Fennoscandia archaeologica XXXIV (2017)
Kristin Ilves STONE FOUNDATION HOUSES OF THE LATE IRON AGE AND EARLY MEDIEVAL ÅLAND AND NEW RADIOCARBON DATES FROM THE SETTLE- MENT OF KULLA
Abstract This article provides a comprehensive overview of the current knowledge of Late Iron Age (AD 550–1050) and early medieval (AD 1050–1300) stone foundation houses and sites on the Åland Islands. More than a quarter of known houses and sites have been subjected to archaeological excavation, most of them over 50 years ago. Many investigations have not been published and research is suffering from speculative and loosely founded generalizations covering long periods of time. In this article, I will compile and compare the excavated and documented stone foundation house sites to each other, examining their differences and shared similarities providing a qualita- tive description. The lack of solid chronology is a major obstacle in drawing conclusions of a more general character. New radiocarbon dates are presented from the seven stone foundation houses at the settlement site in Kulla, reliably anchoring this site and its chronology.
Keywords: The Åland Islands, Kulla, Late Iron Age, settlement archaeology, stone foundation hous- es
Kristin Ilves, Archaeology, University of Helsinki, P.O. Box 59, FI-00014 Helsingin yliopisto, Finland: [email protected].
Received: 28 May 2017; Accepted: 5 April 2018; Revised: 14 May 2018.
BACKGROUND at the site does not match the longer description of the same site.2 It is a dubious task to give an There are over 300 archaeologically registered exact number of archaeologically known Late house foundations on the Åland Islands. Most Iron Age and early medieval house foundations of them are on what today constitutes the main on Åland, given that, for the most part, there is island of Fasta Åland. This is the largest of the no solid chronology. Furthermore, occasionally, nearly 7000 islands, making up the Åland archi- especially on sites with long occupation, archae- pelago situated in the northern part of the Bal- ological investigations have been and are unable tic Sea, halfway between Finland and Sweden. to clarify the precise number of buildings at the According to the data compilation published site. However, emanating from the current reg- by Marita Karlsson thirty years ago (Karlsson ister of ancient monuments on Åland, held by 1987), 175 of these house foundations belong to the Museum of Åland, house foundations today the Late Iron Age (AD 550–1050) and early me- estimated to date to the Late Iron Age and early dieval times (AD 1050–1300). The chronologi- medieval times number over 160 and are distrib- cal estimation is primarily based on their type and uted between 64 sites (see also Table 1; Fig. 1). location close to Late Iron Age burial sites. Yet In addition, there are about 25 areas with no reg- this register also includes a number of undisput- istered house foundations but with layers and/or edly high and late medieval house foundations1 finds indicating dwelling sites that are estimated and, in some cases, the stated number of houses to belong to the same period based on the re-
59
Comment
Publication Kivikoski 1946 Kivikoski Roeck-Hansen 1991 Roeck-Hansen 14C Finds Finds and 14C Primary dating Type and location Type and location Type and location Type and location Type Type and location, Type - - - - + + + + + Re port M. Dreijer A. Vinberg E. Kivikoski H. Antonson B. Cederhvarf B. Cederhvarf B. Cederhvarf Archaeologist B. Roeck-Hansen
1914 1948 1905 1986 1986 1986 Year of Year 1942–5 1904–5, 1904–5, 1906, 1909 1906, excavation 7 3 1 1 7 partial partial partial houses excavation excavation excavation Excavated Excavated 2 1 1 7 5 1 5 1 2 1 1 1 1 2 3 1 1 3 1 1 10 houses Registered Registered Geta Eckerö Eckerö Eckerö Jomala Jomala Jomala Jomala Jomala Jomala Jomala Parish Lemland Finström Finström Finström Finström Finström Finström Hammarland Hammarland Hammarland Torp Torp Torp Torp Kulla Ingby Tjudö Gölby Storby Storby Ytterby Ytterby Grelsby Village Pålsböle Rörstorp Emkarby Östergeta Önningeby Önningeby Önningeby Bartsgårda Fi 2.1 Fi 6.9 Ec 7.3 Fi 9.18 Fi 10.3 Fi 19.9 Fi 13.3 Jo 10.3 Jo 13.2 Jo 32.3 Jo 32.4 Jo 35.2 Jo 35.5 Jo 35.7 Le 16.1 Site ID Site Ec 6.18 Ec 6.23 Ge 16.5 Ha 22.12 Ha 22.22 Ha 22.23 Table 1. The total of registered Late Iron Age and early medieval house sites on Åland with detailed information on the excavated sites and houses. information on the excavated Age and early medieval house sites on Åland with detailed Late Iron 1. The total of registered Table Excavated stone foundation houses marked grey.
60
Comment Turf building Turf Not prehistoric Not Timber structures Timber structure(s)
Publication Karlsson 1984 Karlsson Hackman 1940 Ilves 2015b; 2017b 2015b; Ilves Roeck-Hansen 1991 Roeck-Hansen
14C 14C Finds Location Finds and 14C Finds and 14C Primary dating Type and location Type Type and location, Type - - - + + + + + + + Re port
K. Ilves M. Dreijer M. Dreijer K. Darmark J.E. Tomtlund B. Cederhvarf B. Cederhvarf A. Hackman & C.A. Nordman, C.A. Nordman, Archaeologist B. Roeck-Hansen M. Kampman-Kenttämaa
2011 1905 1966 1986 Year of Year 1913–4 1955–9 1922–4 1982–3 2014, 2016 2014, excavation - 1 1 5 3 2 1 1 vation unclear houses Excavated Excavated partial exca 1 4 1 7 2 1 9 5 3 2 3 2 6 1 1 2 1 1 2 1 many houses Registered Registered Sund Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Saltvik Parish Mariehamn Näs Berg Lavö Sålis Hjortö Bertby Village Ödkarby Kvarnbo Kvarnbo Kvarnbo Rangsby Rangsby Rangsby Rangsby Rangsby Lavsböle Borgboda Lagmansby Mariehamn Främmanby Prästgården Ma 5 Su 1.1 Sa 2.6 Sa 3.8 Sa 6.2 Sa 17.1 Site ID Site Sa 21.8 Sa 22.1 Sa 23.1 Sa 16.2 Sa 14.3 Sa 14.7 Sa 14.9 Sa 11.3 Sa 18.4 Sa 23.2 Sa 23.3 Sa 23.5 Sa 29.2 Sa 35.6 Sa 23.10 Table 1. The total of registered Late Iron Age and early medieval house sites on Åland with detailed information on the excavated sites and houses. information on the excavated Age and early medieval house sites on Åland with detailed Late Iron 1. The total of registered Table Excavated stone foundation houses marked grey.
61
Comment Not prehistoric Not
1992 Publication Dreijer 1955a Kivikoski 1980 Kivikoski Núñez & Lempiäinen Núñez & Lempiäinen
Primary dating Type and location Type and location Type and location Type and location Type and location Type -
+ + + + + + + Re port
M. Núñez M. Dreijer M. Dreijer M. Dreijer K. Lundholm K. Lundholm C.F. Meinander C.F. Archaeologist
1965 1967 1953 1967 1953 Year of Year 1965–6 1985–93 excavation
1 2 1 2 3 1 1 houses Excavated Excavated 1 4 2 1 6 3 9 9 2 1 1 2 2 1 1 3 2 1 2 1 3 2 houses Registered Registered Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Sund Parish Kulla Kulla Finby Finby Finby Träsk Sibby Persby Persby Tranvik Village Gesterby Gesterby Kastelholm Kastelholm Kastelholm Kastelholm Kastelholm Kastelholm Högbolstad Mångstekta Prästgården Brännbolstad Su 7.2 Su 7.3 Su 4.4 Su 6.3 Su 6.4 Su 17.3 Su 17.8 Su 27.2 Site ID Site Su 6.10 Su 18.1 Su 21.9 Su 16.7 Su 10.8 Su 13.6 Su 13.7 Su 12.4 Su 12.7 Su 12.9 Su 28.2 Su 12.16 Su 12.14 Su 12.24 Table 1. The total of registered Late Iron Age and early medieval house sites on Åland with detailed information on the excavated sites and houses. information on the excavated Age and early medieval house sites on Åland with detailed Late Iron 1. The total of registered Table Excavated stone foundation houses marked grey.
