Veget Hist Archaeobot (2005) 14:1–13 DOI 10.1007/s00334-004-0052-9

ORIGINAL ARTICLE

Karl-Dag Vorren

Stone Age settlements at Sørøya, sub-arctic Norway: impact on the vegetation Received: 28 October 2003 / Accepted: 23 September 2004 / Published online: 14 December 2004
Springer-Verlag 2004

Abstract Four pollen diagrams, which together cover the period 10400 cal bp to the present, from an area rich in Stone Age house foundations show that vegetation was increasingly influenced by man in a stepwise process. There was a modest, temporary appearance of anthropogenically-induced vegetation about 8600 cal bp. After 7100 cal bp there was a development towards semi-open birch woodland where the open patches were characterised by natural heath and meadow species and more typical apophytic taxa such as Ranunculus acris and Rumex acetosa types, grasses and juniper. The main changes towards this vegetation seem connected with the most important cultural changes, especially at about 7100 cal bp, when there was a transition to honed slate tools in the regional archaeology, at about 4000 cal bp when pottery was introduced, and at about 3000–2800 cal bp since when the use of stone artefacts gradually ceased. It is suggested that the settlements were mainly summer habitations. Keywords Stone Age · Vegetation · Sub-arctic Norway

Introduction As part of an archaeological project sponsored by the Norwegian Research Council during the period 1962– 1973 (Simonsen 1996), the late curator in archaeology at Tromsø Museum—Professor Povl Simonsen—carried out extensive excavations of Meso- and Neolithic housegrounds at Vatnan and Vatnhamn, Sørøya, Finnmark (Fig. 1) in the years 1967–1969. The results of the Vatnan investigations are available as a manuscript at Tromsø Museum (Simonsen 1999). Simonsen’s main idea in asking me to carry out pollenanalytical investigations at Vatnan was to see if hunterK.-D. Vorren ()) Institute of Biology, University of Tromsø, N-9037 Tromsø, Norway e-mail: [email protected]

gatherers’ activity throughout the period could be traced in the pollen diagrams, and to obtain an idea of continuity or discontinuity in the settlement history. These are also the main aims of the present study. Lack of 14C-dates was the main reason why the present results have not been published earlier. In addition, field and lab methods have developed a great deal since 1969. The pollen profiles were sampled by means of a Hiller corer, an instrument which is viewed with scepticism due to contamination risks. However, Prøsch-Danielsen (1997) has shown that carefully sampled old “Hiller samples” can be dated by the AMS method with great reliability. Pollen-analytical studies in northern Norway on Stone Age man’s impact on the vegetation and the environment in which he lived are scanty. Most relevant are the studies of Nilssen (Hesjedal et al. 1996) at Slettnes, Sørøya, Høeg (2000) who cored mires at several sites in Finnmark, and Jensen (in press) at Melkøya close to Hammerfest. Hicks (1991) has investigated the impact of ancient Lapps in winter settlements in a pine forest area of northern Finland. The archaeological periods of the Stone Age of Finnmark are also thoroughly 14C-dated at Slettnes, 15 km northeast of the present main study area, Vatnan (Hesjedal et al. 1996). According to this chronology, the Stone Age in western Finnmark starts about 9600 bp (10,860– 10,580 cal bp) and ends about 3650 bp (3950 cal bp). The transition between the Mesolithic and the Neolithic period is dated at about 7000 cal bp, when hewn slate tools were introduced. It must be emphasized that the term Neolithic does not imply introduction of domestic animals in this arctic/sub-arctic region. The period 7000–4000/3900 cal bp is divided into three Neolithic periods in accordance with archaeological evidence. The introduction of pottery c. 3950 cal bp marks the end of the Neolithic period and the start of the so-called Early Metal Age, lasting to 1950 cal bp. The gradual decline of cut and hewn stone tools some hundred years before Christ is considered an indirect indication of the presence of metal tools according to Hesjedal et al. (1996). In the present context,

2

however, the entire period prior to 1900–1700 cal bp is considered as the Finnmark Stone Age in accordance with Simonsen (1970).

Physiography of the investigation area Sørøya is an island of 816 km2, about 65 km long and 20 km broad, with many fiords (Fig. 1). Mountains up to 650 m a.s.l. with narrow ridges characterise the inner (south-eastern) part of the island. To the north and west the mountain surfaces are slightly lower with plateaux forming the tops. The bedrock in the investigation area is mainly gabbro (Sigmond et al. 1984). The Younger Dryas Main Line (raised beach) at Slettnes is probably situated at about 26 m a.s.l. (Holtedal 1960), and the Tapes level about 12.5 m a.s.l. The Tapes raised beach is very prominent on aerial photographs. It forms the damming threshold of the Vatnan lake (which is situated 11.5 m a.s.l. through stream erosion of the threshold). According to Simonsen (pers. comm. 1969), mesolithic house grounds occur above18 m above the upper drift wall of the sea shore. House grounds from the Early Metal Age occur down to 9.6 m at Vatnan, and down to 6.2 m above the upper drift wall at Gshopen (Fig. 1) (Simonsen 1991). The climate is oceanic (maritime) with a precipitation maximum in the autumn, though boreal (>10C mean temperature in July) according to Kaj Petersen’s modified version of the Kppen system (Abrahamsen 1977). The nearest relevant meteorological observation stations are Loppa and Hasvik Lufthamn in the south and Hammerfest in the north. Interpolating, the annual mean temperature at Vatnan is between 2.8 and 3.6C (1961–1990 normal period; Aune 1993; Førland 1993), the July mean 11.3– 11.6C, and the January mean between 5.2 and 2.0C. Annual precipitation is about 815 mm. The vegetation below c. 100–150 m a.s.l. belongs to the northern part of the north-boreal vegetation zone (Eurola and Vorren 1980; Moen et al. 1998), whereas “alpine”, or oro-arctic vegetation prevails above 100–150 m a.s.l. Birch woodland climbs up to c.150 m a.s.l. in the most protected parts of the island such as at Vatnan. Moist birch “forest” at Vatnan includes several apophytes: Agrostis capillaris, Anthoxanthum odoratum s.l., Deschampsia cespitosa, Nardus stricta, Allium schoenophrasum ssp. sibiricum, Alchemilla alpina, Leontodon autumnale, and Vicia cracca. The latter (Vicia cracca), however, also seems to occur as a natural element in swamp forests and carrs. Apophytes on drier slopes with birch woodland are: Anthoxanthum, Alchemilla alpina, Rumex acetosa s.l. and

