Introduction
Hunting pits, a significant archaeological feature, are abundant in Sweden, with over 30,000 documented in the National Heritage Boards' database for archaeological sites and monuments (KMR). They are not exclusive to Sweden, as the practice of digging pits to trap animals is well represented in Fennoscandia (e.g. Hvarfner, Reference Hvarfner1965; Myrvoll et al., Reference Myrvoll, Thuestad and Holm-Olsen2011) and is also known elsewhere in the world (Lemke, Reference Lemke2021: 1–2). Such pits could also have been used for several purposes, such as storage pits or hunting blinds (Smith, Reference Smith2013; Lemke, Reference Lemke2021). In a broader context, diverse hunting techniques relying on structures have been and are still being used globally. These structures, made of perishable or more durable material, include kites, drive lanes, hunting blinds, and flags, all designed to guide the animals to a kill site or to allow the hunters to get close to the prey. The type of material used has clearly had an influence on the survival of such structures in the archaeological record (e.g. Stewart et al., Reference Stewart, Keith and Scottie2004; Klaussen, Reference Klaussen2008; Smith, Reference Smith2013; Lemke, Reference Lemke2021). Hunting pits, pitfalls, or trapping pits can be described as oval or circular, more or less steep depressions surrounded by a low soil embankment. The size of the pits ranges from two to six metres in diameter, and the depth can be up to two metres (Figure 1). A common feature is that they are arranged in rows in often large numbers of pits to catch wild reindeer or moose (Mulk, Reference Mulk1994: 160–61, Reference Mulk and Andersen2005; Spång, Reference Spång1997: 51; Halinen, Reference Halinen2005; Klaussen, Reference Klaussen2008: 5–6; Myrvoll et al., Reference Myrvoll, Thuestad and Holm-Olsen2011: 3; Solli, Reference Solli2018: 10–11; Hennius, Reference Hennius2020: 5–6). In northern Fennoscandia, hunting pits are often organized in systems and located near large river valleys, animal migration routes, or narrow passages (Vorren, Reference Vorren1958; Manker, Reference Manker1960; Selinge, Reference Selinge1979; Mulk, Reference Mulk1994, Reference Mulk and Andersen2005; Spång, Reference Spång1997; Klaussen, Reference Klaussen2008). These hunting structures have yielded radiocarbon dates from the Mesolithic and Neolithic (8000–1800 bc) to modern times (Hennius, Reference Hennius2020: 1). Although they are common archaeological features, hunting pits have mainly been documented from a morphological perspective (e.g. Hansson & Rahtje, Reference Hansson and Rathje1999: 23–24), because the lack of dateable material has hampered further research initiatives. However, chronologies have been elaborated over the years (Halinen, Reference Halinen2005; Ramqvist, Reference Ramqvist2007; Hennius, Reference Hennius2020). Furthermore, there is a general lack of studies addressing the hunting pits’ spatial articulation that would contextualize them as material manifestations of landscape organization (Länsstyrelsen i Norrbottens län, 1998: 59; Hansson & Rahtje, Reference Hansson and Rathje1999: 23–24; Klaussen, Reference Klaussen2008: 1–2; Myrvoll et al., Reference Myrvoll, Thuestad and Holm-Olsen2011). Consequently, our general understanding of hunting pits and the societies that use them is limited. The emergence of concepts, such as landscape domestication, and techniques, such as Geographical Information System (GIS), as well as advances in radiocarbon dating are, however, advancing our understanding of the phenomenon.
Two examples of preserved hunting pits.