62 Fig. 1. The distribution of Late Iron Age and early medieval monuments on Åland. The house sites have been divided into four size categories, each subdivided into groups of investigated and un-in- vestigated structures. The shoreline roughly consistent with that of the beginning of Late Iron Age. Illustration: K. Ilves. sults of the archaeological investigations as well and Table 1). Stone foundation houses were as their landscape situation and vicinity to Late constructed with low dry-stone walls set on the Iron Age burial grounds. Considering that there outside of an inner wood-wall house structure, are almost 11 000 burial mounds registered at often of the wattle-and-daub type, but log- and about 450 Late Iron Age burial grounds in the stave-built walls have also been documented. archipelago, it is clear that the number of coe- The outer stone walls were usually built with a val settlement sites was significantly higher, as single row of bigger boulders, which could vary every burial ground presumably had a farm or a lot in size, and were functioning as a founda- farms attached to it. However, some farms prob- tion for the outer turf or peat walls. Sometimes ably lacked cemetery, while some might also there was a packing of smaller stones between have had several cemeteries attached to them. In the boulders and the inner wood-wall structure the present context, it is also relevant to mention supporting the wall posts. The gap between the that there are no calculations made how much boulders and the inner walls has also been ob- destruction by cultivation has affected the pres- served to have been filled with peat (Hackman ervation of Late Iron Age settlement structures, 1940; Kivikoski 1946; Dreijer 1955a; Núñez though it must have been rather extensive. 1994). Such a building method means that the Almost all of the Late Iron Age and early me- width of the walls might vary greatly and there dieval houses registered on Åland are so-called can be a considerable difference between the in- stone foundation houses. Just three sites out of ner and outer dimensions of the house. During 64 have houses of a different type (cf. below the Late Iron Age houses built in such a man-
63 ner is a uniquely Ålandic phenomenon. Fur- the simple data compilation and a section in thermore, the principal method of construction the short descriptive article published by Karls- has been shown to remain largely unchanged son (1987; also 1997), there are only single and until the late medieval times (Cederhvarf 1910; largely site-specific publications on this catego- Dreijer 1955a: 29; Kivikoski 1962: 19).3 This ry mostly originating from the middle of the last clearly renders typology-based house dating a century. Based on these studies, Ålandic stone rather unreliable exercise. The closest parallels foundation houses have for a long time been just to the Ålandic stone foundation houses are to recognized without any further scrutiny. In some be found on the islands of Gotland and Öland, cases, generalizations are made that do not bear in Sweden (known as kämpagravar or jättegra- much relation to the main body of evidence and var), and in some parts of Norway, but these are suffer from ambiguities arising from a non-syn- dated to the earlier period of AD 200–700 (cf. thesizing view. As a part of my previous research Carlsson 1979; Fallgren 2006; Svedjemo 2014). (Ilves 2017b), I have addressed and refuted the In addition to the stone foundation houses, in re- statement that houses around 20 metres in length cent years, there are also timber structure houses are typical for the Late Iron Age Åland. In this from the Late Iron Age registered on Åland. In article, I will compare the excavated stone foun- Hindersböle (Ma 5), within the city limits of dation houses and sites to each other, examining Mariehamn, discovered and rescue-excavated in their differences and shared similarities as well 2011 (Darmark & Ilves 2016), and in Kvarnbo as discussing the chronology problems involved (Sa 14.9), Saltvik parish, at the site discovered in drawing conclusions of a more general char- in 2012 and excavated in 2014 (Ilves 2015a) and acter. New radiocarbon dates are presented in 2016 (Ilves 2017a). There is one more house the examination of internal chronology of the registered and archaeologically excavated in site Fi 10.3 in Kulla, Finström parish, where all 1982–3 that is not of the stone foundation type, of the registered house foundations have been in Borgboda (Sa 3.8), Saltvik. The walls of this excavated. house were entirely of turf. Unfortunately, this building remains undated (cf. Karlsson 1984: AN OVERVIEW OF THE STONE FOUNDA- 66–7). This notwithstanding, it is still commonly TION HOUSE SITES reckoned among the Late Iron Age and early me- dieval houses because of its location close to the Introduction Late Iron Age burial ground. Being a relatively visible type of ancient The 61 Late Iron Age and early medieval settle- monument, the prehistoric stone foundation ment sites with visible remains of stone foun- houses on Åland were excavated already by the dation houses on Åland are not evenly distrib- archaeological pioneers of the beginning of the uted (see also Table 1; Fig. 1). Two-thirds of the 20th century. In Table 1, together with the total sites (n=40) are in the central and eastern part of registered Late Iron Age and early medieval of Åland, spread between the parishes of Saltvik house foundation sites on Åland, a detailed over- (n=17) and Sund (n=23). The remaining third view is given of the excavated sites and houses. (n=21) has an unbalanced distribution, with most Most of the Ålandic stone foundation houses of the sites (n=13) spread more or less evenly that have been subjected to archaeological in- between the parishes of Finström and Jomala. vestigation were excavated in the first half and Approximately half, 26 out of 61 sites have in the middle of the last century. During the past just one building registered. Five such houses 30 years, no prehistoric stone foundation houses have been subjected to archaeological investiga- have been excavated on Åland. But with more tions, but three of these too partially for mean- than 40 houses from 17 sites having been exca- ingful consideration. Excavation documentation vated, more than a quarter of the registered stone from just two sites (Sa 35.6 and Su 12.24) is use- foundation houses and sites have been studied. ful for further scrutiny. 15 sites have two stone This is a considerable amount. It is therefore foundation houses registered, one site has been paradoxical that these sites have not yet been studied, but unfortunately, existing site records analytically examined. With the exception of do not enable further analysis (see, however,
64 Vinberg 1986). Seven sites have three buildings lished. But perhaps more importantly, it is also registered; one site (Su 12.7) has been investigat- an attempt to offer a more comprehensive and ed and has available excavation documentation. systematic approach to the study of the Ålandic 13 Late Iron Age and early medieval settlement stone foundation house sites. Combined, this in- sites on Åland have four to ten stone foundation formation provides a qualitative overview of the houses registered. These are known from all of nature of the Late Iron Age and early medieval the four above-mentioned parishes where house settlement sites on Åland. foundations concentrate. Ten of these sites have been archaeologically excavated. Thus, the ar- Pålsböle chaeological research has clearly targeted the largest house sites in the archipelago. It is unfor- The house site Fi 13.3 in Pålsböle, Finström par- tunate that most of the large sites were the first to ish (Fig. 2), has five stone foundation houses be excavated and there is no available documen- registered and estimated to belong to the Late tation in regard to the excavations from three of Iron Age and early medieval period (Dreijer the ten sites studied. However, with available 1948; Helminen 2012). There is also a sixth documentation regarding seven sites with four or stone foundation house and an earth cellar at this more stone foundation houses (Fi 10.3, Fi 13.3, site, but these belong to the modern era. Three Sa 18.4, Sa 14.7, Su 12.4, Su 12.9 and Su 12.14), house foundations – nr. 49, 51 and 52 – have it is clear that the largest house sites dominate been