Runsa A hilltop settlement during the Migration Period

Runsa A hilltop settlement during the Migration Period
Runsa
A hilltop settlement during the Migration Period
Distinguishing spatiality and organization through analyzing
chemical imprints of daily activities.
Masters Thesis VT 2012
Christoffer Andersson
Supervisors: Sven Isaksson, Malgorzata Wojnar Johansson & Kjell Persson.
The Archaeological Research Laboratory
Stockholm University
Runsa – A hilltop settlement during the Migration Period
Distinguishing spatiality and organization through analyzing chemical imprints of daily
activities.
Abstract
Archaeologists have long noted the striking monumentality and large-scale efforts behind the
Iron Age hilltop settlements. Yet, because of limited excavations, they represent a controversial
part of the Migration Period society and much of their function remains hidden. This paper deals
with questions concerning the inner organization and activities that took place within the Iron
Age hilltop settlement at Runsa. The study is linked to the ongoing project ”Runsa fornborg – En
befäst centralplats i östra Mälardalen under folkvandringstid” which aims to investigate the
socio-political functions of Runsa. In an attempt to establish a nuanced picture and distinguish
space use within the hilltop settlement, a multi-variable approach is used. Alongside more
traditional methods, element analysis by atomic absorption spectrophotometer (AAS) and lipid
analysis by gas chromatography-mass spectrometry (GC-MS) is emphasized.
Keywords: Migration Period, hilltop settlement, spatial organization, geochemistry, lipids, metal
elements, Runsa, soil, vessel-use.
Cover illustration: Model of Runsa from Agaton television, produced for Svt/Vetenskapens värld
Acknowledgements: Writing a thesis is a long-spun process in which a large number of persons
assist you and contributes to your work. Starting from the initial phase, I would like to thank
Michael Olausson for introducing me to the ever so exciting field investigations at Runsa, and
especially for giving me access to unpublished material. Thanks are also directed to the people
who helped me with the logistics: Ola Winter, Magnus Lindberg and Gustav Gonelius Stenvall
my long-time fellow companion and source of inspiration through my years as a student in
human geography and archaeology. My field companions from two excavation seasons at Runsa
deserve an acknowledgement for their patience during the sampling procedures. During my time
at the Archaeological Research Laboratory (AFL) several persons contributed, each with their
own knowledge. Thanks are primarily directed to my supervisors: Kjell Persson for helping me
with the phosphate analysis and GIS-work, Maria Wojnar-Johansson for instructing me and
teaching me about the AAS, Sven Isaksson for his support both regarding the GC-MS and the
writing. Nathalie Dimc deserves a special acknowledgement for her invaluable help, always
taking her time to assist me with the laboratory work, despite being very busy herself. People
outside the archaeological sphere deserves to be mentioned here as well; I would like to thank my
mother for helping me out with otherwise time-consuming events in my private life and finally I
thank my partner Nathalie Rullander for living with an extra noisy refrigerator, storing my soil
and pottery samples, next to the bed.
1
TABLE OF COTETS
1. INTRODUCTION ......................................................................................................................... 2
1.1 Background ............................................................................................................................... 3
1.2 Area and material....................................................................................................................... 3
1.3 Aims and structure ..................................................................................................................... 4
2. RUNSA – A HILLTOP SETTLEMENT IN THE EASTERN MÄLAR VALLEY DURING
MIGRATION PERIOD ...................................................................................................................... 5
2.1 Introduction and surroundings .................................................................................................... 5
2.2 The courtyard ............................................................................................................................ 6
2.3 Earlier excavations – results and interpretations ........................................................................... 7
3. THE IRON AGE HILLTOP SETTLEMENTS ................................................................................. 9
3.4 Definition .................................................................................................................................. 9
3.4.1 Distribution and settlement structure ............................................................................................ 9
3.4.2 Previous excavations and findings .............................................................................................. 10
3.4.3 Interpretations of functions ......................................................................................................... 11
3.4.4 The courtyard - activity areas ..................................................................................................... 13
3.5 Formulation of the research gap ................................................................................................ 13
4. ANALYTICAL TECHNIQUES .................................................................................................... 14
4.1 Geochemistry and vessel use .................................................................................................... 14
4.1.1 Phosphates .................................................................................................................................. 14
4.1.2 Metal elements ............................................................................................................................ 15
4.1.3 Lipids .......................................................................................................................................... 18
4.2 Sampling strategies .................................................................................................................. 22
4.2.1 Sampling in trenches ................................................................................................................... 22
4.3 Source criticism and problems connected to the sampling ........................................................... 23
5. RESULTS ................................................................................................................................... 24
5.1 Excavations 2011 ..................................................................................................................... 24
5.2 Terrace I - geochemistry ........................................................................................................... 27
5.3 Terrace III - geochemistry ........................................................................................................ 32
5.4 Trench X - geochemistry .......................................................................................................... 35
5.5 Vessel use ............................................................................................................................... 37
6. DISCUSSION AND INTERPRETATION ..................................................................................... 40
6.1 The prominent hall-building? .................................................................................................... 40
6.2 Dwelling and crafting or a Harg? .............................................................................................. 43
6.3 Synthesis ................................................................................................................................. 47
7. SUMMARY AND CONCLUSIONS ............................................................................................. 51
8. REFERENCES ............................................................................................................................ 52
2
1. ITRODUCTIO
1.1 Background
To date, the knowledge about Migration Period settlements is both qualitative and quantitative,
however due to archaeological excavations being directed by exploitation, certain sites makes up
missing puzzle pieces – hilltop settlements being one of them. During the early 20th century the
interest in hilltop settlements was flourishing and several research-excavations were undertaken
(Schnittger 1908a, 1908b, 1909, 1913 ATA; Gihl 1918; Hermelin 1929; Nordén 1938). These
however, were of obvious reasons implemented using old archaeological documentary methods,
why the information of find circumstances is limited. By the time of the mid-century, two larger
excavations took place on Swedish ground, whereof Eketorp at the island of Öland, although
somewhat different to mainland hilltops, is the best known site when discussing settlement
features (Borg et al. 1976). The other excavation, at Darsgärde, remains unpublished. Over the
last couple of decades interest once again have grown among archaeologists, leading to the
establishment of two larger projects ´Ett fornborgsprojekt i Rekarnebygden, Södermanland´
(Lorin 1985) and ´Strongholds and Fortifications in Central Sweden AD 400-1100´(Olausson
2009:6). Certainly these projects have contributed to considerable advances regarding dating,
exterior attributes (Damell & Lorin 2010) and the overall character (Olausson 2008, 2009);
nevertheless the knowledge gained around hilltop settlements are seemingly mainly of a
contextual nature where the site is discussed based on the surroundings (Törnqvist 1993, Damell
1993, Wall 2003) or only notifications of whether a settlement can be verified or not (Damell &
Lorin 2010). Although attempts have been made to deepen the understanding of mainland hilltop
settlements (Olausson 2008, 2009), interpretations are based on limited material; thereby still
hypothetical and in need of being evaluated. The main reason for this neglect can be traced in the
fact that the majority of the investigations have been minor, the aim is consequently seldom to
discuss the internal structure (see Damell & Lorin 2010). The most recent project round hilltop
settlements,”Runsa fornborg – En befäst centralplats i östra Mälardalen under folkvandringstid”,
is focused to the hilltop in Runsa, Eds parish, Uppland. The project can be seen as an attempt to
approach the socio-political role of a specific hilltop settlement. The investigation is of an
interdisciplinary character, trying to establish a deepened understanding of Runsa at a micro-level
in order to discuss the site in relation to the hinterland.
This thesis adds the knowledge gained mainly from geochemical methods, which are used to
discuss and identify activity areas. Successful results have previously been reached on Swedish
ground (Isaksson 2000a, Hjulström 2008) admittedly stressing the potential of geochemistry
analyses in deepened settlement studies. The basic archaeological assumption is that human
agency generates deposition of artifacts, however the less visible residues originating from waste
disposal, food preparing, craft production, stabling etc. of both solid and liquid character
(Middleton 2004), have been less observed. These residues are known as chemical imprints in the
soil, and when studied together with the traditional parameters findings, constructions and
location, they are thought to add another dimension in the interpretation of the internal
organization of the Runsa hilltop settlement.
1.2 Area and material
Based on the many settlement traces above ground, Runsa, situated in southern Uppland, is a
good example of a hilltop settlement during the Migration Period. It is currently being
investigated by associate Professor Michael Olausson and the author of this thesis has had the
opportunity to participate at the excavations over the past two years. It is viewed as a part of the
context of stone wall systems within the eastern Mälar Valley, located mainly along the northern
shore of Lake Mälaren. Examples from settlement studies within this area and hilltop settlements
located in southern Sweden are referred to in the text, whereof several of the most important
investigated sites are plotted on the map in fig. 1 and 2.
3
Figure 1. Left: Southern Sweden with selected sites mentioned in text (Bergström 2007:191).
Figure 2. Right: The Mälar Valley with reference points and important sites mentioned in the text. 1) Vendel 2)
Valsgärde 3) Darsgärde 4) Runsa 5) Sanda 6) Gåseborg 7) Stockholm 8) Alby 9) Helgö.
The material in focus in this thesis consists mainly of soil samples and pottery collected during
the excavation at Runsa in 2011. Artifacts and constructions are however used to aid the
interpretations as well, together making up a study of the same phenomenon from multiple
angles. The newly acquired information is mainly obtained from the trenches that were excavated
during 2010 and 2011, but furthermore also put into a larger whole, i.e. the context of the hilltop
settlement.
1.3 Aims and structure
As a contribution to the research of Runsa, and on basis of what was stated above, the aim of the
thesis is to discuss the functions and spatial organization within the hilltop settlement. Using a
traditional archaeological approach combined with laboratory analyses, i.e. geochemical analyses
of soil samples and lipid food residue analyses of pottery, I intend to discern patterns and obtain
information concerning the planning of the hilltop settlement, house arrangement and activity
areas. In particular, I aim to approach the following questions:
(i)
The primary aim is to answer the question of how Terrace I and Terrace III were used.
Is it possible to discuss and identify the functions based on geochemistry, vessel use
and excavation results? Furthermore, can the obtained knowledge be used to
understand how the settlement was spatially organized?
(ii)
How can we understand Runsa in relation to the earlier investigations of Migration
Period settlements? Do the results contribute to a more nuanced picture? If so, does it
infer a separate function for the hilltop settlement of Runsa or does it fulfill a similar
role as the many ordinary and magnate farms in the area?
The thesis is constructed as follows: the first part of the text gives a review of the research of
Runsa and briefly covers the research of hilltop settlements in general. The second part focuses
on the methodological issues and how these will be applied and performed. Lastly the results are
4
presented, evaluated and interpreted, in particular on a micro-level but also to some extent in
relation to the research front that concern Migration Period settlements.
2. RUSA – A HILLTOP SETTLEMET I THE EASTER MÄLAR
VALLEY DURIG MIGRATIO PERIOD
2.1 Introduction and surroundings
“The settlement with the many houses”, is one of the proposed meanings of the place name
Runnhusa (Vikstrand 2011:43). The place name itself might indicate the special position and
characteristics that was connected to the Runsa hilltop during its active phase. Another
characteristic, the monumentality, is still striking today. During the active phase, it most certainly
gave a respectful impression in the landscape. Located to the northwestern point of Eds parish
and according to paleogeographic maps Runsa was an island during the active phase (Risberg
2011:50). However, the historical arable land connected to the medieval village, Runusum, is
about 5 to 7 m a.s.l, why a neck of land, joining the hilltop to the rest of Eds parish, probably was
established during the late Iron Age.
While the land-use directly southeast of Runsa seems to be very sparse throughout the Iron Age,
one of the most intensive land-use areas in the eastern Mälar Valley begins some 5 km southeast
from Runsa. With its focus to the Fresta and Hammarby parishes this area with its characteristic
systems of stone walls seems to have been fully colonized during the Migration Period (Ericson
& Hermodson 1994:28). The area has been subject to several excavations and is one of the
reasons behind today’s knowledge about the social organization during the midmost Iron Age
(see Olausson (eds.) 2008). The same is valid for Norrsunda parish, northeast of Runsa, where the
stratification between the settlements is pronounced, stretching from small simple households to
ordinary farms and magnate farms with rather large-scale crafts and hall-buildings (Renck 2009,
Hamilton & Vinberg 2011:92f).
What distinguishes Runsa from these settlements is not only the fortification and settlement
structure, but the absence of adjacent arable land and pastures as well (Olausson 2011b:15). The
possibilities for resource exploitation are in general very limited at the site. Chisholm (1965:114)
set up five location criteria for an agrarian settlement; water supply, availability of arable and
grazing land, fuel supply and availability of building material. Runsa, however, do not fulfill any
of these criteria completely. A water hole is situated in the center of the hilltop, but its capacity of
supplying the inhabitants is unexplored. Fuel and building material must have been brought to
Runsa from the lands in possession of the surrounding farms. This is demonstrated by the
rampart; a massive construction that have involved an enormous effort, probably not only from
the inhabitants. When it comes to the question about cattle and cultivation, it cannot be fully
excluded. However, it can be said that the settlement was not oriented towards agriculture and the
amount of animal bones found at the site is not in proportion to the possibilities of holding a
livestock at Runsa. Nonetheless, animals may have been grazing in the courtyard or on small
meadows beneath the outcrop. Another possibility for grazing is the leveled areas on top of
Kohagen, east of the hilltop, which obviously, judging from the name, have been used during
modern times.
To define the outer borders of a settlement is usually problematic, but in this case it is broadly a
foregone conclusion. Runsa offers in comparison to other contemporary Iron Age settlements, a
visible outline, the main wall, which in turn together with the topography narrows down the
search for houses and activity areas. This is said with the existence of sites as Hultberget in mind,
where contemporary terraces have been identified just outside the rampart (Damell & Lorin
2010:212). The hilltop is surrounded by steep outcrops in west and east, while the terrain is
somewhat less steep in the north and southeast where two entrances are situated. Outside the
southeastern entrance, a path cleared from stones leads up from the small valley below. Parallel
5
to the main wall, an offshoot leads down to approx. 8 m.a.s.l., which is equivalent to the sea level
during mid-Iron Age. The function of this additional wall is not known, however it has been
suggested as one of the probable locations for a harbor (Olausson 1996:9, Risberg 2011:50).
Judging from the topography, a rather advanced construction must have been necessary to
overcome the steep slope leading down to the inner, protected area of the outshoot. Nevertheless,
an island with a settlement like Runsa most reasonably had a proper landing area with jetties and
activities connected to the shore. Just east of the offshoot is a lower strip of land which has been
connected to the shoreline in the north and south during the Migration Period and holds two more
potential harbor areas, Lilla Borgviken and Stora Borgviken. The former, which include the
outshoot of the wall, is deeper and also holds the findings of a wooden log boat dated to 7th
century, found in a narrow trench intended for a telephone cable (Östmark 1976). The latter on
the other hand, houses the burial grounds (RAÄ 1, RAÄ 3) connected to the inhabitants of Runsa,
signaling property rights. These graves, situated on the former shoreline, were therefore possibly
consciously exposed to foreign ships approaching the Runsa harbor.
2.2 The courtyard
The inner area (fig. 3), protected by the rampart, can be divided into four separate sections.
Immediately within the south entrance is a flat open plateau with several traces of prehistoric
activities above ground (Olausson 1996:9, Olausson 2011a:226). The survey from 1992
suggested that within this area were three parallel terraces which were the remnants of houses.
50m
Figur 3. Runsa hilltop settlement, after Olausson 1996 (modified by the author).
6
Terrace I: Located at the highest point and in the center of this section, is the largest terrace at
Runsa which measures 38 m between the gables marked by larger blocks. The northern gablemarking is somewhat asymmetrical in its position in relation to the southern one. The width is
approximately 11 m and the overall impression is a terraced area with marked boundaries, also
named platåhus (sw.) (Olausson 1996:9, 14, Olausson 2011a:225f).
Terrace II: This terrace, situated east of House I, was originally interpreted as the remains of a
longhouse, measuring approx. 25 x 9 m. However, the excavations during 2009 and 2010 (see ch.
2.3) showed traces of a rather small shed rather than a longhouse.
Terrace III: West of Terrace I, only separated by a small outcrop, is a somewhat smaller and less
exposed terrace. Since only one gable is visible above ground, it is hard to determine the size of a
probable house.
The central section of the inner area does not include terraces; instead it is characterized by
outcrops and a topography unsuitable for housing. The northern part of the hilltop contains two
areas, separated by an outcrop going north-south. The western area is situated on a lower level,
while the eastern area is somewhat elevated in the central parts. The former is occupied by three
defined terraces.
Terraces IV, V, VI: On a separated section lie three rather similar terraces. They are situated
along the rocky outcrop at the central part of the hilltop and occupy the larger part of the area
towards the wall.
The fourth section harbors no obvious indices of housing, although several potential locations can
be spotted. These however, only have one terraced side, making the chance of them being natural
formations considerable.
2.3 Earlier excavations – results and interpretations
The very first proper archaeological excavation at Runsa was implemented in 1902 by prof.
Oscar Almgren (marked O.A in fig. 3). With the intention of examining the age of the rampart, a
25 m2 large trench was dug along the southwestern part of the wall. Two different cultural layers
were discovered, separated by a clayey layer without findings. The bottom layer yielded a
fragmented crucible while most of the findings were retrieved from the upper layer; a dice,
ceramics, crucibles, loom weights, rivets and bones from domestic animals. According to
Almgren, similarities between the findings on Runsa and Birka made it reasonable to date the
hilltop to approximately the same time. However, the spatial proximity to Sigtuna later
contributed to the interpretation of Runsa as a garrison related to the defense around the early
town (Gihl 1918:85, see Olausson 1996:4f).
It would take 90 years before any archaeological fieldwork were undertaken at Runsa again. This
time the work was focused on two areas. A 20 m2 trench was located to the wall at the southern
entrance (marked 1992 in fig. 3), while the larger trench, 52 m2, was located to the largest terrace,
T I, within the courtyard. From the smaller trench it was concluded that the rampart had been
burnt several times, hence rebuilt three or possibly four times (Olausson 1996:10). Directly inside
the wall, findings of ceramics and animal bones, indicating food processing, was found
throughout the trench. The central occupational layer also harbored a floor layer with a hearth,
bordered by a wall. Other findings included loom weights, a polishing stone, carbonized bread, a
whetstone, as well as a dress pin in bronze and a comb both dated to the late 5th century
(Olausson 1996:13). The uppermost layer of the house displayed somewhat less findings in
general, with the addition of slag. Overall, this gave the interpretation of a small Migration period
house situated along the rampart, with activities connected to food processing and other everyday
activities (Olausson 1996:13f). It has been stressed that the trench from 1902 show a rather
similar context, together possibly being the remains of several buildings along the wall (Olausson
2011a:240).
7
The trench within TI was located to the southern gable, stretching north with a width of 2m. An
additional minor test pit was excavated more centrally on the terrace. Remains of a large house
were uncovered, including the southern gable wall and several post holes. The findings were
generally sparse in comparison to the wall-trench mentioned above; they included ceramics, a
few fragments of loom weights and crucibles. However, more interesting finds was uncovered
here as well; a piece of bread and an iron-ring originating from a chain mail. The location, size
and settlement context, led to the preliminary interpretation of this house being a hall-building
(Olausson 1996:16, Bergström 2007:45). During the excavation and later on in the same year,
phosphate mapping was implemented throughout the whole inner area of the hilltop settlement
(Rudin 1992:4). The sampling points were located with 5 m interval and at approx. 15-30 cm
depth, depending on the actual soil depth. The results from the phosphate mapping showed
enhanced values at the western, lower part of the hilltop settlement, peaking at terraces V and VI.
Certain points, directly southwest of the northern entrance also indicate elevated values. Along
the eastern wall, and throughout the central and southern part, the levels were generally low
(Rudin 1992:26). The high phosphate-values on the western side of the settlement area led Rudin
to the suggestion of an eventual stable (V), a crafting area (VI), and a possible area for refuse
deposition (IV) (see Rudin 1992:24). However, these interpretations must be seen as preliminary
since trying to establish specific pre-historic activities based on just phosphate values is
problematic (see ch. 4.1.1).
In 2009 the project round Runsa once again went into a phase of excavations. The investigated
area on Terrace II resulted in a rather complex occupational layer. Remains of a smaller house
were found in the center of the terrace, with several levels with traces of reconstruction
superimposing the oldest phase. Two fragments of blue glass and three combs, found at different
levels within the area used most intensively, can be dated to the time-span 5th to early 6th century
A.D. Within the uppermost layer, two hearths were uncovered and later dated to the late 6th
century. Other artifacts found at Terrace II include, an awl, loom weights, whetstones, a bolt to a
shrine, a knife and a bead. Noteworthy are also the findings of sporadic crucibles, again
indicating the existence of metal crafting within the settlement (Olausson 2011a:232ff).
The uppermost part of the occupational layer in the north end of House I was also excavated
during 2010. A compact and low wall was discovered about three meters in from the stone-built
edge of the terrace. It was found separating the main part of the house from a small northern
gable, suggesting that the actual longhouse ended inside the wall. The use of the outer area was
diffuse; sporadic animal bones, ceramics, iron fragments and slag were found. A more detailed
picture of House I is given in chapter 5.
Several test-pits were also placed in the north and northeastern parts of the hilltop settlement. The
results from these investigations reveal what appears to be a dense settlement with traces of
postholes and clayey floor layers in most of the inner area of the hilltop settlement (see Olausson
2011a:226, 238). Among the concluding remarks from the early as well as the later research is the
manifold of traces of different activities that have been uncovered, along with some distinctive
artifacts, suggesting an elite-settlement during the Migration period (Olausson 1996:20,
2011:242f). In line with this and characteristic for almost all the trenches, except House I, are
also the large amount of unburnt animal bones (Olausson 1996:16), suggesting food handling and
presence of animals to an extent that greatly exceeds the potential of grazing inside and around
the hilltop settlement. When discussing the active phase of Runsa, 14C-datings from the
excavations correspond to the time-span 230-650 AD (Olausson 2011a:241), while the dateable
artifacts emphasize the central part of this period, 450-550 AD, possibly indicating a rather short
period of use (Olausson 2011a:238f).
8
3. THE IRO AGE HILLTOP SETTLEMETS
3.4 Definition
When stone-constructions are found encircling a hilltop, the general term given by archaeologists
is hillfort. Studies of a certain type of settlement, basically a site where house-terraces or other
indications of buildings are found upon a hilltop and surrounded by a rampart, have obtained a
separate title within the hillfort-category; hilltop settlements (sw. höjdbosättning). It should be
stated at this point that it is difficult to tell whether a hilltop have harbored a settlement or not. On
certain sites there are remains of several terraces that indicate that it has been populated at least
occasionally, other sites have a few findings from minor excavations that shows an Iron Age
activity. In line with the aim of this thesis the following review focuses on the hilltops where
there are substantiated reasons to believe it once harbored a settlement, i.e. visible terraces or
finds and datings indicating a hilltop settlement (see Damell & Lorin 2010). Nevertheless, it can
be concluded that the hypothesis given that practically all forts harbored dwellings (Anjou
1935:2ff), is not supported by the archaeological record.
3.4.1 Distribution and settlement structure
The distribution of hilltop settlements in Sweden displays a concentration to southern Uppland
and northern Södermanland where a total of about 20 hilltops with traces of dwellings are
identified. The eastern part of Östergötland is another focal point with approximately 13
examples of hilltop settlements (Olausson 2009:47, 60). Similarities, i.e. accumulation of houses
within a fortified area, is found within contemporary forts on Öland as well. These are however
different regarding their structure and exhibits settlements with a more circular structure called
´ringforts´, located to flat land areas. The disparities may however be partly explained by the
natural conditions (Wegraeus 1976). The location in the landscape differs between the mainland
hilltop settlements as well; central positions within the agrarian landscape are represented as well
as peripheral positions at the borderline between cultivated land and forested areas. Generally
though, they are situated at strategic positions with visual command over the surrounding
territory (Olausson 2009:44).
A characteristic feature regarding several hilltop settlements are the density and extent of houses
upon the yard. The ringforts is a separate category with a well-organized courtyard consisting of
radial house foundations along the wall and a central block within. The mainland hilltop
settlements display a more inconsistent picture; e.g. Gåseborg, Broborg and Darsgärde have a
similar density of terraces as the ringforts (Ambrosiani 1958:168, Löfstrand 1982, Olausson
1996:24; Carlström 2003:8), but not to the same extent and definitely not organized similarly (see
Näsman & Wegraeus 1976). The ringforts should rather be titled fortified villages (see Olausson
2009:47), when considering the fact that at least 12 contemporary farms where discovered at the
Eketorp II fort (Nordström & Herschend 2003:51). Other hilltop settlements however, as
Lundboborg in Uppland (Olausson 2008:30), Mjälleborgen in Jämtland and Männö fort in
Södermanland have a settlement structure that are reminiscent of a single farmstead
(Hemmendorff 1985:238, Olausson 1996:25). Furthermore, the concept also includes sites as
Fållnäs, where the only area suitable for buildings was excavated with the results of only a couple
of postholes and cultural layer being found, maybe indicating a single building (Olausson 2008).
Nevertheless, the density of buildings at several sites is striking and characteristic. Accumulated
building groups have also been found elsewhere, as in arable fields on the lowlands (Appelgren
2002, Östling & Larsson 2007:292ff), these settlements, however, have distinctly discernible
building groups representing different farmsteads. Most likely they have cooperated in some way
regarding the organization of arable land and meadows, but the fact that they are so intertwined
with the agrarian production is probably a major difference in comparison to most of the hilltop
settlements. To find similar planning which is contemporary and somewhat similar also regarding
economy, we must look upon sites as the impressive Helgö-complex. This site shows an
9
accumulated pattern of buildings, however still distinguishable into separate groups (Holmqvist et
al. 1970, Reisborg 1994:17ff). On the other hand the economy seems more specialized on crafts
and the question whether there has been a local agrarian production or not, is not settled as of yet
(see Carlsson 1988). Moreover, the presence of an aristocracy is another joint-factor to the hilltop
settlements. Summarizing the discussion over settlement structure it must be argued that the
organization of the hilltop settlements distinguishes itself in comparison to the known structure of
contemporary settlements on the lowlands. However, the amount of buildings is by no means
standardized, indicating that this category, hilltop settlements, is not uniform.