Fig. 1 a Map with position of the study area (small square), and place names mentioned in the text. b Coring sites at Gshopen (G), Vatnan 1 and 5 (V1 and V5), and Husfjord (H). c Stone Age and Early Metal Age settlement sites at Vatnan (hatched). The numbers indicate the archaeological dates of the sites expressed in Ka cal bp. V1 and V5: Coring sites bp. The grid squares cover 10001000 m. o): Birch woodland

3

Ranunculus acris. A typical feature in grazed, open birch woodlands is the frequency of Hieracium sect. Prenanthoidea and Dryopteris filix-mas. At Gshopen on shell-sand there is a mixture of alpine species such as Dryas octopetala, Silene acaulis and Carex atrata, and hemerophilous plants such as Alchemilla alpina, Achillea millefolia, Leontodon autumnalis, Polemonium caeruleum, Ranunculus acris, Rumex acetosa, Taraxacum sect. Ruderalia, Gentianella aurea, and Carum carvi. Close to the shores Arrhenatherum elatior and Elymus repens occur. Around the farm houses at Vatnahamn, N-and P-plants dominate, such as Anthriscus sylvestris, Rumex longifolius, Ranunculus repens, Achillea millefolia, Stellaria media and Poa annua. The mires are mainly sloping, intermediate mires (Sjrs 1950) with some growth of Betula nana, Salix glauca, S. nigricans ssp. borealis and S. phylicifolia. Important plant communities are Trichophorum cespitosum-Sphagnum subsecundum communities with Equisetum fluviatile or Betula nana, forming the lawns, and Carex aquatilis–C. rariflora–Drepanocladus exannulatus communities forming the carpets. In moister positions Eriophorum angustifolium, Menyanthes trifoliata and Sphagnum subsecundum form carpets. Bordering the flark pools (open water) are Carex aquatilis, Menyanthes, Calliergon giganteum, Sphagnum cf. jensenii and Drepanocladus spp. Small scale farming with a few sheep and cows, and fisheries have been the main ways of living till recently. Most of the woodland has been exploited for pasturing, hay harvesting and by the cutting of trees and branches for fodder and fuel. Sheep and reindeer were the main pasturing animals. The sites Simonsen (1991) characterises Vatnan-Vatnahamn at Sørøya, western Finnmark (70320 N. Lat., 22540 E. Long., Fig. 1) as the site with the highest concentration of Stone Age house grounds in western Finnmark known up until 1988. A total of 120 house grounds are visible on raised beaches over a distance of 1500 m. The house grounds at Vatnan are mainly dated to the periods I–IV (Simonsen 1991), which according to the chronology of Hesjedal et al. (1996) cover the time span 7000–1950 cal bp Only a few mesolithic house grounds occur. Two pollen profiles from Vatnan (Vatnan 1 and Vatnan 5) are included in this paper. In addition a sample from an ash layer (midden) at the wall of a house ground in the westernmost Stone Age settlement of Vatnan (Fig. 1) has been analysed (Simonsen 1999). The Vatnan 1 profile (c. 18 m a.s.l.) was retrieved in a sloping, mesotrophic mire with Salix glauca and S. lapponum, Carex aquatilis, C. rariflora, Equisetum fluviatile, Trichophorum cespitosum, Menyanthes trifoliata, Drepanocladus purpurascens and Sphagnum subsecundum, close to a little mire pond c. 250 m northeast of the lake. The mire is part of a complex of sparsely treed soligenous mire elements, c. 350300 m

in extent. The distance to the nearest Stone Age settlement is about 430 m. The Vatnan 5 profile (c. 31 m a.s.l.) was retrieved from a peat cutting at the southern end of a 350300 m mire complex 20 m from the mineral ground, c. 100 m from the lake margin. Vegetation at the sampling point was characterised by Nardus stricta, Trichophorum cespitosum, Vaccinium uliginosum, Ptilidium ciliare, Sphagnum papillosum and S. subfulvum. The grass (Nardus stricta) does not belong to the natural mire vegetation and should be considered an apophyte spread onto the mire by sheep. The distance to the nearest Stone Age settlement is 75– 100 m. 3.5 km south-west of Vatnan is the site G
shopen (c.12 m a.s.l.). On the two promontories at western Gshopen, 43 house grounds have been recorded (Simonsen 1999). At Markusneset there is an area of aeolian sand. The light soils, south-western exposure and a topography that protects the area from northern winds make Gshopen a potential site for agricultural settlements. Sandy soils and dry grassy slopes confirm this impression. The presence of Arrhenatherum elatior may be an indicator of rather old dry pastures. Stones with rock carvings are found at a low level above the sea (Simonsen 1991). These are not typical hunters’ carvings according to Hagen (1976). The Gshopen core was retrieved from a 2030 m eroded mire at Sætervollen on Markusneset, the most significant promontory. The vegetation cover was dominated by a wet Carex limosa-Drepanocladus exannulatus community. The distance to the nearest known Stone Age settlements is c. 300 m. 5 km north of Vatnan is the head of Husfjord. It is characterised by typical local glaciation topography with cirques and local, frontal moraines. The sampling site is a mire where peat cutting has taken place, on a periglacial forefield about 30 m a.s.l. at the place named Husfjord on the south-western shore of the fiord, and c. 280 m distant from the closest frontal moraine. No archaeological remains or Stone Age housegrounds have been recorded here according to Simonsen’s (1996) map.

