Hearths in the coastal areas of northernmost Sweden , from the period AD 800 to 1950

This paper presents a study of the chronological setting of hearths registered in FMIS (digital register containing records of all known ancient monuments in Sweden) in the provinces of Västerbotten and Norrbotten, Northern Sweden. A total of c. 1500 hearths are known in the area, mainly situated north of the river Skellefteälven. Within a study area of 107 x 94 km, 32 hearths were randomly selected for excavation, each site embracing 1-14 hearths. The sites were scanned using a metal detector and nearly all artifacts found were from the period AD 1600-1900. 14C-datings of charcoal and burned bones corroborated that most hearths were used during this period, with a large number dating to the 19th and 20th centuries. Many hearths contained bones from mature reindeer, indicating that the hearths were related to reindeer herding. We suggest that most hearths are related to nomadic Sami reindeer herders using coastal areas for winter pasture, possibly resulting from the breakdown of the “lappskatteland” (taxation lands) system and an increase in reindeer numbers.


Introduction
In the coastal areas of northern Sweden (the counties of Västerbotten and Norrbotten) sedentary agrarian farms established by the 14 th century at the latest.In AD 1543 there were around 2300 farms (taxation units) in these areas (Nordlander, 1905).The economy included hunting and extensive lake fishing in the interior (träskfisken) and in rivers close to and far away from the farms and at sea (Westin, 1953;Fahlgren, 1963Fahlgren, , 1970:156pp;:156pp;Nordberg, 1965;Bergman & Ramqvist, 2016).These coastal societies were divided into parishes which generally stretched 100-120 km (and as far as 260 km in the Torne river valley) towards the Lappland border (lappmarksgränsen) which was poorly defined at this time.
West of this border the Sami reindeer herders had their land, known as Lappland (mentioned as early as AD 1340 in a letter by King Magnus Eriksson).The Lappland border was finally defined as late as AD 1766 (André, 1998:59).The agrarian colonization of Lappland, including establishment of sedentary farms, began during the second part of the 17 th century in the south (Norstedt, 2011;Göthe, 1929) and somewhat later further north (Bylund 1956).
Numerous remains related to reindeer herding, mostly hearths, occur all over Lappland.The hearths consist of a stone lining and on the inside, in many cases, a packing of stones.They were intended to contain a fire inside a tent or a permanent building, and were constructed in such a way that the fire did not spread on the ground.This source of heat was used to warm the living space, to cook food and to give light (Liedgren & Östlund, 2011;Östlund et al., 2013).
Today, our knowledge of stone-lined hearths west of the Lappland border is relatively extensive (Edbom et al., 2001;Hedman, 2003;Liedgren & Hedman, 2005;Liedgren et al., 2007;Bergman, et al., 2013).They date back to at least the late Iron Age (AD 700-800) onwards, in both the alpine and boreal areas of the region (Storli, 1994;Mulk, 1994;Hedman, 2003;Liedgren et al., 2007;Hedman & Olsen, 2009;Bergman et al., 2013;Hedman et al., 2015).Single hearths or hearths in rows from the late Iron Age are found in locations suggesting that people utilized the landscape during different seasons in accordance with the needs of the reindeer -an animal of central importance (Aronsson, 1991;Bergman, 1991;Hedman, 2003).The time when the reindeer was domesticated in Sweden has been intensely debated (for a general review on the debate see Sommerseth, 2011).This is not an entirely straightforward issue, as Ingold (1980) and Lundmark (1998) have recognized a clear difference between "rennomadism", i.e. living on the surplus provided by tame reindeer herds, and merely owing some tame reindeer.
During the late 19 th and the beginning of the 20 th century both forest Sami and mountain Sami used the Bothnian coast in the wintertime for grazing their animals on the heathlands.The mountain Sami groups travelled as far west as the Norwegian border and beyond in the spring, and all the way down to the Bothnian Bay in late autumn in a cyclical man-ner (Manker, 1953).Today, the coastal area is still used by reindeer herders during the winter.However, knowledge of the utilization of the coastal areas for reindeer herding before the late 19 th century is poor, and only a few hearths have been excavated so far.Thus the objective of the present study was to assess whether the historical records outlined above can be supported by the archaeological record further back in time.
For this purpose we have focused on hearths found east of the Lappland border and adressed the following specific questions.
1) To what extent were coastal hearths connected to reindeer herding?2) Are the registered hearths only a manifestation of late usage (post-18 th century) of the coast for reindeer grazing? 3) Are there any signs that hearths were used during the snow-free period?In addition to addressing these questions we also discuss correlations between the spatial distribution of hearths and geographical place names associated with Sami groups.

