4.3. The excavations

4.3.1.1. Sundbergen 1

The cairn Sundbergen 1 (258) has been erected on a convex, fairly flat top, only a little larger than the cairn itself. The surface is partly of coarse granite, partly of granodiorite, with biotite-plagioclase schist in small inclusions. The area covered by the stone setting consists mostly of granodiorite. The rock surfaces surrounding the grave are fairly even, so the shore processes have evidently swept away soil, stones and boulders from the top some 4500 years ago. Thus, all of the construction material of the cairn must have been transported to the site. The area, measured along the edges of the stone setting, was 33 m. There were 768 boulders weighing at least 0.5 kg, and their total weight was 11.30 tons.

Most stones were rounded and, especially those of the surface layer, rather even-sized. There was a clearly discernible pit in the middle part of the cairn (fig. 11). Below the surface stones, there were a lot of small stones and pebbles, about the size of a fist, which had been loosely wedged in between the larger boulders so that they often fell deeper when the large boulders were being moved. Thus, the average size of the stones and boulders could not be determined on the basis of the surface stones only. Possibly the pebbles were used to fill up the holes between the larger boulders on the surface of the cairn, but due to gravity, a lot of pebbles slid down downwards between the boulders. This gravity effect, the abundance of small stones in the filling layer beneath the surface was also observed by Tapio Seger in the investigation of the cairn 088 at Furunabb, Houtskär, two years earlier.

The soil layer on the ground was covered by litter spilt down from between the boulders. Under the litter, there was a 1–3 cm thick layer of dark humus containing plenty of organic material and angular gravel particles, which were residual originating from weathering. The most gravel was to be found under boulders lying on the granodiorite surface, and in the holes and crevices of the surface. The excavation of the humus layer revealed unburnt bone. On the bottom surface of the pit, burnt human bone fragments were found in an area of 3.2 m (161 g).

Among the fragments of burnt bone, there were two pieces of burnt antler (0.6 g). One of the pieces had been carved into a rectangular and flat artefact. On the smooth-cut 16 mm long face there are two engraved cross-like figures which form part of a longer decorated surface (fig. 12). The rivet hole, too, is visible. The artefact is probably the peg plate of a one-sided comb made up of three parts. This type of comb is rare on the mainland of Finland (e.g. Lehtosalo-Hilander 1982: 65); they are, however, more frequent in in the mound cemeteries of Åland from the Viking Age, such as in Långängsbacken (Kivikoski 1980: 35–38) and, for instance, in Birka, Sweden (Ambrosiani 1981). A comb decorated with a lattice-like figure was found in a Viking Age cairn in Tammenpää, Halikko, on the SW coast of Finland; it is, however, made of bone and two-sided (Hirviluoto 1992: 68).

4.3.1.2. Sundbergen 2

The cairn Sundbergen 2 has been erected on a plane horizontal rock surface inclining slightly downwards from the edges of the stone setting (fig. 13). The rock surface is of smooth-worn granodiorite with criss-crossing granite veins. On the distal side SE, the cairn is bordered by a fracture bank over whose edge boulders have fallen from the stone setting. The area of the cairn, measured along the edges of the stone setting, was 15.8 m. There were 458 boulders weighing at least 0.5 kg, and their total weight was 4.34 tons.

The stone cover consisted of fairly even-sized stones among which there were more angular pieces than in the cairn 1. After the removal of the surface vegetation and the surface layer of stones, a soil-mixed stone layer was visible. In this layer, the large boulders were mainly rounded. The boulders were in one or two layers above the bottom layer, which was composed of humus soil, with scattered edged and rounded pebbles, organic material, roots, fragments of charcoal and weathering residue. The humus layer covered the rock so that most of the stones were not directly lying on the granodiorite surface, but on an intermediate 3–4 cm thick soil layer. In the soil layer, seven iron rivets were found. In two of them, a rhomb-like piece has been preserved, typical of the rivets used in clinker boat-building. The weight of the rivets was 4.4–14.0 g and length 29–42 mm. Further, two artefacts, best characterized as nails, as well as unidentified fragments of iron, 189 g of unburnt and 1.3 g of burnt bone, a fish scale and pieces of shells were found.