Materials and methods The pollen sequences were collected in late August 1969 by means of a Hiller corer. Peat and detritus gyttja samples of about 1 cm3 were carefully sampled from the central part of the core using two thin-bladed spatulae 1 cm wide, and were stored in small glass vials. The corks were sealed with paraffin. The samples were prepared using the KOH + acetolysis method of Fægri and Iversen (1964). Most of the material was analysed by the author in 1971. Due to the more recent possibilities of AMS-dating, pollen profiles were dated during the year 2002 using the material sampled in 1969. Some additional pollen analyses were made at the same time. The pollen and spore diagrams (Figs. 3, 5, 7 and 8) are percentage diagrams. Not included in the pollen sums are ferns and high-percentage mire species (Cyperaceae, Ericales, Potentilla-type and Menyanthes trifoliata), for which the individual sums are added to the pollen sum and their percentages of the total then calculated. Apophytic pollen taxa in this region are listed according to Vorren (1986) based on empirical data from the agricultural region further south in northern Norway. The time scale is calibrated years bp.

4 Table 1 Radiocarbon dates. Calibrated using OxCal v.3.8, middle of range with 68.2% probability limits Pollen profile

Lab. Number

Author’s number

Depth cm

Dated material

yr bp

cal age bp

Vatnan 1

TUa-2958 TUa-2959 TUa-2960 TUa-2961 TUa-3159 WK-10833 TUa-2725 TUa-2726 TUa-3583 T-910

V1-70 V1-155 V1-230 V1-315 V1-347 V5-344 V5-351 V5-361 G-331 Husfjord 1

70 155 230 315 347 50 100 159 45 185–190

Peat, bulk Peat, bulk Peat, bulk Peat, bulk Peat, bulk Peat, bulk Peat, bulk Peat, bulk Peat, bulk Gyttja, bulk

1820€65 2915€70 4540€70 5680€70 8755€80 1297€57 2685€55 4320€60 3730€70 9130€130

1745 2080 5185 6455 9725 1230 2800 4900 4080 10340

Vatnan 5 Gshopen Husfjord

Fig. 2 Zonation of the pollen diagrams Vatnan 1, Vatnan 5 and Gshopen into anthropogenic zones based on the four main apophytic pollen taxa

The diagrams were drawn using Grimm’s (1990 TILIA·GRAPH v. 2.0.b.5. The zonation program CONISS in TILIA (Grimm 1990) was used to distinguish anthropogenic zones. Charcoal particles were not recorded during the earliest analyses. Control analyses of selected samples with respect to charcoal particles were therefore performed for the Vatnan 5, Gshopen and Husfjord series. Percentage of charcoal particles is calculated in the same way as for ferns and high-percentage mire species. Only one 14C-sample, the Husfjord sample, was conventionally dated. The rest of the dates are AMS-dates (Table 1). All except one of them was dated at the AMS-laboratory in Uppsala, Sweden, after being processed at the National Laboratory for 14C-dating in Trondheim. One sample (WK-10833) is dated at the Waikato laboratory, Hamilton, New Zealand. Calibrated dates were calculated using the OxCal v3.8 program (Stuiver et al. 1998; Bronk Ramsay 2001). Size differentiation between the Betula pubescens s.l. type and the Betula nana type mainly follows the criteria of Eneroth (1951). Any Betula pendula long-distance transported pollen is classified as Betula pubescens. The Rumex acetosa type includes Rumex ace-

tosella coll., Rumex acetosa ssp. acetosa and R. acetosa ssp. alpestris. The latter taxon is indigenous, and only weakly apophytic. The Ranunculus acris type includes both R. acris and R. repens. They are both clearly adventitious, and at their distribution limits in Finnmark probably anthropochorous. Based on the geography of plant distribution, I would suggest that insectogamous pollen such as Filipendula pollen in northern Norway is most likely Filipendula ulmaria, Geranium pollen in Finnmark most likely Geranium sylvaticum, Bistorta pollen most likely Bistorta vivipara and Valeriana pollen Valeriana sambucifolia etc. In general, the pollen taxonomy follows Fægri and Iversen (1989). Nomenclature of vascular plants follows Lid & Lid (1994), and that of mosses Nyholm (1954–1967).

5

Fig. 3 Percentage pollen diagram from the Vatnan 1 site, Sørøya, Norway

6

Fig. 4 Age-depth diagram with calibrated dates bp for the Vatnan 1 profile

Results Zonation Numerical zonation of the Vatnan 1, Vatnan 5 and Gshopen diagrams was performed, based on the four main apophytic pollen taxa (Juniperus, Poaceae, Rumex acetosa-type and Ranunculus acris-type) (Fig. 2) in order to create “anthropogenic zones”.