3.4.2 Previous excavations and findings
Initially it should be stressed that several of the mainland hilltop settlements have only been
partly excavated, and only a few sites has undergone more extensive excavations, e.g. Boberget
in Ö Stenby parish, Gullborg in Tingstad parish, Darsgärde in Skederid parish, and Runsa in Ed
parish (Schnittger 1908b, 1909, Nordén 1938:266f, Ambrosiani 1958, Olausson 1996, 2011).
The very first project concerning hilltop settlements with organized excavations were initially
undertaken at Runsa in Uppland and later focused to Östergötland where eleven hilltops to a
varying extent were excavated by Bror Schnittger during 1906-1913. When Arthur Nordén,
summarized his work in 1938, a total of 17 hilltops had seen the efforts of archaeologists; many
of the sites showing traces of settlement. However, Schnittgers and Nordéns research belong to a
period when the archaeological documentary technique was undeveloped, why a major part of the
information over the contexts and constructions are lost from these excavations since determining
any houses or other features was not the same issue as it is today. Nevertheless, the findings from
these excavations are of great value, however reflecting inconsistency and disparities. Three of
the forts have a relatively rich combination of findings. Except a high amount of ordinary
artifacts such as animal bones, ceramics, loom weights etc., they also display findings
characteristic for settlements which are generally defined as magnate farms; Gullborg - gaming
pieces, Roman glass, parts of a sword case, arrow heads, a gold rod, dress-pins and brooches
(Schnittger 1908a & Schnittger 1913 ATA); Boberget - a silver rod, gaming pieces, spearheads, a
brooch, bread (Schnittger 1908b, 1909) and Odenfors - gaming pieces, bread, arrow heads and
brooches (Nordén 1938:308f). Other hilltop settlements such as Brudberget, Braberg and
Borgberget display findings of a more ordinary character (Nordén 1938, Wahlberg 1964). Among
the characteristics are the amount of traces after textile production from hilltops in Östergötland
(Olausson 1987a, Hemmendorff 1992:13), including hundreds of fragments from loom weights in
Onssten and Gullborg (Schnittger 1913 ATA, Olausson 1987b:405). Noteworthy is that among
the reoccurring objects are also the strainers, i.e. perforated ceramics, which are found both at
sites with rich find combinations, and the more ordinary sites (Schnittger 1908b, Schnittger 1913
ATA, Nordén 1938).
Strainers have also been documented in Baldersborg, Södermanland and Gåseborg, Uppland
(Hermelin 1929:95, Carlström 2003). The findings from the hilltop settlements in Södermanland
are otherwise to a greater extent characterized by what archaeologists expect to find in the
ordinary Iron Age settlements; ceramics, unburnt- and burnt bone, loom weights and slag
(Hermelin 1929:93, Lorin 1989, Damell & Lorin 2007, Damell & Lorin 2010:210ff). The
relatively minor excavations within these hilltops might be part of the explanation to this
discrepancy. One of few exceptional findings is the engraving tool from Hultberget, Husby –
Rekarne parish (Damell & Lorin 2010:211).
The excavations of hilltops in Uppland have, besides from Darsgärde, also been very limited to
their extent. Minor excavations have been undertaken in Gåseborg, Järfälla parish and Broborg in
Husby-Långhundra parish (Löfstrand 1982, Carlström 2003). The latter resulted in the finding of
one of very few objects, a blue colored glass bead, with a late Iron Age dating found at hilltop
settlements. However, 14C-datings of a posthole resulted in a time span of 1470 BP ± 105 years
(Fagerlund 2009:16, 19). In Darsgärde the prestigious objects are absent, yet the findings and
10
contexts indicate an upper hierarchical level; a scythe, a ploughshare, a ferrule ax (sw. holkyxa), a
key, some clothing items and fragments from crucibles (Ambrosiani 1958:167).
The excavation of Darsgärde is the largest undertaken at a hilltop on the mainland and resulted in
about 20 houses being documented (fig. 4), the greater part of the archaeological investigation
however remains unpublished. The occurrence of crucibles and moulds for bronze casting as in
Darsgärde are features that seem to be more frequently occurring at hilltop settlements than
among other Migration Period settlements (see Hall 1992:33ff, Kangur 2004:24, Olausson
2008:31). The extent of the metal crafting is difficult to estimate, but so far seems rather limited
(Olausson 1987b:409). The metallurgical ceramics from Gåseborg might be an exception,
indicating production exceeding household needs (Carlström 2003:14ff, Olausson 2009:52) but
cannot be compared to sites as Helgö. Metal crafting though is generally a high status indication
during this period and even more so when traces after manufacturing of prestigious objects can be
identified as again in Gåseborg where traces of gold crafts have been documented (Kangur
2004:24). Within this general picture of the findings from hillfort settlements are sites with
obvious Roman influences; bread, rotary querns, agricultural iron-tools, pyramid-shaped loom
weights, along with other very rare findings as the engraving tool and decorated antler pins (see
Olausson 2009:52)
The question, whether hilltop settlements in general were used during the late Iron Age, has had a
lengthy discussion and was once again raised recently. The prevailing perception during early
20th century placed the Iron Age hilltops mainly within the Viking Age (Gihl 1918:81ff; Almgren
1934:171), this was later questioned with arguments connected to the crisis during the Migration
Period (e.g. Stenberger 1964:537ff), which better corresponded to the interpretations of
Östergötland´s hilltop settlements (Schnittger 1908a, 1908b, 1909, 1913 ATA). Based on the 14Cdatings from hilltops in northern Södermanland, Damell & Lorin (2010:218) claims the time span
for the use of hilltops as dwellings to be pre Roman Iron Age – Early Medieval Period, but with a
main period of use during late Roman Iron Age – Migration Period. This summary generally
seem to fit the other regions with a high density of Iron Age hilltop settlements as well; Uppland,
Öland och Östergötland (Wahlberg 1964, Wegraeus 1976:43, Engström 1984, Olausson
1987a:92, 1987b:401ff). This is a somewhat simplified picture though, since forts on Öland have
an obvious renaissance during the Viking Age and the extent of the continuity in use of hilltops
on the mainland after the Migration Period still is rather unknown. However, the only house
construction linked to a hilltop settlement that has been dated to late Iron Age on the mainland is
located outside the rampart (see Damell & Lorin 2010:218ff).
3.4.3 Interpretations of functions
In the works investigating hillforts and hilltop settlements over
the last century, the dominating interpretation over the reason
behind the construction of them is based on their natural linkage
to war, and therefore its defensive function as retreats in the
periphery (e.g. Schnittger 1908:30f; Hermelin 1929:90). A
connected and reoccurring thought stresses the relation to the
great political turmoil on the continent (Damell & Lorin
2010:206) and the sometimes adjacent place names connected
to the hillfort settlements with the prefix Karl- or Rink- have
been seen as indications of a garrison related to a nearby hillfort
(Hellberg 1975). David Damell and Olle Lorin (2010:218) also
note that some hilltop settlements probably should be
interpreted as if the rampart did have a protective function,
while others do not seem to be defense facilities. A similar
thought is presented by Michael Olausson who claims that
11
Figure 4. The hillfort settlement of
Darsgärde (Olausson 1996).
the fortification was not always its primary function (Olausson 2009:48). Connected to this
discussion is the question whether tree constructions have strengthened and elevated the wall.
This has in fact been proven at several hilltops (Ambrosiani 1958, Engström 1984, Hemmendorff
1992).
Åsa Wall is among the archaeologists who have a conflicting view of the hilltop settlements.
Wall sees them as traces after cultic exercise; the rampart or stone-wall surrounding the hillforts,
including hilltop settlements, should be seen as a border-zone, separating the landscape of the
living from the inaccessible world (Wall 2003). This fits well with the notion that the general
absence of adjacent cemeteries have been seen as a sign of semi-permanent use (Törnqvist
1993:12), i.e. the population using the hilltop settlement are possibly residents from farmsteads or
magnate farms in the hinterland. Furthermore, the general localization to the periphery between
more populated areas has been proposed to be an indication of occasional use. In line with this,
Wall stresses that the occupational traces and the houses are remains after simulations of
activities associated with settlements. She describes these activities, i.e. food processing and
crafting as transformations linked to the cosmology with the farm having an increasing
significance within the cult, why the houses are not traces of proper settlements (Wall
2003:140,141ff, 183).
Michael Olausson on the other hand saw a connection between the hilltop settlements and
centralization of the production and power (Olausson 1987b). He later describes the hilltop
settlements in more detail as possibly a form of magnate farms harboring different
families/groups of people (Olausson 1997:110f). In line with this reasoning terraces have been
observed outside the rampart at a few hilltop settlements (Carlström 2003:8, Damell & Lorin
2010:212), which have been interpreted as a sign of local hierarchies (Skyllberg 1991:12ff, cf.
Olausson 2008:27). Hierarchies on a micro-level are however also found within the ringforts of
Öland (Nordström & Herschend 2003). In Olaussons (2009:38) recent discussions concerning the
origin of the hilltop settlements, he interprets them as a new way of life, inspired by the Roman
and provincial Roman models, where home, specialization of crafts, trade and military protection
is combined (see also Olausson 2008). For instance the relation between bread and magnate
farms has been pointed out as an indication of Roman influences (Bergström 2007). The presence
of hall-buildings within hilltop settlements have also been stressed based on the topographical
situation of certain terraces and the findings in others (Olausson 1987b:405, Olausson 2008:32).
The wealth is also apparent in the studied bone-material from Gåseborg; the recovered bonematerial displays ideal slaughter age and fragments of bones from the edible parts of the animals
dominate the osteological material. These observations indicate slaughter outside the hillfort and
possibility of choosing from different food sources (Olausson 2009:52). The picture that
Olausson sketches in his attempt to interpret the people behind the walls of larger settlements,
describes a family in leadership with an accompanying retinue (sw. här) and a luxury
consumption of meat and crops. Craftsmen must also have been among the people who were
included in this form of household. Furthermore, he reflects over the possibility of these men and
other workers being thralls (Olausson 2008:32). Hence, the hilltop settlements probably overruled
the hinterland to different degrees (Olausson 2009:38, 55).
David Damell notifies a certain type of hilltop settlement with very limited building traces. He
links hilltop settlements as Sunnersta and the earlier mentioned Broborg to similar functions; i.e.
guarding the trade routes on land- and waterways leading to Old Uppsala. The buildings are by
this context interpreted as parts of a garrison and customs location, composing an outer defense
system in the early state formation (Damell 1993).
Obviously, the hilltop settlements are not in themselves a homogenous group. The different levels
of building density and settlement structures upon hilltops have been interpreted as indications of
a hierarchical differences and variations in the status and composition of its inhabitants (Olausson
2008:29, 2009:54f). Magnate farms have different profiles and as Olausson argues (1987b), the
12
concept of specialized crafting includes different groupings. Darsgärde with unusual finds such as
a scythe and a ploughshare made of iron as well as grindstones and pottery in quantities along
with several buildings possibly for storage at its disposal might have been a magnate farm
specialized in agriculture. Sites as Gullborg and Onssten with its characteristic loom weights
might on the other hand be related to extensive textile crafting (Olausson 1987b:405). The
discussion must also take the interpretations mentioned earlier into account as a holistic view
stresses the need for somewhat sliding interpretations; from surveillance forts with few findings
to forts as Sunnersta and Fållnäs with one or two buildings, to settlements more or less
characterized by their aristocratic appearance, specialized crafting and trading functions.
3.4.4 The courtyard - activity areas
As a consequence of limited excavation-areas a more detailed picture of the spatial organization
is difficult to depict. This problem is even more obvious when noting that almost the entire inner
area seem to have been used at several hilltop settlements (Olausson 2009:49). As was briefly
mentioned above, we might expect a hierarchical division within the larger hilltop settlements.
Eketorp on Öland with several more or less ordinary farms together with a prominent building
group including a hall might display a reoccurring pattern when discussing spatial organization
within hilltop settlements. Darsgärde with approx. 20 houses serves as a point of departure.
Analyzing the courtyard with the different buildings, a variation can be detected with one
distinctively larger building in NE (fig. 4), containing two hearths in the western half and most
findings, among shorter longhouses and small houses often with barely any findings and without
hearths. A hypothesis has been proposed by Olausson (1987b:405) where he interprets the large
house on the constructed terrace as a hall-building. In this house and in the surrounding smaller
houses he argues for the possibility of storage-functions as their inner area measure about 1150
m2 in total, which exceeds the area at ordinary farms by far. Certainly an organization of this type
must have included a hierarchical society with inhabitants in power alongside labor.
Recalling all the functions that has been ascribed to hilltop settlements and all the traces of
different activities that were reviewed above, it is somewhat disturbing that so little is known
about the internal organization. Among the repeatedly documented activities are iron smithing,
textile working and bronze casting, while gold crafting and bone working seems to be more
unique (Kangur 2004, Olausson 2009:50). The extent of crafts seem to exceed the individual need
to different degrees (Olausson 1987b:409, 2009:52) and it has occasionally been concentrated to
a limited area (Schnittger 1913 ATA). In Eketorp on the other hand it is obvious that the
production seems to be distributed between households (Rydberg 1995:25f).
3.5 Formulation of the research gap
The question about the meaning behind the hilltop settlements as debated above is according to
my conclusions too wide and general in order to be answered thoroughly by a case study. Instead
the expectation on the following chapters is to narrow down the question at hand and primarily
add knowledge about one specific site, Runsa, in order to acquire a deeper understanding
regarding one hilltop settlement. The fact is that hilltop settlements are often discussed in their
relation to the surroundings and sometimes with the addition of finds without an accurate relation
to their context, but seldom (ringforts excepted) a deeper knowledge is obtained about their inner
organization and the functions that lay therein. With all knowledge that has been gained
concerning the Late Roman Iron Age and the Migration period settlements (e.g. Göthberg (eds.)
2007; Olausson (eds.) 2008), it is exceptional how little is known about these monumental
constructions and their relation to the mid-Iron Age society. In the light of this, it is interesting to
investigate if several/certain crafts or ordinary agrarian functions can be traced at the monumental
settlement in Runsa. Were there separate areas of activity and dwelling areas, or were there a
spatial grouping based on households within the wall? How does the settlement differ from our
knowledge of ordinary and magnate farms? What kind of houses can be traced and how were
13
they used? An eventual specialized production requires dwelling houses, or did the people live
elsewhere? The questions are many and probably they cannot be answered completely in this
study. However, the following sections approaches these questions and hopefully a more nuanced
picture about what these settlements consisted of can be established.
4. ANALYTICAL TECHNIQUES
4.1 Geochemistry and vessel use
The basic thought behind using geochemistry in archaeology is that activities above ground will
affect the composition of the soil. Deposition of solid particles and their by-products resulting
from decomposition, alternatively by complexation and absorption of free ions or molecules
originating from liquid residues, become fixed to the parent material in the soil (Middleton
2004:49). These changes in soil chemistry may also be better preserved than the actual artifacts
and constructions that once were part of the prehistoric society. Unless the sudden event of a fire
or an assault, settlements was seldom abandoned immediately, allowing inhabitants to collect
their possessions. Most likely the domestic floors was also swept and kept relatively clean, thus
obstructing the interpretation by the archaeologists. Chemical residues, however, have the ability
to accumulate in their primary context (Parnell et al. 2002:379, Hutson & Terry 2006:394). Areas
that are considered as more or less empty can therefore be given more interpretable variables. The
soil characteristics that are discussed here are divided into two branches; inorganic and organic.
The former is represented by metal elements and the latter by lipids in this thesis.
Likewise, as regarding the chemical imprints in soil, the matrix of prehistoric unglazed pottery
reflects the contents, e.g. food ingredients or storage material, from mainly the last uses of the
vessel (Craig et al. 2004). Nevertheless, studies of vessel use are thought to reflect food habits
related to diet and functions such as storage, serving and processing of food. The concept has
been used successfully to distinguish vessel use both between sites (Isaksson 2000b, Hjulström et
al. 2008), and between houses/households or activity areas within sites (Isaksson et al. 2005,
Olsson & Isaksson 2008, Dimc 2011). The different methods are briefly presented below,
including a historical review and methodological concerns, before the next section discusses the
on-site sampling strategies.
4.1.1 Phosphates
Phosphate mapping is certainly a part of the inorganic geochemistry together with the other metal
elements, but it has a rather long and individual history in connection to archaeology (see Bethell
& Máté 1989), why it deserves an introduction of its own. The theory behind measuring the
amount of phosphates in soil is based on the fact that all organic material contains phosphorus.
Thus, when organic material is deposited in the soil it is decomposed into phosphate ions. These,
in turn, are distributed into the soil solution or become fixed to the surfaces of minerals.
Phosphate (P) therefore appears in three different fractions, organic, stable inorganic (parent
material) and plant available phosphate, which are held in equilibrium; plants receives their
nutrition P from the available phosphate fractions and are later decomposed, transferring P back
to the soils where they originated (Jahnke 1992:308). However, this can be altered by a
displacement, interrupting this equilibrium, and input from a source containing phosphorus;
therefore it is argued that accumulations occur where the organic input is greater. Hence, because
of its ability to be stored in the soil, phosphates can be used to trace anthropogenic activities
(Brady & Weil 2002:602f). There are different techniques for determining the phosphate
concentration within soils; analyzing the amount of labile inorganic phosphate, i.e. plant
available, is the common implementation, although it has been criticized for not taking the total
phosphate level, i.e. the organic and the stable inorganic into consideration (Johnson 1956). Since
the proportion of the different fractions vary depending on soil properties, analyses of total
phosphate have been argued to be required when inter-site comparability is demanded (Bethell &
Máté 1989:19f). The aim of the mapping therefore decides what method is suitable.
14
Reviewing the applications of phosphate mapping in archaeology, the initial and very important
conclusion was drawn by Olof Arrhenius, where he pointed out the relation between enhanced
phosphate values and prehistoric settlements (Arrhenius 1929). During the following decades
several archaeologists noticed the same relation (e.g. Lutz 1951, Dietz 1957) and also tried to
establish a more accurate mapping strategy in order to distinguish certain constructions (e.g.
Barker et al. 1975). During the second half of the 20th century, phosphate mapping was
frequently used as a standard survey method, both by archaeologists and also historical
geographers. Their aims were often to locate settlements (e.g. Sporrong 1971, Widgren 1983,
Broberg 1990), or harbors, i.e. human activity adjacent to shorelines (Carlsson 2004:19).
Furthermore, phosphate mapping has also been used as a post-excavation analysis, for example to
specifically pinpoint the functions of certain buildings (Ramqvist 1983, Eriksson 1995) or to
determine the position of buried bodies (Barker et al. 1975).
The manifold of activities that results in deposition of phosphorus is though somewhat
problematic when discussing functions and intra-site variations; deposition of organic material
within settlement archaeology, can for example be associated with food handling, i.e. processing,
consumption and disposal (e.g. Proudfoot 1976, Terry et al. 2004), burning of organic material
(Middleton 2004:53) and certain areas of crafting such as processing of wood and bone or
lapidary crafts (Eidt & Wood 1974:44, Middleton 2004). Wilson et al. (2008) summarizes this
complexity with showing how P tends to concentrate to several activity areas, but with the byre
often coinciding with its maximum and others at descending concentrations. On the other hand,
low and relatively depleted values also are interpretable, for example as pathways, sleeping areas
(Terry et al. 2004:1243) or storage functions (Sanchez et al. 1999:56).
Phosphate-analysis
The laboratory method used for this thesis is called the PMB-method (see Persson 2005). Approx.
1 g of each soil sample was dried and homogenized with a grinder. The phosphate ions were then
extracted into solution by adding 5 ml of 2 % citric acid and put on a shaker table over night.
After sedimentation, 2 ml of molybdenum-sulfuric acid and 0,5 ml of sodium sulphite
hydroquinone solution was added. The solution was then diluted with distillated water to the 50
ml mark and put in an oven in 50°C for six hours. The dilute sulfuric acid functions as a reaction
medium; the molybdenum reacts with the phosphate ions and forms phosphomolybdate
complexes. These are then reduced with the hydroquinone and receive a blue color. The degree of
blue color is finally determined using a spectrophotometer. The more phosphate ions within the
sample, the bluer the solution become. As a standard, potassium dihydrogen phosphate (KH2PO4)
was used. 1 ml of the standard solution is equal to 50 Phosphate° (P°). The absorption from the
standard solutions was then divided with the known P° for each of them. The mean value is
calculated and used as a quotient to obtain the P° values for the samples.
4.1.2 Metal elements
As well as for phosphorus, the composition and concentration of other metal elements relies both
on geological aspects and human agency. As a natural consequence to the successful analyses of
phosphates, other elements were soon experimented on as well. In the early 50´s Lutz (1951)
showed how Ca, Zn and Cu values, similar to P, were enhanced in anthropogenic soils. Other
elements were tested, Konrad et al. (1983) showed how Mg could be useful and Bintliff et al.
(1990) added Mn to the metal elements varying with human occupancy. It was also confirmed
that accumulations of Ca, Cu, Mn and Zn also could be detected in settlements in Swedish soils
(Arrhenius et al. 1981, Linderholm & Lundberg 1994). This is not a supreme picture however,
other studies have demonstrated that enrichment of these metal elements on settlement sites is not
obvious (Entwhistle et al. 1998), indicating that deposition varies between sites, alternatively that
soil properties contributes to mobility and loss by leaching (e.g. Pickering 1986). Archaeologists
soon also implemented more detailed studies where correlations and variations between elements
were pointed out within the settlement limitations and described as different activity areas
15
(Konrad et al. 1983:22). During the last decades of the last century, the method became more
frequently used, on a more general level (Aston et al. 1998, Entwistle et al. 1998, James 1999) as
well as on a household-level (Middleton & Price 1996, Wells et al. 2000). The next step was
consequently to shed light on the relationship between specific space use and chemical imprints,
enhancements or depletions of metal elements. Ethnoarchaeological studies were one answer.
These have been accomplished successfully and the results of metal element analysis have led to
observations regarding activity areas and chemical imprints, which have aided in the
interpretation of the prehistoric sites (Middleton & Price 1996, Fernandez et al. 2002, Terry et al.
2004). Similar suites of metal elements, i.e. defined patterns from ethnoarchaeological sites, have
also been observed at prehistoric sites (Middleton 2004:55). One of the remaining problems
however, is the difficulties of interpreting activities that do not exist among modern people and
thus cannot be predicted (Hutson & Terry 2006:394). Following the ethnoarchaeological studies,
metal element multi-analysis gained interest and acceptance and the methodology of sampling
archaeological indoor spaces has recently been emphasized especially on Mesoamerican sites
(e.g. Cook et al. 2006, Middleton 2004, Hutson & Terry 2006, Hutson et al. 2007). The validity
of the method has also been demonstrated in works where discriminant analyses are performed on
samples that are grouped by visible room divisions (Hutson & Terry 2006, Hjulström & Isaksson.
2009). Nevertheless, the methodology is still only rarely used in Swedish settlement archaeology,
works of Isaksson et al. (2000) and Hjulström et al. (2008) still are the only examples covering
intra-site variations.
Metal elements used for interpretation
The metal elements that are analyzed in the current work consist of (K) potassium, (Mn)
manganese, (Ca) calcium, (Fe) iron, (Cu) copper, (Zn) zinc and (Mg) magnesium. These are
equivalent to the elements used by Hjulström (2008), and as can be concluded from earlier works,
these display concentrations both to and within settlements (Konrad et al. 1983, Linderholm &
Lundberg 1994, Middleton & Price 1996, Aston et al. 1998, Wells et al. 2000,).
Reviewing earlier works, the interpretations of relative concentrations of the current metal
elements are obviously somewhat varying and confusing, underlining the inherent challenges
with interpreting distribution of metal elements: Enhanced Ca values has, due to its inclusion in
hydroxy apatite, been interpreted as areas where activities such as food preparation (Middleton
2004:56) bone butchering or deposition of bones where committed (Konrad et al. 1983:26), while
it on the other hand has been stressed that Ca also is a major component in wood ash (Isaksson et
al. 2000, Middleton 2004:56). K seems to be a relatively strong indication of heating activities
connected to hearths as well (Isaksson et al. 2000, Middleton 2004:56). Mg belongs to the same
category which has been pointed out as an indication of intensive burning, and therefore possibly
signaling hearths (Heidenreich et al. 1971, Middleton & Price 1996). When no hearths have been
located through excavation it has been mentioned in association with ash dumping and reduction
of lithics (Konrad et al. 1983:26). Isaksson et al. (2000) also points out potentially elevated Mg
concentrations due to manuring/stabling since it is a moderate component in plants.
Manganese has been interpreted as an indicator of organic refuses (Bintliff et al. 1990). Parnell et
al. (2002:392) have suggested the same source, originating from food processing/consuming, for
Mn and also Cu, since these appear to correlate to P values. Similar thoughts are presented by
Hjulström (2008:18), who links manganese to cereals. The opposite interpretation is put forward
by Hutson & Terry (2006) when investigating plaster floors in Chunchucmil; in what is
interpreted as primary context, Cu, Mn, and Fe are found enhanced together with the lowest
phosphate concentration. This fits the picture drawn by Isaksson et al. (2000), where the
maximum Cu, Zn and Fe values are connected to metal crafting. However, Isaksson et al. (2000),
also mention Fe together with Zn as significantly rich in meat. In a later work with the same
authors, food is interpreted as the source, when elevated Mn, Fe and Zn values are detected
(Hjulström et al. 2008). Elevated Fe values have elsewhere also been linked to crafting (Terry et
16
al. 2004:1244), workshops with worked bone and worked stones (Parnell et al. 2002:391), as well
as to butchering and food processing (Manzanilla 1996, Parnell et al. 2002:391). Wilson et al.
(2008) contributed to this discussion by analyzing post-medieval farmsteads on different soils,
where the activity areas (e.g. byre, hearth, dwelling and midden) were known. Ca tended to have
its maximum round hearths followed by the general dwelling-area. Zn showed similar tendencies.