Over the last decade, archaeologists have developed the concept of landscape domestication as a theoretical framework for understanding human land use, primarily in non-agrarian environments (e.g. Erickson, Reference Erickson, Balée and Erickson2006; Eriksson et al., Reference Eriksson, Ekblom, Lane, Lennartsson, Lindholm, Crumley, Lennartsson and Westin2017). Landscape domestication derives from the insight that in agrarian settings the results of human land use are relatively apparent and straightforward, e.g. cultivation, irrigation, and enclosures (Terrell et al., Reference Terrell, Hart, Barut, Cellinese, Curet and Denham2003: 329–51; Smith, Reference Smith2011; Eriksson et al., Reference Eriksson, Ekblom, Lane, Lennartsson, Lindholm, Crumley, Lennartsson and Westin2017: 155–56). In non-agrarian milieus, such impacts on the landscape are, in comparison, subtler and more difficult to capture. Often, such environments are considered to represent wilderness, which has implications for understanding the communities that live in and manage these landscapes (Bergman et al., Reference Bergman and Hörnberg2015: 57–59; Lindholm et al., Reference Lindholm, Ersmark, Hennius, Lindgren, Loftsgarden, Svensson and Symes2021: 9–11). Landscape domestication aims to bridge the nature–culture dichotomy by capturing the consequences of human practices and identifying related ecological processes that shape non-agrarian landscapes (Terrell et al., Reference Terrell, Hart, Barut, Cellinese, Curet and Denham2003; Erickson, Reference Erickson, Balée and Erickson2006: 235–37; Widgren, Reference Widgren2012: 117–21; Eriksson et al., Reference Eriksson, Ekblom, Lane, Lennartsson, Lindholm, Crumley, Lennartsson and Westin2017: 148–56). This framework has contributed to enhancing our understanding of how low-intensity land use by small-scale societies structured the landscape beyond the procurement of resources (Odling-Smee et al., Reference Odling-Smee, Lala and Feldman2003; Terrell et al., Reference Terrell, Hart, Barut, Cellinese, Curet and Denham2003: 347–50; Widgren, Reference Widgren2012: 122; Eriksson et al., Reference Eriksson, Ekblom, Lane, Lennartsson, Lindholm, Crumley, Lennartsson and Westin2017: 155–56, Reference Eriksson, Arnell and Lindholm2021). Moreover, landscape domestication can also involve structural modifications, including built environments (Erickson, Reference Erickson, Balée and Erickson2006: 244–45; Smith, Reference Smith2011: 837–44; Lemke, Reference Lemke2021: 1–2). Hunting pits are an example of such a landscape modification, prominent in the circumpolar north (Klaussen, Reference Klaussen2008; Myrvoll et al., Reference Myrvoll, Thuestad and Holm-Olsen2011; Hennius, Reference Hennius2020).
The central idea discussed here is that the hunting pit systems should be understood as part of an elaborate hunting technique involving the whole landscape. A detailed understanding of such systems can provide an opportunity to better comprehend the nature of land use and, by extension, the communities that constructed them. Hence my wider concern is to discuss how hunting pits can contribute to a better grasp of landscape domestication by small-scale societies. To achieve this, these features require research approaches that go beyond the pits' morphology and chronology, focusing more on how societies used pitfalls for organizing and structuring their environment. Subtle human land-use practices can be extensive, as demonstrated by the large number, concentration, and distribution of hunting pits. This comprehensive land use and its processes and consequences need to be discussed by addressing the landscape as a whole, to understand how hunting pits embody and manifest aspects of landscape domestication by hunting communities.
I shall first review previous research on hunting pits and follow with a theoretical discussion. Next, I present the spatial data collected from roughly 1500 hunting pits in the Arctic region of Sweden. These data will form the basis of a discussion of extensive land use to characterize landscape domestication in northern Fennoscandia.
Previous Research on Hunting Pits in Fennoscandia
Hunting pits have been of interest to researchers since the first half of the twentieth century. Initially, the primary focus was on documenting their construction and distribution (e.g. Vorren, Reference Vorren1958; Manker, Reference Manker1960; Hvarfner, Reference Hvarfner1965; Barth, Reference Barth1981), followed by morphological aspects and the animals targeted (e.g. Selinge, Reference Selinge1979; Forsberg, Reference Forsberg1985; Hansson & Rahtje, Reference Hansson and Rathje1999). From the early 1980s onwards, the use of radiocarbon dating presented opportunities for exploring hunting pits, their chronologies, and interpretations of the fluctuations in their construction and use over time (Forsberg, Reference Forsberg1985; Mulk, Reference Mulk1994; Furset, Reference Furset1995; Spång, Reference Spång1997; Halinen, Reference Halinen2005; Ramqvist, Reference Ramqvist2007). Societal interpretations range between land use by hunter-gatherers, by Sami societies, or outland use by farmers (e.g. Mulk, Reference Mulk1994; Ramqvist, Reference Ramqvist2007; Wehlin, Reference Wehlin and Carlsson2016). However, a lack of stratified and datable material makes it difficult to date hunting pits, resulting in sparse and inconclusive dates. It is also relatively complex to establish links between hunting pits and other sites in the landscape, such as, for example, dwelling or butchering sites, since only a few are known and dated (e.g. Hansson & Rahtje, Reference Hansson and Rathje1999: 35–37; Klaussen, Reference Klaussen2008; Myrvoll et al., Reference Myrvoll, Thuestad and Holm-Olsen2011: 4; Jonsson, Reference Jonsson2022: 18–19). Despite these challenges, a refined chronology of hunting pits in Sweden using Kernel Density Estimation (KDE) modelling has recently been proposed (Hennius, Reference Hennius2020). Most of the radiocarbon-dated hunting pits in northern Sweden range between the Late Neolithic and the Middle Ages (e.g. Mulk, Reference Mulk1994, Reference Mulk and Andersen2005; Jonsson, Reference Jonsson2022), with a majority of dates in the region under study in the first millennia bc and ad predating the domestication of reindeer (Mulk, Reference Mulk1994: 160–63, Reference Mulk and Andersen2005: 47–49). Since the radiocarbon-dated pits are relatively few at this stage, compared to the overall number of pits in the area, conclusive interpretations represent a challenge. The pits appear to remain fixed in the landscape while fluctuating in time (Widgren, Reference Widgren2012: 123).