perus curve and a low, discontinuous Ranunculus acristype curve. Indigenous meadow and heath plants belonging to the woodland floor increase and occur more or less continuously after c. 7100 cal bp (Geranium, Trollius, Solidago virgaurea-type, including Antennaria). Va1-3a: 4000–1900 cal bp (192–82 cm below surface). This sub-zone is distinguished by the increase of Juniperus to its maximum level. The expansion of an indigenous heath taxon such as Melampyrum and a meadow plant such as Trollius may indicate an increase in open meadows to the cost of Betula pubescens s.l. woodland. The upper part of this zone is characterised by a decrease in Rumex acetosa- and Ranunculus acris-types and a simultaneous increase in grasses. The period 2800–1900 cal bp constitutes the optimum era for apophytes such as Rumex acetosa/R. acetosella, Ranunculus acris/R. repens and juniper. Va1-3b: 1900–1750 cal bp (82–70 cm below surface): The statistical zonation suggests that this onesample event with a marked decline in grasses should be marked as a special (sub)-zone. Va1-4: 1750– (?) cal bp (70–0 cm below surface). The grass curve increases to a high level. Other apophytes stabilise at a moderate level. Juniper, pine, birch and ferns decrease while Bistorta vivipara and Ericales increase. At the end of the zone most apophytes decrease and birch increases. Ferns decline to a very low level at about 1100 cal bp as based on the extrapolated age scale.

Vatnan 1 The chronology in the pollen diagram (Fig. 3) is based on five AMS radiocarbon dates (Table 1), of which one (V1315, TUa-2961) was rejected (Fig. 4). The extrapolated age scale indicates that paludification started prior to 10,000 cal bp. The oldest peat is formed by carr vegetation with Betula pubescens s.l. as the dominant megafossil. Anthropogenic zones The diagram may roughly be separated into five main anthropogenic zones (Va1-1–5) and four sub-zones (Va12a and 2b, Va1-3a and 3b). Va1-1: >10,000–8,600 calbp (370–317 cm below surface). The lack of Ranunculus acris type pollen makes it difficult to say whether the other apophytes (Poaceae, Juniperus, Rumex acetosa-type) are reflecting human activity. Their occurrence may be due to pioneer vegetation on the mire (Poaceae, Juniperus) or may be a reflection of the post-glacial vegetation succession in which Rumex plays an important role (Fig. 8). Va1-2a: 8600–7100 cal bp (317–282 cm below surface). The scanty occurrence of Ranunculus acris and a temporary peak in the grass curve may indicate slight human impact. Va1-2b: 7100–4000 cal bp (282–192 cm below surface). A period with a moderate anthropogenic impact. The sub-zone is characterised by a rise in the Poaceae and Rumex acetosa-type curves, a low but continuous Juni-

Vatnan 5 The mire has been subject to peat-cutting, but was considered to be undisturbed at the sampling point. The profile was 171 cm deep, with the base of the peat resting on shore gravel. Internal vertical distance between most of the samples was 7.5 cm in this peat sequence. Three dates were obtained, indicating the start of paludification before 5000 cal bp (Fig. 6). This means that the anthropogenic zones 2b (upper part), 3 and 4 of the Vatnan 1 diagram (Fig. 3) are represented in the Vatnan 5 diagram (Fig. 5). Anthropogenic zones Va5-1: (?)–c. 4900 cal bp (171–160 cm below surface). The dating of the lowermost peat is uncertain. Closed moist birch forest rich in ferns and tall herbs is characteristic. Apophytes increase from low initial percentages and the percentage level of charcoal dust (7–9%) is comparatively high. Va5-2: 4900–4300 cal bp (160–142 cm below surface). Ferns decline and the grass curve and other apophyte curves expand to a moderate level. Indigenous pollen taxa of certain heath plants such as Melampyrum spp. and Cornus suecica, and meadow plants such as Trollius, Geranium and Apiaceae seem to respond to human impact in the same way as the most important apophytes in this and the next zone.

7 Fig. 5 Percentage pollen diagram from the Vatnan 5 site, Sørøya, Norway

8 Table 2 Pollen content in a charcoal sample from below a house ground wall in house No 7, house grounds group no. 17, westernmost Neolithic settlement at Vatnan (Simonsen 1999), archaeologically dated to c. 5300–4000 cal bp

Fig. 6 Age-depth diagram with calibrated dates for the Vatnan 5 profile

V5-3: 4300–1300 cal bp (142–47 cm below surface). The juniper and grass curves and the Ranunculus acris and Rumex acetosa type curves increase towards a steady and comparatively high level. Birch declines a little at the start of the zone, but maintains a steady level above this. Willow increases temporarily. The low charcoal dust percentages in the lowermost part of the zone and the higher ones in the upper part of the zone are not supported by any correlated oscillations in the apophyte curves. The period between c. 3000 and 2000 cal bp is the optimum for juniper, Ranunculus acris and Rumex acetosa pollen taxa indicating a period with a relatively high and stable anthropogenic impact. Va5-4a: 1300–700 cal bp (47–25 cm below surface). The juniper, Ranunculus acris and Rumex acetosa type curves are unstable and vary considerably. Ericales and ferns increase. Va5-4b: 700–0 cal bp (25–0 below surface). Expansion of Poaceae and Ranunculus acris. Decline in the tree pollen and fern curves. “Culture soil” at House No 7, Group 17, Vatnan A soil sample was taken from a homogeneous, c. 20 cm thick, black soil layer (waste heap) 20 cm below the soil surface at the south wall of a house from the period 5300– 4000 cal bp (Simonsen 1999; pers. comm. 2002). The house foundation had overlapping culture layers from the succeeding period (Early Metal Age). The soil layer contained mainly microscopic remains of charred wood, i.e. ashes. The identifiable charred wood fragments were about 78% from deciduous trees and 22% from conifers, the latter mainly Pinus but also some cf. Juniperus communis. The ashes contained a few pollen grains. The pollen was badly degraded, hence many pollen grains were non-identifiable. Table 2 shows the assemblage of 221 pollens and spores.