Generally elements were found to be elevated at a descending scale, ranging from hearth,
dwelling, byre, and others (e.g. arable fields and gardens), with the exception of P (see ch. 4.1.1).
The passage above addresses the issue that when relative concentrations are measured,
consequently the risk of masking occurs. An activity that deposits a high amount of a certain
metal element may imply other elevations to appear as more moderate in the same population,
e.g. metal crafting may mask the deposits of organic material, which otherwise, in a population
without samples from a metal crafting area, may have been distinguished based on the maximum
values. This is one of the explanations to the discontinuity in the interpretations, even though the
enhanced metal element is the same. Every site is unique, and this is why relative values cannot
be transferred in a general sense from one site to another, at least not without examining the
activity areas and soil environments more closely (Wells et al. 2000). This is also underlined as
the correlations between different metal elements are varying when comparing several studies
where multi-element analyses are performed (see Isaksson et al. 2000, Parnell et al. 2002, Terry
et al. 2004, Hjulström & Isaksson 2009). Different material obviously contains varying amounts
of similar suites of metal elements (e.g. Isaksson et al. 2000:9ff), consequently making the
potential to accurately pinpoint deposition of specific material more complex. The conclusion
from this brief review must be that concentrations of a single metal element may originate from
several potential sources (see Middleton 2004:54). Therefore they should preferably not be used
solely, in regards to the interpretation of functions. Correlations between several metal elements
must be considered and preferably mapping of metal elements should be used as a variable
amongst others. The latter is clear in the works of Isaksson et al. (2000), Terry et al. (2004),
Hutson & Terry (2006) and Hjulström et al. (2008), where the interpretation of activity areas is
dependent on the varying values of metal elements but is still guided and aided by other sources;
artifacts, organic residues, constructions etc.
Extraction & Atomic Absorption Spectrophotometry
The laboratory method used for this paper can be described as extraction using a strong acid
followed by a measurement of the concentration through Atomic Absorption Spectrophotometry
(AAS). Initially each soil sample is grinded with a mortar and pestle and quantitatively weighed
in at approx. 1 gram. The following step, the extraction of the metal elements, has been, as all
analysis of soil chemistry, lively discussed. The question is, as was mentioned above, whether a
weak or strong acid is the best alternative. However it can be concluded that different techniques,
using a weak acid (e.g. Middleton & Price 1996), and using a strong acid (e.g. Entwistle et al.
1998) have yielded archaeologically interpretable and pleasing results. Furthermore, when
applied on the same material, similar results are achieved, although some areas appear to be
detected only by one or the other technique (Parnell et al. 2002). Differences in the parent
material, i.e. the natural background, may however be conclusive for which method is most
suitable (see Parnell et al. 2002:382). Moreover, extraction techniques have previously been
tested at the Archaeological Research Laboratory (ARL) with a reference soil. This was intended
for investigations on Swedish soils and implemented in order to establish the best settings for the
parameters; temperature, digestion acid and time (see Hjulström 2008:26f). It is the results from
this evaluation which are used in this thesis. Hence, the samples were put in a teflon container
and digested with 10 ml of Aqua Regis (nitric acid: hydrochloric acid 1:3 v.v) in a MARSX
microwave oven. Since it has been showed that the exchange level for certain elements were
reduced when the temperature of the oven were set too high (Hjulström 2008:27), the maximum
temperature were set to 175°C. This was achieved after automatically increasing the temperature
17
during 20 minutes. After filtration, a stock solution, diluted with deionized water to 25 ml, was
created for each sample. Since the concentration of metal elements by experience differs a lot, the
samples were once more diluted, depending on what metal element studied, in order to keep the
measurable concentration within the detection interval.
The measuring technique executed by the AAS, in this case a Z-5000 Polarized Zeeman Atomic
Absorption Spectrophotometer, then analyzes each sample with assistance from an auto sampler.
The liquid solution is injected into a spray chamber, where the flame, fueled by acetylene,
vaporizes the particles into gaseous molecules. These are further dissociated into free atoms.
Simultaneously, a radiation beam is emitted by a cathode lamp, containing the element that is
measured. Within the chamber, the electrons of the specific metal absorb energy, depending on
the wavelength that is unique for each element, and becomes excited. This affects the output of
energy, measured by the detector, and the absorbance, i.e. the concentration can be calculated
since the original input of energy is known.
4.1.3 Lipids
Lipid is a generic name of a subgroup within the organic compounds, which in turn cover a
manifold of compounds. It is often entitled as a synonym to fats, while a more proper but still
simplified description views them as “…fatty acids and their derivatives, and substances related
biosynthetically or functionally to these compounds…” (Christie 1987:42). Furthermore, this
classification can be separated into two divisions; neutral and polar lipids, where the former
composes the main material for this thesis, including: fatty acids, n-alkanols, triacylglycerols,
sterols, and long-chain ketones. These compounds cover several functions among living
organisms; energy storage, insulating material, structural components, signaling molecules,
protection etc. (Brown & Brown 2011:54), while their joint factors are their origin from living
organisms and general insolvability in water. These two factors hold great potential for using
lipids as an archaeological tool, since amorphous and invisible organic residues, otherwise lost
for the archaeologists, can be identified (Evershed 2008a).
The interpretative step when analyzing organic residues involve the concept of biomarkers. As
Evershed (1993) stated, the concept of biomarkers is basically about fitting observed chemical
imprints into known constituents of organisms likely to have been present during the period in
question. In certain cases an observation of a single compound is enough, in other cases the
mixture of components is characteristic for a specific origin, such as beeswax (e.g. Heron et al.
1994) and ruminant fat (e.g. Dudd et al. 1999). Moreover a measurement of a ratio of different
components in order to distinguish between slightly different origins can be used (e.g. Evershed
& Bethell 1996). Furthermore, the anthropogenic transformation of lipids as caused by e.g.
heating (see Hayek et al. 1990, Evershed et al. 1995, Hjulström et al. 2006, Evershed
2008a:901ff) needs to be taken into account when interpreting activities once consciously
performed by the prehistoric man. Another important alternation of the lipids occurs once
deposited in the soil when the decomposition of these compounds sets in. Diagenesis certainly
causes a loss of information, especially in the early stages after deposition, e.g. β-oxidation of
fatty acids (Isaksson 2000a:34), however mapping of degradation products in order to trace the
origin further backwards might be a possibility (Isaksson 2000a:34f, Hjulström 2008b:25).
Altogether, analyzing the lipid residues, i.e. tracing the origin of certain compounds as described
above, is often a key to obtain knowledge of for example space- and vessel use. However, it is of
importance to stress that although biomarkers are identified, the results are still archaeological
interpretations (Brorsson et al. 2007:422).
The background of lipid analyses originate from successful studies of visible organic remnants
(Brorsson et al. 2007:421) in combination with the introduction of chromatographic methods
during the mid-century (Evershed 2008a:896). As was mentioned above concerning the metal
elements, lipids, due to their physical characteristics, can be stored in different materials. This
may occur in the soil matrix after diagenesis of the organic material initially deposited, or in the
18
ceramics pores when fluids enter the matrix. This opened up for research on invisible organic
residues by using a solvent to extract potentially bound lipids. The early investigations were very
experimental but yet successful; Condamin et al. (1975) proved the presence of olive oil in
ancient amphorae, Lin et al. (1978) showed survival tendencies of steroids in coprolites, Knights
et al. (1983) discussed the possibility of using sterols to evaluate diets of past societies. Others
had a more critical approach where the decomposition of lipids in for example soils (Moukawi et
al. 1981, Bridson 1985, Bull et al. 2000), bogs (Evershed & Connolly 1994), and pottery (Heron
et al. 1991, Malainey et al. 1999), was emphasized. However, the field has grown stronger and a
multitude of archaeologists studying different fields, agriculture (e.g. Bethell et al. 1994,
Evershed et al. 1997), diet (e.g. Morgan et al. 1984, Dudd & Evershed 1998, Isaksson 2000b)
wood tar (e.g. Hayek et al. 1990, Aveling & Heron 1998), found the use of lipid residue analysis.
These are all examples of key works which paved the way for the research on lipids in different
archaeological materials and consequently also the background to the works recently
implemented on Swedish grounds, which forms the methodological backbone for this thesis
(Isaksson 2000a, 2003, Isaksson et al. 2005, Hjulström 2008).
There are two types of source material where lipid analyses have been applied for this paper; soil
and pottery. Concerning the lipids in soils, whereof several occur naturally; these are in most
cases measured relatively in order to distinguish areas with enhanced or depleted values, which in
turn can be argued to depend on human activities if sampled from a cultural layer. Other lipids
are as stated above, often direct evidences of certain activities. Results from earlier similar studies
have proven the usefulness when discussing different settlement issues; Isaksson (1998) and
Hjulström & Isaksson (2009) showed that lipids in the soil can be used to identify activity areas,
and Isaksson (2003), Isaksson et al. (2005) and Hjulström et al. (2008) indicated the potential to
distinguish different food habits on a micro level alternatively house functions by comparing
lipids on potsherds statistically.
Lipids used for interpretation
Fatty acids are carboxylic acids with an adjoining carbon chain. When the compounds are not
attached to other molecules, they are referred to as free fatty acids. The length of the aliphatic tail,
the carbon chain, is described by the number following the chemical symbol C. These are the
most common lipid compounds and occur as saturated or unsaturated, referring to an eventual
double bond between the carbon atoms in the carbon chain. The number following the colon
describes the number of double bonds. (Brown & Brown 2011:55). When these are found in soil
samples or in the ceramic matrix, they are originating from different sources; degraded
triacylglycerides originating from adipose tissue or vegetable oil, biological waxes both from
plants and animals, membranes etc. Even though degraded, there are ways to interpret the prehistorical origin, whereof the two calculations used for this thesis are presented shortly below.
Palmitic acid (C16:0) and stearic acid (C18:0) usually make up the largest part of fatty acids
among the lipid residues in pottery. The ratio of C18:0/C16:0 has been shown to be an indicator
of input from adipose tissue from terrestrial animals in relation to plants or fish (Isaksson 2000b).
Since stearic acid is abundant in animal adipose tissue, a calculated ratio of 0,5 or more is
suggesting an input from animal origin (Hjulström et al. 2008). This tool was originally used on
ceramics but has repeatedly shown results also for soil samples (Rogge et al. 2006:41, Hjulström
et al. 2008, Isaksson 2009a).
The ratio of C17:0branched/C18:0straight has been shown to be indicative of the input from ruminant
animals and milk lipid residues in pottery (Dudd et al. 1999:1480). This ratio has been shown to
correlate well to the more established method using stable isotopes (δ13CC16:0 – δ 13CC18:0). A ratio
of ≥ 0.02 has been suggested to indicate an input from ruminant animals or milk while ≤ 0.0077 is
thought to show no traces of such input (Hjulström et al. 2008:11).
19
The overall picture is also defined by the distribution of fatty acids as calculated by ACL, CDI
and CPI. An input of mainly short carbon chains, i.e. an input of mainly animal origin, results in
low average chain length (ACL) and vice versa. The CDI value is a measurement of the diversity,
meaning that there might be fatty acids of different lengths dominating the sample. A high score
is indicating a large diversity. Lastly, CPI is describing the relation between odd versus even
numbered carbon chains; a measurement of the age, i.e. the degree of diagenesis of the fatty
acids. A number closer to 1 indicates that the diagenesis has proceeded farther.
n-Alkanols differs from fatty acids by the exchange of a carboxyl acid at the benefit of a hydroxyl
group. Likewise they occur in different lengths, where long-chained alkanols are characteristic of
decomposed plant material (Hjulström 2008:21). Similarly as for the fatty acids the amount of nalkanols are measured for ACL, CDI and CPI.
Triacylglycerols (TAG) are formed by ester bonds between glycerol and three fatty acids and are
a major component in vegetable oils and especially in adipose tissues of animals (Campbell &
Farell 2006). These are however more seldom found intact; instead the TAG’s are decomposed
into free fatty acids. However, when they occur intact, they could be used as another indication of
that the lipid composition derive from ruminant animals/milk fats (Dudd et al. 1999). This is
based on the fact that short carbon chains are common in fat from ruminants and in milk products
rendering in a broader range of TAG´s. However, since short chained TAG’s are decomposed
faster than long chained TAG’s, a lack of these compounds does not prove the absence of short
chained TAG’s originally (Brorsson et al. 2007:423).
Sterols are parts of the steroid group. Because of the differences in biosynthesis between
organisms, they are found in animals as cholesterol, in plants as e.g. campesterol, stigmasterol or
sitosterol and in fungi as ergosterol (Christie 1989:22). This distinction open up for the possibility
to discuss sterols as biomarkers. As a measurement of the sterol input respectively, hence
discussing handling of meat and dairy products contra plants, two different sterol-ratio
calculations have been used for soil samples. Isaksson (1998) successfully analyzed the
derivatives 5α-cholestan-3β-ol and 5α-campestanol while Hjulström & Isaksson (2009) and
Hjulström et al. (2009) used the compounds as they occur before reduction and also included the
other two main phytosterols. The former was thought to better measure the ancient input
(Isaksson 1998:46), although the latter also arguably is depending on human activities (Hjulström
& Isaksson 2009:7) and has shown promising results (Hjulström et al. 2008:19). When analyzing
potsherds, just identification of one or the other is taken as a signal of animal products or
vegetables.
Another biomarker among the sterols is coprostanol or 5β-cholestan-3β-ol. The difference from
5α-cholestan-3β-ol, the derivative from cholesterol reduction in soil, is that this stanol is
produced by biohydrogenation in the gut among animals with a high trophic level, and hence
used to locate faecal material (Bethell et al. 1994). In the guts of ruminant animals, having a
herbivorous diet, the reduction products are mainly 5β-campestanol and 5β-stigmastanol. A few
projects have dealt with this particular biomarker, for example a Roman ditch was interpreted as a
latrine drain due to detection of coprostanol (Knights et al. 1983). After analyzing feaces from
cows and sheeps, the ratio of coprostanol/5β-stigmastanol was suggested to detect an input of
faecal material from ruminants (Evershed & Bethell 1996). The ratio in ruminant faeces is
estimated to 0,25 (Bull et al. 2002). 5β-stigmastanol is also known as 24-ethylcoprostanol and
was successfully detected in the stable of a reconstructed Iron Age-house (Hjulström & Isaksson
2009).
Ergosterol (5, 5, 22-ergostatrien-3β-ol), originating from fungi, could possibly be inherent in
pottery shards due to mould growth. However it has been suggested as a biomarker for yeast and
alcohol fermentation (Isaksson et al. 2010). It was successfully identified in studies of Bronze
Age/Early Iron Age pottery, while it was absent in Neolithic pottery (Isaksson et al. 2010).
20
Terpenoids are a modification of the carbon structure among terpenes, caused by oxidation for
instance. These cover a vast group of organic compounds all with the basic isoprene unit (C5H8)n
as a backbone (Brown & Brown 2011:60), however the substances that are targeted here includes
triterpenoids (six units) and diterpenoids (four units) and their function as a component in resin.
The former has been successfully detected in floor layers as in the case of betulin (Isaksson 2007,
2009b) a compound indicative of birch bark and together with lupa-2,22(29)-dien-28-ol also
indicative of smoke. These two compounds are also major components in birch tar. The ratio
lupa-2,22(29)-dien-28-ol / betulin has in turn been experimentally proven as a useful tool when
trying to distinguish birch tar from birch bark, where a ratio > 0.2 often indicates tar production
(Aveling 1998:91ff). Diterpenoids are also components in resins, although mainly from conifers.
It may be found as abietic acid or dehydroabietic acid in soils where conifers are growing,
alternatively indicating smoke (Isaksson 2009b:5). The by-product methyl dehydroabietate which
is produced when the resinoic acids reacts with methanol, however functions togheter with the
diterpenoids mentioned above as a biomarker for distillation of resinous conifer wood, i.e. tarproduction (Hjulström et al. 2006).
Ketones are compounds characterized by their carbonyl group, i.e. a carbon atom with a double
bond to an oxygen atom (C=O), and by two more carbon atoms bonded to the carbonyl group.
Ketones may be formed by oxidation of free fatty acids, hence detection of long chained ketones
with an uneven number of carbon atoms (C29-C35) has been argued to demonstrate that a vessel
has been heated, i.e. through cooking (Evershed et al. 1995). Recent experiments however, have
questioned this hypothesis indicating that a temperature over 350°C is ideal when forming longchained ketones, why vessels that displays a lipid distribution including these compounds have
been suggested to demonstrate cooking failure (Evershed 2008b). However, there are other ways
of preparing food in a vessel than boiling which might cause very high temperatures, why
examining this factor still is fruitful (Isaksson pers. comm. 25/5 2012). Another potential
indication of heated lipids are the ω -(o-alkylphenyl)alkanoic acids (Matikainen et al. 2003),
which are produced when polyunsaturated fatty acids with three double bonds are heated.
Extraction, derivatization & Gas Chromatography - Mass Spectrometry
The separation and characterization of lipids is a method conducted in several steps. Before
running the samples through the GC-MS, the samples need to be extracted and derivatized. The
soil samples were therefore dried in an oven in 45°C over night and then grinded and sieved, and
a small sample of each potsherd was pulverized using a tile grinder. Approx. 2 g and 0.5 g of the
soil and ceramic samples were weighed in respectively.
The extraction of the lipid residues were carried out using chloroform and methanol (2:1) by the
aid of ultrasonication. The analyte were then derivatized through trimethylsilylation by adding
BSTFA, bis(trimethylsilyl)trifluoracetamide (90%), and chlortrimethylsilane (10%), as a reagent.
The analyte transferred into separate vials were then injected into the gas-chromatograph via an
autosampler with a syringe. The liquid sample is volatilized due to an increasing temperature. An
inert gas, Helium (He) was then used as a carrier gas, transferring the analyte through a column in
a mobile phase. The motion was hindered as the analyte molecules sticks chemically to the
coating of the column walls, i.e. the stationary phase. Depending on the structure and volatility of
the molecule, the constituents forming the analyte were progressing forward towards the end of
the column as the temperature was rising. The time to complete the process is known as retention
time and since the components possesses different properties deciding the absorption strength to
the stationary phase, these are separated by retention time (Kitson et al. 1996, Brown & Brown
2011:63f).
Once separated the components requires identification, which is done in a mass spectrometer.
The stream of separated compounds is initially guided into an ion source. Electrons are here
emitted which in turn ionizes the compounds and transforms them into different ion fragments.
21
Thereafter the ions are fed onwards by a potential, entering a mass-selective filter, in this case a
quadropole, which selectively allows ions to continue to the detector. This is controlled by the
use of oscillating electrical fields, which allows a certain range of ions to pass at a time,
depending on an alternating radio frequency along with the variation of magnetic fields (Kitson et
al. 1996: Brown & Brown 2011:64f).
Each sample generates a total ion chromatogram, established in the detector of the mass
spectrometer, displaying abundance (Y-axis) and time (X-axis). The individual mass spectrums
along the time-axis are then searched for ion compositions, characteristic for the biomarkers in
question. Concerning the ACL, CPI and CDI calculations for n-Alkanols and fatty acids, an ion
chromatogram for the characteristic ions m/z 103 and m/z 117 was used respectively. The peaks
originating from the separated compounds are then integrated to obtain a percentage ratio. The
same was implemented for the sterols with the characteristic ion m/z 129 in order to exclude
other co-eluting compounds.
The equipment and settings used for the analysis in this thesis consists of a Hewlett Packard
model 6890 GC supplied with a SGE BPX5 (15m x 220 µm x 0,25 µm) fused silica capillary
column. The injection of the samples was performed by a pulsed splitless technique (pressure
17.6 Psi) at 325°C via a Merlin Microseal TM High Pressure Septum using an Agilent 7683B
Autoinjector. The oven was programmed with an initial isotherm at 50° C successively increased
with 10°/min until 350°C was reached, with a subsequent terminating isotherm of 15 min.
The adjoining mass spectrometer used is a HP 5973 Mass Selective Detector. The fragmentation
of the separated compounds was carried out through electric ionization (EI) at 70eV, and the
temperature at the ion source was 230°C. The mass filter had a temperature of 150°C and was set
to scan in the range of m/z 50-700 providing 2.29 scans/sec.
4.2 Sampling strategies
4.2.1 Sampling in trenches
The aim of the phosphate, lipid and metal element analysis is firstly to discern patterns of spaceuse, especially within House I and III. Secondly, these analyses are used as an interpretative tool
in order to identify the function of the observed areas. An initial geometric grid was applied to T I
and III, with the aspiration to collect samples from layers identified during the excavation. The
grid was constructed with sampling points every second meter in the longitudinal direction of the
terraces. One meter separated these sampling-columns with a shifting starting position of one
meters displacement, at every other column in order to cover the ground more thoroughly.
However, these grids were adjusted as structures were discovered, and sampling points were in
many cases relocated from border areas, e.g. from walls to inside/outside the activity areas.
Additional samples were also taken around features, i.e. the hearths in T I, T III and T X, and the
“oven” in T III, in order to establish a pattern on a smaller scale.
When executing the sampling of houses, the samples should preferably be collected from an
actual surface, i.e. the floor, instead of sub- or above floor fillings (Middleton 2004:50). This
advice was strictly followed wherever possible. No preserved floor layer was found in the
uppermost occupational layer in T III, the samples from this layer were therefore taken in level
with the topography of the terrace. However, both layer 2 and layer 3 had preserved floor
remains, most pronounced centrally around the hearth and with diminishing occurrence further
north. Layer 2 (34 samples) was only partially identified along the longer eastern and western
walls; where no floor layer was identified, the samples were collected in level with the overall
angle of the floor. The same principle was followed at the areas interpreted as outside the houses.
Layer 3 (18 samples) was identified throughout most of the excavated parts; however the
occupational layer outside what was interpreted as the northern wall is considerably thinner.
22
Trying to discern between levels without stratigraphic differences or references in the proximity
was simply not possible, why no samples specific for layer 3 were collected outside.
Concerning T I, no continuous floor surface was found here either. However, the bottom of the
layer containing most of the finds coincided with a stratigraphic transition point in the soil
texture. Darker soil with finer grains, typical for occupational layers, was replaced by smaller
stones and gravel. The samples were collected from this transition-level. Regarding the north
gable (see ch. 5), the samples were taken within the limit of the constructed gutter (see fig. 5). A
total of 97 samples were collected from T I.
In terms of T X, a more adaptive sampling was performed. The two uppermost non-dump layers,
labeled L2 A and B were sampled with the aim to cover most of the excavated area (11 + 8
samples). No actual floor layers were discovered, however the soil properties, i.e. color and
presence of charcoal and soot made the distinction easy. The lower stratification did turn out to
be very complicated though, why only a few samples per layer were collected (see appendix).
These were taken from the bottom of the different layers, visible in the profile.
As Hjulström (2008:34) concluses a larger set of samples, in this case all of the above metioned
soil samples, can be fruitfully analyzed by their concentration of different metal elements in order
to discern certain space use areas. Secondly, the more time-consuming and expensive but also
qualitative technique, lipid analysis, can be performed on a smaller set of samples from the
observed groupings to identify the areas. In this thesis, the second step of this process was
implemented by randomly selecting samples from the identified groupings based on the
concentrations of metal elements. Regarding T I the samples which were considered to be
responsible for the division in groupings were pinpointed and from this pool 25 samples were
randomly selected. Concerning T III all the potential samples in the random selection (6 samples
in each layer) were restricted to the area which was considered to be active (see ch. 5.3).
The pottery shards were sampled from all trenches, with the initial aim to make an intra-site
comparison of vessel use. Due to the complicated stratigraphy in T X, and lack of suitable shards
from House I, most of the analyzed shards, 12/15, are collected in House III. Rim shards
generally hold a higher amount of lipids (Charters et al. 1993) why these were prioritized (9/15).
Furthermore, an assessment to identify shards without damages on the interior and from unique
vessels was implemented. This was done in order to enhance the chance of preserved lipids as
well as to minimize the risk of analyzing shards from the same vessel several times.
4.3 Source criticism and problems connected to the sampling
Three problems needs to be stressed in this section; the natural spatial variability that is inherent
in the soil matrix, the risk of contamination during excavation and laboratory work and the
diachronic variation in space-use. Beginning with the natural conditions, these can be divided
into two problem areas. Firstly, the prerequisites in the soil vary, affecting the possibilities for
inorganic material to be incorporated in the soil together with the parent material; the inner
structure of primary and secondary minerals determines whether ions of different size fits
(Eriksson et al. 2005:54), likewise the grain size determines the binding ability, i.e. the amount of
inorganic and organic material that can be accumulated in the soil. Moreover, the moisture of soil
affects whether mono- or multivalent ions are primarily bonded to the mineral (Eriksson et al.
2005:128). Secondly, the processes following the initial deposition are affecting the composition
today. These include different factors linked to degradation, leaching and weathering processes
such as activity from microorganisms, temperature, pH, moisture, mobility of groundwater, relief
and red-ox potential (Eriksson et al. 2005:91ff, 197ff, 204f). Observations have in line with this
been made considering post-depositional altering of the imprints (Ottaway & Matthews 1988). It
is therefore of importance to evaluate these factors, especially when making comparisons
between areas. The quantities may be affected, leading to misinterpretations of what seems to be
a higher concentration, when it is in fact dependent on post-depositional processes (Hutson &
23
Terry 2006:394). Hereby it is vital to measure relative values within the site and not perform
inter-site analyses (Terry et al. 2004). Absolute values differ due to soil properties but relative
concentrations may better correlate between sites of similar use.
Contamination from modern sources is in most cases a manageable risk, in other words possible
to minimize. In order to meet this requirement, all analyzed samples were taken with clean tools
and put into plastic bags. Samples intended for lipid analyzes, including the pottery, were
wrapped in aluminum foil. Before laboratory analyzes, all samples were stored deep-frozen. To
avoid contamination from modern fingerprints (see Dimc 2011) minimum handling of shards
have been performed in field, while in laboratory environment gloves have been used at all times.