The archaeological setting
The National Heritage Board (NHB) in Sweden started landscape inventories for ancient monuments in 1938.The surveys were connected to the production of economic maps (Hyenstrand, 1978).In the province of Västerbotten and Norrbotten surveys were started in the 1950s (Jensen, 1997).During this period the focus was mainly on stone settings, cairns, ruins related to seal hunting (tomtningar) and labyrinths (stone linings forming a spiral towards a center point, commonly associated with sea fishing), and to some extent on historically known Sami settlement places in coastal areas (Thålin, 1962).These inventories were mostly based on old records and information collected from local people.Later more systematic searches were conducted during the 1980s and 1990s (Persson, 1985;Klang, 1986).In-  ventories were mostly conducted during the summer by archaeology students, each student having a certain area to cover during a set time frame.Every site was described and the location was plotted on an aerial photo map.All monuments were checked by an experienced archae-ologist (platsledare) to guarantee quality.All of this material is now digitalized and presented in FMIS (in English ASIS = Ancient Sites Information System). Figure 1 displays the areas that were surveyed during 1980-1995.The map shows that most areas up to 50 km from the coast-line have been inventoried, and in the Torne River valley coverage extends up to 200 km inland.The most and best covered areas are situated in the county of Norrbotten.
The FMIS-register also includes a number of features generally connected to reindeer herding such as: remnants of Sami settlement sites (visten), huts (kåtor), reindeer fences and corrals (gärdor).Occasionally, information acquired from oral tradition, maps and old records was registered.Despite the somewhat uneven coverage of the landscape, the register offers a relatively accurate picture of the dispersion of features that might have been connected to reindeer herding during the 19 th century and earlier.About 1500 hearths (which are protected by Swedish law) had been registered in the counties of Västerbotten and Norrbotten by 2012.The numbers of hearths per site vary from one to more than 10.A few hearths had been excavated in this area prior to the excavations reported here, some of which yielded artifacts datable to the 18 th and 19 th centuries.In addition, a silver buckle dated to the medieval period has been found close to one hearth in Gråträsk in Piteå parish (Fossum, 2003).