excavated, all of these just partially. The among the sites with available material. All in investigated buildings are rectangular, oriented all, although of greatly varying quality, there is in a north–south direction, with maximum (out- applicable excavation documentation regarding er) measurements of 9 x 6.5 m, 7–10 x 4.5 m ten house sites and 29 stone foundation houses and 5 x 3 m. These three houses are situated in a that are estimated to belong to the Late Iron Age north-west–south-easterly line at some 30–40 m and early medieval period. The data set is heav- distance from each other. Finds were unearthed ily biased towards the sites containing more than from two larger houses (nr. 49 and 51), both in- three houses. terpreted as simple log cabins. Finds are few and From a distributional point of view, half of non-diagnostic, consisting mainly of burned and these ten sites are located on the lands of the vil- unburned bones, burned clay, fragments of pot- lage of Kastelholm, in the parish of Sund. This tery and nails. Also, two similar segmented glass must be considered very favourable for the study beads were discovered, one from each house, of settlement development in this particular mi- as well as small fragments of an iron kettle, a cro-region, while making it harder to assess the piece of flint, a fire steel and a knife. The small- validity for the rest of Åland. Of the remaining est stone foundation house without finds is inter- half, three sites are located in the parish of Salt- preted as an outbuilding. vik and two in Finström. The site is situated directly south from and Despite the large number of registered stone adjacent to a Late Iron Age burial ground with foundation houses and sites that have been ex- 48 registered burial mounds; one unexcavat- cavated, the data set is meagre when it comes to ed house foundation is in-between the burial offering possibilities for answering questions on mounds. There is another burial ground with the nature and dynamics of these sites. The same 30 visible burial mounds about 150 m to the is true for the Ålandic Late Iron Age and early east. Not far, there are also several Bronze Age medieval settlement in general. Most of the sites (1500–500 BC) or Early Iron Age (500 BC–AD with available documentation were excavated 550) burial cairns. The house site of Fi 13.3 was more than 50 years ago, and most of the attention not directly on the shore. It was more than 300 m then was given to individual house plans. I find to the north-east from the shoreline in the begin- it necessary to give a concise summary of these ning of the Late Iron Age. At the end of this pe- ten sites, their position in the landscape and the riod, the shoreline was at a distance of about 500 results of the excavations. Such a presentation metres.4 In regard to the waterways crisscrossing is relevant in a purely empirical sense because Åland, the site had a dead-end location. But as most of the excavations have not been pub- there are traces of the Late Iron Age and early
65 Fig. 2. The house site Fi 13.3 in Pålsböle, Finström parish. Insets are showing the site’s location on the present-day map of Åland and in relation to the Late Iron Age and early medieval shorelines, as well as the archaeological monuments in the vicinity. Illustration: K. Ilves.
66 medieval settlement up to 2 km in all directions Lagmansby from the site, it was a populated area at that time. The medieval church of Finström is just 1.2 km At the house site Sa 18.4 in Lagmansby, Saltvik west–north-west from the site. (Fig. 4), nine stone foundation houses were reg- istered in the beginning of the 20th century. The Kulla houses are about 5–15 m from each other. Five houses have been archaeologically excavated, At the house site Fi 10.3 in Kulla, Finström but only two have been published (Hackman (Fig. 3), all seven stone foundation houses reg- 1940). Both are rectangular and in north-west– istered have been archaeologically studied and south-east direction. The largest house (nr. A/7) published by Ella Kivikoski (1946). As this site measures 24 x 6–7.7 m. It has a partially sanded has been discussed in detail by Kivikoski, but floor, wattle-and-daub walls, several fireplaces also by most authors who have written or com- and internal transverse walls; also, it showed mented on the Ålandic stone foundation houses signs of reparation. A second stone founda- (e.g. Liedgren 1992; 1994; Karlsson 1997; Vu- tion house (nr. B/9) measures 14 x 6 m and has orinen 2009), I will not repeat the house-con- stave-built inner walls. Finds are rather scarce structional description here. I will, however, and similar in both houses, consisting mostly of turn to the specifics of this site and its finds that burned bones, loom weights, whetstones, grind- are dating the site to the Late Iron Age in my ing stones and pottery as well as some glass later discussion. beads. In addition, a comb fragment and some At the beginning of the Late Iron Age, the knives were unearthed from the larger house site was situated close to the rear end shore and a fragment of an arrowhead from the other of a sheltered shallow bay running in a north- one. Clearly diagnostic finds include a bird-of- east–south-west direction. It had a hidden loca- prey brooch and a decorated bronze bracelet tion and in order to reach this site, one had to from the largest house, and a decorative iron venture off the natural course of the waterways pin found in the area between these two houses; intersecting Åland. Because of the flat terrain all these objects date to the Merovingian period in this area, the site lost its immediate access (AD 550–800). Several other diagnostic finds to water quite rapidly. The nearest shore was at from the Late Iron Age have been discovered at about 1 km distance in the end of the period. this house site, including a fragmentary Anglo- Two Late Iron Age burial grounds with 114 Saxon penny of King Athelstan (AD 924–939) burial mounds in total are situated about 100 m (Talvio 2002: 206), different brooches and a to the north-east from the house site. A smaller Thor’s hammer-ring. However, these finds were burial ground with 14 burial mounds is situated made during the investigations that lack pub- 100 m to the south-west from the house founda- lished site documentation. tions. About 500 m to the east, there is a house During the Late Iron Age, in the beginning site with one stone foundation house registered of the period, the site was situated about 300 m adjacent to the Late Iron Age cemetery with 30 north–north-west from the rear end shore of a burial mounds. This small farmstead, that was sheltered shallow bay. There is a Late Iron Age also shore-bound in the beginning of the Late burial ground (Sa 18.3) with ten burial mounds Iron Age, has three Bronze Age–Early Iron Age directly to the east, adjacent to the house founda- burial grounds in the vicinity. In addition, not tions. Two larger Late Iron Age cemeteries with far from that site and 1 km from the house site in total 104 registered burial mounds are situated Fi 10.3 in Kulla, a treasure consisting of 92 between the house site and the shoreline as it was Arabic silver coins struck between AD 829 and in the beginning of the period. Similarly to the Fi 958 has been unearthed. Three Late Iron Age 10.3 site in Kulla, the site in Lagmansby had a cemeteries about 1.5 km to the north-east from hidden location, when waterways are considered the house site of Fi 10.3, marking the entrance – in order to reach the site, one had to venture to the former bay, indicate coeval settlement in off the natural course of the nearby, west–east that direction, but generally speaking, the house directed waterway. The entrance to the bay was site in Kulla had a secluded position. marked with Late Iron Age cemeteries. As a fur-
67 Fig. 3. The house site Fi 10.3 in Kulla, Finström parish. Insets are showing the site’s location on the present-day map of Åland and in relation to the Late Iron Age and early medieval shorelines, as well as the archaeological monuments in the vicinity. Illustration: K. Ilves.
68 Fig. 4. The house site Sa 18.4 in Lagmansby, Saltvik parish. Insets are showing the site’s location on the present-day map of Åland and in relation to the Late Iron Age and early medieval shorelines, as well as the archaeological monuments in the vicinity. Illustration: K. Ilves.