Superimposed imprints may occur in the event of changed space use. A floor level in a household
or activity area is often subject to changing activities and with that a gradual accumulation of
different residues. This problem is as Middleton (2004) stresses however general for settlement
archaeology. The contemporary spatiality is always an archaeological interpretation and the
composition of superimposed signatures is an important conclusion as any. However, the limited
time-span of dateable artifacts and 14C-datings at Runsa does decrease the necessary for caution
regarding the interpretations of contemporary activities. Likewise the historical and modern
activity seems negligible, causing no observed disturbances to the sampled layers. One recent pit
has been identified however, located in the eastern central part of House I. Overlapping
signatures may also derive from differences in the parent material, i.e. the constructed floor layer
might differ in its composition within or between houses (Middleton 2004:55). Therefore, it is of
importance to identify eventual variations in the surface.
5. RESULTS
5.1 Excavations 2011
House I
Parts of the terrace were as earlier mentioned unearthed during 1992 and 2010 (Olausson 1996,
Olausson 2011a). During 2011, almost the entire Terrace I was excavated, with the exception of
the northeastern part of the wall line (fig 5). The house located on the terrace has a convex shape;
widest in the center and narrowing towards the gables. The total length of the longhouse is
approx. 29-30 m, while an additional 3 m are found between the north gable of the house and the
stone constructed terrace. Of the registered postholes, 15 can be ascribed to the roof-supporting
construction of trestles, forming the nave. These are distributed over 8 trestles, whereof two,
number 3 and 8 counted from the south, are single posts. The reasons why no paired posts are
found at these spots are twofold; the west post of the northernmost pair must have stood on very
shallow soil, therefore leaving no traces. The west posthole in the third trestle is probably
superimposed by a hearth. A misplaced stone, larger than what a lone person can lift, intrinsic to
the hearth, suggests that the hearth is part an activity subsequent to the active phase of the house.
Perhaps this can be interpreted as a kind of activity connected to the abandonment of the house,
were the former hall-building was ritually locked. Hence the hearth in the southern part of the
house should be overlooked.
Furthermore, the small postholes between the 1st and 2nd trestle are probably not included in the
supporting structure; their size infers a weak supporting capability. Similarly the accumulation of
postholes at the center of the building most likely is the result of reconstructions, leading to the
interpretation that trestle 5 and 6 are replacing each other. Hence, the position of and distance
between trestles, as counted from the middle of each posthole, can be interpreted in two ways
(see tab.1). Additional posts, which are not part of the roof-supporting construction, are primarily
found within the central and southern parts. Parallel with the 3rd trestle and between the 3rd and
4th trestle, smaller postholes, within the aisles were discovered. These form pairs recessed to and
parallel to the wall line. Opposing breaks in the row of wall posts were documented in line with
24
these, supporting the interpretation of entrances with adjoining portal posts. Recessed posts are
also found in line with the 5th trestle, these are further discussed in the last chapter.
The wall lines encircling the house on the long sides are very unique in comparison to what is
known round construction-techniques in the eastern Mälar Valley. The double rows of wall posts
and an inner trench with traces after burnt clay suggests a massive wall structure. The walls were
ended by larger gable posts, labeled Hörn 2 (Ulväng 1992), typical for the mid-century AD (see
Hjulström 2008). Among other features, the hearths are multiple; four of different size are located
centrally within the mid-axis, while two is slightly offset, positioned in line with the western line
of the trestles. As was mentioned above, the southern one of the latter must be relegated to a later
stage.
Table 1. Approx. distances in meters between post groups in House I. H= hearth.
Jorth gable
3
Single 8
3,5 HH
Group 7
5 HH
Group 6
1,5 H
Group 5
3,5
Group 4
4
Single 3
3
Group 2
2,5
Group 1
2,5
South gable
Figure 5. Distribution of finds in House I. The figure is incomplete and therefore preliminary. Only selected finds
from the excavations of 2010 and 2011 are plotted. The finds from 1992 are missing (including the piece of bread
and ring from a chain mail).
Within the postholes of the 6th trestle, a gold gilded cover plate of bronze, once part of a swordhilt, and larger fragments of at least three loom weights were found. Similarly, within the 1st
trestle several loom weights along with a spindle whorl were found. A thin ring of bronze was
25
found within the posthole in the 8th trestle and a lancet-shaped arrowhead was discovered in the
wall line close to the northern gable. In the wall line between the 6th and 7th trestles, a small piece
of Snartemo-glass, characteristic for magnate farms during the Migration period, was found. The
find material is apart from the above mentioned very sparse, consisting of a few iron rivets,
ceramic shards, fragments of loom weights, a sinker, a comb and knifes. Interestingly, an
accumulation of ceramics and material indicating household activities is seen in the southern end
of the house (fig. 5). The bone material from House I is minor. Certainly, objects might have
survived within the active part of the society over centuries, but the more or less dateable findings
in House I, the comb, the sword-hilt, the arrowhead and the glass fragment, all point to the 5th and
early 6th centuries AD, implying a rather short period of use (Olausson 2011a:238f).
The earlier analyses of the material from House I includes a soil sample from the left posthole of
the 2nd trestle (Olausson pers. comm. 15/3 2012) which was examined for macrofossilzed seeds.
The analysis showed occurrence of arable weeds, which was already threshed.
House III
Terrace III was only partly excavated during 2011. From the stone construction, forming the
north end of the terrace, and southwards slightly less than 100 m2 were excavated. The features
within Terrace III is generally more diffuse than within Terrace I, however a preliminary draft of
the northern part of a house can be proposed. Centrally within the trench a semicircular
construction of stones was discovered. It was extended southwards along the edges of the trench,
suggesting that is might be a wall with rounded corners. No certain postholes were documented
within the inner area of this feature though, somewhat questioning the construction of this
potential house. Alternatively the area should be interpreted as a platform. However, in this thesis
it is referred to as House III, which by extending the trench might be falsified in the future.
The occupational layer within the house was deep, complicating the question round stratigraphic
layers. The uppermost occupational traces were very diffuse to their character, leaving the
impression of randomly distributed refuse material. Within the major part of the occupational
layer, two distinct timehorizons, i.e. floor surfaces, was identified. The more recent, which
formed the bottom of layer 2 (L2), was characterized by a round stone construction at the
southern end of the trench. The feature had a flat stone slab at its base, with stones of different
sizes laid on top of the edges. It was preliminary interpreted as a baking oven or a cooking area.
Most pronounced in the eastern proximity of this feature, and stretching northwards, was a
compact grayish layer of clay. The extent became more diffuse towards the west and north,
occurring there as white-grayish spots. Within the mid-section of the tentative house, an
elongated hearth was discovered. The uppermost part of the hearth probably correlates with layer
2, while it was initially used contemporary to layer 3 (L3).
The find material within L2 is abundant and holds a great diversity. Ceramics and unburnt bone
material are plentiful. Common are also the waste products from iron work, i.e. slag as well as
waste and semi-manufactured products from bone work. Concentrations of these traces are found
primarily northwest and east of the hearth (fig. 6). The bone work activity is exemplified by
material from all the production stages; bones with cut marks – worked bone – semi-finished
combs. Dispersed findings of loom weights, beads and whetstones were also made. Belonging to
the more unusual findings in settlement contexts are the polyhedral dress pin, a gilded belt
buckle, a crossguard to a knife or small sword and three fragments of bread. Regarding the dating
of L2, the combs are similar to the ones found in T II and T I, decorated with edge-lines. The
polyhedral dress pin also belongs to the general time-interval of 5th – 6th century AD (Waller type
II:1 1996:48, 117).
The floor layer, which was burnt in proximity of the hearth, laid firmly on top of a rectangular
boulder situated close to the slab in L2. South of this block several flat stones were arranged at
the same level, forming a small platform. Immediately north of it, the elongated hearth was
26
located, and surrounding the platform were layers of ash. This context wsa found resting on the
floor layer labeled L3, an orange burnt layer of compact clay stretching throughout the area south
of the semi-circular wall. The occupational layer upon L3 was dominated by bone material and
ceramics. The more spectacular finds came towards the bottom of the hearth, a bone needle and a
carved figurine with the appearance of a horse. However, as mentioned in 4.2.2 L2 was best
preserved in the eastern parts and not identified along the walls making a separation of
occupational layers difficult. Hence, it is somewhat tricky to assign the find material to either L2
or L3.
Figure 6. Distribution of finds and features at Terrace III Layer 2.Ceramics and unworked bone material are not
included.
Trench X
Situated along the northeastern part of the wall an area characterized by hearths and surrounding
soot layers was excavated. The topography is very rough and the area best suited for a house is
located directly west of the trench. The excavated area can be seen as an activity area in relation
to this house. The uppermost layers in the western part of the trench, which superimposed the
sampled layers, contained a small equal-armed brooch, an artifact often found in graves in
combination with the simple polyhedral pins, as was found in T III (Waller 1996:72). The general
dating relates this find to the transition period/early Vendel period (see Waller 1996:72 and ref.
therein). The upper layers may be dump layers while the lower layers (L2 A/B, L4, L3/6 and L9)
were all interpreted as occupational layers. L2 is characterized by soot, hearths and bone material
while L4, also containing abundant bone material, probably functioned as filling material to level
the area. The oldest layers, L3/6 and L9, are more diffuse in character, whereof the former occurs
in patches, hence the split labeling. It is noteworthy that a mould was found within the lower
layers, indicating metal crafting. It is mainly L2 that was sampled and discussed in this thesis.
5.2 Terrace I - geochemistry
Although eight different elements where analyzed the degree of interpretability differs markedly
(fig. 7). The occupational layer of T I is mainly characterized by four major patterns
corresponding to the longitudinal direction. However, Ca more or less reflects variations within
these observed areas while other metal elements occur concentrated in what must be interpreted
as rooms. Beginning from the north, six samples were collected from the area encircled by a
gutter and separated from the northern gable of the longhouse by the low and compact stone
construction. The samples are characterized by relatively high concentrations of Fe, Zn, Cu, Mn
and P, while Mg, K and Ca values are constantly low.
27
The inner area of the northern part of the house constitutes of a second identified activity area
that reaches from the northern gable to the 5th trestle, counted from the south. This area is
visualized by the distribution of K, that display significantly elevated values. There are similar
tendencies in the distribution of Mg, however more complex and probably deriving from different
sources (see ch. 6). Additionally, other metal elements within this area display an intrasite
variation. Zn, Ca, Mn and P values are enhanced in the nave, in proximity to the hearths, while
lower in the aisles.
28
Figure 7. Distributions of metal elements across Terrace I. The symbols are graduated by natural breaks based
solely on the samples from TI and the internal variation of each element respectively.
Between the 3rd and 5th trestles, excluding Mg, the concentrations of metal elements are generally
low, probably reflecting a separate area with depleted values. In detail, the values seem to be
depleting from west to east. In the eastern and central parts there is only one sampling point for
two elements (Zn and P) that deviate from the overall pattern of low values. Furthermore, the
floor at the western entrance and the ground just outside the house show rather high
concentrations of Mn, Ca, P, Zn, and Cu, indicating a different concept and use of the two
29
entrances. The fourth group identified through elemental analyses reaches from the 3rd trestle to
the southern gable. The suite of enhanced metal elements is rather similar to those enhanced in
the northern gable. Cu, Fe and to some extent also Mn and Zn seem to be accumulated from the
3rd trestle and southward. The difference between the northern gable and this southern part of the
house, regarding metal elements, is instead pointed out by the P concentrations, which are
generally low within the latter.
0.25
0.2
0.15
Sterol ratio
0.1
0.05
0
House I House I House I
House I House III House III
Trench
SR
Ent
CR
NG
L2
L3
X
Figure 8. Bar-chart displaying the sterol ratio (cholesterol / [stigmasterol + campesterol + β-sitosterol]). Samples
within House I are displayed from south to north with the start at the very left in the chart. SR= Southern room, Ent=
Entrence, CR= Central room, JG= Jorthern gable.
Figure 9a-b. a) Left, presence of methyl dehydroabietiate. b) Right, plan of sterol ratios and C18/C16 ratios.
30
Hence, based on the distribution of metal elements a tentative division of the terrace into four
activity areas was implemented. Three of these areas were located within the actual house, while
the context of northernmost area is more diffuse; the construction and relation to the building is
not clear. In an attempt to further separate and identify the different areas, entitled southern room,
entrance, central room and northern gable, soil samples were selected for lipid analyses.
Respectively, seven (7) southern room, three (3) entrance, nine (9) central room and six (6)
northern gable samples were selected for the analysis. The samples were compared based on four
targeted variables, sterol- and C18/C16 fatty acid ratios, alkanoic acids and n-alkanols, as well as
eventual identifications of other biomarkers.
The sterol ratio (cholesterol / [stigmasterol + campesterol + β-sitosterol]) as well as the
C18:0/C16:0 ratio regarding alkanoic acids are both indicative of the input of material with
animal origin in relation to plant material. The sterol ratio exhibits clear variations (fig. 8); the
most distinct difference is found between the southern room, which have low values, and the rest
of the house, where the ratio is generally exceeding 0,1. Hence, the input of cholesterol is
significantly higher within the central room and northern gable. The samples from the entrance
area are rather few, but seem to display a transition-zone between these two sections. Both ratios
are illustrated in fig. 9b, with a high sterol ratio set to > 0,1 (cf. Hjulström et al. 2008:19). Sample
points with coinciding positive indications are found mainly within the northern gable and the
central room, while the southern room only displays sporadic indications among an overall
picture of negative results. The strong correlation between these two variables, as seen in both
fig. 9b and table 4, strengthens their validity as indications of activities associated with handling
of animal products.
Another biomarker found at a limited area was methyl dehydroabietiate. The samples with signals
from this compound were all but one restricted to the northern gable, indicating a use of pine-, or
less probable spruce tar within this small area (see fig 9a).
Table 2. Factor loadings for the two most influential factors in the factor analysis. AA = Alkanoic acids. AL = nalkanols. Ch= Cholesterol. Ph = Phytosterols.
Variable
Fe
Factor 1
Factor 2
-0.032790 0.815146
Ca
0.551366
0.434819
K
0.147580
-0.199763
Cu
-0.377885
0.828113
Mg
0.206639
-0.635316
Mn
-0.128167
0.591982
Zn
0.010343
0.712507
P
-0.757244
0.365118
ACL AA
-0.817345
0.338294
CDI AA
0.109554
0.381813
CPI AA
0.549173
-0.445675
ACL AL
0.797456
-0.090955
CDI AL
0.725248
-0.259303
CPI AL
-0.852157
0.125554
Ch/Ph
-0.421844
-0.379420
C18/C16
-0.696168
-0.264057
Variation
31.92%
17.69%
Figure 10. Scatter plot based on the factor analysis for all metal elements, the lipid indices ACL, CDI and CPI, the
sterol ratio and the C18/C16 ratio from the observed activity areas on Terrace I.
The disparities between these areas are not as pronounced regarding the ACL, CPI and CDI
calculations for alkanoic acids and n-alkanols. A rather homogenous result generates no
31
possibilities to discern between the tentative rooms, notable though is the somewhat surprising
score considering the northern gable, where the ACL’s for alkanoic acids are slightly higher,
opposing the results of the sterol- and C18/C16 fatty acid ratios. On the other hand the ACL’s for
n-alkanols are slightly lower than the rest of T I, underlining the less input of plant material.
In order to visualize and weigh all variables together, a data reduction method was used, i.e.
factor analysis. This method enables description of the variability among the observed variables
and searches for joint variations. Hence, co-varying variables are weighed together in so called
latent dimensions as caused by underlying non-observed variables. For example, an elevated
value of three different metal elements might have a common cause, i.e. a non-observed variable.
These, also known as factors are responsible for a certain amount of the variation within the
population. Here the two factors which are mostly responsible for the variation are plotted in a
two-dimensional diagram. In the legends, each variables contribution to the variation is
described; a value close to zero equals a low contribution, while a value close to 1 or -1 equals a
high ratio/enhanced value and low ratio/depleted value respectively.
The samples which were selected for lipid analysis were compared in a factor analysis where all
the measured variables, the metal elements, the lipid indices ACL, CDI and CPI, the sterol ratio
and the C18/C16 ratio, were used. The scatter plot (fig. 10) displays the distribution of the
samples from the four observed areas. The tentative division is more or less verified in the scatter
plot; the south room, central room and the northern gable are all significantly separated. The
entrance area, as suggested above, displays an overlapping pattern indicating a transition zone.
The northern gable is primarily separated from the actual longhouse by factor 1, responsible for
31,92% of the total variation; high negative loadings of P, ACL AA, ACL AL, CDI AL and CPI
AL accounts for the major variance. However, the scores from the calculations of carbon lengths
might be somewhat delusive (see ch. 6.2). Factor 2 separates the northern gable from the central
room and also divides the central room from the south room; high positive loadings of Zn, Cu
and Fe accounts for the major variance (tab. 2).
5.3 Terrace III - geochemistry
The samples from three different horizons within the occupational layer were analyzed for metal
elements. However, due to the stratigraphic difficulties, i.e. no distinguishable floor layer, the
uppermost layer was not prioritized in this thesis. The two older layers were both identified as
floor layers based on deviations regarding texture and color, hence suitable as sampling horizons.
The concentrations from layer 2 are illustrated in fig. 11 and 13, and layer 3 is visualized in fig.
11. Hence, the latter is not illustrated regarding intrasite variations; instead it is presented as a
whole, only separating the in- and outside.
The spatial distribution pattern within layer 2 is very constant over the different metal elements;
the semi-circular line of stones, located centrally within the trench, seems to form the northern
limit of the concentrations of most metal elements (Cu, Zn, Mn, Ca, K, P, Fe), suggesting a
division of the trench in two activity areas, whereof the southern area display intensive use, while
the northern area is less used. However, minor intrasite variations are seen regarding some
elements; Fe values are found very high in the central western part of the trench, while Fe
together with P are found low around the hearth. The metal element that deviates completely
from this picture is Mg, which forms an almost complete contradistinction to the other elements,
there are low values in the southern part and high concentration within the northern part.
Regarding the distributions of metal elements from layer 3, these correspond generally well to the
pattern observed for layer 2 (fig. 13). The observed division of the trench into two areas (fig. 11)
is valid also for layer 3. Likewise, the concentrations of Cu, P, and Ca are similarly enhanced in
comparison with the northern area, while Mg is again considerably lower, underling a similar
space use over time. However, the only significant differences are found in the concentrations of
Zn and Mn, which are lower in the layer 3.
32
Soil lipid analysis was performed on six samples from layer 2 and 3 respectively. These were all
selected from the area with generally higher concentrations of metal elements, encircling the
hearth. The sterol ratios are for both layer 2 and 3 very low and falls short of even the ratios
measured for the southern room within T I, only exceeding 0,1 in one case (fig 8, Table 3).
Similar results are obtained from calculations of C18/C16 fatty acids (Table 4), suggesting a
rather low input of lipids of animal origin in comparison to plant material. The calculations of
ACL’s for the alkanoic acids are more or less confirming this conclusion, generally exceeding the
results from T I. Judging from the n-alkanols the input from plant material seems to be rather
similar or slightly lesser than in T I (Table 5). The biomarker betulin was detected within most
samples from both layer 2 and 3. However, no traces of lupa-2,22(29)-dien-28-ol was detected
excluding a positive identification of birch tar.
33
Figure 11. Distributions of metal elements within Layer 2 across Terrace III. The symbols are graduated by natural
breaks based solely on the samples from TIII and the internal variation of each element respectively. The excavation
planes (see fig. 5) are simplified in order to visualize the concentrations.
A factor analysis including samples from all trenches where all variables have been investigated
show a similar factor 1 score between both floor layers in T III and the northern gable within T I
(fig. 12). The major difference is observed in factor 2 where the Ca, Mn and Mg concentrations
together with the sterol ratio accounts for the variation (tab. 3). The input from Ca and Mn is
considerably higher in T III while the sterol ratio is lower.
To sum up, the two layers in T III show rather similar chemical imprints when compared
internally; the organic material show very few traces of animal origin and the metal elements
34
display equal concentrations in generally. However, there is a discernable difference between a
northern and southern activity area based on the elemental analysis.
Table 3. Factor loadings for the two most influential factors in the factor analysis. AA = Alkanoic acids. AL = nalkanols. Ch= Cholesterol. Ph = Phytosterols.
Variable
Fe
Factor 1
Factor 2
0.353342
0.074742
Ca
0.250660
0.872162
K
0.248077
0.265525
Cu
0.785743
0.479198
Mg
-0.250766
-0.833205
Mn
0.578397
0.613876
Zn
0.644500
0.573401
P
0.841094
-0.020874
ACL FS
0.935235
-0.110649
CDI FS
-0.423798
0.677041
CPI FS
-0.780187
-0.085685
ACL AL
-0.764526
-0.124868
CDI AL
-0.844164
-0.253922
CPI AL
0.808512
0.206171
Ch/Ph
0.345766
-0.633764
C18/C16
0.537863
-0.219502
Variation
43.68434
18.35658
Figure 12. Scatter plot based on the factor analysis for all metal elements, the lipid indices ACL, CDI and CPI, the
sterol ratio and the C18/C16 ratio from the observed activity areas on T I, T III and Trench X.
5.4 Trench X - geochemistry
The samples from TX were collected from layers with a complicated stratigraphy within a limited
trench, hence a deepened understanding regarding space use cannot be reached. Nonetheless, it is
interesting to discuss the results in comparison to, or as a reference material to the results from T
I and T III. Considering the metal elements, the concentrations are similar to the results from
layer 2 in T III (fig. 12). The maximum of Cu is however worth noting, a few samples within
each layer exceed the values from T I and T III. Two of the samples are even in parity with the
levels from the area characterized by handicraft in copper in Vendel (Isaksson et al. 2000:13).
Slight differences in the inorganic material are also found in a few samples with very enhanced P
values.
The organic material show greater disparities; as the sterol ratio and C18/C16 fatty acid ratios
show (fig. 8, Table 3), the input of lipids with animal origin seem far greater in T X than in T III.
Again the ACL’s does not support this conclusion, instead the values are the greatest measured
(Table 5). Additionally, the only samples with traces of coprostanol were documented within T
X. These belong to the early phases on the terrace, layer 4 and 9. The traces were accentuated
within layer 4, suggesting a leak downwards. No traces of 24-ethylcoprostanol could be detected,
indicating that the coprostanol originates from human faeces. This observation is however most
interesting in relation to T III and T I where no coprostanol were detected, hence inferring that
these compounds can be preserved in these soils and a general absence of faeces as caused by
diagenesis, should not be taken for granted.
35
Figure 13. Box and whisker plots of mean values, quartiles and max/min values, all samples included. The samples
are separated by the tentative activity areas. SR= South room TI. E=Entrance TI. CR= Central room TI. JG=
Jorthern gable TI. T3L2O= Terrace III, layer 2, outside. T3L2= Terrace III, layer 2. T3L3= Terrace III, layer 3.
TX= Terrace X.
36
Table 4. Sterol and C18/C16 ratios for T III and TX.
Table 5. ACL, CDI and CPI for alkanoic acids and n-alkanoles within T I, T III and TX. Cmax for Alkanoic acids.
ACL = Σ ([Ci]*i/ Σ [Ci]; CDI = 1/ √ ([Ci/100] 2); CPI for alkanoic acids = Σeven (C8-C36)/Codd(C9-C35) CPI for nalkanols = Σeven (C8-C36)/Codd(C9-C35).
Ci = the relative abundance of each carbon chain in percent. i = The carbon number.
5.5 Vessel use
The analyzed shards were collected from T I, three shards (3), and T III, twelve shards (12). The
shards from T I derive from the entrance and southern room. The shards from T III are divided
between the two floor layers L2 and L3, whereof the shards from the former were selected in the
vicinity of the slab/oven, while the shards from the latter comes from the area east of the hearth.
Three (L2) and nine (L3) shards were analyzed for the two layers respectively. A general
distinction regarding the vessels can be seen in the thickness of the ware; slightly finer vessels
with less temper size characterize the L2 and T I areas while coarser ware with a greater temper
size was obtained from L3. This pattern suggested a variation in vessel use over time and
between the terraces. Considering T III, the older phase could be dominated by cooking activity,
while the findings of bread and the construction interpreted as a potential baking oven in layer 2
led to the assumption of parts of this area being used for baking. Hence one of the targeted
biomarkers was ergosterol in order to evaluate a vessel use linked to the fermentation process.
However, no such traces were found, giving no support for this hypothesis. Also opposing the
initial working hypothesis, the shards from both layers in T III show a uniform pattern, while the
37
results of the three shards from T I are more difficult
icult to interpret archaeologically.
archaeologically All shards
from T III show traces of lipids of animal aand plant origin; cholesterol, phytosterol and wax
residues were detected in all shards, while the C18/C16 fatty acid ratio exceeded 0,5 in all
al cases,
suggesting a considerable contribution from terrestrial animals. One of the shards from T I
acceded to this pattern, while the other two shards were documented as empty,
empty meaning that the
peaks did not reach a measureable level in the ion chromatog
chromatogram (tab. 6).
Table 6. Lipid distribution and interpretation of each shard. DT=Diterpenoids. A=Animal. V=Vegetable.
I=Ruminant. M=Milk.
Sample
F10228
F10231
F10114
F10277
F10278
F10274
F10356
F10341
F10346
F10348
F10323
F10347
F9005
F9046
F9120
Terrace
Layer
TIII L2
TIII L2
TIII L2
TIII L3
TIII L3
TIII L3
TIII L3
TIII L3
TIII L3
TIII L3
TIII L3
TIII L3
TI
TI
TI
C18/
C16
0.96
0.83
0.50
1.11
1.00
0.92
1.52
1.00
0.88
0.94
0.52
0.76
0.74
-
C17br/
C18str
0.006
0.014
0.009
0.015
0.010
0.014
0.011
0.033
0.018
0.009
0.018
0.020
0.030
-
Intact
Triacylglycerols
40-54
42-52
40-54
40-54
40-54
40-52
40-54
42-54
-
Cholesterol
X
X
X
X
X
X
X
X
X
X
X
X
X
-
Phytosterol
X
X
X
X
X
X
X
X
X
X
X
X
X
-
Ergosterol
-
Longchained
ketones
-
ω -(oalkylphenyl
)alkanoic
acids
-
Terpenoids
DT
DT
DT
-
Interpretation
AV
AV
AV
IMV
IMV
IMV
IMV
IMV
IMV
AV
IMV
IV
IMV
-
Figure 14. Bar-charts
charts showing the pottery use at Runsa in relation to fiv
fivee other sites. The number of shards
sha
are
shown within brackets. The shards
rds from Vendel are collected from two areas
areas,, displayed in combination “Vendel”
and as “Vendel without storage”. A= Terres
Terrestrial
trial animals; V= Vegetables; I= Ruminant animals; K= Cooking.