Sampling strategy and field investigations
Our objective was to obtain an unbiased view of the coastal hearths, thus we chose to focus on a randomly selected sample, of sufficient size for statistical analysis of the hearths spatial and temporal distributions.The northern part of Västerbotten and the southern part of Norrbotten seem to have most hearths and were therefore probably frequently used over a long period of time (see Figure 2).This area is also fairly well documented by surveys.
A rectangular area covering roughly 107 x 94 km (NE-SV) was laid out to define the study area.This area was further divided into four zones, each covering about 27 x 94 km in size (see Figure 3) in order to detect possible differ-ences along an east-west gradient.There are to 341 hearths in each of these zones and about 1000 in total.
The elevation ranges from sea level in zone 1 to over 600 m above sea level (a.s.l.) in zone 4, and the highest marine limit stretches from southwest to northeast in zone 2 and 3 at ca 230 m a.s.l (Berglund, 2012;Näslund, 2015).The soils are mainly sediments along the river valleys, and morainic till dominates at higher elevations (www.sgu.se).All rivers have their source areas in the western forests except for the Pite River that starts further west in the Scandinavian Mountains.Further, all zones are located in the boreal forest where Scots pine (Pinus sylverstris), dwarf shrubs (Calluna vulgaris, Empetrum nigrum and Vaccinium vitis-idaea) and reindeer lichen (Cladonia species) dominate dry sediment soils, while mixed forests of Scots pine, Norway spruce (Picea abies), dwarf shrubs (Vaccinium myrtillus and V. uliginosum) and feather mosses (Pleurozium schreberi and Hylocomium splendens) characterize mesic sites (Engelmark & Hytterborn, 1999).
We randomly selected eight hearths in each zone; the minimum required for detecting statistically significant differences among hearths in different zones (see Zar, 2010), as follows.The number of hearths in each zone was divided by eight which gave a sampling interval.A random starting number was attributed based on the total number of hearths in each zone.From these two numbers eight excavation objects could be selected in each zone.Thus, hearths were randomly selected in total, 3.2 % of all hearths in the study area or 2 % of all known hearths in the coastal areas of Västerbotten and Norrbotten.Due to problems during field work (locating some hearths proved difficult) five initially selected hearths had to be replaced by random resampling.In three cases the selected hearths proved to be natural features, and in these cases the closest hearths on the respective sites were excavated and dated.During our field work 18 previously unregistered hearths were found.
All 32 sites (Figure 4) were mapped using measuring tapes or GPS.All hearths at each of the sites were described, and measured and tested with an earth probe for stone packings and red burned soil.Only half of each selected hearth was excavated (Liedgren, 2013 a-c,;2014).All excavated material was sieved (mesh size 3 mm).Samples of charcoal were collected for 14 C-dating, mostly from close to the rim of the constructions.This part of a hearth normally contains more charred wood than the central part because wood in the center is usually completely consumed by the fire, leaving only ash.All bone fragments (mostly retained by the sieve) were also collected.In one case, as detected bone fragments were very small, nearly all material from the hearth was scrutinized under a magnifying glass in the laboratory.
The area within a radius of 10-15 m around all hearths on each site (32 sites, containing 112 hearths in total), was scanned with a metal detector and all metal responses were checked.Each artifact found was photographed and mapped.During this process some burned and unburned bones were found.
Bone fragments were cleaned and subjected to osteological analysis.Generally, c. 1 g samples of undetermined bones found in the sieve were used for 14 C-dating.The use of bones for dating poses several problems that were raised by Lanting et al. (2001) and have been heavily discussed since then (see, for example, Van Strydonck et al., 2009;Van Strydonck et al., 2010;Zazzo et al., 2009;Hüls et al., 2010;Olsen et al., 2013;Taylor & Yosef 2014;Philippsen 2015;Ekholm 2015).However, the soils in the boreal study area are mostly podzols, which are acidic and do not contain dissolved carbonates, thus minimizing one potential source of errors.Furthermore, we used only hard and compact bones from extremities to minimize potential problems arising from impurities leaching into the bones.The bones present in hearths in our study area probably originated from boiled meat from terrestrial animals.Traditionally all bones were chopped through, to gain access to the marrow, using an ax.When chopping the bones splinters popped away and some landed in the fire.These pieces were probably free from meat.However, most of the bones were probably disposed of outside the living quarters when the meal was finished.Based on the red color of soils beneath the hearths a large share of them have probably been fired many times and/or maintained for long times at temperatures up to 700-800 C ̊ (Liedgren et al., 2016).Thus, many bone fragments probably reached high temperatures in the middle, and lower temperatures towards the stone lining.We chose as large fragments as possible (mean number: 3.1) to minimize amounts of impurities in the samples originating from, for example, the fuel.
Charcoal samples collected for dating were cleaned with a scalpel and only hard pieces were selected for analysis.Each of these pieces was split into two parts, one of which was dated while the other was used to determine the tree species it originated from.Both charcoal and bone samples were dated when the hearths contained enough bone and charred wood material for such analysis to assess correlations between the dates, and utility of the two materials for dating.Like bone, charred wood poses some dating problems, partly because old pine wood was frequently used as fuel and partly because charcoal is produced during forest fires.

