Dogs,ducks, deer and diet: New stable isotopeevidence on Early Mesolithic dogs from theVale of Pickering, North-East England. Journalof Archaeological Science 29, 327—333.

Schulting, R. J. & Richards, M. P. 2002b Findingthe coastal Mesolithic in southwest Britain:AMS dates and stable isotope results on humanremains from Caldey Island, south Wales.Antiquity 76, 1011—1025.

Schulting, R. J. & Richards, M. P. 2002c The wet,the wild and the domesticated: The Meso-lithic—Neolithic transition on the west coast ofScotland. European Journal of Archaeology 5,147—189.

Schurr, M. R. 1997 Stable nitrogen isotopes asevidence for the age of weaning at the Angelsite: A comparison of isotopic and demo-graphic measures of weaning age. Journal ofArchaeological Science 24, 919—927.

Schwarcz, H. P. 1991 Some theoretical aspects ofisotope paleodiet studies. Journal of Archaeolo-gical Science 18, 261—275.

Schwarcz, H. P. & Schoeninger, M. J. 1991 Stableisotope analyses in human nutritional ecology.Yearbook of Physical Anthropology 34, 283—321.

Sealy, J., Armstrong, R. & Schrire, C. 1995Beyond lifetime averages: Analysis of differentcalcified tissues from archaeological humanskeletons. Antiquity 69, 290—300.

Sealy, J. C. 2001 Body tissue chemistry andpalaeodiet. In Handbook of ArchaeologicalSciences (ed. D. R. Brothwell & A. M. Pollard),pp. 269—279. Chichester: John Wiley & SonsLtd.

Serpell, J. 1995 From paragon to pariah: Somereflections on human attitudes to dogs. In TheDomestic Dog: Its evolution, behaviour, andinteractions with people (ed. J. Serpell), pp.245—256. Cambridge: Cambridge University

35

norm and difference

Press.Short, R. V. 1984 Breast feeding. Scientific

American 250, 23—29.Steele, D. G. & Bramblett, C. A. 1988 The

Anatomy and Biology of the Human Skeleton.College Station: Texas A&M University Press.

Strassburg, J. 2000 Shamanic shadows: Onehundred generations of undead subversion insouthern Scandinavia, 7000—4000 BC.Stockholm Studies in Archaeology. Stockholm:Arkeologiska institutionen.

Stuart-Macadam, P. 1995 Biocultural perspectiveson breastfeeding. In Breastfeeding: BioculturalPerspectives (ed. P. Stuart-Macadam & K. A.Dettwyler), pp. 1—37. New York: Aldine deGruyter.

Stuiver, M. & Braziunas, T. F. 1993 Modellingatmospheric 14C influences and 14C ages ofmarine samples to 10,000 BC. Radiocarbon 35,137—189.

Taylor, R. E. 1992 Radiocarbon dating of bone: Tocollagen and beyond. In Radiocarbon AfterFour Decades: An Interdisciplinary Perspective(ed. R. E. Taylor, A. Long & R. S. Kra), pp.375—402. New York: Springer-Verlag.

Taylor, R. E., Stuiver, M. & Reimer, P. J. 1996Development and extension of the calibrationof the radiocarbon time scale: Archaeologicalapplications. Quaternary Science Reviews 15,655—668.

Thornhill, N. 1994. Why are we different? In G.Burenhult (ed.): The Illustrated Encyclopediaof Humankind.

van Klinken, G. J. 1999 Bone collagen qualityindicators for palaeodietary and radiocarbonmeasurements. Journal of ArchaeologicalScience 26, 687—695.

White, C. D., Pohl, M. E. D., Schwarcz, H. P. &Longstaffe, F. J. 2001 Isotopic evidence forMaya patterns of deer and dog use at PreclassicColha. Journal of Archaeological Science 28,89—107.

White, C. D. & Schwarcz, H. P. 1989 AncientMaya diet: As inferred from isotopic andelemental analysis of human bone. Journal ofArchaeological Science 16, 451—474.

Wiedemann, F., Bocherens, H., Mariotti, A., vonden Driesch, A. & Grupe, G. 1999 Methodo-logical and archaeological implications ofintra-tooth isotopic variations (δ13C, δ18O) inherbivores from Ain Ghazal (Jordan, Neo-lithic). Journal of Archaeological Science 26,697—704.

Wilson, C. S. 1980 Food taboos of childbirth: TheMalay example. In Food, Ecology and Culture:Readings in the Anthropology of DietaryPractices, vol. 1 (ed. J. R. K. Robson), pp. 67—74. New York: Gordon and Breach SciencePublishers.

Wright, L. E. & Schwarcz, H. P. 1998 Stablecarbon and oxygen isotopes in human toothenamel: Identifying breastfeeding and weaningin prehistory. American Journal of PhysicalAnthropology 106, 1—18.

Wright, L. E. & Schwarcz, H. P. 1999 Correspond-ence between stable carbon, oxygen andnitrogen isotopes in human tooth enamel anddentine: Infant diets at Kaminaljuyú. Journal ofArchaeological Science 26, 1159-1170.

Zvelebil, M. 1996 The agricultural frontier andthe transition to farming in the circum-Balticregion. In The Origins and Spread of Agricul-ture and Pastoralism in Eurasia (ed. D. R.Harris), pp. 323—345. London: UCL Press.