However, there are significant differences detected regarding the two floor layers within TIII.
Targeting the intact TAG’s and calculating the C17:0br/C18:0st ratios the ion chromatograms were
searched for indications of an eventual contribution from ruminant fat or milk. A broad range of
TAG’s, i.e. a total of carbon atoms ranging between 42
42-52 to 40-54,
54, was found in 7 out of 9
vessels from layer 3, while being absent within the samples from llayer
ayer 2. Likewise, a ratio of
C17:0br/C18:0st, exceeding 0,02 was documented in two of the shards from layer 3 and in none
from layer 2. Another, but less informative disparity was observed concerning the diterpenoids,
dehydroabietic acids originating from Pinaceae was found in small amounts within three shards
from layer 3 indicating soot or smoke (see Brorsson et al. 2007:423).. However, no traces of
heating, as indicated by long-chained
chained ketones (Evershed et al. 1995) or ω-(o--alkylphenyl)
38
alkanoic acids (Matikainen et al. 2003) were found in any sample. Even though the long-chained
ketones have been questioned as a suitable variable, these are often found within vessels
originating from culinary contexts (fig. 14). The fact that none of the samples within the
population demonstrates these indicators suggests that the analyzed vessels have not been used
for cooking.
Put in a larger context the pottery use at Runsa can be compared to similar studies from Iron Age
settlements (fig. 14). Suitable sites are the settlements in Vendel, Valsgärde (Isaksson et al.
2000:5f), Alby (Hjulström et al. 2008) and Tuna in Alsike (Hjulström & Isaksson 2005, Forsgren
2007) which all have been subject for analysis of vessel use. Pinpointing unique vessels from
joint contexts, certain shards from these sites must be selected (cf Hjulström et al. 2008:11f). The
shards from the manorial site Valsgärde are collected from buildings labeled as outbuildings
below a terrace with a hall. The shards from Alby originate from a prominent hall building. The
analyzed shards from Vendel are collected from two separate settlements upon the Vendel-ridge;
Vendel 1 and Vendel 28. The shards from the former are selected from a multifunctional house
with a storage area combined with a dwelling/representational area, while the latter represents an
ordinary farm, maybe dependent to the former (Isaksson 2000b, Isaksson pers. comm.). Lastly the
shards from Alsike are chosen from two adjacent trenches at the historical farm Mellangården.
The occupational layers, L60 and L97, were separated in time, but both dated to mid Iron Age.
Deeper knowledge about possible house-constructions is unknown, but the findings within L97
suggest an adjacent manorial site, while L60 is associated with food preparation perhaps linked to
a representational building (see Hjulström & Isaksson 2005:22ff, 36f, 44, Isaksson pers. comm.).
The interpretations of vessel use from these sites are based on the material presented by
Hjulström et al. (2008), Isaksson (2000b), Hjulström & Isaksson (2005) and Olsson & Isaksson
(2008). Allocating a percentage of the total distribution of lipids among seven different
classifications of vessel use, it is possible to discern certain patterns. Cooking pots are arguably
found only at Alsike and Vendel 28. All vessels from T III in Runsa and from Alby display traces
of terrestrial animal fats, while vessels with lipids originating from vegetables exclusively are
found at all sites except Alsike L60, Alby and Runsa. However, separating the observed division
of the Vendel-house, a similar pattern is obtained for the dwelling/representational area.
Although, a variation concerning the lipids with animal origin is observed; more shards with
traces of milk/ruminant fats are found at Runsa than at the other sites. This fits the picture given
by the osteological report; mostly bones form cattle and sheep were detected, while pigs were
rare but represented by young animals (Olausson 1996:17).
Table 7. The Euclidean distance between seven sites (Vendel 1:1 is separated into two areas) based on the seven
categories of pottery use shown in figure 14.
Alby
Alsike 97
Valsgärde
Vendel 1
Vendel 1 without storage
Runsa
Alsike 60
Vendel 28
Alsike
Vendel ,,,,,,,Vendel 1 ,,
Alby ,,,97
Valsgärde .....1
without storage
0
0
76
105
90
0
76
87
47
0
49
102
108
64
0
60
99
88
85
41
37
40
94
77
72
53
64
89
57
49
Runsa
0
33
70
,,Alsike
,,,,,60
Vendel
,,,28
0
46
Measuring the qualitative distance between the settlements, based on seven categories of vessel
use, the Euclidian distance was calculated (Table 7). Frequencies in percent were compared
between Runsa and the other sites, resulting in resemblance primarily to Alsike 60 and Alby.
Similarly when Runsa is compared individually with each site, no statistically significant
difference is seen between Runsa and Alby, Runsa and Alsike L60, and neither between Runsa
and Vendel 1 without storage area (Table 8). Summarizing the vessel use at Runsa, T III, where
39
0
the majority of the samples are collected, is characterized by the contents of animal origin but
without traces of cooking. Hence, showing great resemblance to the Alby population and to some
extent also to Alsike L60 and Vendel 1(without storage).
Table 8. Results of χ2 tests of pottery use distributions. When pottery use-frequencies are equal to zero regarding the
same classification for two sites, these have been excluded, resulting in differing degrees of freedom.
Runsa vs Alby
Runsa vs Alsike 97
Runsa vs Vendel 1
Runsa vs Valsgärde
Runsa vs Vendel 1 without storage
Runsa vs Vendel 28
Runsa vs Alsike 60
Degrees of Freedom Pearson Chi-Squ
2
3.68
5
11.16
3
16.58
3
12.96
3
7.80
6
15.37
4
4.36
Probab. P
0.15869
0.04830
0.00086
0.00473
0.05022
0.01755
0.35881
6. DISCUSSIO AD ITERPRETATIO
6.1 The prominent hall-building?
As is true when studying most cultures, the majority of the remains are from the people who
could afford to manifest themselves. During the Migration Period it is argued that a certain kind
of house, a hall, should be seen as an indication of high status. The definition is frequently used
and also applied to a wide range of houses in archaeological contexts, ranging from central places
with regional influence to local magnate farms (Hamilton 2008b:199) and as well as from being
incorporated in multifunctional houses to separate buildings. The very meaning is thereby a bit
diffuse but it is agreed of being a symbol in the upper strata of the Iron Age society. Hallbuildings functioned as manifestations of the social hierarchy between the chieftain, his retinue
and the people with lower status. It was here that the religious activities were practiced and the
social gatherings involving gift exchange were held (e.g. Olausson 2009:54). It is also notable
that food preparations are not linked to the representative room (Hultgård 1996), while the actual
eating and drinking are. The food ideal for the elite, gathered in these representative buildings,
was meat and dairy products and beer to drink (Montanari 1994, Isaksson 2000a:55). It was of
great importance for the reputation to be part of these feastings and for the host to be generous
with food (see Isaksson 2000a:17ff and refs. therein), why we might expect an accumulation of
animal products within the hall-buildings (see Hjulström et al. 2008). Frands Herschend (1998)
defined different criteria for a building to be classified as a hall (translation by the author):
1. They belong to large farms.
2. They consist of a room with a minimum of posts.
3. They distinguish themselves by its location on the farm.
4. The hearths are neither used for cooking or crafts.
5. The finds in these houses are different from those found in the dwelling house.
Building I in Helgö constitutes a proper example of a hall, with rich findings and a typical
construction (Herschend 1995). Another example is found in Sanda, Fresta parish, where the
settlement context, findings of glass and gold items (Olausson 1996) along with the position are
distinct indications of a hall (Åqvist 2004:58). Nevertheless, these criterions are more or less
relative, which is obvious when looking at other sites interpreted as housing a hall. A prominent
settlement such as Lilla Sylta in Fresta parish has been interpreted to consist only of two houses
with one of them being a hall even though the characteristic findings within the actual house are
absent (Edenmo et al. 2005:31,160). The interpretation of a hall in Skrävsta, Botkyrka parish, is
based on the size (50m) of the house, the arrangement of posts as well as a rich weapon-grave
assumed to be contemporary to the house (Bratt & Werthwein 1999). House 4 in Arlandastad,
Norrsunda parish, and house XV in Skäggesta, Litslena parish has been classified as a hall based
40
on the dimension and distribution of the post holes over three trestles (Göthberg et al. 1996:98,
Andersson 2001:43). This house also illustrates the fact that a hall function may be located within
a multifunctional building, in those cases named sal (Thompson 1995, see also Hjulström 2008).
House V in Skäggesta, Litslena parish, is a more typical multifunctional building including a hall
function. Also this house lacks prestigious findings but has got several hearths interpreted as light
sources rather than cooking pits, suggesting representative functions (Göthberg et al. 1996:101f,
Herschend 1993:175ff).
As mentioned in chapter 3.3, House I in Runsa was considered to be a potential hall-building, as
it fulfils criteria 1 and 3 mentioned above. The excavation and geochemical analyses was
implemented to evaluate this hypothesis, and give a somewhat more detailed picture. Interpreting
the metal elements, the elevated K values within House I are especially interesting. These are
aggregated to a defined area, labeled the central room. The distribution pattern also appears to be
spatially connected to the five hearths within this part of the longhouse. Since wood-ash is rich in
K (Isaksson et al. 2000) this linkage is both exciting and illustrative. However, the extent
somewhat contradicts the ethno-archaeological results (Middleton & Price 1996), where the K
concentration is very tied to the immediate surfaces around the hearth. Parallels are on the other
hand found in Vendel, Uppland (Isaksson et al. 2000) where the distribution of enhanced K
values seems to be spread not only to the vicinity, but throughout the whole rooms where hearths
are found. The enhanced Mg values within T I can be assigned a partly similar interpretation; the
metal is obviously connected to burnt wood tissue (Middleton & Price 1996:678), but has also
been connected to manure and activities such as stabling, owing to its presence in chlorophyll
(Isaksson et al. 2000:9). Enhanced K and Mg values might be indications of stables (see Isaksson
et al. 2000, Hjulström & Isaksson 2005) but the total absence of coprostanol as well as 24ethylcoprostanol within the house indicate the contrary. The detection of coprostanol in Trench X
shows that the possibility of total decomposition and leaching, therefore no traces, is negligible.
The distribution of enhanced Mg values overlaps the increased K values, but furthermore it also
seems to be oddly linked to the wall line. This might have a twofold explanation; burning activity
within the central room, and inclusion of manure within the wattle-and-daub construction of the
wall. The latter interpretation is possible since no samples from the wall line where analyzed for
their lipid distribution. However, this hypothesis needs to be evaluated by analyzing the lipid
distribution within the wall-samples before any conclusions can be drawn.
Monovalent ions as K+ can fluctuate in soils by the degree of moisture and tend to aggregate and
bond to the grains where the moisture is low (Eriksson et al. 2005:127). The central room is
located at the highest point of the terrace, hence theoretically being the driest spot, but the
coherence to both the bivalent ion Mg2+ and the archaeological features are too obvious to be
explained as a pattern caused by natural effects. Examining the roof-supporting construction, if
understanding the proximity between the 5th and 6th trestles as the result of reconstruction, the
longer spans are found within the northern part of the house, thus correlating to the traces of
wood-ash, but also extending southwards until the 3rd trestle. However, the abrupt decrease in
concentration regarding K, found in line with the 5th trestle must be interpreted as a wall-division.
Supporting this thought is the recessed posts that were documented in the eastern side-aisle, also
in line with the 5th trestle, which might be the actual construction traces of the identified wall.
Adding the sterol- and C18/C16 fatty acids -ratios to this picture, these were concluded to be
significantly higher within the central room. Relatively enhanced values have earlier been
documented at what have been labeled culinary (Isaksson 1998), dwelling (Hjulström & Isaksson
2009) and feasting areas (Hjulström et al. 2008). In Frands Herschends (1995:225) summary of
findings from several hall-buildings, he concludes a reoccurrence of certain items; glass,
weapons, loom weights, figural gold foils, whetstones and knifes, symbolizing luxury, power and
domestic crafts. Studying the find distribution within the central room, the scarcity is marked; the
findings characteristic for culinary and dwelling areas, i.e. cracked stone, ceramics, bone material
41
and household utensils are absent (Göthberg 2000:22f: Hamilton 2008a:77f). Instead the material
consists of finds more typical for aristocratic contexts; an arrowhead, pieces of drinking vessels,
gilded weapon details and a gaming piece. Additionally, a few findings linked to domestic crafts
are noticeable as well; loom weights, knifes and a polishing stone. The presence of lipids linked
to animal products may therefore originate from spilling connected to eating and serving. The
overall impression of a tidy floor in House I, established on the basis of find-distribution, is valid
also for the metal elements. Apart from the wood-ash indications, several other elements (Zn, Ca,
Mn and P) are low in the side-aisles, but somewhat enhanced in proximity to the hearths. A
pattern with most of the elements elevated close to the hearth has been observed elsewhere
(Middleton & Price 1996:676, Wilson et al. 2008) and might be caused by heating activities
alternatively sweeping patterns and subsequent waste disposal in the hearths.
The overall impression of the central room suggests an area with several light sources but no
traces of metal crafting or cooking although the presence of food is demonstrated by lipid
distribution in the soil. Biomarkers linked to animal products, i.e. meat, together with the tidy
room, settlement context, and artifacts, supports an interpretation of a representative hallbuilding, or in this case since being well delineated by a wall-division, a sal.
The spans between the 3rd and 5th trestle are also rather long but are separated both
archaeologically and geochemically from the central room. No hearths were found in this area,
and the findings are rather scarce and without characteristic artifacts. The location of two
opposite entrances between the 3rd and 4th trestle does lead to the labeling entrance, of this area.
The depleted values within this section supports this idea, as earlier works covering house floors,
doorways and entrances seem to display constant and relatively low values (Middleton 2004,
Barba 2007, Hjulström et al. 2008). Concerning Runsa this is only half the story since the values
seem to be depleting from west to east. What must be viewed as the front door, when considering
topography, relations to surrounding buildings and the entrance through the rampart, display
repeatedly low concentrations of metal elements. Wear and tear resulting from frequent usage is
the probable explanation, which subsequently have led to erosion and depletion of the metal
elements. The deviating element is Mg, which does not fulfill the statement above. Interestingly
this is the exact same conclusion that was drawn from the results from the Alby house, Botkyrka
parish (Hjulström et al. 2008). The floor at the western door and the ground just outside on the
contrary, show rather high concentrations (Mn, Ca, P, Zn, Cu). Parallels are found in other
investigations where sweeping patterns are detected from the kitchen and consumption areas, as
well as crafting areas, with high concentrations following deposition outside the living area
(Parnell et al. 2002:386, 394). A similar pattern is found regarding the artifacts, with a clear
accumulation around the western entrance (fig. 4). Most elements (P, Ca, Cu, Zn, Mn, Fe) have
their concentrations located to the south room or the north gable, as well as in the proximity of
the hearths in the central room. These patterns suggests tidying of the central room and entrance,
with the consequence of enhanced values following deposition of refuse close to the hearths and
at what must be considered as the backdoor. It can be concluded that the entrances were used
differently, the eastern entrance were probably perceived as the main door which was exposed to
the people entering the settlement, while the western entrance were used for disposal of wastes
originating from the feastings.
The south room is delineated primarily by the concentrations of Cu and Fe, which are
corresponding well to the probable inner wall indicated by the portal posts in line with the 3rd
trestle. Sporadically enhanced levels of Mn and Zn are also documented within this area, however
not delineating it. A section of short spans in combination with no hearth are often argued to be
indications of a stable (Olausson 1998:40ff, Göthberg 2000:22f, Hamilton 2008a:81). High
phosphate values from manuring is another strong variable (Hamilton 2008a:77, Wilson et al.
2008). The absence of the latter, coprostanol and 24-ethylcoprostanol suggests that stabling was
not implemented here. A soil sample collected from the western posthole of the 2nd trestle during
42
the excavation of 1992, and examined for macro fossilized seeds, may lead to the answer of how
this room functioned. Seed distribution has occasionally been used to locate and discuss functions
within Iron age houses; cultivated seeds have been stressed as an indication of the dwelling area,
alternatively storage (Ramqvist 1983:155, Liedgren 1992:154f, Viklund 1998:113) while
documentation of wet- and grassland plants, associated with grazing and foddering, have been
used to locate barns and stables (Ramqvist 1983:155, Viklund 1998:127). The sample from the
southern room contained mainly threshed grains of barley, wheat, and oats (Olausson 1996,
Bergström 2007:82, Olausson pers. comm.), i.e. cultivated seeds. Re-examining the metal
elements characteristic for this area it is notable that Cu, Zn and Fe are highly present within both
viscera and cereals (Isaksson et al. 2000:11) while Mn is abundant in especially cereals
(Hjulström 2008:18). Certainly this suite of metal elements is characteristic for metal crafting as
well (Middleton & Price 1996, Parnell et al. 2002), but the levels are much lower than what could
be expected from that activity (see Isaksson et al. 2000). This is also underlined by the results
from House III and Trench X where traces of metal crafting have been found. Furthermore, the
absence of elevated P values and bone material limits the possibility of a space used for food
handling (e.g. Proudfoot 1976, Sanchez et al. 1999, Terry et al. 2004). The metal elements
indicating burning activities, Mg, K and Ca, are also low within this area, supporting the
interpretation of the hearth at the 3rd trestle as remains from a later stage and possibly only a
single use. The organic material, or rather the lack of traces originating from animal adipose
tissues in comparison to the central room and northern gable, also supports the hypothesis that
the inorganic substances does not derive from meat or viscera. Hence, the most plausible
explanation is to understand the southern room as used as a storage area for grains. The findings
from this area (fig. 4), i.e. pottery, loom weights and a spindle whorl, together with the absense of
bone material as well as light sources underline the notion of a storage room (see Bennett
1984:45, cf. Hamilton 2008:86). The extra span with smaller posts, alongside the accumulation of
posts in the eastern side-aisle of this room might be the remains of a loft (see Liedgren 1992:149).
Hence, the grains might have been stored in ceramic vessels and divided over two levels.
The area between the longhouse and the northern stone terrace, labeled northern gable, is
displaying an enhancement of a similar suite of metal elements, with the addition of P. Neither
this area show traces of burning activities, neither archaeologically or geochemically. The lipid
analysis suggests a high input of animal adipose tissues, which together with the relatively high P
values suggests that the concentrations of Cu, Zn and Fe are traces from an activity area linked to
meat storage. A similar distribution of P is visible in the hall at Arlandastad, Norrsunda parish,
with elevated values in one of the gable rooms (Andersson 2001:40). How the traces of methyl
dehydroabietiate should be understood within this context can be discussed. Dehydroabietic acid
occurs naturally within the soil but the methylated compound is only acquired by reaction
between the abietic acid and methanol under anoxic conditions (Hjulström et al. 2006:284). It is
possible the pine tar was used as a wood sealant either for the construction around the northern
gable or of the tentative storage vessels containing the meat within this area. Alternatively the
traces of pine tar should be seen as by-products from an anthropogenic or natural activity; a less
likely interpretation suggests that the gable was used as a small smokery for the meat products,
thereby accumulating the methyl dehydroabietiate. This thought is however contradicted by the
absence of charcoal within the area. Dry distillation might also have occurred in relation to the
burning of the house or the hearths in the central room, where roots may have functioned as the
dugout log in a tar dale, channeling the tar into the lower area of the northern gable.
6.2 Dwelling and crafting or a Harg?
The results from the analyses of T III are far more ambiguous than the ones obtained from T I.
The consistency within the distribution of artifacts and metal elements suggests an inner and
outer area divided by the semi-circular feature found centrally within the trench. The southern
part of the trench might in fact be the northern part of a house, delineated by a stone-built
43
foundation of a gable wall. The most characteristic find material at the terrace, i.e. the abundance
of bone material and ceramics within both L2 and L3, suggests that food processing was at least
one of the activities here. This is also supported by the hearth located centrally between the walls
since workshops and cooking activities often requires a hearth, either for heating or as a source of
light. Areas characterized by a hearth, enhanced phosphate values and findings of household
utensils are often labeled “cooking houses” (e.g. Hamilton 2008a:78, Göthberg et al. 1996:90,
93). Furthermore, dwelling areas are closely related to these cooking areas; sections within the
multifunctional house are often combined workspaces and living areas with findings as grinding
stones, loom weights and kitchen-ware (e.g. Myhre 1980:258, 323, Hamilton 2008a:82f).
Northwest of the hearth within the house on T III, the traces of both iron slag and bone work are
relatively abundant. More scattered throughout the house, findings of other craft-related artifacts
as loom weights, whetstones and iron objects, were made, inferring a multitude of activities
within House III. The traces of iron work are also separated by the concentrations of Fe, which
are peaking within this area.
The bone work is not signaled by the metal elements; instead these signals are drowned in the
major deposition of unworked bone material throughout the southern part of the trench. This
input of hydroxy apatite, with its inclusion of Ca, is probably affecting much of the results
obtained from the analysis of metal elements. Highly calcareous soils in the Maya region in
Mexico, due to the regions carbonate geology, have been designated as very advantageous
considering geochemical methods. This is because of the calcium ions and soil alkalinity which
render metallic ions insoluble (Terry et al 2004:1238). The calcium-rich soils found on the Baltic
islands often generates good conditions for preservation of bone-material as well, while the
generally acidic calcium-poor podzols characteristic for large parts of Sweden is a far worse
condition. A calcium-rich soil as caused by the extensive bone-deposition in House III probably
have the same effect; the soil matrix become saturated with Ca rendering in a slower
decomposition of the hydroxy apatite caused by fewer H+ ions and more Ca2+ ions within the
solution. Furthermore, the generally higher concentrations considering other metal elements at
Terrace III, both within and outside the house, in comparison with House I, does not necessary
derive from a more extensive input, rather the soil alkalinity might have led to fewer bonds
between metal ions and grains being dissolved. In order to evaluate this hypothesis, the pH levels
were tested for three soil samples from around the hearth in L2 and from the central room in T I
respectively. The results showed around 4-5 pH for T I, which is in line with the results, 2,8 –
5,2, from the small trench dug in the entrance and southern room 1992 (Rudin 1992:21), while
the pH for T III was generally close to 7. This must be considered a large intrasite variation since
only approx. 10 meters are separating the terraces. Hence, the concentration of metal elements
might appear greater on Terrace III, making studies of intrasite variations between T III and T I
more complex. The impact from the Ca2+ ions is visible in the intra-distribution to the Mg2+ ions;
the higher values of Mg form an almost reversed picture to the concentration of Ca. A great input
of a cation as Ca leads to an exchange of the ions bonded to the grains. The concentration of Ca
becomes increased within the inner solution, displacing other cations to the outer solution, where
they are exposed to the risk of leaching (see Eriksson et al. 2005:143). The pattern of enhanced
values regarding seven out of eight metal elements within the house, while Mg appears to be
enhanced outside the house, is therefore explained by ion exchange and binding ability. Hence,
the Mg values outside the house are not enhanced, rather it is the Mg values within the house that
are depleted, following a deposition of most analyzed elements but with less input of Mg. The
observation in field where the unburnt bone material appeared to be better preserved within the
western part of the house than especially the northeastern part, coinciding with the Ca
concentrations, also supports this interpretation.
Most artifacts derive from the upper floor layer (L2), whereof clothing objects, the crossguard
and the three fragments of bread are not the kind of artifacts expected in ordinary dwelling
contexts. The bread could be related to food preparations. Sporadic observations of a certain type
44
of construction, interpreted as baking ovens, has been made within some cooking houses (e.g.
Holmqvist 1969:34, 42, Tesch 1972:26ff). However, the interpretation of the stone-construction
within L2 as a baking oven is very vague, and is not supported by the lipid analysis performed on
the shards from the direct vicinity. No traces of ergosterol, which would indicate fermentation,
questions if the bread actually were processed on T III as my initial hypothesis suggested. An
absence of traces within a few sherds is however no proof for an absence within the whole
population. Neither the aspect of diagenesis should be neglected why any final conclusions
cannot be drawn. However, in general, carbonized bread is almost exclusively found in graves,
which are belonging to the upper hierarchy (Bergström 2007:200f, 217). Regarding settlement
contexts bread is rare, but is closely linked to the hilltop settlements (Bergström 2007:217). The
excavations of Boberget, Odensfors and Börsås are as stated above very old and difficult to
approach more closely. In more detail, it is noteworthy that bread has been discovered in hallbuildings (Bergström 2007:45, 49). Recalling the only contemporary settlement within the Mälar
Valley where more than one find of carbonized bread have been made, Helgö, it is of great
interest to approach the different contexts that are not interpreted as dump layers. Within building
group I, bread has been found in an oven (Holmqvist 1969:35) and close to the sunken floor
houses, probably being remains from processing and consumption (Bergström 2007:36, 70).
Within the prominent building group II, single findings of bread were documented outside the
hall, outside the controversial building at Terrace III and next to the wall of the house on Terrace
IV (Bergström 2007:36). However, the major parts of the bread findings are from graves (Sander
1997:77f, Melin 2001:76) and the ritual context upon foundation IV (see Zachrisson 2004:143).
The ritual aspect of bread is also underlined by the observation of bread baked especially for the
burial (Bergström 2007:217). Excluding the contexts of graves and processing of bread it is
interesting to evaluate the possibility of a ritual context regarding T III in Runsa.
