Early Neolithic human bog finds from Falbygden, western Sweden New isotopic, osteological and histological investigations

Main Article Content

Karl-Göran Sjögren Torbjörn Ahlström Malou Blank T Douglas Price Karin Margarita Frei Hege Ingjerd Hollund


Departing from two recently dated finds of human bones in wetlands from the area of Falbygden in western Sweden, this paper describes the finds in detail and sets them in a wider context of depositional practices in the south Scandinavian Early Neolithic. The two finds are both of girls in the age of 15-20 years, and in one of the cases it is probable that she was bound and possibly intentionally drowned in shallow water.

The Falbygden finds have clear parallels in Denmark and in southern Sweden in the same period. The Danish material indicates that a particular segment of the population was treated in this way, primarily young individuals around 15-20 years. Some of them show signs of disease or deformations, many have signs of trauma, and a couple have been found with cords around the neck, suggesting violent deaths. In a number of cases, they were also found in pairs.

One of the Falbygden finds, the so-called Hallonflickan, was also unusual in that her Sr isotope ratio indicates that the was born far away, probably in Scania in southern Sweden. Whether other finds of this kind were also long distance movers is not known.

Direct bone datings show that this practice is established at the same time as the introduction of agriculture and of the Funnel Beaker culture in Scandinavia, around 4000 BC, and continues until ca 3000 BC after which it disappears for ca 1000 years. Deposition of humans in wetlands is paralleled by deposition of animals, pottery and stone tools, and we suggest that these form parts of a complex of ritual practices established over a large area at the time of the earliest Scandinavian agriculture.


In this paper, we will describe two finds of human bones from Falbygden that have recently been 14C dated to the Early Neolithic. The two finds are so-called bog finds, and consist of two females. One is the so-called Hallonflickan ("the Raspberry girl") and the other is the girl from Härlingstorp. At present, these are the earliest Neolithic individuals known from Western Sweden. They are therefore of considerable interest in relation to the Neolithization of the region and because they seem to follow a practice of ritual deposition in wetlands otherwise best known from Southern Scandinavia.

Falbygden is a ca. 30 x 40 km large area of Cambro-Silurian sedimentary bedrock, well-known for its concentration of at least 255 megalithic tombs from the Middle Neolithic period, built ca. 3300--3000 cal BC. In the last decades, several investigations have focused on these graves and aspects of the society associated with them, so that this period may be regarded as comparatively well studied (Persson/Sjögren 2001; Sjögren 2003; Ahlström 2009; Axelsson 2010).

For the preceding Early Neolithic (EN), ca. 4000--3300 BC, the situation is very different. Rescue-driven fieldwork has concentrated on the west coast in the provinces of Bohuslän and Halland, where a number of Early Neolithic settlement sites have been excavated (Sjögren 2003; 2013). Dates on carbonized cereals in this area document cultivation from shortly after 4000 BC, along with finds of EN TRB materials (Sjögren 2013). Faunal remains are limited to scatters of burnt bones, enabling the determination of domesticated species and wild fauna in some cases. No quantitative evaluation is possible, however.

Inland areas, such as Falbygden, have been much less subject to rescue excavations. Pollen evidence suggests the practice of cultivation and husbandry in these areas (Fries 1958), but the archaeological material is largely limited to finds of axes, in most cases without any contextual information.

Wetland finds in Western Sweden

Due to its calcareous bedrock, Falbygden is one of the areas in Sweden where human and animal bones are often preserved. Finds from wetlands were recorded in particular during the period 1920 -- 1950, when peat cutting was a common activity, and local antiquarians were actively registering such finds. Finds of human and animal bones are now stored at the natural history museum in Gothenburg, as well as in the Falbygden museum in the town of Falköping. Other finds are scattered in various places such as at the Natural History Museum in Lund and the Statens Historiska Museum in Stockholm (SHM).

There are a number of finds of human bones in these museums, some of which have recently been subjected to 14C dating. The human bone finds in the museums are, however, only a fraction of the number of skulls and other bones actually found over the years. There are several reliable accounts of skulls or skeletons found in wetlands, which have since been lost or misplaced. For example, 3--4  skeletons were found ca. 1915 during peat cutting in a small bog in Slöta parish1. The skeletons were placed by the side of the trench, but their fate is unknown. From Lake Hornborga, there are several accounts of human crania and other bones found on the lake bottom in periods of low water. A particularly interesting account describes the find of a cranium, a femur and some animal bones associated with a stone concentration on the lake bottom in 1934, during a period of low water. The site was well documented, and the human bones were sent to SHM in Stockholm2, while the animal bones were sent to Lund. Unfortunately, none of the bones can now be found, and only the documentation remains in the museum archive in Stockholm.