Results
Most sites were located on or in the vicinity of heathland, but many were also situated close to lakes or mires.Hearth numbers at each site varied from one to14.The investigations showed that out of 113 hearths in total on the 32 chosen sites, 67 % were oval, 20 % were rounded and 13 % had other forms.Their mean length and width were 1.20 m and 0.95 m, respectively.About 75 % of the hearths had stone packings.Eleven sites included damaged hearths or damaged areas around hearths, mostly due to soil scarification carried out for planting trees.
On most sites, metal objects were found using the metal detector.These artifacts were mostly of utility items, such as knives, iron needles, thimbles, buttons, iron nails and lead bullets (see Table 1 and Figure 5).Most of the artifacts were found within a few meters of the hearths.Artifacts from 17 sites can be fairly accurately dated to the historical period between the 17 th and 19 th centuries (see Table 1).Ten copper and silver coins were found on five sites, issued between 1832 and 1888.

Excavated hearths
Of the 32 excavated hearths, 61 % were oval, some 32 % were round or rounded and 7 % had other forms.Their mean length and width was 1.17 m and 0.98 m, respectively.Around 90 % of the hearths had a stonepacking, of varying prominence (Figures 6-7), and the rest only had stone linings (Figure 8).Most constructions had a brown/-dark brown hearth fill, but in some only charcoal was found.About 70 % contained small patches to well-defined areas of red burnt soil in the bottom, some of which were bowl shaped.Red soil beneath a hearth surface usually indicates strong heat, and thus use in winter (Liedgren et al., 2016).
Only 10 hearths contained materials other than burned and unburned bones, and six contained only charred wood (see Table 1).The artifacts found were: an iron awl, porcelain, pieces from glass bottles (sometimes partly melted), lead, a piece of a chalk pipe, a whet stone, and metal (tin and bronze) plates.Flint was found in only two cases.The artifacts from within the hearths are thus sparse and give few overall indications of the hearths' ages.

Osteological analysis
Fragments of bones were found in 27 hearths (see Table 2).Burned bones were detected in 24 hearths and bones with little or no signs of burning in seven.Most bones came from hearths in zones 1 and 3. Reindeer bones were identified in 20 hearths, mostly from animals' extremities, but trunk, rib and hip bones were also identified (Vretemark, 2012;2014).All bones came from adult reindeer except for one juvenile that was 6-12 months old when Figure 6.Rounded hearth with a pronounced stone packing from area 3, NHB 358:5.Bones from the hearth are 14 C-dated to the period AD 1480-1660.Artifacts found in the hearth show that it can not be older than the 17 th century.slaughtered.Fish bones were identified in only two hearths one of which contained nearly 12 000 fragments in total, including substantial numbers pike, perch, whitefish, vendace and carp (Vretemark, 2014).Three hearths contained bones from birds, and in one of these goose and black grouse/grouse bones were identified.Two hearths, located within 15 km from the coast, contained sheep/goat bones.