69 ther similarity to the Fi 10.3 site in Kulla, the site urements of 9 x 5.5 m. House nr. 2, measuring in Lagmansby lost its fairly easy access to water 12 x 7 m, is immediately adjacent and parallel quite rapidly. The nearest shore was at about 1 to house nr. 1. About 10 m to the south of these km distance during the end of the period. The lays another pair of houses, nr. 4 and 6. House largest treasure ever found on Åland, consisting nr. 4 was documented to have been built on top of at least 859 Arabic silver coins struck between of older building(s). It measured 8–9 x 5 m in its AD 739 and 874/875 and buried in an Oriental earlier stage but was re-built into a square house jug (Talvio 2002: no. 108) has been documented measuring 5 x 5 m. It is noteworthy that this lat- just 400 m to the south-east from the house site. er phase of house nr. 4 has an, at least partially The house site Sa 18.4 in Lagmansby shares stone-built furnace in the corner of the building.5 many similarities with the house site Fi 10.3 in Adjacent and parallel to house nr. 4, house nr. Kulla. Unlike Fi 10.3, that had a relatively se- 6 is situated. This structure, measuring 9 x 5 m cluded position in the landscape, the house site in in total, might actually consist of two separate Lagmansby however, is in an area of very dense square structures. It is very probable that in this Late Iron Age and early medieval settlement. Its area more constructions might have been de- surroundings are dotted with the remains of buri- stroyed while building the houses that were de- al grounds and settlement sites. Among these, fined during the archaeological investigations. there is a large cemetery (Sa 2.4) with 140 burial The destruction of older structures is clearly the mounds, about 500 m to the east. This cemetery case with the house nr. 5 situated about 10 m to was excavated in its entirety during 1957–60 by the east from houses nr. 4 and 6 – archaeological Kivikoski (1963). Many sites with stone founda- features indicating older buildings were docu- tion houses estimated to belong to the Late Iron mented in the areas both west and south, adja- Age and early medieval period are also in the cent to the stone foundation of the house nr. 5. neighbourhood. Of special interest is the cluster It seems more than justified to assume that at of house foundation sites up to 1 km to the south the site Sa 14.7 there are several settlement phas- and south-west, in Rangsby. To the west from the es on top of each other. This is also supported by house site Sa 18.4, at a distance of 1.7 km, there the thickness of the cultural layers around and is the house site Sa 14.7, in Kvarnbo-Kohagen, in-between investigated house constructions, es- where five out of seven registered stone founda- pecially by house nr. 5, and by the richness of the tion houses have been excavated. archaeological find material unearthed. In addi- tion to an abundance of the material commonly Kvarnbo-Kohagen found at Late Iron Age and early medieval house sites on Åland, i.e. burned and unburned bones, The real number of houses at the site of Sa 14.7 burned clay, loom weights, whetstones, grind- in Kvarnbo-Kohagen, Saltvik (Fig. 5), was cer- ing stones, flint, cereal grains, glass beads and tainly higher than the registered seven. There is pottery, there is a much higher number of dif- information about removal of several rows of ferent metal tools and objects, including knives, stones directly to the west from the area with the needles, scythes, arrowheads, harpoons, drills, investigated foundations due to land reclamation nails and rivets, as well as iron and bronze rings. for agriculture (Dreijer 1955b). Also, at least Furthermore, a fair and varying amount of diag- two of the studied house foundations (nr. 4 and nostic artefacts were discovered, such as several 5) were documented to have been built on top of different kind of brooches, ornate silver finger- older houses (cf. Dreijer 1960; 1963; Dreijer, S. rings, bracelets, faceted carnelian pearls, Ara- 1961). The excavated and defined house struc- bic silver coins and oriental belt rivets, as well tures at the site are built in a north–south (nr. 1, as Thor’s hammer-rings. To complement the 2 and 5) and north-west–south-east direction (nr. artefact-based chronology that has a clear em- 4 and 6). The buildings are mostly rectangular, phasis on the Viking Age (AD 800–1050) and only the largest house (nr. 5), measuring 18 x the beginning of the medieval period (although 7.5 m, has a slightly convex form. Houses nr. with few Late Merovingian exceptions), there 1 and 2 are situated farthest to the north of the are also three radiocarbon dates from house nr. 5 studied structures. House nr. 1 has outer meas- that cover the period of AD 820–1280.6
70 Fig. 5. The house site Sa 14.7 in Kvarnbo-Kohagen, Saltvik parish. Insets are showing the site’s loca- tion on the present-day map of Åland and in relation to the Late Iron Age and early medieval shore- lines, as well as the archaeological monuments in the vicinity. Illustration: K. Ilves.
71 Although not directly on the shore, Kvarnbo- a modern lock, clearly indicate a period much Kohagen site was close to the settlement-con- more recent than the Late Iron Age. It is very temporary shoreline and hence close to one of the surprizing that this house is considered among main waterways intersecting the whole of Åland prehistoric buildings. I will resume this matter in south–north direction throughout the Late later when discussing the chronological prob- Iron Age and in the early medieval period. It has lems related to the study of the Ålandic stone three Late Iron Age burial grounds in its vicinity. foundation houses. Just 50 m to the east, there is a cemetery with 33 registered burial mounds and 100 m to the north, Kastelholm a burial ground with 31 mounds. About 150 m to the north-east, there is burial ground with 17 The five archaeologically excavated house sites registered burial mounds and one stone founda- in Kastelholm, Sund (Fig. 6), are situated on a tion house. Most notably, however, the Kvarn- land mass created by two large waterways that bo-Kohagen house site is just 350 m to the north fork from the large bay of Lumparn and head from the settlement and house site of Sa 14.9 – a through Åland in a northerly direction. potential hall-farm that is argued to have formed The southernmost site in that area, Su 12.4, a central place of intraregional importance (cf. had a sheltered location and was approach- Ilves 2015b; 2017b). In addition to the nearby able from the western waterway. According to and extensive settlement layers, on top of which the principal investigator Matts Dreijer (1954), the medieval church of Saltvik is situated, the there are six house foundations at the site; two largest Late Iron Age cemetery on Åland with have been subjected to archaeological excava- 180 registered burial mounds is neighbouring tion and published (Dreijer 1955a).7 The ex- the hall-farm site. At the hall-farm site, three cavated house nr. I is situated somewhat apart, well defined and a number of undefined timber- about 100 m to the north-east from the other constructed houses have been documented. The archaeological house remains. This rectangular site is dated to cover the whole Late Iron Age building is in north–south direction, measures (Ilves 2015a; 2017a). It is clear that the Kvar- 14 x 9 m and has a central row of internal post nbo region was rather heavily populated during holes. It has been used as a foundation for more the Late Iron Age and with clear signs of social modern structures, but there are also indications stratification. of an older building under (cf. Dreijer 1954). There is a drainage ditch associated with the Ödkarby structure. The excavations unearthed pottery of both medieval and modern period but also a The stone foundation house at the site Sa 35.6 great amount of decorated loom weights of Late in Ödkarby, Saltvik, was discovered among Iron Age character. Unlike this house that has 18 Late Iron Age burial mounds situated at been used for a long period of time, the other about 200 resp. 400 m distance from the Late building excavated at the site – house nr. II – was Iron Age and early medieval shoreline. How- interpreted to have been used for a rather short ever, according to the site report (Dreijer 1966), period of time. This building also has a rectangu- there are actually traces of two more houses at lar stone frame and is situated in a north–south that site. Two small rectangular buildings were direction. Similarly to house I, there is a drain- documented lying on top of burial mounds, age ditch associated with the structure. House II and determined to belong to the modern era has outer measurements of 15 x 8 m, a partially following the account of the local inhabitants. sanded floor and just two pairs of internal post The stone foundation house, that in the current holes, about 6 m apart. This creates a large open register of ancient monuments on Åland stands space in the middle of the house with a central as belonging to the Late Iron Age and early fireplace close to the northernmost set of inter- medieval period, was also documented to have nal post holes. The house incorporated a byre. been built on top of the burial. Furthermore, Finds, which concentrated around the fireplace, the few finds made in that house, such as post- are few – burned bones, loom weights, nails and medieval pottery, porcelain and a fragment of pottery, but also some fragments of melted glass,
72 Fig. 6. The five archaeologically excavated house sites in Kastelholm, Sund parish in relation to the Late Iron Age and early medieval shorelines as well as the archaeological monuments in the vicinity. Inset is showing the sites’ location on the present-day map of Åland. Illustration: K. Ilves.