Based on the description, we may suspect that this is a find from the Mesolithic period, perhaps similar to the recently excavated site at Motala Kanaljorden (Eriksson et al. 2016).

Fig. 1. The location of the research area with finds of human skeletons in Falbygden. The Bredgården and Nossamaden finds are outside the map (see Fig. 15).
Fig. 1. The location of the research area with finds of human skeletons in Falbygden. The Bredgården and Nossamaden finds are outside the map (see Fig. 15).

More well-documented finds come from several prehistoric periods. Two Mesolithic wetland finds are known from Hanaskede not far from Härlingstorp and from Bredgården some 35 km to the south of Falbygden (Vretemark 1996; Borrman et al. 1996). From the Neolithic, in addition to the two finds discussed here, a Late Neolithic skull was found at Nossamaden, some 35 km to the west of Falbygden (Hellgren 2007). There are no remains from the Bronze Age, but three finds belong to the Early Iron Age3 (Fig. 1).

In addition to the human bone finds, there is also a considerable number of finds of animal bones. Counting only finds of larger mammals (reindeer, red deer, cattle, aurochs, elk, etc.), around 50 sites are known from Falbygden.

In the coastal area, the only known Neolithic find is that from Rolfs-åker in Northern Halland (Sarauw/Alin 1923). This is the complete skeleton of a male, found during the digging of a canal in 1920. The skeleton was lying in the bottom part of a layer of clay gyttja, above an oyster shell layer. The position was not documented in situ, but canal workers reported that it had been lying on the left side, with its face downwards. Beside his head, a Neolithic stone axe of unusual type was found, and in the same layer there were also a few animal bones (red deer, cattle and beaver). Flint artefacts dating to the Mesolithic were found in the shell layer and in the underlying sand layer. The bones were examined by Carl Magnus Fürst, who concluded that they belonged to a male, ca. 25--30 years of age, of short but sturdy build (Sarauw/Alin 1923; Lindälv 1968). The man had suffered sharp force trauma on his left ulna and his right tibia, presumably from a stone axe. Pollen analyses of gyttja and clay samples indicated an environment of mixed oak forest and the absence of spruce and beech (Erdtman 1921).

In 1992, a small excavation was made by Nordqvist in order to test the hypothesis of a burial (Nordqvist 1999). Two direct 14C dates have been made. A conventional date on the left humerus yielded a date of 3516--2925 cal BC (4530±100 uncal BP, St-2488; Lindälv 1968) and an AMS date resulted in 3329--2916 BC (4430±70 uncal BP, Ua-7836) with the AMS δ13C value of -20.87 ‰ (Nordqvist 1999). More recently, a red deer humerus from the gyttja layer was dated to 3264--2895 cal BC (4370±45 uncal BP, GrA-17911, previously unpublished). Light stable isotopes in collagen were measured by Liden et al. (2004, see appendix).

A renewed osteological and odontological study was undertaken around 2000 by Leif Jonsson and Helene Borrman (Borrman 2004; Jonsson et al. 2002). This confirmed earlier sex determinations, and showed his dental health to be good, while the age estimate was reduced to ca. 20--25 years.

The original interpretation was that the man had drowned after being attacked in a boat (Sarauw/Alin 1923; Lindälv 1968). Another view has been proposed by several authors (Oldeberg 1952; Moberg 1960; Nordqvist 1999; Jonsson et al. 2002), who suggested that it was a burial connected to a settlement on the nearby small island. This is, however, difficult to reconcile with existing shoreline displacement curves, according to which the sea level at that time was at least 10 m above present, i.e. about 4 m higher than the level of the skeleton (Påsse 1987). A third possible interpretation could therefore be that we are dealing with a wetland deposit/sacrifice similar to the other finds discussed here.

Hallonflickan (the Raspberry girl)

This skeleton was discovered in May 1943 during peat digging in the Rogestorp bog, Luttra parish4, part of a large bog complex called Mönarps mosse. A preliminary report was published by Axel Bagge (1947) and a more detailed account by Gejvall et al. (1952). The find was also discussed by Ahlström/Sten (1995) and by Jankavs (2011) in connection with a facial reconstruction (Fig. 3 and 4).