C-datings
In total 50 datings were obtained from charcoal (n=31) and burned bone (n=19) samples (see Table 3).Of the dated charcoal pieces, 87 % (n=27) are from pine and 6 % (n=2) from spruce (Danielsson, 2014).).Dates obtained for 14 of these charcoal samples (from all four zones) are in the period AD 810-1670: six from the Middle Ages or earlier, and eight from the late Middle Ages up to AD 1670.The earliest dates are both for samples from zone 4 (NHB 280 and 284).However, there was no significant difference (P < 0.05) in mean charcoal 14 C-dates between zones, and the only significant difference in this respect was that dated bones from zone 3 were significantly older than bones from zone 1 (two-sample t-test, 0.02 ˂ P 0.05: mean dates ±SE, 328±62 and 112±14, respectively).
Nineteen dates were obtained from heated, mostly undetermined bones, to meet the weight requirements for datings.However, most of the bones were probably from reindeer.Five bone datings are in the period AD 1280-1660, all for bones from zone 3; the rest are from later periods.
Datings of both charcoal and burned bone samples were obtained for 18 hearths.As shown in Table 3, according to t-tests nine pairs of the dates do not significantly differ (at the P < 0.05 level), and eight do significantly differ.One pair (from hearth NHB 166) could not be tested.Seven of eight dates from charred wood are older than dates obtained for the bone ma-terial and the other is younger.The mean value of the seven datings ( 14 C age) from charred wood is 210 years older than the mean date for the bones.The difference between charcoal and bone datings is illustrated in Figure 9.It should also be noted that the overall mean date for charred wood samples is c. 90 years older than the mean date for the bone samples.