Pollen taxon

Pollen count

% of SP

Pinus Betula pubescens Alnus Betula nana type Ericales Poaceae Stellaria type Taraxacum type Onagraceae Hordeum type Geranium sylvaticum Cichorioideae indet Solidago type NAP indet Sphagnum Filices p.p Gymnocarpium type Lycopodium annotinum Huperzia selago Polypodium vulgare

2 141 4 2 7 3.2 1 7 21 2 4 8 7 39 1 28 5 3 2 1

0.7 50.5 1.4 0.7 2.5 11.5 0.3 2.5 7.5 0.7 1.4 2.8 2.5 14.0 0.3 9.1 1.7 1.0 0.6 0.3

G
shopen The diagram from Sætervollen, Markusnes at Gshopen (Fig. 7) is based on a short (50 cm) peat sequence. The upper 15 cm of the profile is unconsolidated, fibrous peat with a fresh moss carpet and modern roots, and is therefore omitted from the analyses (and the diagram - Fig. 7). The 45 cm level is 14C-dated to 4165–3955 cal bp The Picea pollen increase at between 20 and 25 cm below the surface is probably the regional one, which is dated to about 2800 cal bp at Vatnan. If this estimation is right, the basal layer may be dated to about 5300 cal bp by extrapolation, and the top (15 cm below surface) to about 2400 cal bp The start of the paludification thus seems to be synchronous with that in the Vatnan 5 profile. The lack of Sub-Atlantic sediments may be due to erosion. No marks of peat cutting in the mire were observed. Anthropogenic zones Ga-1: C. 5000(?)–4100 cal bp (65–47 cm below surface). The first indication of human impact on vegetation seems to be a decline in birch pollen with an expansion in grass pollen and two of the indigenous pollen taxa (Cichorioideae and Apiaceae). Ga-2: C. 4100–3000(?) cal bp (47–33 cm below surface). Start of the continuous Rumex acetosa-type curve. Rumex and Poaceae characterise this zone. In the upper part birch declines further while Geranium, Trollius and Solidago-type expand. Ga-3: C. 3000 (?)–2400 (?) cal bp (33–15 cm below surface). Start of continuous Ranunculus acris curve. Rumex, Ranunculus, Juniperus and Poaceae characterise the zone.

9

Fig. 7 Percentage pollen diagram from the Gshopen site, Sørøya, Norway

Fig. 8 Percentage pollen diagram from the Husfjord site, Sørøya, Norway

10

Husfjord

Sedentary or non-sedentary settlements

On a regional scale, the diagram (Fig. 8) reflects periglacial vegetation in its basal layers, dated to 10340 cal bp This is indicated by the high Artemisia values at the base, the succeeding high Oxyria curve, high Ericales values and low arboreal pollen percentages. A possible anthropogenic feature in the diagram is the fairly high charcoal dust curve. Charcoal particles of about 100 mm in the lowermost strata were recorded. It is doubtful whether the Rumex longifolius pollen type (=EuRumex in Fig. 8) should be ascribed to Rumex longifolius.

Simonsen (1999) discusses the extent to which the users of the Stone Age settlements at Vatnan were sedentary. He suggests that people stayed there from late summer to early winter, and argues that there were traces of snow avalanches and ravines with soil and boulders covering the westernmost settlement area during the periods of habitation. Also, the wood resources must have been limited in the area. Simonsen (1999) thinks that people may have moved between Gshopen and Vatnan in order to save the wood resources. Simonsen does not consider the Vatnan settlements as part of a larger movement pattern—a rotation between moving inland in the winter and to the coast in the summer, such as practised by the Lapps in east Finnmark and north east Finland in historical times (viz. the Skolte Sami: Tanner 1929). The analyses of ashes from the period c. 5300– 4000 cal bp may throw some light upon the seasonal occupation of Vatnan. The ashes imply that about 20% of the fuel wood was from conifers (mainly pine). However, whereas birch is represented by more than 50% of the total pollen sum, pine is represented by less than 1 percent (Table 2). The over-representation of tree-birch pollen may be due to over-representation of birch twigs cut in the spring, when the anthers are not emptied. The old technique of søyring, which is still used in northern Norway, involves the cutting of birch in early summer (June), when leaves have not yet developed to a full size and flowering has just started. The birch is then left to dry out on the ground during the summer. In the autumn and winter it used partly for fuel and partly for sheep fodder (the inner bark of the branches). The de-barked branches are later used for fuel. Birch twigs were probably also used as an underlay for beds. On the basis of the ash contents it may be suggested that people were present at Vatnan at least during the early summer, during the flowering period for birch.

Discussion Dates The conventional Husfjord date T-910 (9130€120 bp, c. 10340 cal bp) is a bulk date from the transition between a grey gyttja and a 5 cm thick chocolate brown layer of gyttja. It should be noted that dates of bulk gyttja published by Sepp (1996) from the northernmost county of Finnmark (Norway) indicate a date for birch forest expansion of around 11200 cal bp However, a date from a peat profile from the northern coast of Finnmark (Høeg 2000) indicates establishment of birch forest in coastal Finnmark (at Mehamn, Fig. 1) at around 10250 cal bp corresponding better to the Husfjord date. The date from the gyttja from Husfjord may also be too early because of the general reservoir effect in bulk gyttja. However, the periglacial vegetation with Oxyria and Artemisia and the small temporal difference between the Husfjord and the Vatnan basal layers do not indicate that the Husfjord date should be more recent. One of the AMS-dates, TUa-2961, from 315 cm below the surface in the Vatnan 1 profile (Figs. 3 and 4), yielded 5680€70 bp (6470 cal bp) instead of an expected age close to 7500 bp or 8350 cal bp based on the date the final rise of pine. During a later examination of the samples just above and below this level, it was found that these consisted mainly of Equisetum fluviatile roots and stolons. It was therefore reasonable to believe that the actual sample TUa-2961 also consisted of the same type of peat. Equisetum fluviatile may have its stolons and roots creeping down to 1–3 m below surface and thus it seems that these remains constituted a contamination source for the TUa2961 date which is therefore too recent (see age-depth diagram Fig. 4). The pine expansion in another diagram from a mire on the coast of Finnmark (Momyra at Mehamn, Fig. 1) (Høeg 2000) is dated to c. 8350 cal bp. At Melkøya, some 30 km to the northeast, the event is also dated to 8350 cal bp (Jensen, in press). The very early dates for the immigration and expansion of Pinus in the region (10400–9200 cal bp), reported by Sepp (1996) and Sepp and Hammarlund (2000) may be due to a hardwater effect in bulk sediments.