Lars Jørgensen has observed a certain type of building adjacent to hall-buildings at magnate
farms in Tissø, Lejre and Gudme in Denmark, which he suggests had a function connected to
pagan rituals (Jørgensen 1998:242ff, 2002:234) (fig. 16). An interpretation which has been
applied to buildings at magnate farms on Swedish grounds, e.g. Uppåkra (Larsson 2006), Sanda
(Åqvist 1996) and Lunda (Skyllberg 2008). The ritual context at Helgö, located in proximity to
the hall-building within building group II, might be understood similarly. These constructions are
linked spatially to the hall-building and they are distinguished from surrounding farm-buildings
by the findings. The cultic practices are assumed to partially been performed within the hall, but
not entirely, and as Åqvist (1996) and Andrén (2002:315f) stresses, these constructions might be
understood as a harg or haergtræf, a cultic house or area known from the Old Norse literature.
Findings of both tools and material related to crafting are frequently documented in the
aristocratic environment both within and in proximity to the hall-buildings. This is however not
surprising since it appears to be a close relation between power, specialized crafting and religion
within the Iron Age society (Söderberg 2005:221ff). The “side-building” in Borg at Vestvågøya,
Lofoten, held except obvious ritual depositions also findings from the whole production process
associated with ironwork (see Söderberg 2005:234 and refs. therin). Perhaps it is in this view the
traces from the whole production process of bone work in House III, should be seen. Likewise,
the accumulated deposition of ironwork and bone material as in House III has a parallel in the
fenced area and “side-building” next to the hall in Järrestad (Söderberg 2005:212, 236). The latter
has been associated with leftovers from ritual meals and the use of bones in ironwork. However,
the connection between food processing, bone material and ceramics in House III is questioned
by the lipid analyses. Both sterol and C18/C16 fatty acids ratios suggests a very limited input of
animal adipose tissue, which is very surprising considering the amounts of bone material that
were unearthed. Two potential explanations can be formulated; either the soft tissues were
removed somewhere else and the bones later deposited on T III, or the burning of the floor layers
(L2 and L3) and consequently oxidation of the organic material have altered the proportion of
sterols and free fatty acids. Theoretically, oxidation should render a more random pattern with a
45
mixture of signals, the results are however uniform. Perhaps a more recent input of plant material
following the burning could generate a distorted picture which is incomparable to unaffected
areas as T I. This question will however remain unsettled.
If viewing the signals as not significantly altered, i.e. the proportion is reflecting the prehistoric
input, then food processing must be excluded from the activities that were performed upon T III.
The result obtained from TX, where the bone material is as extensive as in T III, on the other
hand suggests a large input from animal adipose tissues. These are however questioned by the
ACL calculation. The results of the latter might however be affected by the recent vegetation. In
an analysis of reference soils from different vegetational biotopes, differences in Cmax, i.e. the
dominant free fatty acids, were detected; a dominance of C22 och C24 were characteristic for the
coniferous forest while C28 and C26 were abundant in the deciduous forest (Hjulström & Isaksson
2007:259). Looking at the distribution for the terraces respectively C24 is dominating all samples
within the longhouse on T I, while C28 and C24 dominates the samples within T III L2 and the
northern gable. Within T III L3 and T X C28 is the dominating free fatty acid in all samples but
one (Table 5). This pattern reflects the present vegetation very well, with three pine trees
enthroning T I while deciduous trees and shrubs are characteristic for TX. Hence intracomparisons regarding ACL´s between the terraces cannot be implemented.
Figure 15. Bar-charts showing the relation between grave contexts (black bar) and settlement contexts (white bar)
regarding vessel use. Fm= fish/marine, T= empty. Modified by author after Forsgren 2007.
Figure 16. The Hall and the related fenced side-building in Tissø, phase 2. Modified by author after Jørgensen 2001.
Emphasizing the results from the lipid analysis of the sampled shards they support the thought of
no food processing at T III, neither during the use of L3 or L2. No shards displayed traces of
while residues from both animal and plant material were registered in all shards. Although traces
after cooking can be difficult to detect within the lipid distribution and perhaps only occasionally
will leave any signatures, the total absence of the ω-(o-alkylphenyl)alkanoic acids or longchained ketones infers that no vessel in this context was used for cooking. The food content, i.e.
the mixture of meat and vegetable products must be explained differently. Studying the vessels
from T III in relation to the contexts of other sites, it can be concluded that the vessel use is most
similar to the hall-building in Alby, and the dump layers interpreted as related to the food
preparation to an anticipated hall, in Alsike 60 and 97. They are on the other hand distinguished
from the dwelling/food preparation area within the ordinary farm Vendel 28, the farm-buildings
from the manorial site Valsgärde and the multifunctional house in Vendel 1. The separating
variables are the pure vegetable content as well as the cooking vessels. In a comparative study it
was shown that these two categories were among the main distinctions between settlement
contexts and grave contexts (fig 15, Forsgren 2007:31ff). An observation which has earlier been
applied to the Alby site, where the vessel use, similar to T III, was stressed as indicative of a
ritual context (Hjulström et al. 2008:13). Hence, the shards from T III can be understood as more
in line with the vessel use expected from a grave context rather than settlement contexts.
46
When combining the results from the soil lipid analyses, the location, vessel use, the exclusive
findings, as well as the recovered fragments of bread, it is tempting to link the construction on T
III to the observations of cultic areas in proximity of the hall-buildings.
6.3 Synthesis
To be able to point out the uniqueness and characteristics of the hilltop settlements and Runsa in
particlur, it is of importance to briefly review the types of settlements that are known to
archaeologists. The farmstead was the most common type of settlement during the Migration
Period. The basic combination of houses, which have been discovered at archaeological
excavations in the Mälar Valley, consist of a longhouse together with a smaller separated farm
building (e.g. Hamilton 1995, Bratt & Lindström 1997:9ff, Frölund & Larsson 1997:15ff,
Häringe Frisberg 1998). Furthermore, there are settlements that seem to consist of three or more
contemporary buildings. Since a single house seldom can be dated more precisely than within a
time span of 100 years or more there is always a question of whether the houses really are
contemporary or not. This problem can be pointed out in examples where the buildings cannot be
separated in time through 14C-dating even though they are overlapping stratigrafically (e.g.
Andersson 2001:41). This is certainly a problem concerning the interpretation and it gets even
more complicated since the lifetimes of a house on a farmstead have been approximated to a very
wide time span, 30 – 300 years (Göthberg 2000:108f). It is nonetheless obvious at certain sites
that the buildings are planned in relation to each other and that there must have existed large
settlements (e.g. Tesch 1972, Göthberg et al. 1996:120f). These larger settlements, not
universally though (see Göthberg 2000:101, Andersson 2010:25ff) as well as settlements with
certain representational buildings, specialized crafts or exclusive finds are thought to demonstrate
the upper strata of the hierarchical division among settlements.
To express the exclusivity and afford the luxury characteristic for the upper social strata, a
production generating a surplus must have been necessary, which in turn requires an effort
probably exceeding the labor and resources of an ordinary farmstead (see Widgren 1998:291).
The consequence of this need is seen in the larger farms, the magnate farms, where the space
requirements for storage have led to auxiliary buildings (Hamilton 2008b:201). To achieve this
position in the mid-Iron Age society the prestige and economic wealth brought home from
trading trips and primarily services in armies on the Continent, was probably vital. Therefore the
ideal of warlords, as manifested in rich graves, with status expressions inspired from continental
rulers, is significant (Grönwall 2008:128). In the densely populated stone-wall regions, the
different settlement-types as described here, have been shown to occur in an intertwined pattern
with larger and smaller settlements by turns, indicating a social organization with magnate farms
in power of the smaller dependent farms (Grönwall 2008:123ff, Hamilton 2008a:101f, Renck
2009:II17ff). Runsa on the other hand is, as mentioned earlier, located at the periphery of these
stone-wall regions characterized by agrarian production, but strategically positioned in relation to
waterways.
Looking at Runsa from a wider perspective and in comparison to other settlement forms, judging
from the excavated areas the hilltop settlement has similarities to the ringforts regarding the
accumulation of buildings, presence and expressions of aristocracy and the separating and
defensive wall. The ringforts however display both a close linkage to the Continent as well as the
background of wealth; byres, ordinary dwellings, extensive craft material and storages are
densely built in a village-like community surrounded by low-lands suitable for livestock farming,
creating an ideal foundation for a surplus-production. Looking at magnate farms at different
hierarchical positions in the vicinity of Runsa such as Vendel, Lilla Sylta, Arlandastad and Sanda,
these are situated within the agrarian landscape with dependent farms or households within its
properties and in the nearby systems of stone walls. Alternatively, they are characterized by
extensive production of craft material as the unique site Helgö. In the same sense, the hilltop
settlements of Östergötland, especially Onssten and Gullborg, has been noted for their many
47
loom weights, which in combination with a potential orientation towards stockfarming within the
surrounding farms, make up indications for extensive wool-production and textile-craft (Olausson
1987b:406). However, the term specialized crafts as has been associated with the hilltop
settlements does not necessary imply an extensive production as documented at these sites. The
characteristic traces from specialized craftwork in form of crucibles for bronze casting and iron
smithing in Darsgärde, Gullborg, Boberget and Braberg and gold smelting at Gåseborg has been
argued to be the remains of an attempt to monopolize the administration and exchange of
prestigeous metals (Olausson 2009:51). Nevertheless, as Olausson stated earlier (1987b:409) “so
far there is no proof of forts where metal-craft has been performed in a larger scale than to cover
household needs”. In line with this statement it is difficult to see the so far scattered traces of
crafting material in Runsa as being parts of a manufacturing with the purpose of extensive export
and trade; nonetheless their ideological and symbolical significance is important, perhaps only
meant for the present elite.
North gable – Meat storage
Central room – Hall
Entrance
Southern room – Granary, storage
Figure 17. Left: House V, Skäggesta, Litslena parish (Göthberg et al. 1996). Right: House I in Runsa, Ed parish.
Similarities regarding the inner construction and interpreted functions are found between the sites.
What does Runsa represent then? Undeniable, the manifest and solitaire position in the landscape
and the impressive wall suggests that Runsa was a focal point within the mid-Iron Age society
(Olausson 2011b:12). Material aspects from within the site, the hall-building, the traces of
trading/tributes and crafting now underline this notion. However, at the moment it is mainly the
flat plateau directly within the southern entrance, or subarea A (see Olausson 2011a:226), that
can be described in more detail. Upon the central terrace, a 30 m long multifunctional house was
located. A sal, a representational area where feastings were held and agreements were reached,
along with a room intended for grain storage characterize this prominent building. A separate
room/area for handling of meat products, probably a storage function, was discovered directly
outside the northern gable. The building is by no means unique; hall-buildings which display
areas linked to the household activities are not uncommon; cooking and storage functions,
probably connected to the food that was served at the feastings are occurring at sites as Lunda
(Skyllberg 2008:21), Eketorp (Herschend 1992) and Arlandastad (Andersson 2001:43). A close
parallel regarding the inner construction and disposition are found in e.g. House V, Skäggesta,
Litslena parish (fig.17), where the wider spans are located to the central part along with traces of
reconstruction. This house is slightly longer and has an extra section of short spans. The northern
part lacks hearths and has been suggested to have functioned as a storage area alternatively a
stable, while at least six hearths, all in the nave, are found in the southern part (Göthberg et al.
48
1996:66, 102). The southern part of the house have been suggested as a hall-building (Herschend
1993:175ff) corresponding to the central room. Hence, similarities to the manner of construction
is found among the upper strata in the region.
The occurrence of a granary within a hall-building situated on an island might at firsthand seem
like a far-fetched thought. The importance of cultivated seeds is however reoccurring in different
ways at hilltop settlements. Except for the relation between bread and hilltop settlements which is
already mentioned, also rotary querns have been found at Boberget, Odensfors, Brudberget and
Börsås. Additionally, grain storages have been documented at Odensfors (Nordén 1938:336) and
probably also Börsås (SHM 14560 see Bergström 2007:191). The abundant findings of cultivated
seeds within the ringforts of Gråborg (Hansson 2004) and Eketorp (Helbæk 1979), that exceed
the amount found at ordinary farms is exciting in this context. The former also contained findings
of bread, while the latter held unique findings of seeds meant for beer-production (Hansson &
Bergström 2008:62). If the hilltop settlements were able to display a surplus of cultivated seeds,
while having a disadvantageous position in relation to the arable land, this surely must have
signaled their influence over the hinterland. Likewise, the processed products of beer and bread
were luxury products closely related to power (see Bergström 2007:190ff). Cereal grains could
also be used for e.g. frumenties and pottages. The presentation of a granary within the public
building intended for social gatherings might therefore be a demonstration of power, control of
land and external surplus-production (cf. van der Veen & Jones 2006:225f). Similarly, the power
as centered around the Germanic warlord is manifested in the sal. The findings of a gilded piece
of a sword hilt within the western posthole of the 6th trestle can be understood as pars pro toto,
i.e. a fraction representing a concept. The weapon may symbolize the power of the chieftain and
his retinue. Hence, House I can be paralleled to the main influences from the Roman empire,
adopted by the Germanic aristocracy (Andersson & Herschend 1997:47).
In proximity to the prominent building a less exposed house characterized by some personal
clothing objects along with findings of bread and vessels containing already prepared food, was
situated. Whether this area was intended for dwelling and crafting or ceremonial/ritual activities
related to the hall-building where bread was shared among the participants and meals were eaten,
can be discussed. However, the link between bread, the sharing of bread and the chieftain has
been stressed earlier (Bergström 2007:206) and is vital here, why I suggest a ceremonial/ritual
space. In addition of the bread discussion also the reoccurring findings of perforated vessels at
hilltop settlements, often interpreted as used for curdling (cf. Skyllberg 2008:56f), is interesting
in relation to the relatively high representation of lipids indicating ruminant and milk fats within
the vessels from T III. The assumption of the relation between curdling and perforated vessels
however have not been verified in any lipid food residue analysis (Isaksson pers. comm 25/52012). Nevertheless, the suggestion of hilltop settlements as sites with specialization of cookingrelated activities (see Bergström 2007:193) might be a clue to the greater understanding of how
they functioned.
The disposition of the northern part of the inner area is still rather unknown. However, the many
traces of buildings suggests that several functions remains to be uncovered. Furthermore, the
suggested accumulation of organic material in the western area as indicated by phosphate
concentration (Rudin 1992), was not detected in this study. A more detailed mapping of House I
and III along with a renewed sampling upon Terrace VI (not included in the thesis) revealed
phosphate concentrations to similar extents within T VI and both T III and the northern gable.
Although a large area remains to be excavated, the presence of the aristocracy is obvious in
Runsa. The two buildings at the southern plateau can be interpreted as incorporated in a wider
complex of activities centered to the southern part of the courtyard (fig.18). The houses excavated
along the southern wall might have functioned as cooking houses linked to the prominent house,
while the findings within T II are more or less reversed, probably indicating a dump layer.
However, postholes indicates a small house also at T II. As has been observed regarding some
49
hall-buildings and their immediate surroundings, they seem to be separated from the rest of the
settlement area (Söderberg 2005:192). The massive stone terraces exposed towards the northern
area of the courtyard fulfill this role at Runsa, distinguishing a ritual and official space from an
area expectedly characterized by more private and daily activities. Yet, the household and
crafting activities associated with the so far excavated areas within the settlement are more or less
fragmentary. Only judging from the scattered findings of crucibles, loom weights, slag and
pottery within T II, T III, T X, the trenches along the wall, as well as the test pits throughout the
inner area, it is difficult to identify locations where the activities were performed continuously.
The traces from crafts rather seem to be the remains of occasional use.
Figure18. Interpretation of the space use for the southern plataeu within the hilltop settlement of Runsa.
The extent of inhabitants and the number of dwellings and subordinate households is also
unknown. Certainly, several potential houses and activity areas remain to be excavated but I
would like to stress the interpretation of Runsa as primarily a demonstration of political power.
The wall, the sal accompanied by a granary, along with the ceremonial/ritual context are all
expressions of power and wealth, neither is a basis for wealth. How this power was obtained is
difficult to capture. Basically and somewhat simplified, wealth, ideology and threats of violence
can be important factors in creating this dependent hierarchy. These are all present within the
hilltop settlement of Runsa manifested in the ceremonial activities where the order was
normalized, the luxury of food habits and the need for defense and findings of weapons.
However, the question of how the hierarchical relations originally were created demands further
research beyond the range of this thesis. My point here is the conclusion that the wealth was
seemingly not produced within the hilltop settlement of Runsa or the immediate surroundings, as
is the case for magnate farms in the system of stone walls as well as the ringforts; the extent of
workshop material in form of textile- or metal production is too small and the possibility to
maintain efficient farming in the vicinity is excluded. Likewise, the 14C-dating from the lower
part of the wall (230-580 AD) (Olausson 2011a:241) more or less precedes the active phase of the
houses as dated by the archaeological context (450-550 AD) (Olausson 2010, 2011). The massive
construction of the wall must surely have been engineered and supervised by an already
established elite. As is true with the ringforts of Öland (see Nordström & Herschend 2003:51),
50
the hilltop settlements were probably spatially planned from the beginning. This is evident from
the outline of the wall, designed to harbor a certain amount of activities. My suggestion is that the
wealth and social status of the people who used and lived within the hilltop settlement Runsa was
created elsewhere, either by early formation of manors and dependent farms or possibly by
payment of Roman warriors, plundering and trade. Regardless of the source, the contact with the
Continent is indicated by the access to novelties and acquiring of precious metals. The
establishing of Runsa is thereby primarily a consequence of a need to maintain and strengthen the
prevailing hierarchical order. The disagreement with this social organization is seen in the
reoccurring fires that have devastated the fort (see Olausson 1996:10).
Situated in a peripheral position in relation to the agrarian settlements, but strategically positioned
in terms of trading routes Runsa might in a wider perspective have functioned as an over-regional
port for magnates settled within the dense populated areas northeast and southeast of the hilltop
settlement. The organization, alliances and agreements among the parties involved were restored
by social and ritual gatherings within the houses upon the southern plateau. The exclusiveness
was manifested in the food culture; abundances of meat and refined products from the cultivated
seeds, both tributes or gifts from the farms in dependency to the magnate and his retinue.
Osteological analyses performed on the material from the hilltop settlement Gåseborg indicated a
seasonal use, primarily during summer time (Olausson 2009:53), perhaps a consequence from
social gatherings and feastings primarily held during the time of the year when the waterways
were passable. As has been stressed earlier, the hilltop settlements are not a uniform group
(Olausson 1987b, 2011b:19) and variations within the organization and subsistence surely
existed. The wealth might have been created or increased within sites as Onssten and Darsgärde
by textile production and agriculture, while sharing functions similar to Runsa as well. The
question whether Runsa was used occasionally or inhabited regularly remains unanswered.
Likewise, the matter if Runsa should be seen as an initiative performed by a lone chieftain or a
collective interest among the magnates in the hinterland is unsettled.
7. SUMMARY AD COCLUSIOS
In an attempt to widen the knowledge of the inner organization and activities within the hilltop
settlement of Runsa, soil samples and ceramic sherds were collected from the trenches excavated
during 2011. The soil samples were analyzed for inorganic matter, metal elements K, Ca, Fe, Mn,
Mg, Fe, P, Cu and Zn, as well as lipid distribution while the ceramic sherds were sought mainly
for lipid food residues. These variables were furthermore analyzed in combination with the
archaeological context in order to establish an understanding of the two terraces, T I and T III.
Regarding the metal elements four major signatures were distinguished within T I, and two
within T III. All regocnized activity areas were also supported by the archaeological features,
together discerning room divisions as well as in- and outsides of the houses. Intrasite variations
within the rooms were interpreted as sweeping patterns, and accumulation in secondary contexts.
Moreover, the space use within T I were convincingly identified by the aid of soil lipid analyses
alongside distribution of artifacts and fossilized seeds. The correspondence between the
concentrations of K, Mg, the considerable input of animal products, hearths as purely sources of
light and the archaeological material characterized by exclusive finds led to the identification of a
representational room within House I. Depleted values within the center of the house were linked
to the main entrance at the eastern long side of the house while the enhanced values of Cu, Fe,
Zn, and Mn in the southern part of the building were linked to the lack of hearths, but unilateral
occurrence of threshed cultivated seeds, alongside ceramics and loom weights, indicating a
granary/storage room.
The chemical imprints within T III resulted in discussions round the influence of a major input of
Ca as well as the possible altering of signals caused by burnt floor layers. Two potential
interpretations seemed relevant, a dwelling/crafting area alternatively a ritual context. However,
the contents of the analyzed vessels alongside the fragments of bread found within the house
51
supported the thought of a ceremonial side-building related to the activities within House I. Based
on the available archaeological records a discussion of the properties significant for the hilltop
settlement of Runsa can be outlined:
(i) The rampart and the location within the landscape.
(ii) The focus on social gatherings/feastings and the specialized food.
(iii) The traces of metal crafting.
(iiii) The technical innovations (documented at hilltop settlements in general), novelties within
the mid Iron-Age, as well as the Roman influences and imported precious metal.
None of the identified characteristics separates Runsa or hilltop settlements in general fully from
the magnate farms. The combination of properties however suggests a close linkage between the
Runsa and the Continent and implies a function as a port to the external world where innovation
and influences are first embraced. The position might therefore have been advantageous in an
attempt to create a contact point between the trade routes and the hinterland. The ruler and his
craftsmen in Runsa might hereby have functioned as intermediaries, processors and distributors
of the prestigious material (cf. Olausson 2009:52). Hence, the venue, i.e. the hilltop, was possibly
a secured staging point situated in the borderland, where the desirable material was received and
divided. The ceremonial activities and social gatherings, e.g. sharing of bread and feasting were
the important and informal relations between the chieftain, his retinue and the guests. These were
performed in the representational and exposed southern area of the yard before the more formal
agreements were reached. The use of, and the extent of the traced crafts performed in the more
private northern side within Runsa is still questionable.
The construction of hilltop settlements have been a most conscious act through which the upper
level of the hierarchy managed to organize the landscape, both in the meaning of surveillance but
more importantly through manifesting the social strata and landowning and furthermore control
the exchange of goods and refined products.
8. REFERECES
Almgren, O. 1934. Sveriges fasta fornlämningar från Hednatiden. Föreningen Urds skrifter I, Uppsala.
Ambrosiani, B. 1958. Darsgärdekomplexet. In: Fornvännnen. Stockholm, pp. 161-176.
Andersson, C. 2010. Gårdar, ättebackar och social status - En studie av hierarkier i östra Mälardalens
bebyggelsestruktur ca 200 – 600 e.Kr. Magisteruppsats i arkeologi. Stockholms universitet.
Andersson, G. 2001. Arkeologi i Arlandastad – järnåldershus och stenåldersskärvor. Arkeologisk undersökning,
UV Mitt rapport 2001:27. Riksantikvarieämbetet.
Andersson, K. & Herschend, F. 1997. Germanerna och Rom. Occasional Papers in Archaeology 13. Uppsala.
Andrén, A. 2002. Platsernas betydelse. Norrön ritual och kultplatskontinuitet. In: K. Jennbert, A. Andrén & C.
Raudvere (eds.) Plats och praxis – studier av nordisk förkristen ritual. Vägar till Midgård 2. Lund.
Anjou, S. 1935. Fornborgarnas betydelse ur etnografisk synpunkt, RIG, Stockholm pp 1-12.
Appelgren, K., ilsson, A. & Perming, A. 2002. Hus och gård vid Lida äng. Arkeologiska förundersökningar och
undersökningar. UV-Mitt, rapport 2002:5. Riksantikvarieämbetet.
Åqvist, C. 1996. Hall och harg. Det rituella rummet. In: K. Engdahl. & A. Kaliff (eds.) Religion från stenåldern till
medeltid. Riksantikvarieämbetet. Arkeologiska undersökningar. Skrifter 19. Linköping.
Åqvist, C. 2004. Sanda – en gård i södra Uppland. Arkeologisk undersökning. UV-Mitt, rapport 2004:15. RAÄ.
Arrhenius, O. 1929. Die phosphatfräge. Zeitschrift für Pflanzenernährung, Dungung, und Bodenkund, 10, pp 185194.
Arrhenius, B., ordahl, E., Slytå, K. & Sundlin, H. 1981. Spårämnesanalyser av organiskt material från
arkeologiska undersökningar. Rapport från Stockholms Universitets Arkeologiska Forskningslaboratorium, Nr. 1.
Aston, M. A., Martin, M. H. & Jackson, A. W. 1998. The potential for heavy metal soil analysis on low status
archaeological sites at Shapwick, Somerset, Antiquity, 72, 838–47.
Aveling. E. M. 1998. Characterization of natural products from Mesolithic sites in Jorthern Europe. Ph.D thesis,
University of Bradford.
Aveling, E. M. & Heron, C. 1998. Identification of birch bark tar at the Mesolithic site of Starr Carr, Ancient
biomolecules 2, pp 69-80.
Barba, L. 2007. Chemical residues in lime-plastered archaeological floors. Geoarchaeology 22, pp. 439-452.
Barker, H., Hughes, M. J., Oddy, W. A. & Werner, A. E. 1975. Report on phosphate analysis carried out in
connection with the cenotaph problem. In: R. L. S. Bruce-Mitford (eds.) The Sutton Hoo Ship Burial Vol I, pp 550572. British Museum. London.
52
Bennett, A. 1984. Karleby och Gärtuna. Bebyggelse och gravar från bronsålder och järnålder i Östertälje socken
Södermanland. Riksantikvarieämbetet rapport UV 1984:29.
Bergström, L. 2007. Gräddat. Brödkultur under järnåldern i östra Mälardalen. Theses and paper in Scientific
Archaeology 9. Stockholm.
Bethell, P. & Maté, I. 1989. The use of soil phosphate analysis in archaeology: a critique. In: J. Henderson (eds.)
Scientific Analysis in Archaeology, Monograph No. 19, Oxford University Committee, Oxford, pp. 1-29.
Bethell, P. H., Goad, L. J. & Evershed, R. P. 1994. The study of molecular markers of human activity: The use of
coprostanol in the soil as indicator of human faecal material. Journal of Archaeological Science 21, pp. 619-632
Bintliff, J. L, Gaffney, C., Waters, A., Davis, B. & Snodgrass, A. 1990. Trace element accumulation in soils in
and around ancient settlements in Greece. In: S. Bottema, G. Entijes-Nieborg & W. van Zeist (eds.) Man’s Role in
the Shaping of the Eastern Mediterranean Landscape, Balkema, Rotterdam, pp. 159-172.