Discussion
Most excavated hearths were oval and many had stone packings, in some cases constructed with larger flat stones and in other cases with rounded stones.Some of the hearths did not have any pronounced stone lining, with just a layer of stones in evidence.Some of these features are closely related to hearths typical of the interior area, but not all.For example, up to a third of the excavated hearths were rounded, which is a type seldom found in Lappland (Hedman, 2003:133pp).
Most of the excavated hearths contained bones.Nearly all of the bones from mammals have been identified as coming from reindeer strongly suggesting that people connected to reindeer used the hearths.They may have been reindeer herders or hunters of wild reindeer (Fahlgren, 1963:163pp;1970:160p;Ekman, 1983:8pp;Lundström, 2008:72p).According to Ekman (1983:13) the wild reindeer seems to have been hunted to extinction in all coastal areas of the country apart from the Torne river valley during the 18 th century.
Almost all the mammalian bones also came from mature adult reindeer, and have been splintered, probably for consumption of bone marrow.Several artifacts, such as forks (to take up meat from a kettle) and hooks for kettles, indicate that the meat was boiled.In some cases there is evidence (osteological or from finds of cast lead bullets and pieces of lead) of hunting, trapping and/or fishing.The large number of reindeer bones shows that meat from this animal was the main food, suggesting that many of the excavated hearths were used by Sami as boiled reindeer meat formed a large part of the traditional diet of the Sami in Lappland (von Düben, 1977:134pp;Drake, 1979:136pp;Högström, 1980:116pp;Graan, 1983;Rheen, 1983;Tornaeus, 1983;Hedman, 2003).
The artifacts found in and around the hearths are mostly pieces of domestic items that were used in every day life.We did not find any fire steels and just a few pieces of flint, possibly because in some cases matches were used to start fires (the use of matches in Sweden started in the middle of the 19 th century).
Most sites are located in the vicinity of heath lands, suggesting occupation during winter when the reindeer traditionally feed on lichens.This hypothesis is corroborated by the distinct red burnt soil found under many hearths (c.half of all excavated hearths) indicating strong heat (Liedgren et al., 2016).In some cases these hearths must have been heavily fired, possibly continuously for several days and probably during very cold weather.Six hearths, in zone 1, had red burned soil as deep as 17˗22 cm below the hearth surface.
Evidence of residence during the summer is rare.However, the goose bones from one hearth (NHB 358:5, zone 3) indicate spring or summer use (Vretemark, 2014), and constructions at one site (NHB 81, zone 1) also indicate summer occupation.At this particular site a hut ruin, traces of a building for storing food and a small cellar were found (see Figure 10).This kind of small cellar could be used for storing milk during summer and autumn.At the same site some hearths lie so close to each other that they can be interpreted as smoke hearths, used to generate smoke to protect the reindeer from insects and facilitate gathering of the animals for milking.
Most artifacts, found in and around the hearths, cannot be dated with any accuracy.However, one single item discovered might be older than the 17 th century (NHB 624), while  Many 14 C-dates show a wide range due to the calibration curve or the "Suess effect" (Taylor & Bar-Yosef, 2014:62pp).Fourteen of the charcoal datings are in the period AD 810-1670, but only five bone datings are in the period AD 1280-1660, all for samples from zone 3. The earliest bone dates were for samples from a hearth that had a packing of flat stones, and substantial amounts of dark smeary soil mixed with soot, charcoal and thousands of small burned fish bones (NHB 396).This hearth may have been connected to lake fishing (träskfiske) by the coastal farmers in the late Middle Ages (Berggren, 1995;Bergman, 2014).However, the oldest mean ages were obtained for the samples from zones 3 and 4 (331 and 347 years, respectively; more than 100 years older than mean dates for samples from zones 1and 2).Similarly, although the mean datings of charcoal samples from the four zones did not significantly differ, dated bones from zone 3 were older than dated bones from zone 1 (too few were found in zone 4 to test).The datings of bones from zone 3 and the two very old charcoal datings from zone 4 suggest that these areas were the earliest to be used.It is possible that dates for some of the bones were slightly too old due to contamination with old wood.
It is not surprising that charred wood and burnt bones give divergent results.The preferred wood for fuel in the coniferous forest was dry pine, which traditionally came from naturally seasoned trees that were abundant in olden times, especially in the interior parts of Norrland (see Siitonen 2001;Josefsson et al., 2010).Accordingly, we obtained older dates for c. 40 % of the charcoal pieces we analyzed than for dated bones from the same hearth.However, a later dating was obtained for charred wood than for bone samples from one hearth (NHB 1068:2), possibly due to later use or perhaps later forest fires (see Drobyshev et al., 2015).Thirteen hearths could only be dated with charcoal.Although some of the early charcoal datings could be correct, the data are insufficiently accurate for us to claim this with confidence.The historical record The purpose of this chapter is to give a historical background against which the archaeological results could be valued.Several sources can be used to describe the development in Lapland and in the coastal areas during the 16 th to 19 th centuries: Landskapshandlingar (Västerbottens handlingar och Norrlands lappmarker), cameral records, judicial records, cadastral maps and also place names.Information about the northern part of Sweden is also given by clergymen (such as Graan, 1671;Rheen, 1671;Lundius, 1675;Högström, 1747) working in the area and by various travelers who also published their experiences (such as Linné, 1732;Outhier, 1742;Hülphers, 1758).
During the Medieval period private ownership of reindeer developed in Lappland and specific areas started to be used by family groups (Bergman et al., 2013).In historical times, from the late 16 th century, these areas became known as "lappskatteland" (taxation lands), and the families had to pay taxes to the Swedish state for their land.The historical development of this fiscal institution has been thoroughly investigated (Holmbäck, 1922;Hultblad, 1968;Lundmark, 2006).The "lands" were constructed in such a way that they included enough resources to meet most needs of each family (Josefsson et al., 2009).The Sami communities in the interior were divided into two main groups: forest Sami and mountain Sami.The mountain Sami groups were heavily committed to reindeer herding and lived mostly from the production of meat and milk products from tame reindeer (Lundius, 1983:11;Rheen, 1983:19p).The forest Sami relied mostly on fishing and hunting, and had relatively few reindeer (Rheen, 1983:20;Graan, 1983:36).For example, at the beginning of the 17 th century the forest Sami in southern Lapland only had about 15 reindeer per family, so most of their subsistence came from sources other than tame reindeer (Norstedt, 2011:37).
In 1614 complaints were raised against tax pressure on the holders of tax lands.Some families had difficulties supporting their livelihoods and had to start a new life among the farmers by the coast (Lundmark 1998:39).In addition, numbers of reindeer gradually increased in Lappland, due (according to Lundmark) to a shift from taxation on fur and hides (in the 16 th century) to taxation of food products at the beginning of the 17 th century.This could have led to intensified hunting for wild reindeer, which became rare, and an increase in numbers of tame reindeer.With time the "lands" became too small to support each herd (Lundmark 2006:193).The "lappskatteland" system started to break down during the 18 th and 19 th centuries, beginning in the north (Holmbäck 1922;Lundmark 2006), mainly because each family tried to maximize their herds (Ingold 1980:78;Kvist 1989:162).A key requirement for every herd was access to winter pastures, which became increasingly limited.Thus, the reindeer herders looked for other areas for pasture during the winter.They were not allowed to use the land east of the Lappland border for pasture until the latter part of the 18 th century (according to Lundström 2008:70), when the Lappland border was fixed (1766).From 1770 onwards the Sami were allowed to use the forest owned by the Swedish state (during hard winters), but not privately owned areas (André 1998:59p).New laws, introduced in 1886 and 1898, enabled nomadic Sami to use the coastal areas during October to April without paying any compensation to landowners (Holmbäck 1922:65, SFS 1886:38, SFS 1898:66).
Evidence of a movement of Sami groups, from the interior to the coast, in Västerbotten and Norrbotten can be detected in the historical records (mainly from zone 1) from the 17 th century onwards (Lundström 2001:151).During the 18 th century Sami groups were noticed in the coastal areas in Hälsingland up to Torne river (de la Motrayes 1988:188;Outh-ier 1982:28;Linné̓ 2003:12p, Lundström 2008:84;Hülphers 1978:127).
During 1730s several Sami were contracted as farmhands and maids in Skellefteå parish (Lundström, 2001:151) and during the latter part of the 18 th century there were relatively large Sami settlements in the western part of Skellefteå parish (Lundström, 2008:51p, 84).During the second half of 18 th century there was a significant increase in numbers of Sami living in the Luleå area (Nordberg, 1965:84).In this area, as in other places, the Sami generally lived in the outskirts of the villages (Nordberg, 1965:87).During the 19 th century the number Sami living in the coastal areas multiplied in Norrbotten and Västerbotten.For example, the number in Skellefteå roughly doubled, from about 50 to 100, between the year 1800 and 1850 (André, 1998:67) and in Luleå there were more than 150 Sami in 1820 (Nordberg, 1965:83).
Place names provide numerous examples of name elements connected to the Sami population in the coastal areas.The prefix Lapp-(Sami-) is widely used in names of places scattered across the whole of Norrland, but most commonly east of the Lappland border, from Medelpad up to Sangisälven (Figure 11).
Names with the prefix Lapp-(Sami) are documented as early as the 16 th century on cadastral maps, for example "Lappviken", which was a small village in Lövånger recorded in 1543.Several "Lappviken" can be found on cadastral maps from the first half of the 18 th century, from Lövånger to Luleå parishes.Several records of this kind also indicate that Sami groups had huts in the coastal areas: "Lappkåtamyran" (1661, Skellefteå parish), "Lappkåtagärde" (1661, Burträsk parish), "Lappkåtamyran" (1803, Skellefteå parish) and "Lappkåtasjön" (1822, Bygdeå parish).There are more than 600 names with the prefix Lapp-in the province of Västerbotten and Norrbotten in the present database for place names (LMV).Most of these are combinations of the prefix Lappand the name of a geographical feature, such as heath, lake, river, mountain or mire (see also Bergman, 2010:175, fig. 1).Some also combines the prefix Lapp-and the name of some type of building.This category is dominated by the combination Lapp-and -kåta (hut).In the county of Västerbotten there are 82 such places and 19 in Norrbotten.These names can refer to permanent timber framed huts, probably of forest Sami origin.
Interestingly, the distribution of place names with the prefix Lapp-does not correspond to the distribution of known hearths in the same areas.This initially suggests that the distribution of hearths may be incorrect.However, many places with Lapp-names are (for example) known in Medelpad (Figure 11), but only one to two hearths of the types that were excavated in Västerbotten and Norrbotten, all others are connected to Stone and Bronze Age sites.Thus, the hearths in Västerbotten and Norrbotten may generally be connected with nomadic reindeer herders who lived in tents.In contrast, many names with the prefix Lapp-were presumably associated with Sami groups that settled in the coastal areas working (perhaps as craftsmen), or practicing a kind of reindeer management that has left few or different traces.
As shown above the movement of Sami towards the coast can be understood by the changes in the conditions for reindeer herding in the interior, which resulted in using the coast for grazing but also in people searching work and subsistence in the coastal areas.