Pollen rain as evidence of hunter-gatherer habitations in northernmost coastal Norway The increase of apophytes (especially the Ranunculus acris and Rumex acetosa types) to a stable and high level about 3000–2800 cal bp at Vatnan may reflect a more sedentary habitation during the second part of the socalled Early Metal Age. This percentage level of these apophytes in more southern regions of northern Norway is normal in farm habitation contexts. However, there are special features at Vatnan. Here, the comparatively low and stable grass curve, and the high and stable fern curve indicate that there were no freely roaming domestic animals. Nitrogen- and phosphorus-demanding plants such as Urtica, Rumex longifolius, Stellaria media, Achillea millefolia are hardly present, even at waste heaps (Table 2). Until a radiocarbon date for the ash is available, it is difficult to discuss the occurrence of the Hordeum type

11

pollen. It might derive from Leymus arenarius, which occurs locally on the island, but does not have any natural habitats at Vatnan at present. The pollen-analytical evidence from Vatnan 5 suggests the introduction of domestic animals at about 700 cal bp From that time there is archaeological evidence of a sedentary settlement (Simonsen 1999). The most reasonable interpretation of the anthropogenic features of the post-7100 cal bp period at Vatnan seems to be that the pollen diagrams reflect the open spaces of a hunter-fisher village with no domestic animals (except dogs, Olsen 1975). The vegetation might have been a grass-rich meadow with apophytes such as Juniperus communis, Rumex acetosa and Ranunculus acris/R. repens, but also with natural meadow plants such as Anthriscus sylvestris, Geranium sylvaticum, Trollius europaea, Campanula rotundifolia, Solidago virgaurea and Hieracium spp. This picture is also the case in Høeg’s (2000) pollen diagram from Momyra (at Mehamn, Fig. 1) some 130 km to the northeast. Simonsen (1999) suggests that there might have been about 25 inhabitants at Vatnan as it was in historical times. Helskog (1984) suggests that 20–60 inhabitants constituted similar Stone Age settlements in eastern Finnmark. During the period 4000– 2000 cal bp, houses at Vatnan were located within a distance of 200–300 m from the sampling site of the Vatnan 5 profile. A summer habitation at Vatnan for a community of more than 20 people would be able to establish an area with apophytes and natural meadow and heath plants of more than one hectare around and between the houses, this being reflected in the diagram. In addition, the need for fuel and building materials would have led to a thinning of the surrounding woodland and created niches there for a light-demanding flora, characterised by juniper and more abundantly flowering indigenous herbs such as Geranium sylvaticum, Trollius europaeus and Anthriscus sylvestris in meadow woodland, and Solidago virgaurea, Melampyrum pratense and Cornus suecica in heath woodland. Pollen rain at winter settlements of hunter-gatherers in the inland part of northern Fennoscandia Hicks (1991) investigated a turf wall section of a house in a Lappish (Skolte Sami) winter settlement in the pine forest region of northern Finland, at the lake of Inari. The settlement was dated to the period from the 14th century to ad 1800. The inhabitants were living at Inari in the winter and on the Norwegian coast in the north in the summer (Tanner 1929). In the Inari results charcoal particles (>40 mm) are clearly the most important anthropogenic indicator. Charcoal and Ranunculus are well correlated, showing the importance of this pollen taxon as an indicator of human impact in northern Fennoscandia. The curves of Poaceae and Rumex acetosa/acetosella are low—compared to the Vatnan and Gshopen coastal sites. Indigenous heath forest herbs such as Solidago type seem correlated with anthropogenic activity at the Inari winter