73 a glass bead, an ornamented comb fragment and is a burial ground with 137 mounds (Su 12.7) a probable weight. The houses were close to the located centrally in this area; this cemetery was shoreline in the beginning of the Late Iron Age excavated during 1962–7 and in its entirety by and at about 400 m distance in the end of the Kivikoski (1980). A smaller burial ground with period. Adjacent to the house foundations, there just seven registered burial mounds is also part are three mounds of fire-cracked rock, which are of this complex. This settlement area was situ- estimated to be coeval with the settlement. Two ated on the shore during the Late Iron Age and burial grounds with 70 registered burial mounds has actually not lost its easy access to water up in total are situated about 50 m to the south and to this day. 100 m to the east from the settlement area. Some Two excavated houses directly to the west 1.7 km to the north-east, as the crow flies, is the of the large burial ground Su 12.7 are both in house site Su 12.9 that was also approachable north–south direction. House nr. 1 is rectan- from the western waterway. gular, 12 x 6.5 m in size and has a partially The site of Su 12.9 (Fig. 7) had immedi- stone-paved floor. Finds are few and consist ate access to water throughout the Late Iron of pottery, burned clay, burned nutshells, loom Age. The closest Late Iron Age cemetery, with weights, two glass beads, a knife, some nails 116 registered burial mounds, is about 200 m and iron slag. House nr. 2 has a slightly convex north of the site. Nine stone foundation houses form and measures 15 x 8 m; there is a smaller have been registered at the site, in a double row building or an annex measuring 4 x 4 m related along the slope toward the shore. The houses are to this house. There were more finds made from about 5–10 m from each other. Two buildings – house nr. 2, however, even these of a rather low identified as nr. 6 and 7 – have been excavated chronological value. There were many frag- (Meinander 1954; Lundholm 1966). House nr. 6, ments of loom weights documented outside and situated in north-east–south-west direction, has a along the western wall of the house. Within the slightly convex form and outer measurements of house, burned bones, burned clay, whetstones, 17 x 8.5–10 m. There is a drainage ditch associ- knives, nails, iron slag and pottery as well as ated with the structure. It seems likely that the two glass beads were unearthed. Further to house saw several phases of construction. It had the west from these houses, there are remains a partially stone-paved floor and is interpreted to of a historic farmstead, but more prehistoric have incorporated a byre. Also, house nr. 7 has a buildings are also hypothesized at the site (cf. slightly convex form measuring 10–15 x 7.5 m. It Kivikoski 1980: 57). Only some 100 m to the is situated in west–east direction at the southern south from the excavated houses at Su 12.7, end of house nr. 6 and was somewhat destroyed there is the house site Su 12.24, which should by the latter. However, house nr. 7 is itself docu- be seen as part of the same complex (also mented on top of an older building. Finds are few Núñez & Lempiäinen 1992). At the site Su and similar in both houses, consisting mostly of 12.24, in addition to extensive settlement layers pottery, burned clay, burned nutshells and burned and an excavated stone foundation house, there bones. Fragments of whetstones and some iron are strong indications of an additional set of objects were also found. Two glass beads were house foundations (cf. Núñez 1994). The mea- unearthed from house nr. 7. surements of the investigated house are about About 500 m to the east of Su 12.9, by the east- 18 x 9 m. The building is in north–south direc- ern waterway, there is one of the spatially most tion, has a partially stone-paved floor, a central extensive Late Iron Age settlement complexes hearth and internal roof-bearing post-holes in documented on the Åland Islands (see also Fig. pairs. The inner wood-walls of this house were 6). Two sites with stone foundation houses, Su interpreted to have been built using a split-log 12.7 with three registered and excavated houses technique. There is a drainage ditch associated and Su 12.24 with one house foundation regis- with the structure. The investigations revealed tered and studied, are part of this complex. In an abundance of the usual material found at the addition to the stone foundation houses, exten- Late Iron Age and early medieval house sites sive settlement layers have been documented on Åland. There were lots of pottery and loom within an area covering about 35 hectares. There weights discovered as well as iron slag and
74 Fig. 7. The house site Su 12.9 in Kastelholm, Sund parish. Insets are showing the site’s location on the present-day map of Åland and in relation to the Late Iron Age and early medieval shorelines, as well as the archaeological monuments in the vicinity. Illustration: K. Ilves.
75 burnt clay. In lesser numbers, burned bones, CHRONOLOGY flint, fragments of whetstones and grinding stones as well as burned nutshells and cereal The number of reliably dated stone foundation grains were recovered. Also, a number of iron house sites and individual houses from Åland is artefacts were unearthed, knives, nails and riv- small. At present, Fi 10.3 in Kulla and Sa 18.4 ets, even an arrowhead and a spearhead. The in Lagmansby can with certainty be identified as latter is considered to point towards the second Late Iron Age dwelling sites and the houses at half of the Viking Age. the site of Sa 14.7 in Kvarnbo-Kohagen as be- About 1 km further north from the described longing to the Late Iron Age and the beginning complex, following the waterway, the house site of the medieval period. Despite claims of Late Su 12.14 was documented on a very prominent Iron Age and early medieval houses at the other position: on a narrow isthmus running north to locations studied, these are the only locations south into the waterway (see also Fig. 6). This where absolute dates and/or the reliable bulk site of Su 12.14 has four registered and three of diagnostic artefacts recovered clearly sup- studied stone foundation houses as well as 37 port a Late Iron Age origin. However, four of adjacent burial mounds. It had a position excel- the five studied sites in Kastelholm (Su 12.4, Su lently suited for control over the waterway dur- 12.7, Su 12.24 and Su 12.9) are also probably of ing the entire period of Late Iron Age and me- Late Iron Age origin based on the great amount dieval times. House foundations at this site are of loom weights and single diagnostic objects in the northern (houses II and III) as well as in discovered. The house site Fi 13.3 in Pålsböle the southern part (house I) of the burial ground. might have roots in this period. However, it Also, settlement layers were discovered while seems more than probable that the registered and investigating burial mounds – close to mound excavated houses at the sites Su 12.14 in Kastel- nr. 7, which is somewhat detached from oth- holm and, especially, Sa 35.6 in Ödkarby are of ers, there were layers that yielded an abundance medieval and post-medieval origin. of rich finds dating to the medieval times. The The stone foundation house sites on Åland mound itself is not a burial but a probable fur- have been interpreted as farmsteads that were nace of a house. The investigations of this area in use for a long time. Furthermore, a prehis- were, however, discontinued when the discov- toric origin is readily seen. It is easy to see how ered find-rich layers were estimated to belong this understanding arose. The stone foundation to the medieval times (Dreijer 1964). Some house sites investigated show similar character- metres to the east, closer to the burial ground, istics in terms of house plans. The bulk of mate- traces of another house (house IV) were docu- rial is made up of chronologically