Conclusions
This study provides a well-founded view of the locations, construction and chronology of hearths in the study area, and the types of artifacts found in their vicinity.Fewer hearths appear to have been constructed in the area south of the river Skellefteälven than further north in the counties of Västerbotten and Norrbotten.Artifacts and bone material excavated from the hearths and their surroundings clearly suggest that most of the coastal hearths have been used in connection with reindeer herding.
The 14 C-dates suggest that the earliest hearths, situated in the western part, may be as old as the Middle Ages, but most of them seem to have been used during the 18 th and 19 th centuries.There are only a few indications, in the archaeological material, that the coastal area was used for reindeer herding during the summer.The dates mainly support connections between the increase in herd size, breakdown of the "lappskatteland" system and use of the coastal area for winter pasture at around the time when most of the hearths were used.A consequence of larger herds was that fewer families could live on reindeer herding, so some were marginalized and had to seek their fortune in the coastal areas among the farming communities (Nordberg, 1965:82pp;André, 1998;Lundström, 2008:339pp).Some place names dating back as early as the 16 th century imply Sami presence from Late Medieval times.The prefix Lapp-is common in place names from Medelpad in the south to the river Sangisälven in Norrbotten, but hearths are scarce in the south.This indicates that in most cases the prefix emanate from the presence of Sami groups that were settled but not involved in nomadic life.The background of these people, in Västerbotten and Norrbotten, is still obscure and requires further study.Knowledge of Sami presence in coastal areas during the Medieval period and earlier, is also weak and calls for further investigations.