Borg, K., äsman, U. & Wegraeus, E. (eds.) 1976. Eketorp. The Monument. KVHAA. Lund.
Brady, . C. & Weil, R. R. 2002. The Jature and Properties of Soil. New Jersey.
Bratt, P. & Lindström, J. 1997. E18 : genom forntiden i Skederid. Arkeologisk undersökning av RAÄ 106, 189
och 190, Skederid socken. Rapport 1997:8. Stockholms läns museum.
Bratt, P. & Werthwein, G. 1999. Hallen i Skrävsta. Arkeologisk för- och delundersökning av fornlämning RAÄ 36,
Botkyrka socken och kommun, Södermanland. Stockholms läns museum, Rapport 1999:16. Sthlm.
Bridson, J. . 1985. Lipid fraction in forest litter: Early stages of decomposition. Soil Biology and Biochemistry 17,
pp. 285-290.
Broberg, A. 1990. Bönder och samhälle i statsbildningstid. En bebyggelsehistorisk studie av agrarsamhället i Norra
Roden 700 – 1350. Rapporter från Barknåreprojektet III. Stockholms Universitet.
Brorsson, T., Isaksson, S & Stenbäck, . 2007. Stil, gods och kärlanvändning – neolitisk keramik från E4:an
undersökningarna i norra Uppland. In: Stenålder i Uppland. Uppdragsarkeologi och eftertanke. Arkeologi E4
Uppland – Studier. Volym 1, 2007, pp. 409-438.
Brown, T. A. & Brown, K. 2011. Biomolecular Archaeology – An Introduction. Chichester.
Bull, I. D., van Bergen, P. F., ott, C. J., Poulton, P. R. & Evershed, R. P. 2000. Organic geochemical studies of
soils from the Rothamsted classical experiments – V. The fate of lipids in different long-term experiments. Organic
Geochemistry 31, pp 389-408.
Bull, I. D., Lockheart, M. J., Elhmmali, M. M., Roberts, D. J. & Evershed, R. P. 2002. The origin of faeces by
means of biomarker detection. Environment International 27, pp 647-654.
Campbell, M. K. & Farrell, S. O. 2006. Biochemistry. Belmont.
Carlsson, D. 1988. Helgö – central place or farmstead? In: Lundström, A (eds.) Thirteen studies on Helgö. The
Museum of Jationel Antiquities, Stockholm. Studies 7.
Carlsson, D. 2004. Vikingahamnar på Gotland. Ingår i Gotland vikingaön. Gotländskt arkiv.
Carlström, J. 2003. Gåseborg. Arkeologisk undersökning och kartering av en uppländsk fornborg. CD-uppsats,
Stockholms universitet.
Charters, S., Evershed, R. P., Goad, L. J., Heron, C. & Blinkhorn, P. W. 1993. Quantification and Distribution
of Lipids in Archaeological Ceramics – Implications for Sampling Potsherds for Organic Residue Analysis and the
Classification of Pottery Use. Archaeometry, vol. 35, pp 211-223.
Christie, W.W. 1987. High-Performance Liquid Chromatography and Lipids. Oxford.
Christie, W.W. 1989. Gas Chromatography and Lipids – A Practical Guide. The Oily Press. Ayr..
Chisholm, M. 1965. Rural settlement and land use. (3rd ed.). London.
Condamin, J., Formenti, F., Metais, M.O., Michel, M. & Blond, P. 1976. The Application of Gas
Chromatography to the Tracing of Oil in Ancient Amphorae. Archaeometry 18, pp 195-201.
Cook, D.E., Kovacevich, B., Beach, T. & Bishop, R. 2006. Deciphering the inorganic chemical record of ancient
human activity using ICP-MS: a reconnaissance study of late Classic soil floors at Cancue´n, Guatemala. Journal of
Archaeological Science. 33, 628-640.
Craig, O. E., Love, G. D., Isaksson, S. Taylor, G. & Snape, C. E. 2004. Stable carbon isotopic characterisation of
free and bound lipid constituents of archaeological ceramic vessels released by solvent extraction, alkaline hydrolysis
andcatalytic hydropyrolysis. Journal of Analytical and Applied Pyrolysis 71. Amsterdam. pp. 613-634.
Damell, D. 1993. Hillforts and State Formation. In: G. Arwidsson, A-M. Hansson, L. Holmqvist Olausson, B. M.
Johansson, M. Klockhoff, K. Liden, & H-Å Nordström (eds.) Sources and Resources. Studies in Honour of Birgit
Arrhenius. PACT 38.
Damell, D. & Lorin, O. 2007. Provundersökning av husgrundsterrasser invid Hultbergets fornborg i HusbyRekarne socken, Södermanland.Rapport. (ATA).
Damell, D. & Lorin, O. 2010. Fornborgsundersökningar i norra Södermanland. In: R. Edberg & A. Wikström
(eds.) situne dei. Årsskrift för Sigunaforskning. Sigtuna museum.
Dietz, E. F. 1957. Phosphorus accumulation in soil of an Indian habitation site. American Antiquity 22, pp 405-409.
Dimc, . 2011. Pits, pots and Prehistoric Fats. A Lipid Food Residue Analysis of Pottery from the Funnel Beaker
Culture at Stensborg, and the Pitted Ware Culture from Korsnäs. Master´s Thesis, Archaeological Research
Laboratory. Stockholm University.
Dudd, S. . & Evershed, R. P. 1998. Direct demonstration of milk as an element of archaeological economies,
53
Science, 282, pp. 1478-1481.
Dudd, S. ., Evershed, R. P. & Gibson, A. M. 1999. Evidence for Varying Patterns of Exploitation of Animal
Products in Different Prehistoric Pottery Traditions Based on Lipids Preserved in Surface and Absorbed Residues.
Journal of Archaeological Science, vol. 26, pp 1473-1482.
Edenmo, R., Hamilton, J. & Stjerna, . 2005. Kocktorp. Gård och grav från järnåldern samt historiska
bebyggelselämningar. Arkeologisk undersökning UV Mitt 2005:6. Riksantikvarieämbetet.
Eidt, R. C. & Wood, W. I. 1974. Abandoned Settlement Analysis: Theory and Practice. Field Test Associates.
Engström, J. 1984. Torsburgen – tolkning av en gotländsk fornborg. Diss. Aun 6, Uppsala Universitet.
Entwistle, J. A., Abrahams, P. W. & Dodgshon, R. A. 1998. Multi-element analysis of soils from Scottish
historical sites: interpreting land-use history through the physical and geochemical analysis of soil. Journal of
Archaeological Science 25, 53–68.
Ericsson, A. & Hermodsson, Ö. 1994. Förändrad fornlämningsbild i Fresta. In: O. Kyhlberg (eds.) Arkeologi i
Attundaland. Riksantikvarieämbetet.
Eriksson, J., ilsson, I. & Simonsson, M. 2005. Wiklanders Marklära. Lund
Eriksson, T. 1995. Hus och gravar i Görla. Arkeologisk undersökning.Uppland, Frötuna socken, RAÄ 23. UVStockholm, rapport 1995:40. Riksantikvarieämbetet.
Evershed, R. P. 1993. Biomolecular Archaeology and Lipids. World Archaeology, vol 25, pp 74-93.
Evershed, R. P. 2008a. Organic Residue Analysis in Archaeology: The Archaeological Biomarker
Revolution. Archaeometry, vol 50, pp 895-924.
Evershed, R. P. 2008b. Experimental Approaches to the Interpretation of Absorbed Organic Residues
in Archaeological Ceramics. World Archaeology, vol. 40, pp 26-47.
Evershed, R. P. & Bethell, P. H. 1996. Application of multimolecular biomarker techniques to the identification of
fecal material in archaeological soils and sediments. In: M. V. Orna (eds.) Archaeological Chemistry: Organic,
Inorganic and Biochemical Analysis. American Chemical Society, Archaeological Chemistry Symposium Series.
Evershed. R. P. & Connolly. R. C. 1994. Post-mortem transformations of sterols in bog tissues. Journal of
Archaeological Science 21, pp 577-583.
Evershed, R. P., Mottram, H. R., Dudd, S. ., Charters, S., Stott, A. W. & Lawrence, G. J. 1997. New Criteria
for the Identification of Animal Fats preserved in Archaeological Pottery. Jaturwissensschaften, vol. 84,
pp. 402-406.
Evershed, R. P., Stott, A. W., Raven, A., Dudd, S. ., Charters, S. & Leyden, A. 1995. Formation of
Long-Chain Ketones in Ancient Pottery Vessels by Pyrolosis of Acyl Lipids. Tetrahedron Letters, vol. 36,
pp. 8875-8878.
Fagerlund, D. 2009. Fornborgen Broborg. Arkeologiska undersökningar 1982 och 1983. Upplandsmuseet.
Fernandez, F. G., Terry, R. E., Inomata, T. & Eberl, M. 2002 An ethnoarchaeological study of chemical residues
in floors and soils of Q’eqchi Maya houses at Las Pozas, Guatemala, Geoarchaeology 17, pp. 487-519.
Forsgren, A. 2007. Dödsgott med käk i kistan – En GCMS- och FTIR-analys av keramik från ett vikingatida
gravfält i Alsike hage, Alsike sn, Uppland. C/D uppsats i laborativ arkeologi. Stockholms universitet.
Frölund, P. & Larsson, L-I. 1997. Skämsta - bosättning och gravar i norra Uppland. Arkeologisk undersökning.
UV-Uppsala, rapport 1997:67. Riksantikvarieämbetet.
Gihl, G. 1918. Upplands fornborgar. UFT XXXIII.
Grönwall, R. 2008. Kammargravar i stensträngsbygd. In: M. Olausson (eds.) Hem till Jarlabanke.
Göthberg, H., Franzén, B-M., Holm, J. & Åkerlund, A. 1996. Skäggesta, Fiskvik och Prästtorp. Arkeologi på väg
– undersökningar för E 18. Riksantikvarieämbetet. Arkeologiska undersökningar. UV-Uppsala Rapport 1996:54.
Göthberg, H. 2000. Bebyggelse i förändring, Uppland från slutat av yngre bronsålder till tidig medeltid. Uppsala.
Göthberg, H. 2007. (eds.) Hus och bebyggelse i Uppland. Delar av förhistoriska sammanhang. Arkeologi E4
Uppland. Vol 3. Uppsala.
Hansson, A-M. 2004. Rapport – Analys av fossila cerealier och andra växtrester från Gråborg, Algutsrum sn, Öland.
Hansson, A-M. & Bergström, L. 2008. Sädesförråd under äldre järnålder i Gråborg. In: G. Tegnér. (eds.) Gråborg
på Öland. Om en borg, ett kapell och en by. KVHAA.
Hamilton, J. 1995. Hus och gravar i Knivsta – Forsby. Arkeologisk undersökning. UV-Stockholm, rapport
1995:152. Riksantikvarieämbetet.
Hamilton, J. 2008a. Gårdsbebyggelse under folkvandringstid. In: Gården tingen graven. Riksantikvarieämbetet
arkeologiska undersökningar, skrifter nr 71. (eds.) J. Anund.
Hamilton, J. 2008b. Storgården vid Sylta och den uppländska stensträngsbygdens boplatser. In: M. Olausson (eds.)
Hem till Jarlabanke.
Hamilton, J. & Vinberg, A. 2011. Små gårdar och hushåll i romersk järnålder och folkvandringstid – om de
obesuttna i Mälardalens stensträngsbygd. In: M. Olausson (eds.) Runnhusa, bosättningen på berget med de många
husen. Skrifter från projektet Runsa borg, Eds socken, Uppland nr 1.
Hall, B. 1992. Grävningar i Djupedals fornborgar – vad och varför? In: S, Andersson. B, Hall. & G, Öborn. (eds.)
Borgar från forntid och medeltid i Västsverige. Arkeologi i Västsverige 5, Göteborgs Arkeologiska Museum. 28-35.
Hayek, E.W. H., Krenmayr, P., Lohninger, H, Jordis, U., Moche, W. & Sauter, F. 1990. Identification of
54
archaeological and recent wood tar pitches using gas chromatography/mass spectrometry and pattern recognition.
Analythical Chemistry 62, pp 2038-2043.
Heidenreich, C. E., Hill, A. R., Lapp, D. M. & avratil, S. 1971. Soil and environmental analysis at the Robitaille
site. In: W. M. Hurley & C. E. Heidenreich, (eds.) Palaeocology and Ontario Prehistory. Department of
Anthropology, University of Toronto, Research Report 2, 179-237.
Hemmendorff, O. 1985. The Hillfort at Frösön in a North Scandinavian Perspective. In: Archaeology and
Environment 4. Institutionen för Arkeologi, Umeå Universitet, pp. 235-246.
Hemmendorff, O. 1992. En omstridd bastion i det norrlänska inlandet. Populär Arkeologi, 1/1992, pp. 11-13.
Helbæk, H. 1979. Agronomic practices in migration period Eketorp. In: U. Näsman. & E. Wegraeus. (eds.) Eketorp,
Fortification and Settlement on Öland/Sweden. The Setting. Stockholm, pp. 115-127.
Hellberg, L. 1975. Ortnamnen och den forntida sveastaten (presentation av ett forskningsprojekt.) [Tillägg:] Några
namnteoretiska synpunkter på ”Ortnamnen och den forntida svea - staten”. Inledningar till JORJA:s fjärde
symposium Ortnamn och samhälle på Hanaholmen 25–27.4.75. 1. Namngivning s. 92–110 resp. 3.
Hermelin, E. M. 1929. Om Södermanlands fornborgar. Fornvännen 1929:2.
Heron, C,. Evershed, R. P. & Goad, L. J. 1991. Effects of Migration of Soil Lipids on Organic Residues
Associated with Buried Potsherds. Journal of Archaeological Science 18, pp. 641-659.
Heron, C., emcek, ., Bonfield K. M., Dixon D. & Ottaway, B. S. 1994. The Chemistry of Neolithic
Beeswax. Jaturwissenschaften 81, pp. 266-269.
Herschend, F. 1992. Beowulf and St. Sabas: The tension between the individual and the collective in Germanic
society around 500 AD. Tor, vol. 24. Uppsala.
Herschend, F. 1993. The origin of the hall in southern Scandinavia. In: Tor - tidsskrift för nordisk fornkunskap.
Societas archaeologica Upsaliensis 25.
Herschend, F. 1995. Hus på Helgö. Fornvännen 90, Stockholm
Herschend, F. 1998. The Idea of the Good in Late Iron Age Society. Occasional papers in Archaeology 15. Uppsala.
Hjulström, B. & Isaksson, S. 2005. Tidevarv i Tuna. Arkeologiska undersökningar i Tuna by, Alsike sn, Uppland.
Rapporter från Arkeologiska forskningslaboratoriet 3.
Hjulström, B. & Isaksson, S. 2009. Identification of activity area signatures in a reconstructed Iron Age house by
combining element and lipid analyses of sediments. Journal of Archaelogical Science, 36. pp. 174-183.
Hjulström, B., Isaksson, S. & Hennius A. 2006 . Organic geochemical evidence for pine tar production in middle
Eastern Sweden during the Roman Iron Age. Journal of Archaeological Science 33, pp 283-294.
Hjulström, B. & Isaksson, S. 2007. Dolda spår av forntida verksamhet. Geokemiska analyser i samband med E4undersökningarna sträckan Uppsala-Mehedeby. In: H. Göthberg (eds.) Hus och bebyggelse i Uppland. Delar av
förhistoriska sammanhang. Arkeologi E4 Uppland. Vol 3. Uppsala.
Hjulström, B. 2008. Patterns in Diversity – Geochemical Analyses and Settlement Changes During the Iron Age –
Early Medieval Time in the Lake Mälaren region, Sweden. Theses and Papers in Scientific Archaeology 11. Sthlm.
Hjulström, B., Isaksson, S., Karlsson, C. 2008. Prominent Migration Period Building. Lipid and Elemental
Analyses from an Excavation at Alby, Parish of Botkyrka, Södermanland, Sweden. ACTA Archaeologica,
vol. 79, pp 62-78.
Holmqvist, W. 1969. Helgö – den gåtfulla ön. Stockholm
Holmqvist, W., Lamm, K. & Lundström, A. 1970. Excavations at Helgö III. KVHAA. Stockholm
Hultgård, A. 1996. Fornskandinavisk kult – finns det skriftliga källor. In: K, Engdahl & A. Kaliff (eds.) Religion
från stenåldern till medeltid. Riksantikvarieämbetet. Arkeologiska undersökningar. Skrifter 19. Linköping.
Hutson, S. R. & Terry, R. E. 2006. Recovering social and cultural dynamics from plaster floors: chemical
analyses at ancient Chunchucmil, Yucatan, Mexico. Journal of Archaeological Science 33, pp 391-404.
Hutson, S. R., Stanton, T.W., Magnoni, A., Terry, R. E. & Craner, J. 2007. Beyond the buildings:
formation processes of ancient Maya houselots and methods for the study of non architectural space. Journal of
Anthropological Archaeology 26, 442–473.
Häringe Frisberg, K., Frölund, P. & Göthberg, H. 1998. Boplatser i Danmarks socken – undersökningar vid
Myrby och Bärby. Arkeologisk undersökning RAÄ 156, 157, Danmarks socken, Uppland. RAÄ UV Uppsala.
Isaksson, S. 1998. A kitchen entrance to the aristocracy – analysis of lipid biomarkers in cultural layers. Laborativ
Arkeologi 10-11, pp 43-53.
Isaksson, S. 2000a. Food and rank in Early Medieval time. Theses and Papers in Scientific Archaeology
3. Stockholm University. Stockholm.
Isaksson, S. 2000b. Food Culture in Early Medieval Time. A Pottery-Use Perspective. In: S. Isaksson (eds.) Food
and rank in Early Medieval time. Theses and Papers in Scientific Archaeology 3. Stockholm University. Sthlm.
Isaksson, S. 2003. Vild vikings vivre. Om en tidigmedeltida matkultur. Fornvännen 98. Sthlm, pp. 271-288.
Isaksson, S. 2007. Analys av organiskt material i jordprover från Tjärby södra, Tjärby sn, Laholms kn, Hallands
län. Arkeologiska forskningslaboratoriet. Uppdragsrapport 74.
Isaksson, S. 2009a. Analys av organiska lämningar i jordprover från Gränna. ARL Uppdragsrapport 143.
Isaksson, S. 2009b. Analys av organiskt material i jordprover från Stafsinge utanför falkenberg i Halland.
Arkeologiska forskningslaboratoriet. Uppdragsrapport 122.
Isaksson, S., Karlsson, C., Eriksson, T. 2010. Ergosterol (5,7,22 –ergostratrien-3β-ol) as a Potential Biomarker for
55
Alcohol Fermentation in Lipid Residues from Pre-historic Pottery. Journal of Archaeological Science, vol. 37, pp.
3263-3268.
Isaksson, S., Olsson, M. & Hjulström, B. 2005. De smorde sina krås. Spår av vegetabilisk olja i keramik från
yngre järnålder. Fornvännen, vol. 100, pp 179-191.
Isaksson, S., Wojnar-Johansson, M. & Hjulström, B. 2000. The Spatial Organisation of Subsistence at an Early
Medieval manor: An Application of Soil Chemistry at a Settlement in the Vendel Parish, Uppland, Sweden. In: S.
Isaksson, (eds.) Food and Rank in Early Medieval Time, ARL, Stockholm University.
Jahnke, R. J. 1992. The phosphorus cycle. In: S. S. Butcher., R. J. Charlson., G. H. Orians. & G. V. Wolfe. (eds.)
Global Biogeochemical Cycles. New York: Academic Press.
James, P. 1999, Soil variability in the area of an archaeological site near Sparta, Greece, Journal of Archaeological
Science, 26, 1273–88.
Johnson, A. H. 1956. Examination of soil from Corrimony chambered cairn. Proceedings of the Society of
Antiquaries of Scotland 88, pp 200-207.
Jørgensen, L. 1998. En storgård från vikingetid ved Tissø, Sjælland – en foreløpig presentation. In: L. Larsson & B,
Hårdh. (eds.) Centrala platser, centrala frågor. Samhällsstrukturen under järnåldern. Uppåkrastudier 6. Acta
Archaeologica Lundensia, Series in 8°. N° 39. Stockholm.
Jørgensen, L. 2001. From tribute to the estate system, 3rd – 12th century. A proposal fpr the economic development
of the magnate’s residence in Scandinavia based on settlement structure from Gudme, Tissø and Lejre, Denmark. In:
B. Arrhenius (eds.) Kingdoms and Regionality. Transactions from the 49th Sachsensymposum 1998 in Uppsala.
Jørgensen, L. 2002. Kongsgard – kultsted – marked. Overvejelser omkring Tissøkomplexets struktur og funktion.
In: K. Jennbert., A. Andrén. & C. Raudvere (eds.) Plats och praxis. Studier av nordisk förkristen ritual. Vägar till
Midgård 2.
Kangur, M. 2004. Metallhantverket på Gåseborgs fornborg, Järfälla sn i Uppland – en studie av metallurgisk
keramik, CD-uppsats i laborativ arkeologi. Stockholms universitet.
Kitson, G. F., Larsen, S. B. & McEwen, . C. 1996. Gas Chromatography and Mass Spectrometry: A Practical
Guide. New York.
Knights, B. A., Dickson, C. A., Dickson, J. H. & Breeze, D. J. 1983. Evidence concerning the Roman military diet
at Bearsden, Scotland, in the 2nd century AD. Journal of archaeological Science 10, pp. 139-152.
Konrad, V. A., Bonnichsen, R. & Clay, V. 1983. Soil chemical identification of ten thousand years of prehistoric
human activity areas at the Munsungun lake thoroughfare, Maine. Journal of Archaeological Science, 10. pp. 13-28.
Larsson, L. 2006. Hall, harg eller hof. Ett kulthus I Uppåkra. In: M. Anglert., M. Artursson. & F. Svanberg (eds.)
Kulthus och dödshus. Det ritualiserade rummets teori och praktik. RAÄ. Stockholm.
Liedgren, L. 1992. House and farm in Hälsingland. A study of agrarian settlement and settlement development in
northern Hälsingland AD 1-600. Studia Archaeologica universitatis Umensis 2. Umeå.
Lin, D. S., Conner, W. E., apton, L. K. & Heizer, R. F. 1978. Steroids of 2000-year-old human coprolite.
Journal of Lipid Research 19. 215-221.
Linderholm, J. & Lundberg, E. 1994. Chemical characterization of various archaeological soil samples using main
and trace elements determined by Inductively Coupled Plasma Atomic Emission Spectrometry. Journal of
Archaeological Science, 21. pp. 303-314.
Lorin, O. 1985. Ett fornborgsprojekt i Rekarnebygden, Södermanland. Rapport 1985:6. RAÄ och SHM.
Lorin, O. 1989. ”Fuskröse” eller fornborg? Provundersökning på Ärna fornborg 1989. Rapport. (ATA).
Lutz, H. J. 1951. The concentration of certain chemical elements in the soils of Alaskan archaeological sites.
American Journal of Science 249, 925-928.
Löfstrand, L. 1982. Broborg. En förhistorisk befästning vid Långhundraleden. Rapport över en arkeologisk
undersökning 1982. Upplandsmuseets arkiv.
Malainey, M. E., Przybylski, R. & Sherriff, B. L. 1999. The Effects of Thermal and Oxidative Degradation on the
Fatty Acid Composition of Food Plants and Animals of Western Canada: Implications for the Identification of
Archaeological Vessel Residues. Journal of Archaeological Science 26, pp 95-103.
Manzanilla, L. 1996. Corporate groups and domestic activities atTeotihuacan. Latin American Antiquity 7, 228–246.
Matikainen, J., Kaltia, S., Ala-Peijari.M., Petit-Gras.., Harju, K., Heikkilä, J., Yksjärvi, R. & Hase, T. 2003.
A study of 1,5-hydrogen shift and cyclization reactions of an alkali isomerized methyl linolenoate. Tetrahedron 59,
pp. 567-573.
Melin, J. 2001. Cemetery 118 and Building Group 7 and Cemetery 115. In: H. Clarke. & K. Lamm (eds.)
Excavations at Helgö XIV. KVHAA, Stockholm.
Middleton, W. D. & Price, T. D. 1996. Identification of activity areas by multi-element characterization of
sediments from modern and archaeological house floors using inductively coupled plasma-atomic emission
spectroscopy. Journal of Archaeological Science 23, pp. 673- 687.
Middleton, W.D. 2004. Identifying chemical activity residues on prehistoric house floors: a methodology and
rationale for multi-elemental characterization of a mild acid extract of anthropogenic sediments. Archaeometry 46,
pp. 47-65.
Montanari, M. 1994 . The culture of food. Blackwell. Oxford.
Morgan, E. D., Titus, L., Small, R. J. & Edwards, C. 1984. Gas-chromatographic analysis of fatty material from a
56
Thule midden, Archaeometry 26, pp 43-48.
Moukawi, J., Fustec, E., Jambu, P. & Jacquesy, R. 1981. Decomposition of lipids in soils: Free and esterified
fatty acids, alcohols and ketones. Soil Biology and Biochemistry 13, pp 461-468.
Myhre, B. 1980. Gårdsanlegget på Ullandhaug I. Arkeologisk museum i Stavanger - skrifter 4.
ordén, A. 1938. Östergötlands Järnålder, del 1och 2. Stockholm.
ordström, K. & Herschend, F. 2003. Det ideologiska inslaget i väven. In: T. Lindqvist & J. Myrdal (eds.) Trälar:
ofria i agrarsamhället från vikingatid till medeltid.
äsman, U. & Wegraeus, E. 1976. Eketorp. Fortification and Settlement on Öland/Sweden. The Setting. KVHAA.
Olausson, M. 1987a. Upplands fornborgar. In: Svensk fornborgsforskning, RAÄ. Dokumentationsenheten
arbetshandlingar, Stockholm. 85-93.
Olausson, M. 1987b. Hillforts, Subsistence and economic centralisation A.D. 350-500 in Eastern Middle Sweden.