The iron rivets indicate that the cairn 2 was probably built in the Merovingian or the Viking period (Anderson 1963; Raike 1996). As indicated by the boat finds, the Nordic tradition of using iron rivets in clinker boat-building was continued even longer. For instance, the clinker-built boat from the castle of Turku has been radiocarbon dated to the Middle Ages (640 80 uncal BP) (Forssell 1982). In other Nordic countries, there are rivet finds from as late as the 16th century (Westerdahl 1989: 39–42; c.f. Weski 1999).

As far as I know, these are the first finds of iron rivets in the burial cairns of the archipelago. In Sweden, rivets are known in medieval burial cairns in Misterhult (Norman 1993: 112–113); in Klinta, Öland, iron rivets have been found in a cairn dating back to the Viking Age (Schulze 1987: 58–60).

Observations made during the excavations indicate that a layer of soil has at first been brought on the rock, with edged and rounded pebbles combined with some finer mineral material and some organic material. The rivets and burnt bones were buried in this layer. On the top of the bottom layer, there was a heap of rounded and edged stones and boulders of different sizes, and possibly some soil, too. If the cairn had been built on a mineral soil, in a more fertile, humid and grazed place than the top of rocky Sundbergen, the earth layer on the ground would probably have favoured the vegetation to spread out on the cairn so that the stones would gradually have become covered. The result would have reminded greatly of the mounds on Åland with a flat stone setting inside (Kivikoski 1963; Museibyrån 1984). According to Helena Edgren (1983), there is no unambiguous limit between burial cairns and earth mounds on Åland. In many places, they occur side by side in the younger Iron Age. I regard it as probable that, as to the cairns and mounds, it is partly a question of the same type of construction, on which in different ecological habitats different humus and vegetation covers have been deposited during the past millennium.

4.3.1.3. Osteological analysis

The osteological analysis of the bone samples was made by Tarja Formisto, University of Stockholm. The results show that a young person had been buried in cairn 1 (table 2). The human bones were found in the humus soil layer on the bottom, mixed with bird and fish bones, and bones of a small mammal. Nothing indicated that the animal bones would be secondary finds, but in a burial cairn, the possibility of contamination cannot be entirely eliminated (c.f. Okkonen 1993: 6). The animal bones in the filling layer, however, are not necessarily connected with the interment.

In cairn Sundbergen 2 an adult person had been inhumed (table 3). The weight of the bone material was 44 g, mostly originating from a broken humerus. It had been placed on the ground of the construction, into the humus soil layer, near the surface of the rock, and the S edge of the cairn. The rivets and nails had also been placed mainly in the bottom layer, next to the rock surface, but they formed a scattered group about 1.5 m to the NW from the human bones.

In addition to artefacts evidently set in the humus soil layer, bones were discovered in cairn 2 under and between the stones the cairn was filled with. Among them there were five unburnt teeth which have belonged to a person of 30–40 years of age (Juha Varrela, pers.comm.[3]), 1.7 g of other human bones and 201 g of unburnt bones which have belonged to cattle, pig, sheep/goat, cat, bird, fish, and a small rodent. This leads to the question, whether the animal bones were a part of the interment. The human and animal bones could not be stratigraphically separated in situ from each other. Large animal bones lying under and between the boulders indicated that they were stratigraphically no more secondary than the human bones. Thus the animal bones are probably part of the interment and of the same age.

The cattle bones advert to the name Nötö, which refers to cattle (Sw. nöt). Kurt Zilliacus has demonstrated that the name Malmö in the so-called Danish itinerary, a medieval list of place-names along the Swedish and Finnish coasts (e.g., Cederlund 1988), refers to the present island of Nötö. According to him, the name Nötö came to be used after the 13th century when the people of Nagu started to utilize the island as grazing land for their cattle (Zilliacus 1994: 68). The cattle bones found in the cairn indicate, however, that the use of the area as a pasture probably started even earlier.