settlement. Epilobium occurs in a “cultural layer” and also in the upper layers when the settlement is abandoned. This latter is obviously Epilobium angustifolium, which colonises burnt areas, waste ground, rubbish heaps and decaying turf walls in this region. The similarity with the Epilobium occurrence in the Vatnan ash layer (Table 2) is evident. Pollen from Betula trees in the winter village responds to human activity by decreasing, indicating that it was the preferred source of fuel even in this pine region. However, in the inland regions, the regeneration and productivity of all trees are far better than at the coast, so there was certainly no shortage of fuel for the resident Lapps. Høeg (2000) investigated an inland site in the pine region of Finnmark, Norway, and found a period of high charcoal deposition between c. 3800 and 1500 cal bp. At the start of this period he found cereal-type pollen. The pollen picture of this period shows similarities to that of Hicks (1991) from Inari: A low grass curve and very little Rumex with a stable and small representation of Ranunculus correlated with the charcoal curve. It should be emphasised that Leymus arenarius, which may be confused with cereal pollen of the Hordeum type, does occur inland. Thus it remains open as to whether Høeg’s (2000) diagram from this continental region is mirroring anything else than a winter settlement of a half-nomadic hunter-gatherer culture. Continuity The Vatnan 1 diagram seems to exhibit a more than 5000 year long (7100–1900 cal bp) history of permanent use of the Vatnan-Vatnahamn area, with a fairly stable anthropogenic impact on the vegetation both in the field layer and the tree layer. As suggested by Simonsen (1999), this stability may be achieved by a shift between two or more different habitation sites in a 70–100 year cycle, as was done by the half-nomadic Lapps (Skolte-Sami) in eastern Finnmark at their summer locations on the coast (Tanner 1929). Since the temporal resolution of the Vatnan diagrams is low, c. 200 years in the upper parts of the diagrams, such shifts might be difficult to detect. The incident in the Vatnan 1 diagram, dated at about 1900 cal bp, where the grass and sorrel curves decline to zero as birch increases, reflects an abandonment period of 150 years or more. This is not reflected in the Vatnan 5 diagram, which may be explained by the low temporal resolution. More detailed analyses are needed to elucidate the “habitation shift” question. However, the Gshopen pollen diagram does not support the idea of a regular settlement shift between Vatnan and Gshopen. The Rumex acetosa type curve is a rather late feature (from 4100 cal bp) at Gshopen, and the Ranunculus acris type curve occurs even later. The stable state (equilibrium) of birch woodland and apophyte vegetation during the periods 7100–1900 cal bp (Vatnan 1) and c. 4900–1300 cal bp (Vatnan 5) implies a stable human impact on the vegetation even if this included a cyclical pattern of movement.

12

Size of area reflected in the pollen diagrams There are great similarities in the percentages of apophytes in the Vatnan 1, Vatnan 5 and Gshopen diagrams. However, the Vatnan 5 profile, sampled closer to the settlements than the Vatnan 1 profile, has slightly higher grass percentages and clearly higher Juniperus and Ranunculus acris type percentages than has the Vatnan 1 diagram. Meadow and heath plants are also more abundant in the Vatnan 5 diagram. The Vatnan 5 pollen site is only c.100 m distant from the nearest Stone Age settlement which existed prior to 4000 cal bp The settlements closest to the Vatnan 5 profile after 4000 cal bp are some 200–300 m distant. However, the anthropogenic features in the pollen picture of the post-4000 cal bp sediments in the Vatnan 5 profile are more evident than are those of the pre-4000 cal bp period. This implies that man-influenced vegetation within a radius of at least 200–300 m is reflected in the Vatnan 5 profile. The similarities between the Vatnan 5 and Vatnan 1 diagrams indicate that pollen produced in and around the settlements is also deposited at the Vatnan 1 site. This in turn implies that southern winds may have distributed the pollen northwards over the lake surface and further over the open mires between the lake and the Vatnan 1 pollen site for a distance of 450–600 m. Also the high proportions of mineral ground pollen in both diagrams indicate that the “relevant pollen source area” (RPSA) of Sugita et al. (1999) is at least 200–300 m, because it has to extend outside the mire margins. For lakes with a diameter of c.100 m Sugita et al. (1999) suggest a relevant pollen source area up to 800– 1000 m; however lakes act as large pollen traps, which “homogenise” the pollen deposits over time and space. Pollen samples from terrestrial sites such as mires, or even from mosses on mineral ground reflect a different history. Considerable differences occur in results from systematic investigations of the RPSA reflected by pollen samples from terrestrial sites. Moss samples from an area with a diameter of 0.5 m in a semi-open landscape in southern Sweden are estimated to cover an RPSA with a radius of about 200 m (Brostrm et al. 2002), whereas Bunting (2002) finds that the RPSA around a sampling spot of similar size in Scottish heaths covers an area with a radius of 2 m. There is not enough known to fix relevant pollen source area limits, at least as far as peat samples of 1 cm2 or less are concerned. Local terrain formation, vegetation, climate variables, pollen productivity, background loading, spreading by insects and wind, pollen history after deposition etc. have to be considered in a complex relationship when using quantitative approaches to solve the RPSA question on a general basis. To date I consider the pollen trap results achieved by Tauber (1965) are still adequate for pollen-analytical investigations dealing with peat sediments, if a wide range of local modifications is taken into account. Acknowledgements My sincere thanks go to the late professor in archaeology, P. Simonsen, Tromsø Museum, who was a formidable polyhistorian and an inspiration, not only to archaeologists. They

also go to A.K. Hufthammer, University of Bergen and E. Engelstad, University of Tromsø for information on animal bone artefacts in Finnmark, to the referees S. Hicks, Oulu, B. Aaby, Copenhagen, and J. Daniell, Cheltenham for valuable comments and suggestions, to C. Jensen and E. Elverland, University of Tromsø, for construction of the present diagrams, to T. Midtun for drawing of the map and to A. Karlsen and E. Elverland, for preparation of pollen samples. Lastly, my very sincere thanks go to the National Laboratory for 14C-dating in Trondheim