non-diagnos- mented. This building turned out to be on top tic finds assigned a relative date, and since many of a cremation burial and is interpreted as a po- of the Late Iron Age forms differ little from their tential smithy. All of the few finds made are of medieval counterparts, the date range is wide. medieval and modern character (Dreijer 1969). Few diagnostic artefacts unearthed indicate that Houses II and III were the only ones visible at such sites were and could have been used both the site before the investigations, but allowed in the beginning of the Late Iron Age and in the hardly any constructional interpretations when end of the period as well as during the medieval excavated. House II is estimated to be in a period. Moreover, house sites are all located north–north-west – south–south-east direction close to Late Iron Age burial grounds – inves- and to measure about 6 x 5.5 m; house III has tigations conducted at the cemeteries show use an irregular form and its expanse remains un- throughout the period. Despite the fact that a di- clear. Finds are few and non-diagnostic: burned agnostic artefact type, especially when scant in and unburned bones, burned clay, some nails numbers, merely identifies presence at a certain and flint. Thus, the medieval and modern set- time, generalizations are made that are covering tlement at the site of Su 12.14 is supported by long periods of time. Also, I would not exclude the empirical evidence, but it remains an open the possibility that single artefacts may well rep- question whether there is a prehistoric phase at resent heirloom items. In this connection, it is the site as well. of interest to note that a wheel/leaf-headed pin
76 dating to the first part of the Roman Iron Age houses 3 and 6 are discussed to be unsuitable for (AD 1–400) has been found from the house site a permanent housing. No definite interpretation in Lagmansby during investigations that lack is made in regard to houses nr. 7 and 8, though site documentation (Forsberg & Taffinder 1984). the first is more inclined to be seen as a dwelling Such pins originate from the Upper Volga-Oka- house, the oldest one at the site (cf. Kivikoski Dnepr region and single finds have been also 1946: 65). Through a few diagnostic artefacts made in Estonia, Lithuania, Belarus and Sweden unearthed, the site is dated from the 7th century (Lang 2007: 137). It is unreasonable to argue for to the beginning of 11th century. The principal an Early Iron Age settlement phase in Lagman- investigator Ella Kivikoski never argued for all sby based on a single artefact; instead, I find it of the seven stone foundation houses studied to very conceivable that this wheel/leaf-headed pin be coeval and/or older houses necessarily left to might rather represent an heirloom item. ruin when new houses were built. She is rather When were sites established? When were suggesting a chain of development in the use of they abandoned? What was the length of the oc- the houses. New houses take over the functions cupation and the reasons for that? How many from the older ones that, in turn, change in func- houses were contemporaneous at a house site? tion. Also, several of the houses have been used How many house sites were contemporaneous? in combination. The site in Kulla is interpreted Why would Åland have its own house build- as a large farm belonging to a family of a higher ing tradition in comparison to neighbouring social rank. regions? In looking at the available material, The find material from the site, though not either separately or as a whole, it is clear that rich in numbers, aligns well with these inter- as yet it does not seem possible to answer the pretations. There is the expected material re- range of questions related to the development of lated to settlement activities, such as burned the Late Iron Age and early medieval settlement bones, loom weights, whetstones, flint, cereal on Åland. In order to start to unravel this situa- grains, beads, pottery and metal tools. These tion, I chose to examine the house site Fi 10.3 are more numerous in the constructionally in Kulla. This site was chosen since all of the similar houses nr. 1 and 2, but whereas house seven registered stone foundation houses at the nr. 2 had been burned down (also, houses nr. 3 site have been excavated, very well-documented and 6 have evidently suffered a fire) house nr. and published – furthermore, this site has for a 1 showed no signs of fire. At the same time, long time been the ‘flagship’ of the Ålandic Late all the foundations studied are characterized Iron Age settlement. by common settlement finds. Notably though, The seven houses at the site (see also Fig. 3) loom weights are missing in houses nr. 3, 6 and vary greatly in size with the largest (house nr. 8; house nr. 6 further distinguishes itself from 1) measuring 22 x 10 m and the smallest (house the others through the presence of iron slag. nr. 3) 6.5 x 5 m. The houses do not have a uni- From a chronological point of view, an equal- form direction in regard to the cardinal points: armed brooch dated to the 11th century, as well house nr. 1 is the only one in north–south direc- as a white melon bead found from house nr. 2 tion, houses 2, 3 and 6 are in east–west direc- indicates a later use-phase than, for example, tion and houses 5, 7 and 8 are in north-west– in the case of house nr. 7 where a horse-shaped south-east direction. There is no structural or brooch dating to the 7th century was found. constructional overlap; the houses are about Faceted carnelian beads from house nr. 8 point 10–15 m from each other, occupying an area of towards the Viking Age, and a shield pendant about 0.3 hectares (the area between the houses dating to the 10th–11th century as well as a large has not been investigated). The largest house is mosaic bead were discovered from house nr. 5. the only one that has a slightly convex outline, Kivikoski (1946: 52–3) concludes that houses 1 while the others are rectangular. These seven and 7 are the earliest at the site, with house nr. 7 houses are argued to represent at least three being the oldest, and houses nr. 2 and 5 belong different types of house construction. From a to the final stages of the settlement; house nr. 8 functional point of view, houses nr. 1, 2 and 5 is dated to the Viking Age. A few medieval and are seen as year-round dwelling houses, while post-medieval finds – two shreds of glazed red
77 Lab-index BP ± calAD (2σ) Context Material unearthed, radiocarbon date falls Ua-53500 1326 27 650–770 House nr. 7 Burned bone into the end of the Merovingian period and the first half of the Ua-53497 1251 30 670–870 House nr. 1 Cereal Viking Age – taken together, Ua-53498 1225 29 690–890 House nr. 5 Burned bone a longer period of usage must Ua-53499 1179 27 760–950 House nr. 6 Burned bone be hypothesized. The smallest houses at the site, nr. 3 and 6, be- Ua-53495 1128 27 770–990 House nr. 3 Cereal ing of the outbuilding type, seem Ua-53496 1081 27 890–1020 House nr. 8 Cereal to have been used at the same Ua-53494 1080 27 890–1020 House nr. 2 Cereal time, from the very end of the Merovingian period, but mostly during the Viking Age. The con- Table 2. Radiocarbon dating results from Kulla. All datings ducted radiocarbon analysis is have been calibrated using OxCal 3.10. now firmly anchoring the site of Kulla in absolute chronology, al- earthenware from two different vessels from lowing site-specific patterns to be aggregated houses nr. 2 and 6, and an iron padlock from and less hypothetical inferences to be drawn. house nr. 5 – are considered to point towards But for a fuller understanding of the site, the some post-occupational disturbance. area between and surrounding the stone foun- Based on the house-plans, some aspects of dation houses should be studied as well. Fur- the internal chronology of this site have been ther, targeted dating (of house nr. 7) is needed briefly, but on a holistic level discussed by