Another noteworthy domestic animal among the samples is the cat. Cat bones occur in younger Iron Age contexts, for instance in every tenth of the mounds in Kvarnbacken, Sund, on Åland (Núñez & Lempiäinen 1997; c.f. Hatting 1992[4]). As to the smallest animal bones there is a possibility of contamination. The fish bones, fish scales and shells may have found their way – long ago – into the cairn carried by fish- and mussel-catching sea birds.

What little is left of the inhumation in cairn 2, derives from the humerus and the teeth. Far advanced decomposition might be the explanation for the scarce remains, but it is also possible that the corpse when buried was not complete but in pieces. Parts of the corpse might have ended, as could be concluded from the excavation observations, up in between and mixed with the boulders, not only under the lowermost boulders.

4.3.1.4. Soil analysis

From the spots where human bones were found, samples were taken in order to determine the occurrence of rounded mineral particles, the particle size distribution and the occurrence of pollen grains[5]. The roundness of the particles was determined both from the samples taken for grain analysis and water sieved samples. In the particle size fractions 8, 4 and 2 mm, particles worn by water or wind were searched using a stereo microscope. Particles containing plenty of mica and other dark minerals were recorded since they might indicate the difference of composition between granite and granodiorite. The aim was to determine whether the mineral material was a weathering product originating from the stones and the rock surfaces or whether it had been brought from elsewhere as such or with other materials. In the 8 and 4 mm fractions, no rounded particles were found.

For cairn 1, a sample of 4661 g was screened. It contained, in addition to the mineral material, parts of plants, decaying wood, decomposed organic material, chitin shells, 2.0 g of unburnt bones of rodents, fish and bird, burnt bone in small particles (table 2) and 0.3 g of burnt clay. The sample for cairn 2 (3407 g) contained, in addition to parts of plants, decaying wood, decomposed organic material and chitin shells, bone, iron in magnetically recognizable particles, a fish scale and parts of a mollusc shell. From the water-sieved samples, ca 30 g were examined to determine the portion of rounded particles (table 3). The latter samples are not comparable to the amounts of rounded particles in the grain analysis because organic material was not removed from the water sieved samples.

When excavating the stone settings, soil samples were collected from under the lowermost boulders to identify pollen grains and to examine whether it was possible to find recognizable pollen in layers where the soil has repeatedly become dry, and oxidation has been very efficient. If pollen were found, the composition of the pollengrains could yield some evidence of the source of the soil in the cairn. Because the amounts of pollen were small and some of the pollen grains were poorly preserved, the results are presented qualitatively. Especially in cairn 2 the pollen grains were poorly preserved, but the spores, instead, were of better quality. Because the endurance of pollen grains of different species varies, the results of frequency counts would have been difficult to interpret. Positively identified groups are presented in table 4.

Differences in the grain composition of the cairn soil samples were slight but discernible. If shore sand has been brought into cairns as such or among other material, it’s proportion is larger in the more assorted sample of the cairn 2. Cairn 1 contained plenty of granite, the surface of which weathers in coarse particles whereas cairn 2 contained more of granodiorite weathering in finer products. This difference is probably reflected in the grain size distributions of the cairns.

Concluding from the roundness and mineral composition of the particles, the mineral material of the soil samples has been mainly produced through the weathering process of the boulders and the underlying rock surface. In both cairns, there was a small proportion of round-worn particles of approximately the same size that have ended up in the cairns being most probably attached to the boulders or with the soil brought into the cairns. Especially shore material can easily have remained stuck on the boulders when they were carried wet from the shore or from a wet rock cavity. In the dry sieved samples, the mineral composition is somewhat different so that the greater percentage of granodiorite in cairn 2 is reflected as a greater proportion of dark minerals. In the water sieved samples, the mean weight of rounded particles was greater than that of angular ones. The difference can be partly be explained by the fact that rounded particles pass through the water sieve more easily than angular ones. The lower percentage of humus in cairn 2 may indicate that sand has been brought from elsewhere into the cairn, even though different proportions of minerals contribute to different amounts of weathering products.