References Abrahamsen, J. (1977). Naturgeografiske regioner i Norge. Nordiska Utredningar B, 34, 1–137 Aune, B. (1993). Temperaturnormaler, normalperiode 1961–1990 (Air temperature normals, normal period 1961–1990). DNMIrapport 02, Oslo, pp 1–63 Bronk Ramsay, C. (2001). Development of the radiocarbon Program OxCal. Radiocarbon, 43(2A), 355–363 Brostrm, A., Sugita, S., Gaillard, M.-J., Pilesj, P. (2002). Estimating special scale of pollen dispersal in the cultural landscape of southern Sweden. Lundqua Thesis 46, (Lund University, Sweden), pp101–115 Bunting, J.M. (2002). Pollen-vegetation relationships in non-arboreal moorland taxa. Review of Palaeobotany & Palynology, 125, 298 Eneroth, O. (1951). Underskning rrande mjligheterna att I fossilt materiale urskilja de olika Betula-arternas pollen. Geologiska freningen i Stockholm frhandlingar, 73, 434 Eurola, S., Vorren, K.-D. (1980). Mire zones and sections in North Fennoscandia. Aquilo Ser Bot, 17, 39–56 Førland, E.J. (1993). Nedbørnormaler, normalperiode 1961–1990 (Precipitation normals, normal period 1961–1990). DNMIrapport 39, Oslo, pp 1–63 Fægri, K., Iversen, J. (1964). Textbook of pollen analysis. Munksgaard, Copenhagen Fægri, K., Iversen, J. (1989). Textbook of pollen analysis. Wiley, Chichester, UK Grimm, E. (1990). TILIA and TILIA·GRAPH: PC spreadsheet and graphic software for pollen data. INQUA Subcommission on Data-Handling Methods. Newsletter, 4, 5–7 Hagen, A. (1976). Bergkunst (Rock art). University Press, Oslo Helskog, K. (1984). Younger Stone Age settlements in Varanger. Acta Borealia, 1, 39–69 Hesjedal, A., Damm, C., Olsen, B., Storli, I. (1996). Arkeologi p Slettnes. Dokumentasjon av 11.000 rs bosetning (Archaeology at Slettnes. Documentation of 11000 years of settlement). Tromsø Museums Skrifter, 26, 1–246 Hicks, S. (1991). Ancient Saami in Finnish Lapland and their impact on the forest vegetation. In: Butlin, R.A., Roberts, N. (eds) Ecological relations in historical times. Humans impact and adaption. Blackwell, Oxford, UK, pp 193–205 Holtedal, O. (1960). Geology of Norway. Norwegian Geological Survey (NGU), 208, 1–540 Høeg, H.I. (2000). Pollenanalytiske undersøkelser i Finnmark, Nord Norge (Pollen-analytical investigations in Finnmark, northern Norway). AmS-Varia, 37, 53–97 Jensen, C. (in press) The vegetation history of a coastal stone-age and iron-age settlement at 70N, Norway. Vegetation History and Archaeobotany Lid, J., Lid, D.T. (1994). Norsk Flora (Flora of Norway). 6th edn. Det norske samlaget, Oslo Moen, A., Odland, A., Lillethun, A. (1998). Vegetasjon. Nasjonalatlas for Norge (Vegetation. National Atlas for Norway). Statens kartverk, Hønefoss Nyholm, E. (1954–1967). Illustrated moss flora of Fennoscandia. II. Musci, Lund Olsen, H. (1975). Varanger-funnene. Pattedyr (The Varanger finds. Mammals). Report unpublished. Zoological Museum, University of Bergen

13 Prøsch-Danielsen, L. (1997). New light on the Holocene shore line displacement curve on Lista, the southernmost part of Norway. Based primarily on Professor Ulf Hafsten’s material from 1955–1957 and 1966. Norsk geografisk Tidsskrift/ Norwegian Journal of Geography, 51, 83–101 Sepp, H. (1996). Post-glacial dynamics of vegetation and treelines in the far north of Fennoscandia. Fennia, 74 (1):1–96 Sepp, H., Hammarlund, D. (2000). Pollen-stratigraphical evidence of Holocene hydrological change in northern Fennoscandia supported by independent isotopic data. Journal of Palaeoclimatology, 24, 69–79 Sigmond, E.M.O., Gustavson, M., Roberts, O. (1984). Geological map of Norway. M 1:1 million. Norwegian Geological Survey Simonsen, P. (1970). Fortidsminner nord for polarsirkelen (Ancient memorials north of the Arctic Circle). Universitetsforlaget, Tromsø-Oslo-Bergen Simonsen, P. (1991). Fortidsminner nord for Polarsirkelen (Ancient memorials north of the Arctic Circle). Universitetsforlaget, Oslo Simonsen, P. (1996). Steinalderbosetningen i sandbukt p Sørøya, Vest-Finnmark (Stone age settlement at Sandbukt, Sørøya, western Finnmark (Norway)). Tromsø Museums Skrifter, 27, 1–258

Simonsen, P. (1999). Vatnan og Vatnahamnen p Sørøy, VestFinnmark: Steinalderbosetning og landskap under forandring (Vatnan and Vatnahamnen at Sørøy, western Finnmark: Stone age settlement and landscape in change). Manuscript Tromsø Museum, Tromsø Sjrs, H. (1950). Regional studies in North Swedish mire vegetation. Bot Notiser, 2, 173–222 Stuiver, M., Reimer, P.J., Bard, E., Beck, J.W., Burr, G.S., Hughen, K.A., Kromer, B., McCormac, G., van der Plicht, J., Spurk, M. (1998). INTCAL98 radiocarbon age calibration, 24,000–0 cal. bp. Radiocarbon, 40 (3):1041–1083 Sugita, S., Gaillard, M., Brostrm, A. (1999). Landscape openness and pollen records: a simulation approach. Holocene, 9, 409– 421 Tanner, V. (1929). Skolt-lapparna. Fennia, 49(4):11–28 Tauber, H. (1965). Differential pollen dispersion and the interpretation of pollen diagrams. Danmarks Geologiske Undersøgelse IIR, 89, 1–69 Vorren, K-D. (1986). The impact of early agriculture on the vegetation of northern Norway. A discussion of anthropogenic indicators in biostratigraphical data. In: Behre, K.-E. (ed) Anthropogenic indicators in pollen diagrams. Balkema, Rotterdam, pp 1–18