to shed light on the establishment of the site (cf. Dreijer (1955a: 48ff). While he agrees with the Ilves 2017b). house nr. 7 being probably one of the oldest at To use radiocarbon dates is a standard archae- the site, house nr. 1 is argued to belong to the 11th ological practice. However, there are less than century. However, it is clear that without more hundred radiocarbon ages available for the en- solid chronology only tentative conclusions can tire Iron Age of the Åland Islands (500 BC–AD be drawn. In order to bring clarity to and have 1050), many originating from uncertain con- a more reliable basis for the discussion on the texts. Although there is now a much more solid chronology of the Fi 10.3 house site, I decided chronology for the house site Fi 10.3 in Kulla, in to send seven samples, one from each house, to order to move from a site-specific approach to radiocarbon dating (Ångström Laboratory, Up- the general trends characterizing Ålandic settle- psala University, Sweden). The results of radio- ment during Late Iron Age and early medieval carbon dating and calibration are summarized in period, a solid chronology for a number of set- Table 2. tlement sites is essential. Only upon an unam- The results of radiocarbon dating align with biguous and reliable chronology is there a pos- the interpretations made by Kivikoski. The site sibility to interpret events and cultural processes has been in use for several hundreds of years, on a wider scale. from the late 7th until the beginning of the 11th century. Not all houses were in use at the same CONCLUSIONS time, but a chain of development where new houses take over the functions from the older Today’s main island of Fasta Åland, which in fact ones, as well as the idea of houses also being consists of a few islands connected by bridges, in use at the same time, is very likely. House was a broken archipelago during the prehistoric nr. 7 can definitely be argued to be the oldest times and the nature of the settlement on Åland at the site and house nr. 1 was probably taken has always been more or less maritime. All of into use shortly after the first one. Houses nr. the known dwelling sites estimated to belong to 2 and 8 belong apparently to the final stage of the Late Iron Age and early medieval period lay the settlement, both dating to the second half of within a distance of 5 km from the coast today, the Viking Age. In case of house nr. 5, where but due to shore displacement resulting in raised a 10th–11th century shield-shaped pendant was shorelines, these were originally much closer to
78 site-contemporary shores. As is demonstrated in longer occupation. Yet, there are studies, based this study, it is obvious that there were no certain mainly on the Scandinavian houses from the end types of locations uniformly preferred in regard of the Early Iron Age, that show a rather long to shoreline or waterways. Dwelling sites are life span of a house – most commonly around located immediately by the shore as well as at 100 to 150 years, 50 to 75 years before the need some distance from it. The sites also vary when of reparation (Göthberg 2000: 108–9 with refer- it comes to access to transportation routes, with ences). The life span for stone foundation houses some being adjacent to waterways and others sit- has been suggested to be even longer (Svedjemo uated in more secluded locations. There is little 2014: 64). In that sense, the fact that the house evidence that sites shifted location in response to site in Kulla is so ‘clean’ in regard to house plans changing shorelines. It is, however, justified to is not so puzzling as such. However, the site is suggest that the location close to the shore was also rather scant in finds. The latter aspect has important for the sites established in the very be- been explained by the controlled abandonment ginning of the Late Iron Age. During the initial of the site, while in case of Kvarnbo-Kohagen a establishment phase, the topographical situation, violent end to the settlement has been proposed primarily immediate access to water, was impor- (cf. Karlsson 1997: 91–2; Tomtlund 1999: 28). tant in deciding where to create new settlements. Although both these sites have suffered a fire, I Subsequently, however, these settlements and am not excluding the possibility of, for example, the cultural landscape they formed, seem to have an attack to Kvarnbo-Kohagen settlement lead- become a more important localizing factor than ing to many things left behind. But I do think the receding shoreline. that the ‘messy’ nature of the site is also indicat- This overview of the stone foundation houses ing a much more intense settlement of a differ- and sites that have been archaeologically exca- ent nature than at the single and rather secluded vated has shown that the sites, grouped together farm in Kulla, which seemed to have belonged by the existence of low dry-stone walls marking to one family for generations. Obviously, this house foundations, also differ in many aspects. remains just a hypothesis as long as Kvarnbo- The presence, amount and character of the find Kohagen site’s internal chronology is not clari- material bear obvious witness to the diversity, fied; the same is relevant for other house sites. both functional and social. There are likewise Although a large number of Ålandic Late variations in house plans and construction meth- Iron Age and early medieval house sites has ods, as well as in house sites’ spatial relation to been subjected to archaeological excavation, the burial grounds and to other remains in the vicin- archaeological record has a very uneven char- ity that are estimated to indicate coeval dwell- acter. There are some well-studied and -pub- ing or signify region’s previous importance (i.e. lished houses and sites, but also many sketchily Bronze Age burial cairns). Yet what I find most excavated ones with poor or absent documen- interesting is the fact that there are stone founda- tation. This must be considered as one of the tion houses showing hardly any major signs of main reasons why the Ålandic stone foundation repairs, rebuilding or overbuilding at sites with house sites have not been analytically examined. long continuity, such as in Kulla. At the same Another fundamental, but solvable problem is time, there are really ‘messy’ sites, where many the lack of chronology, which is seriously ham- houses are built on top of older structures and pering a meaningful discussion on the transfor- overlapped by younger ones, such as in Kvarn- mations occurring on Åland (cf. Ilves 2017b). bo-Kohagen that is also one of the richest sites. Radiocarbon dating of samples from what can Still, overlapping stone foundation houses are undoubtedly be described as an abundance of also evident from sites that have yielded very already excavated contexts is a suitable, feasible few finds, as is the case with two excavated and not-so-expensive first step to take. There is houses at Su 12.9 in Kastelholm where a num- a huge potential in such study. Anchoring sites ber of construction phases were documented in a finer chronology and building up a base of in a limited area just as in Kvarnbo-Kohagen. knowledge would open up the possibility for From an archaeological point of view, sites with finding answers to questions related to- under complex stratigraphy are often associated with standing the nature and dynamics of Ålandic Late
79 Iron Age and early medieval settlement. That, in 4 The results of shoreline calculation for the turn, would enable to break the prolonged hiatus whole of Fasta Åland in the beginning of the in the settlement archaeological research and de- Late Iron Age are spanning from 8.5±0.4 m asl. marginalize Åland within the field of Late Iron in the south-east to 9.8±0.4 m asl. in the north- Age and early medieval studies. west; in the end of the period, the values are spanning from 5.5±0.3 m asl. to 6.4±0.3 m asl., ACKNOWLEDGEMENTS respectively (Ekman 2017).