When found, only the skull and a few other bones were collected, then the work was halted and the archaeologist and geologist Karl-Esaias Sahlström was called in for inspection, together with the local antiquarian Hilding Svensson. They documented the stratigraphy, took pollen samples and extracted the remaining skeleton and surrounding peat as a block for later excavation indoors, together with a few loose bones found in the peat trench. The block was transported to Statens Historiska Museum in Stockholm and was excavated later the same year by the osteologist Elias Dahr. During this excavation, a concentration of raspberry seeds in the stomach region was noted, assumed to be part of her last meal. She was therefore assumed to have died in late summer.

The stratigraphy of the find spot was determined as ca. 1.15 m of marsh peat, underlain by chalk gyttja containing freshwater molluscs. The position of the skull could be seen in the trench wall, ca. 1.2 m below the surface, corresponding to the lowermost part of the peat. The environment at the time of deposition was suggested to have been a chalk marsh with stands of sedge (Cladium mariscus), some shallow open water between the sedge tufts, and a seasonally fluctuating water level (Gejvall et al. 1952).

In the same peat trench, a flint blade arrowhead of MN type had been found three years earlier. The location of this arrowhead was ca. 6 m to the north of the skeleton, some 1.2 m deep in the peat.


Several attempts have been made to date this skeleton. A dating to the Middle Neolithic (the "passage grave period") was suggested by Gejvall et al. (1952) on the basis of pollen analysis and by Bagge (1947) by association with the blade arrowhead.

Labno BC cal, 2 s BP uncal δ13Cams δ13Cdiet δ15Ndiet C/N Sample
St-2861 3354-2697 4350±100 - - - Not measured unknown
Ua-3962 3328-2881 4360±65 -20.48 - - Not measured Left humerus
UBA-30518 3928-3651 4964±42 - -20.6 8.4 3.24 Left femur
Tab. 1. The 14C datings and light stable isotope values from Hallonflickan.
Fig. 2. Calibration plot of the dates from Hallonflickan.
Fig. 2. Calibration plot of the dates from Hallonflickan.

Three 14C dates have been measured on bone collagen from the skeleton (Tab. 1, Fig. 2). A first 14C dating was made at the Natural History Museum in Stockholm in 1969 with the conventional method (Oldeberg 1976). It is unclear which bones were used, but the ribs mentioned by Sahlström are now missing so this is a possibility. In 1994, an accelerator dating was done in Uppsala on the left humerus (Ahlström/Sten 1995). Finally, a second accelerator dating was undertaken in Belfast in 2015 on a sample from the left femur.

The two first dates agree, while the Belfast date is considerably older. This difference may be due to different analytical methods (conventional vs. accelerator), but more likely to developments in pre-treatment and collagen extraction protocols. Collagen extraction at the Stockholm lab consisted only of acid treatment and washing in distilled water (Sellstedt et al. 1966). A similar method was used in the Uppsala lab (dating certificates from Uppsala). The Belfast lab employs more recent methods with a series of acid and base cleaning steps including ultrafiltration. Furthermore, they measure the C/N ratio and the δ13C and the δ15N values, which are important parameters to monitor collagen quality and sea/freshwater reservoir effects. We therefore assume that the Stockholm and Uppsala dates are affected by remnant contaminants, such as humic acids, and we regard the Belfast date as more reliable.

Calibration of this date provides the range of 3928--3651 cal BC (95 %), corresponding to the early or middle part of the Early Neolithic. A low δ15N value suggests low freshwater fish consumption and thus also a marginal impact from the freshwater reservoir effect. This re-dating indicates that the Raspberry girl is some 700 years older than previously assumed, and she is at present the earliest known Neolithic human from Western Sweden.


Fig. 3. Hallonflickan, photo of cranium (photo by Gunnar Creutz).
Fig. 3. Hallonflickan, photo of cranium (photo by Gunnar Creutz).

Elias Dahr, the first anthropologist to analyse the skeletal remains from the Rogestorp bog, assessed the skeleton as a young female (Bagge 1947). However, it was Gejvall, Hjortsjö and Sahlström (Gejvall et al. 1952) who produced a proper anthropological investigation and description of the skeleton. Ahlström/Sten (1995) added an x-ray of the maxillary third molar. The sexual assessment of the skeleton has been constant through the years. The skeleton is gracile, with diminutive secondary sexual characters of the skull, suggesting a female (Fig. 3). With respect to the pelvis, the right ilium has a broad incisura ischiadica major, with a double arc composé suggesting a female. The left ilium is not so well preserved, but a broad incisura ischiadica major is present. Thus, apart from the gracile cranium, we also have evidence for an assessment of a female based on the pelvis.