In: Studies in Honour of Mats P Malmer. BAR International Series 366.
Olausson, M. 1996. Runsa fornborg – En befäst stormansgård från folkvandringstid. Skrifter om Väsby 5.
Olausson, M. 1997. Hus och tomt i Uppland och Södermanland under yngre bronsålder och äldre järnålder. In: O.
Kyhlberg (eds.) Hus och tomt i Norden under förhistorisk tid. Bebyggelsehistorisk tidskrift nr 33.
Olausson, M. 1998. ”Säg mig hur många djur du har...” Om arkeologi och stallning. Fähus från bronsålder till
idag. Stallning och betesdrift i långtidsperspektiv. In: K. Viklund., R. Engelmark. & J. Linderholm (eds.) Skrifter om
skogs- och lantbrukshistoria 12. Nordiska museet. Stockholm
Olausson, M. 2008. När aristokratin flyttade upp på höjderna. In: V. Edman & M. Geijer (eds.) Nya perspektiv på
borgar och befästningar. Bebyggelsehistorisk tidskrift Jr 56.
Olausson, M. 2009. At peace with walls – Fortifications and their significance AD 400 – 1100. In: L. Holmquist
Olausson & M. Olausson (eds.) The Martial Society: Aspects of warriors, fortifications and social change in
Scandinavia. Theses and Papers in Archaeology B:11. Stockholm University.
Olausson, M. 2011a. Undersökningen av Runsa borg 2010. In: M. Olausson (eds.) Runnhusa, bosättningen på
berget med de många husen. Skrifter från projektet Runsa borg, Eds socken, Uppland nr 1.
Olausson, M. 2011b. Inledning. Runsa och andra befästa höjdbosättningar i östra Mellansverige. In: M. Olausson
(eds.) Runnhusa, bosättningen på berget med de många husen. Skrifter från projektet Runsa borg, Eds sn, Up. nr 1.
Olsson, M. & Isaksson, S. 2008. Molecular and isotopic traces of cooking and consumption of fish at an Early
Medieval manor site in eastern middle Sweden. Journal of Archaeological Science 35, pp. 773-780.
Orton, C. 2000. Sampling in Archaeology. Cambridge: Cambridge University Press.
Östling, A & Larsson. F. 2007. Skit i byn. In: H. Göthberg (eds.) Hus och bebyggelse i Uppland. Delar av
förhistoriska sammanhang. Arkeologi E4 Uppland. Vol 3. Uppsala.
Östmark, K. 1976. Schaktkontroll 1973 intill Runsa fornborg, fornlämning 2, Runsa, Eds sn, Uppland. RAÄ.
Ottaway, J. H. & Matthews, M .R. 1988.Trace element analysis of soil samples from a stratified archaeological
site. Environ. Geochem. Health 10, 105-112.
Parnell, J.J., Terry, R.E. & elson, Z. 2002. Soil chemical analysis applied as an interpretive tool for ancient
human activities in Piedras Negras, Guatemala. Journal of Archaeological Science, 29. pp. 379-404.
Persson, K. 2005. Integrated Geophysical-Geochemical Methods for Archaeological Prospecting. Licentiate Thesis,
KTH, Stockholm.
Pickering, W. F. 1986. Metal ion speciation—soils and sediments (a review). Ore Geology Review 1, 83–146.
Proudfoot, B. 1976. The analysis and interpretation of soil phosphorus in archaeological contexts. In: D.A.
Davidson. & M.L. Shakley (eds.) Geoarchaeology, Duckworth, London, pp. 93-113.
Ramqvist, P. 1983. Gene. On the origin, function and development of sedentary Iron Age settlement in northern
Sweden. Archaeology and environment 1. Umeå.
Reisborg, S. 1994. Building Groups 1, 4 and 5 Structures and finds. In: H. Clarke (eds.). Excavations at Helgö XII.
KVHAA. Stockholm.
Renck, A-M. 2009. Gården som upphörde. Om en uppländsk järnåldersbygds förändring i mellersta järnålder.
Rapporter från Arkeologikonsult 2009:2095.
Rogge, W. F., Medeiros, P. M. & Simoneit, B. R. T. 2006. Organic marker and compounds for surface soil and
fugitive dust from open lot dairies and cattle feedlots. Atmospheric Environment 40, pp 27-49.
Risberg, J. 2011. Landskapets förändringar vid Runsa fornborg de senaste 3000 åren. In: M. Olausson (eds.)
Runnhusa, bosättningen på berget med de många husen. Skrifter från projektet Runsa borg, Eds sn, Uppland nr 1.
Rudin, G-B. 1992. En studie av Runsa fornborg. Funktionsuppdelning med utgångspunkt från fosfatanalys. Uppsats
i påbyggnadskurs. Stockholms universitet.
Rydberg, M. 1995. Jålar, sländtrissor på Eketorps borg. Arkeologiska institutionen Lunds universitet.
Sanchez Vizcaıno, A. & Canabate, M. L. 1999. Identification of activity areas by soil phosphorous and organic
matter analysis in two rooms of the Iberian sanctuary ‘”Cerro El Pajarrillo”. Geoarchaeology 14, pp. 47-62.
Sander, B. 1997. Cemetary 116. In: H. Clarke (eds.) Excavations at Helgö XIII, Cemetery 116. KVHAA, Sthlm.
Schnittger, B. 1908a. Östergötlands fornborgar. Ett utkast till deras undersökning. Meddelanden från Östergötlands
fornminnesförening, Linköping, pp. 29-32.
Schnittger, B. 1908b. Undersökningen på Boberget 1908. Meddelanden från Östergötlands fornminnesförening,
Linköping, pp. 32-33.
57
Schnittger, B. 1909. Östergötlands fornborgar, grävningsnotiser. Meddelanden från Östergötlands
fornminnesförening, Linköping, pp. 14-17.
Schnittger, B. 1913. Utgrävningen på Onssten, Häradshammarssocken. ATA.
Skyllberg, E. 1991. Fornborgar och stensträngssystem: En studie av boplatshierarkier på västra Vikbolandet. Cuppsats, Institutionen för Arkeologi, Stockholms Universitet.
Skyllberg, E. 2008. Gudar och glasbägare- järnåldersgården i Lunda. In: G. Andersson & E. Skyllberg. (eds.)
Gestalter och gestaltningar – om tid, rum och händelser på Lunda. RAÄ Arkeologiska undersökningar skrifter 72.
Sporrong, U. 1971. Kolonisation, bebyggelseutveckling och administration. Studier i agrar
kulturlandskapsutveckling under vikingatid och tidig medeltid med exempel från Uppland och Närke. Meddelande
från Kulturgeografiska institutionen vid Stockholms universitet. B23. Lund.
Stenberger, M. 1964. Det forntida Sverige: svensk förhistoria i korta kapitel. Almqvist & Wiksell, Stockholm.
Söderberg, B. 2005. Aristokratiskt rum och gränsöverskridande. Järrestad och sydöstra Skåne mellan region och
rike 600-1100. Skrifter No 62. Riksantikvarieämbetet.
Terry, R. E., Fernandez, F. G., Parnell, J.J. & Inomata, T. 2004. The story in the floors: chemical signatures of
ancient and modern Maya activities at Aguateca, Guatemala. Journal of Archaeological Science 31. pp 1237-1250.
Tesch, S. 1972. Kring en järnåldersgård i Turinge sn, Södermanland. Uppsats för 3 betyg i nordisk fornkunskap.
Skånel.
Thompson, M. 1995. The Medieval Hall. The Basis of Secular Domestis Life, 600-1600 AD. Aldershot.
Törnqvist, O. 1993. Fornborgarnas betydelse under ekonomisk expansion, politiskt paradigmskifte och social
segregering. En studie med utgångspunkt från Vikbolandet, östra Östergötland. Bachelor thesis in archaeology,
Stockholms Universitet.
Ulväng, G. 1992. Mälardalens hustyper. En studie av hustyper I Mälardalen under sten-, brons- och järnålder.
Master-thesis. Uppsala universitet: Institutionen för arkeologi
van der Veen, M. & Jones, G. 2006. A re-analysis of agricultural production and consumption: implications for
understanding of the British Iron Age Vegetation History and Archaeobotany 15, pp. 217-228.
Viklund, K. 1998. Cereals, Weeds and Crop Processing in Iron Age Sweden. Methodological and interpretative
aspects of archaeological evidence. Archaeology and Environment 14.Umeå universitet.
Vikstrand, P. 2011. Ortnamnet Runsa. In: M. Olausson (eds.) Runnhusa, bosättningen på
berget med de många husen. Skrifter från projektet Runsa borg, Eds socken, Uppland nr 1.
Wahlberg, M. 1964. Fornborgen vid Tångstad samt några problem kring de östgötska fornborgarna.
C-uppsats, Institutionen för Arkeologi, Uppsala Universitet.
Wall, Å. 2003. De hägnade bergens landskap. Om den äldre järnåldern på Södertörn. Stockholm Studies in
Archaeology 27.
Waller, J. 1996. Dräktnålar och dräktskick I östra Mälardalen. Kontinuitet och förändring under folkvandringstid
och vendeltid. Aun 23. Uppsala.
Wegraeus, E. 1976. The Öland Ring-forts. In: K, Borg., U. Näsman. & E, Wergraeus. (eds.) Eketorp, fortification
and settlement on Öland/Sweden. The Monument.
Wells, E. C., Terry, R. E., Parnell, J. J., Hardin, P. J., Jackson, M. W. & Houston, S. D. 2000. Chemical
analyses of ancient anthrosols in residential areas at Piedras Negras, Guatemala, Journal of Archaeological Science,
27, pp. 449–62.
Widgren, M. 1983. Settlement and farming systems in the early iron age. A study of fossil agrarian landscapes in
Östergötland, Sweden. SSHG. 3. Stockholm.
Widgren, M. 1998. Kulturgeografernas bönder och arkeologernas guld – finns det någon väg till en syntes? In;
Centrala platser centrala frågor. Samhällsstrukturen under järnåldern. Acta archaeologica Lundensia. Series in 8.
No 28.
Wilson, C. A., Davidson, D. A. & Cresser, M.S., 2008. Multi-element soil analysis: an assessment of its potential
as an aid to archaeological interpretation. Journal of Archaeological Science, 35. pp, 412–424.
Zachrisson, T. ”Det heliga på Helgö och dess kosmiska referenser”. 2004. H. Clarke. K. Lamm (eds.) Excavations
at Helgö XVI, Exotic and Sacral Finds from Helgö, KVHAA, Stockholm.
Personal comments
Michael Olausson 2012-03-15.
Sven Isaksson 2012-03-30, 2012-05-25.
Appendix I. Location, concentrations of metal elements (ppm), sterol ratio (C/P), presence of coprostanol (Cop) and methyl
dehydroabietiate (MeDHA) and abundance of the fatty acids C18:0 and C16:0 of each soil sample discussed in the thesis.The
coordinate system used is ST74.
Sample
1
2
3
4
5
6
7
X
6605736.85
6605738.7
6605740.45
6605742.2
6605744
6605745.35
6605747.45
Y
140315.5
140314.4
140313.4
140312.3
140310.6
140310.3
140309.3
Area
TI
TI
TI
TI
TI
TI
TI
P
80
83
155
146
176
116
300
K
2507
2286
2420
3003
2987
2714
2517
Ca
959
1010
1327
1116
1603
1711
1493
Zn
89
80
96
88
96
94
100
Fe
18207
19119
19076
18642
20928
19568
19756
58
Cu
11.1
9.72
20.9
13.9
28.7
14.3
28.8
Mg
820
815
812
822
798
798
800
Mn
452
472
412
508
630
647
516
C18:0
C16:0
C/P
1334739
2471037
0.05
Cop
MeDHA
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
6605736.8
6605738.55
6605740.4
6605742.15
6605745.6
6605747.4
6605749.25
6605734.95
6605736.75
6605738.5
6605739.9
6605733.15
6605734.85
6605736.65
6605738.35
6605750.65
6605752.45
6605754.2
6605756.2
6605757.65
6605757.8
6605756.1
6605754.3
6605752.55
6605759.65
6605757.7
6605756.7
6605753.9
6605752.7
6605759.8
6605758.1
6605756.2
6605754.55
6605752.75
6605750.95
6605742.05
6605743.8
6605746
6605747.35
6605749.1
6605750.95
6605734.5
6605736.3
6605738.05
6605739.75
6605739.7
6605743.3
6605745.2
6605750.55
6605742.05
6605743.65
6605745.1
6605746.1
6605749.05
6605733.3
6605735
6605736.7
6605738.45
6605740
6605741.85
6605743.6
6605745.45
6605747.25
6605748.9
6605750.7
6605752.4
6605743.75
6605745.5
6605759.3
6605755.4
6605755.1
6605753.05
6605751.4
6605759.4
6605760.9
6605759.5
6605761.5
6605762.3
6605761.4
6605762.9
6605762
6605763.15
6605761
6605761.7
6605763.4
6605765
140313.1
140312.1
140311.1
140310.1
140308.1
140307.1
140306.3
140312
140311
140309.9
140309.2
140308.6
140307.5
140307
140306
140298.3
140297.3
140296.6
140295.3
140294.3
140296.1
140297
140298.5
140299.5
140297.3
140299
140299.1
140299.7
140301.8
140300
140301
140301.7
140303
140304
140305
140307.0
140306.9
140305.9
140304.8
140303.8
140302.7
140309
140308
140307.3
140305.9
140304.7
140303.9
140302.8
140300.6
140303.7
140302.4
140301.6
140300.9
140299.2
140306.7
140305.7
140304.3
140303.1
140302.3
140301.2
140300.2
140299.1
140297.1
140298.2
140296.2
140295.2
140313.1
140312.1
140298.8
140298.9
140296.9
140300.5
140301.5
140293.3
140292.4
140295.1
140294.2
140293.9
140296.7
140295.8
140298.7
140297.7
140300
140301.1
140292.9
140292.4
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
TI
111
237
97
156
115
93
100
124
176
150
113
107
119
120
172
85
125
114
89
122
108
83
101
128
134
110
172
315
303
179
1153
124
119
151
137
191
208
139
119
157
251
133
172
122
156
283
171
143
175
430
266
228
216
159
173
144
186
210
217
311
213
136
131
147
181
174
188
135
378
183
185
169
236
197
219
178
256
167
196
305
292
253
274
133
360
225
2368
3250
2694
2293
2460
2501
2343
1828
2624
2803
2876
2722
2681
3158
3570
3176
3741
2296
3745
2488
4632
3102
3644
5664
3617
3105
2931
4088
3673
3412
4227
3688
4219
3153
2100
2067
2715
2698
2908
3148
4456
2895
3172
2697
2885
3266
3526
2222
3551
2729
2726
3220
3094
4243
2618
2609
2445
2990
2913
2814
3236
2908
3088
2996
2456
2150
2661
2901
3636
3327
2910
3791
3783
3384
3554
3746
3504
2470
3819
3292
3351
3935
3537
2700
2569
2792
2415
1861
1354
1092
1531
1346
1351
730
1345
1468
2012
1676
1646
1360
1447
1520
1378
919
992
1433
781
940
1147
1532
918
1185
2268
2450
3138
1213
1629
1535
1486
1273
1299
1450
1883
1382
1679
1302
1445
2313
1133
1231
1237
2001
1292
1291
1388
1607
4028
1635
1793
1405
1371
2135
1134
1126
1574
1800
1894
1195
1533
1728
1123
980
1089
1049
1233
1345
1715
1216
1648
631
663
799
865
616
1224
987
1104
1045
1233
1129
965
1158
104
99
94
95
97
95
88
94
96
108
103
107
102
99
99
94
104
85
103
82
100
105
100
97
95
104
108
105
113
100
104
107
99
93
99
112
117
113
97
100
108
109
87
100
93
96
101
101
101
114
107
98
104
94
102
102
95
105
99
104
103
103
95
85
84
97
100
95
105
107
97
99
106
99
96
94
102
83
95
99
96
96
94
86
100
98
20074
30133
20914
22651
19394
18106
17657
17816
23172
23866
25386
19496
20725
24764
22361
18859
18786
17411
19003
18004
19321
19206
19438
20538
17875
19828
19336
22202
20.88
19568
18830
18526
18468
17730
17238
26341
21884
19336
19293
19119
19597
17484
20784
21609
27630
21015
21927
21290
17947
18873
22130
19090
18439
19076
17715
17542
16803
18236
19553
16905
19553
18323
19394
17368
15979
18974
19148
18902
20943
20248
20046
19090
20813
18453
22448
20827
20103
17411
20740
17064
19481
20017
21768
18077
21681
21522
59
14.7
23.9
20.7
23.5
15.7
11.6
10.2
12
19.6
27.8
21.6
13.8
19.8
20.9
24.1
13.2
13.4
11.3
16.6
13.9
22.2
15.4
13.4
12.6
15.8
13.1
15.9
16.1
23.5
39.4
19.8
15.9
12.6
12.5
11.9
34.2
24.9
20.3
16.2
16.6
19.9
13.9
21.4
21.6
25.3
23.2
21.7
18.7
14.9
34.8
23.3
22.9
21.4
12.8
12.3
11.6
12.4
26.5
18.8
18.6
23.8
16.7
12.9
12.3
12
16.5
14.8
10.1
25.5
16.3
14.9
14.5
18.8
19.9
26.7
16.8
22.6
17.4
23.8
21.8
22.5
21.7
39.9
10.6
28.5
21.3
782
731
802
778
762
816
816
860
798
737
730
798
787
747
797
810
792
813
827
824
806
797
790
739
783
760
727
735
680
767
807
792
788
798
798
746
761
783
783
751
771
801
762
756
611
817
767
773
785
772
771
767
767
708
776
770
799
815
752
798
793
760
768
801
822
801
807
794
763
757
781
767
748
799
748
746
751
801
734
812
776
763
757
773
740
741
689
717
363
405
442
356
417
300
500
418
810
642
856
528
402
502
432
304
317
547
262
463
488
604
526
538
687
583
851
576
557
545
629
405
465
478
752
615
682
573
565
840
411
493
455
740
352
475
483
510
861
767
822
550
641
899
652
738
707
668
668
646
659
518
431
344
387
364
386
540
725
407
524
436
593
351
504
470
607
386
511
492
529
387
598
626
1744008
2845541
0.07
1902800
1959363
3987442
3500152
0.07
0.12
1427725
1620793
4898685
3984193
0.09
0.05
1442458
2864941
0.15
1433280
1176364
2808526
3209937
0.14
0.12
1739085
4622148
0.12
1348114
2232593
0.23
1474339
1217694
3587119
2590856
0.13
0.09
1216340
2434747
0.11
1139285
2536986
0.13
1363374
2905282
0.12
1510305
3155756
0.08
1894701
3284917
0.12
1118017
2349127
0.08
X
1412786
1679633
2057127
3514430
0.16
0.09
X
X
94
95
96
97
1
2
3
4
5
6
7
8
9
10
11
12
13
13b
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
31
32
33
34
3
4
6
9
12
15
17
18
20
21
22
23
24
25
26
27
28
29
1
2
3
4
5
6
7
8
9
10
11
1
2
4
6
12
13
202
203
204
209
212
213
214
219
6605764.85
6605766
6605765.65
6605766.7
6604762
6604760
6604758
6604752
6604761
6604759
6604762
6604760
6604759
6604763
6604761
6604759
6604752.15
6604762
6604760
6604758
6604763
6604753
6604753
6604753
6604756
6604754
6604757
6604755
6604756
6604754
6604757
6604755
6604756
6604754
6604757
6604758.8
6604752.15
6604763
6605758
6605752
6605759
6605758
6605759
6605758
6605753
6605753
6605756
6605754
6605757
6605755
6605756
6605754
6605757
6605755
6605756
6605754
6605829.85
6605829.36
6605828.84
6605829.25
6605828.71
6605828.2
6605826.6
6605826.78
6605827.63
6605827.96
6605828.89
6605829.85
6605829.36
6605829.25
6605828.2
6605827.35
6605827.02
6605827.23
6605827.23
6605827.23
6605827.23
6605827.5
6605827.5
6605827.5
6605827.5
140294.7
140294
140297.1
140296
140284
140284
140284
140284
140285
140283
140282
140282
140282
140281
140281
140281
140282.8
140280
140280
140280
140279
140281
140283
140282.3
140284
140284
140283
140283
140282
140282
140281
140281
140280
140280
140279
140278.4
140279.7
140283
140284
140284
140283
140282
140281
140280
140281
140283
140284
140284
140283
140283
140282
140282
140281
140281
140280
140280
140297.7
140296.8
140295.9
140298.1
140297.2
140296.4
140294.9
140298.4
140295.5
140294.6
140294.7
140297.7
140296.8
140298.1
140296.4
140296.9
140297.9
140297
140297
140297
140297
140297.4
140297.4
140297.4
140297.4
TI
TI
TI
TI
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L2
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T III L3
T X L2A
T X L2A
T X L2A
T X L2A
T X L2A
T X L2A
T X L2A
T X L2A
T X L2A
T X L2A
T X L2A
T X L2B
T X L2B
T X L2B
T X L2B
T X L2B
T X L2B
T X L2B
T X L3/6
T X L4
T X L9
T X L2B
T X L3/6
T X L4
T X L9
389
283
467
286
49
168
353
433
264
305
192
368
374
170
216
366
253
332
263
494
271
158
192
118
506
380
390
458
471
187
390
161
343
371
407
377
435
147
332
517
312
372
366
254
278
504
412
474
390
416
197
356
331
282
442
241
174
202
262
234
182
219
319
401
280
419
293
209
260
527
323
291
489
386
715
478
478
414
702
379
442
4220
2547
2129
2851
3696
3779
3323
3639
4303
2648
3613
2740
2976
3521
3183
2731
4339
2886
2992
3374
2931
3790
3647
4149
3240
3300
2977
3426
3152
3253
4096
3707
3638
3723
3228
3316
3475
3173
4064
2262
4642
2344
3612
2834
2015
2137
4177
2435
2680
4631
4517
3485
5140
4130
1590
4231
2533
3438
3091
3001
2886
2990
3700
3127
3166
3360
3154
3059
3153
3397
3168
4097
3230
4768
4470
3682
6221
3099
3118
3588
3125
1025
1221
1088
1071
1302
1044
1308
11015
1047
2172
1786
2877
3581
1821
2839
3124
14465
1382
3301
5525
953
14154
14603
15447
2765
2735
4151
7602
12410
14520
11185
14647
14078
8564
7484
3183
5736
1918
1089
3909
1297
3884
5131
3174
9233
7563
1349
2266
1992
2816
14529
9451
10698
10957
8093
14609
5558
4911
5832
6259
6104
5540
5284
9674
2894
3718
3973
5467
7453
7703
4496
4144
10682
3865
6341
2678
3010
4425
7368
2727
3302
107
104
103
107
105
107
110
113
109
108
108
112
115
107
110
111
118
110
116
111
96
118
118
115
107
116
111
116
117
117
116
116
117
116
116
111
115
109
103
108
105
109
113
104
111
110
108
114
103
103
109
111
107
101
105
118
112
105
113
111
102
103
114
118
106
115
115
103
112
115
113
107
118
111
118
112
117
116
118
115
114
24764
21942
20552
22897
24476
25774
23197
23375
24058
23974
24603
23907
26779
23654
23378
23252
22303
25424
25300
27077
24850
17992
24325
16859
27257
25124
24575
25225
23651
20875
26275
20808
27642
25699
21547
47150
25875
62087
24900
21300
30600
19975
26213
23450
28796
17964
24576
24527
26546
24800
23553
29004
26675
22931
25125
25649
22390
26978
25184
22578
24909
25922
27195
22043
22955
26761
27572
29453
24532
43825
25386
30090
23201
26964
28570
22766
25674
24185
20914
19655
19765
60
32.9
24.4
29.9
31.5
26.2
16.1
35.8
40.1
33.8
29.2
25.3
28.5
38.1
24.7
26.8
29.4
40.4
31.2
35
36.7
23.4
63.9
50.2
58.6
36.5
33.8
40.2
51.7
50.1
47.9
35
52.9
50.2
46.4
45.2
35.9
38.7
27.2
49.4
29.2
36.4
33.0
43.8
29.4
40.9
23.9
52.5
38.4
36.9
41.7
53.7
50.9
41.9
35.7
30.4
51.3
22.9
29.2
31.9
25.8
25.4
31.5
42.9
138.6
38.5
46.5
37.6
28.7
38.7
38.9
46.2
39.2
49.4
37.9
46.2
56.7
66.5
64.2
66.4
62.5
70.1
725
728
750
739
727
761
754
660
718
732
732
753
712
730
739
740
664
759
719
699
792
496
605
404
685
665
680
659
640
566
668
424
649
644
707
732
661
712
700
743
645
813
691
744
544
739
699
738
750
669
638
521
622
625
619
531
786
718
735
750
734
750
677
734
732
748
698
729
719
701
735
617
715
670
589
709
593
734
766
701
776
680
735
623
697
607
495
634
939
483
839
743
931
1074
848
899
929
861
571
1055
833
408
1347
1141
1378
745
796
1012
1132
1183
1329
584
1281
1017
1173
1133
925
947
1092
445
477
556
518
806
556
432
424
376
545
512
436
639
424
527
443
544
966
836
639
846
725
541
546
789
1080
825
825
913
542
836
836
861
559
1010
705
1672
1088
1046
1083
1065
816
1004
1615267
1961248
1071056
1472720
2618470
3372603
2053358
2874992
0.12
0.14
0.11
0.1
1388767
3524789
0.06
1415642
2376634
0.04
1118041
2558132
0.04
1339559
1428632
3365207
3292976
0.05
0.03
1102869
1913024
0.06
970400
2113776
0.05
1585694
4764797
0.1
741586
1542321
0.04
1673659
1382853
3852433
4518433
0.03
0.07
444445
1204074
0.02
1078610
980974
2258307
2011166
2229714
4052890
2614375
1791372
1055282
1359446
4043625
4083251
2542273
5697279
4324518
3084590
0.09
0.17
0.22
0.09
0.10
0.15
0.23
0.15
X
X
X
X
X
X
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Related manuals

Download PDF

advertisement