Funding and acknowledgement
This study was supported financially by the Bank of Sweden Tercentenary Foundation.We wish to thank Sees-editing Ltd for improving the language.

Figure 1 .
Figure 1.Survey of ancient monuments in the coastal areas of the county of Norrbotten and Västerbotten, made later than 1979 (based on Jensen 1997).

Figure 2 .
Figure 2. Distribution of hearth sites classified as ancient monuments in FMIS in the coastal areas of the county of Västerbotten and Norrbotten.

Figure 3 .
Figure 3. Study area divided into four rectangular zones, i.e. 27 x 94 km in size.Dark dots mark hearth sites.

Figure 5 .
Figure5.Selection of metal artifacts found in connection with hearths on site NHB 694: F 1 iron eyelet; F 2 iron fragment; F 3A flat iron nail; F 4 and 6 iron knives; F 7 iron fragment; F 8A flattened rounded lead bullet; F 9 iron buckle; F 13 iron hook for hanging kettles; F 15A ornamental bronze button; F 16A sheet of bronze (from bronze kettle?).

Figure 7 .
Figure 7. Rounded hearth with a packing of flat stones, from area 2 NHB 564.Bones in the hearth are dated to the period AD 1670-1940.Artifacts connected to the site suggest the latter part of the 19 th century.

Figure 8 .
Figure 8. Rounded hearth without stone packing from area 1, NHB 166:10.Bones in the hearth are dated to the period AD 1660-1950.Artifacts found on the site suggest the 19 th century.

Figure 9 .
Figure 9. Pairs of bone and charcoal dates.The datings are arranged with a start in the latest bone dating with its counterpart made on charcoal.

Figure 11 .
Figure 11.Distribution of place names with the prefix Lapp-(Sami) in Västerbotten and Norrbotten.Data from LMV.

Table 3 .
AMS-datings made by the Ångström Laboratory in Uppsala, Sweden.Datings made on charred wood and bones from excavated hearths.