We can assume that the preconditions of deposition and preservation of the pollens of various species are approximately the same in both cairns. However, there are certain differences in the samples. Pollen of pine (Pinus) did not occur in the sample of cairn 2, even if the graves are located in the middle of a pine forest. Even if the pollen grains of cairn 2 were in a weaker condition than those of cairn 1, it is difficult to believe that the pine pollen grains would have been destroyed entirely. The stratification of cairn 2 indicates that the rock surface was at first covered with soil, probably containing pollen grains, and subsequently the soil layer was covered with stones. The soil can have been brought from an adjacent rock dell, the deposits of which would represent an earlier phase of land uplift when the island of Nötö was still a rocky skerry in the outer zone with no pine forest. This is further confirmed by the lack of spruce (Picea) and juniper (Juniperus) as well as the presence of fragments of shellfish. Probably the soil sealed the layer from external pollen precipitation. The studies by Irmeli Vuorela and Terttu Lempiäinen on the Late Neolithic site of Kotirinne, Turku, also suggest that the pollen stratigraphy may not be strongly affected by contamination in soil profiles (Vuorela & Lempiäinen 1988: 36–39). The preservation of the soil layer from contaminations supports the preceding idea that the unburnt human bones, nails, rivets and animal bones belong to the same interment.

The soil layer of the cairn 1 seems to have gone through a different formation process. No man-made soil layer was observed, and the present soil is probably the result of natural humus formation and weathering products.

The preceding analyses support the observations made during the excavation. The results suggest that a young person was buried in cairn 2 during the Late Iron Age, endowed with rivets, nails and domestic animals or parts of them. The cairn 2 has been better preserved from later external impacts than cairn 1. In the cairn 1, an adult person was buried in the Late Iron Age, equipped – at least – with a comb.

More evidence of the formation processes of burial cairns would be achieved by a further investigation into the composition of the building material and the traces of natural processes, for instance pollen stratigraphy. As to problems concerning mortuary rituals, it is essential to learn more about when and to what extent artefacts and ecofacts can be regarded as primary constituents of interments and not related to subsequent external impacts, of human or other origin. The classic problem in studies on burial cairns is the openness of the stone setting. Although this has not been investigated in Finland, I think it is safe to state that under the first centuries after the burial, oxygen, water, water-soluble compounds, and solar radiation were able to alter and destroy the remains of the burial. After the composition of humus soil on the ground or the rock surface, the remains were probably better protected.

It is conceivable that artefacts and ecofacts fallen down between the boulders may have been mixed with the remains of the interment. Animal bones found in cairns have often, indeed, been regarded as secondary (c.f. Salo 1981 b: 178). However, if possible, it should be investigated separately in each case whether bone remains are secondary or not, which requires a pedantic excavation technique. The excavation of a burial cairn involves a risk of contamination. Therefore, care should be taken that turf, litter, roots, and recent soil layers are carefully removed at the very moment of picking up each stone.

Layers of sand and gravel under and between the stones have been documented in other areas of Finland, too. Soil has been used to fill in the cavities between stones and to heap up a layer where the deceased was buried. The soil filling has been considered to be a characteristic that goes back to the Bronze Age (Edgren 1999: 330), and it seems to appear in Bronze Age and Early Iron Age contexts. For example, the Bronze Age cairns in Råckers, Karis, and the Roman Iron Age cairns in Ekeberga, Sjundeå, and Ketohaka 1, Salo, contained layers of sand and gravel (af Hällström 1948; Keskitalo 1979: 14, 36; Schauman-Lönnqvist 1989: 47–50). According to Oiva Keskitalo, the soil filling was a prevailing feature in the Late Roman Iron Age in Ostrobothnia (Keskitalo 1979: 129). The cairns in Piiloinen, Vehmaa (Early Roman Iron Age), contained soil layers, too (Salo 1968: 67–69). Thus, the soil filling in Sundbergen seems to represent a conservative characteristic in the burial tradition.