Much appreciated financial support for the prep- 5 In the present connection, I would like to aration of this article was provided by Berit Wal- draw attention to Björn Cederhvarf’s investiga- lenbergs Stiftelse and Societas Archaeologica tions at the house site of Jo 22.4, close to the Upsaliensis Forskningsfond. I would also like Late Iron Age burial ground, but also to the to thank the two reviewers for their insightful medieval church in Prästgården, Jomala (Ce- remarks. derhvarf 1910). One stone foundation house of the 17 houses was excavated and firmly dated NOTES to the end of the 14th century–the beginning of the 15th century. The excavated and dated house 1 Such as 17 house foundations at the site of was considered to be one of the two oldest stone Jo 22.4 in Prästgården, Jomala (cf. Cederhvarf foundation houses at the site. These two houses 1910), and 6 at the site of Sa 3.5 in Borgboda, had their stone-built furnace positioned centrally Saltvik (cf. Tomtlund 1986). For the discussion in the building, while others have it in the corner in this article, it is noteworthy that both these of the house. The central position of the furnace sites are immediately adjacent to the Late Iron is discussed to reflect a more archaic building Age cemeteries. tradition, while its position in the corner is seen as a more modern characteristic. Even in the case 2 Cf. Jo 35.7 in Önningeby, Jomala, or Sa 11.3 of another excavated house foundation, in Borg- in Hjortö, Saltvik. boda, dated to originate, at the earliest, from the medieval times, stone-built furnace was docu- 3 The continuity in construction of stone foun- mented to have been positioned centrally in the dation houses is one of the significant arguments building (cf. Dreijer 1955a: 51–2). against the early medieval settlement disconti- nuity theory that has been advocated based on 6 All three 14C dates from house nr. 5 in the history and etymology of place names almost Kvarnbo-Kohagen were obtained from car- throughout the history of scholarly research on bonized cereals and dated in 2006. The results Ålandic settlement (for the most recent insight, cover the transition period from the prehistoric see Sjöstrand 2014). Today, it is generally not to medieval times (data from the Museum of questioned that the majority of Ålandic place Åland, ÅM): 1115±35 BP (Ua-33352; from in- names are not of prehistoric origin, but younger. vestigation unit A5, find ÅM 328:43), 885±35 This has been explained – especially, since Lars BP (Ua-33354; from investigation unit D9, find Hellberg’s classical study of the place names on ÅM 328:258) and 830±35 BP (Ua-33353; from Åland (Hellberg 1980) – that Åland, having had investigation unit C3, find ÅM 328:157). a rather dense population during the Late Iron Age, was depopulated in the early 11th century 7 According to the current register of ancient only to have witnessed a rapid re-colonization monuments on Åland, there are only two stone about 150 years later. This explanation, howev- foundation houses registered at the site. In this er, is in contrast with the archaeological as well article, I have chosen to follow the account of as paleobotanical material. Dreijer.
80 REFERENCES Ilves, K. 2015a. De arkeologiska under sökningarna vid Kvarnbohallen år 2014. Re- Unpublished sources search report, Museum of Åland, Mariehamn. Ilves, K. 2017a. De arkeologiska under Darmark, K. & Ilves, K. 2016. Nyupptäckt bo- sökningarna vid Kvarnbohallen år 2016. plats från yngre järnålder. Arkeologisk under- Research report, Museum of Åland, Marie- sökning av boplatslager vid Ma 5, Hinders- hamn. böle, Mariehamn. Research report, Museum Lundholm, K. 1966. Rapport över arkeolo- of Åland, Mariehamn. gisk undersökning av husgrund 7, Storängs Dreijer, M. 1948. Berättelse över arkeologiska backen, Kastellholm, Sund sn, Åland, under grävningar å gravfältet i Finströms gårds tiden 1–17/7 1965, 14/6–8/7 1966. Research kyrkbol år 1948. Research report, Museum of report, Museum of Åland, Mariehamn. Åland, Mariehamn. Meinander, C.F. 1954. Sund, Kastelholm, Dreijer, M. 1954. Berättelse över utgrävningen Storängsbacken. C.F.Meinanders utgrävning av hustomtning I invid Tjudnäs gård i Kastel- av Husgrund 6. Research report, Museum of holms by i Sunds socken år 1953. Research Åland, Mariehamn. report, Museum of Åland, Mariehamn. Núñez, M. 1994. Archaeological investigations Dreijer, M. 1955b. Berättelse över utgrävningen of a Late Iron Age house in the Kastelholm av hustomtning 1 i Kohagen i Kvarnbo by Kungsgård 1985–93. Research report manu- av Saltviks socken år 1955. Research report, script, Museum of Åland, Mariehamn. Museum of Åland, Mariehamn. Tomtlund, J.-E. 1986. Rapport över 1984 års Dreijer, M. 1960. Berättelse över en arkeologisk undersökningar i Borgboda, Saltviks sn. Sa undersökning av hustomtningarna 4 och 6 i 3.5 Gammelgårdshägnaden. Research report, Kohagen i Kvarnbo år 1959. Research report, Museum of Åland, Mariehamn. Museum of Åland, Mariehamn. Vinberg, A. 1986. Mangelbo – en husgrunds Dreijer, M. 1963. Berättelse över en arkeolo- lokal i Grelsby, Finströms sn på Åland. Del I: gisk utgrävning på Kohagabacken i Kvarnbo kartering och preliminär tolkning (B-uppsats, 1962. Research report, Museum of Åland, Stockholms universitet). Manuscript, Muse- Mariehamn. um of Åland, Mariehamn. Dreijer, M. 1964. Berättelse över den arkeolo- giska utgrävningen på Stenhagens gravfält i Literature Kastelholm, 1963. Research report, Museum of Åland, Mariehamn. Carlsson, D. 1979. Kulturlandskapets utveck- Dreijer, M. 1966. Rapport over utgrävnin- ling på Gotland: en studie av jordbruks- och gar på gravfältet mellan Ollas och Mickels bebyggelseförändringar under järnåldern. ekonomibyggnader i Ödkarby by av Saltviks Stockholm: Stockholms universitet. socken 1966. Research report, Museum of Cederhvarf, B. 1910. Fynd i en medeltida hus- Åland, Mariehamn. tomtning på Jomala kyrkobacke. Finskt Mu- Dreijer, M. 1969. Rapport över en arkeologisk seum 17 (1910): 91–103. utgrävning på forngravfältet i Stenhagen, Dreijer, M. 1955a. Hustomtningarna i Tjudnäs. Kastelholm, Sund, 1967. Research report, Åländsk Odling 1955: 28–69. Museum of Åland, Mariehamn. Ekman, M. 2017. Calculation of Historical Dreijer, S. 1961. Berättelse över en arkeologisk Shore Levels back to 500 A.D. in the Baltic undersökning av ett 64 m2 stort område strax Sea Area due to Postglacial Rebound. Small väster om hustomtning 5 på Kohagabacken i Publications in Historical Geophysics 31. Kvarnbo, Saltvik, 1960. Research report, Mu- Fallgren, J.-H. 2006. Kontinuitet och föränd seum of Åland, Mariehamn. ring: bebyggelse och samhälle på Öland Helminen, M. 2012. Finström, Pålsböle, 3:16. 200–1300 e Kr. Aun 35. Arkeologisk inventering och provunder Forsberg, L. & Taffinder, J. 1984. En hjulhuvud- sökning. Research report, Museum of Åland, nål från Åland. Åländsk Odling 1984: 39–48. Mariehamn.
81 Göthberg, H. 2000. Uppland från slutet av yngre Liedgren, L. 1992. Hus och gård i Hälsingland. bronsålder till tidig medeltid. Occasional Pa- En studie av agrar bebyggelse och bebyggel- pers in Archaeology 25. seutveckling i norra Hälsningland Kr.f. – Hackman, A. 1940. Två åländska husgrunder 600 e.Kr. Studia archaeologica Universitatis från yngre järnåldern. Finskt Museum 47 Umensis 2. (1940): 67–84. Liedgren, L. 1994. Kalaschabrännan och andra Hellberg, L. 1980. Ortnamnen och den svenska sedentära bebyggelselämningar från järnålder bosättningen på Åland. Ortnamn och sam- i Finland. In K. Gullberg (ed.) Järnålder i hälle 2. Mittnorden: ett symposium kring nya arkeolo Ilves, K. 2015b. Fynd efter en yngre järnålders giska och ekologiska forskningsrön: 33–42. aristokrati i Kvarnbo, Åland. Finskt Museum Vasa: Scriptum. 120–122 (2013–2015): 65–80. Núñez, M. & Lempiäinen, T. 1992. A Late Iron Ilves, K. 2017b. The Kvarnbo hall: reconsidering Age farming complex from Kastelholms the importance of the Late Iron Age Åland Is- kungsgård, Sund, Åland Islands. PACT 36: lands. Journal of Island & Coastal Archaeolo- 125–42. gy.
82