Fig. 4. Facial reconstruction of Hallonflick-an by Oscar Nilsson (after Jankavs 2011).
Fig. 4. Facial reconstruction of Hallonflick-an by Oscar Nilsson (after Jankavs 2011).

While the sexual assessment of the skeleton is seemingly straightforward, the aging is not. Gejvall and Hjortsjö assessed the skeleton as an adult, 20--25 years of age, based on epiphyseal fusion, eruption of third molars (the lower third molar being erupted and the upper third molar in the process of eruption), and synostosis of endocranial sutures. In fact, with respect to the attrition of the teeth and the eruption of the third molars, they even suggested an age of 25--30 years. Ages for epiphyseal fusion were based on Ruckensteiner's (1931) roentgenological assessment of the skeleton, claiming fusion of the epiphysis much later than modern synthesis based on osteology (see below). One adult feature of the skeleton, the fusion of the synchondrosis sphenooccipitalis, was not commented upon by them.

Bone and side Epiphysis fused Epiphysis unfused
Fibula, unsided Proximal (12-17 yrs), Distal (12-15 yrs)
Left femur Trochanter major (14-16 yrs) Caput femoris (12-16 yrs), Distal (14-18 yrs)
Right femur Trochanter major (14-16 yrs) Caput femoris (12-16 yrs), Distal (14-18 yrs)
Right pelvis Ischiadicum (16-18 yrs) Crista iliaca (20-23 yrs)
Left radius Distal (14-17 yrs)
Right radius Capiut radii (11,5-13 yrs) Distal (14-17 yrs)
Left ulna Distal (15-17 yrs)
Right ulna Distal (15-17 yrs)
Right humerus Distal (11-15 yrs) Caput (13-17 yrs)
Tab. 2. Fusion of epiphyses in the skeleton of Hallonflickan. Female fusion age estimates after Cunningham et al. (2003) and Krogman/Iscan (1986).

It is the opinion of the present author (TA) that the age estimate provided by Gejvall et al. (1952) was too high. Some epiphyses that could be documented are fused (Tab. 2), and others are unfused. With respect to the unfused epiphyses, the granular appearance of the subchondral bone is evident, with no signs of osseous bridges connecting the epiphysis with the diaphysis. Thus, complete fusion is not at hand for several of the epiphyses. Furthermore, the utility of the synchondrosis sphenooccipitalis has been questioned. A meta-analysis published by Krishan/Kanchan (2013) questions whether the fusion of this site can be used for aging at all, as it has been showed to vary considerably. Ahlström/Sten (1995) provided an x-ray of the erupting third molars of the maxillae and demonstrated that the root was not fully developed. Thus, weighing the age indicators present, it is apparent that the skeleton belongs to the age category 15--20 years, and probably the later span of this age category, 18--19 years, given the eruption of the third molars.


Changes in the bone material after death (bone diagenesis) reflect post-mortem events such as funerary treatments and fluctuations in the burial conditions. Some of these changes can be observed and assessed in histological thin-sections. As such, post-mortem histories can be partially reconstructed by so-called histotaphonomy (Turner-Walker/Jans 2008; Booth 2016; Hollund et al. 2012; 2013). A transversal bone section taken from the mid-shaft of the left femur was prepared for histological characterization of diagenetic alterations. The thin-section was studied in normal and polarized transmitted light. Alterations observable at a microscopic scale include bioerosion, inclusions, infiltrations (staining), microcracks and birefringence. Birefringence is an optical property of bone reflecting preservation of the bone protein (Jans 2005; Grupe/Dreser-Werringloer 1993; Hedges et al. 1995).

Fig. 5. Overview of the sample, normal (left) and polarized (right) light (photo by Hege Ingjerd Hollund).
Fig. 5. Overview of the sample, normal (left) and polarized (right) light (photo by Hege Ingjerd Hollund).

The results of the diagenetic characterization of the Hallonflickan bone are summarised in Tab. 3. The microstructure of the bone is well-preserved with no bioerosion and displays the typical birefringence of fresh bone when viewed in polarized light, i.e., alternating light and dark bands (Fig. 5). The birefringence is normal in most parts of the section, but reduced in a dark orange-brown stained band along the surface suggesting some loss of protein in this area. When viewed in normal light, the bone displays a light orange-brown stain across the whole depth, probably due to infiltration by humic factors, iron and/or manganese compounds.