The most thorough account of exploring the spatial distribution of hunting structures, strategic hunting, and the interplay between different hunting strategies and societies has been compiled for Norway (eg. Klaussen, Reference Klaussen2008; Myrvoll et al., Reference Myrvoll, Thuestad and Holm-Olsen2011; Solli, Reference Solli2018). Consequently, researchers have identified a substantial need for contextual studies concerning societal and spatial organization all over Fennoscandia (Länsstyrelsen i Norrbottens län, 1998: 59; Hansson & Rahtje, Reference Hansson and Rathje1999: 23–25, 40; Klaussen, Reference Klaussen2008: 83–87; Sommerseth, Reference Sommerseth2009: 248; Myrvoll et al., Reference Myrvoll, Thuestad and Holm-Olsen2011: 4).
Theoretical Framework
Hunting pits are the residue of an elaborate hunting technique involving the entire landscape. Here, I propose that one way to capture their spatial extent and multitemporal nature is to apply the concept of ‘landscape domestication’ (LD). It is a conceptual framework comprising various human–environmental interactions that recognizes human investments as resulting in more productive and worked landscapes (Terrell et al., Reference Terrell, Hart, Barut, Cellinese, Curet and Denham2003: 325–30; Erickson Reference Erickson, Balée and Erickson2006: 235–36; Widgren, Reference Widgren2012: 117–24; Eriksson & Arnell, Reference Eriksson and Arnell2016: 79). LD can be defined as ‘all nongenetic, intentional and unintentional practices and activities of humans that transform local and regional environments into productive, physically patterned, cultural landscapes for humans and other species’ (Erickson, Reference Erickson, Balée and Erickson2006: 241). It represents processes of land use that enhance subsistence similar to those in an agricultural landscape in terms of maintenance, social and cultural dynamics, and labour aggregation, but in a subtler form (Terrell et al., Reference Terrell, Hart, Barut, Cellinese, Curet and Denham2003; Erickson, Reference Erickson, Balée and Erickson2006; Smith, Reference Smith2011; Widgren, Reference Widgren2012; Clement, Reference Clement and Smith2014; Lindholm et al., Reference Lindholm, Ersmark, Hennius, Lindgren, Loftsgarden, Svensson and Symes2021). Furthermore, a domesticated landscape is connected to resource management relying on particular skills (Smith, Reference Smith2011, Reference Smith2013: 10–11; Eriksson et al., Reference Eriksson, Ekblom, Lane, Lennartsson, Lindholm, Crumley, Lennartsson and Westin2017, Reference Eriksson, Arnell and Lindholm2021: 2; Hatlestad et al., Reference Hatlestad, Wehlin and Lindholm2021). Within this framework, pit systems embody a hunting land use that targets wild game.
Additionally, we can understand LD within the wider practice of niche construction (NC), that is, processes where a reciprocal relationship between organisms and their environment alters their own and other species' niches (Eriksson & Arnell, Reference Eriksson and Arnell2016: 78). Niche construction has enhanced our understanding of how organisms actively transform their environments and how human land-use practices can contribute to adaptive processes. The processes of both adaptation and transformation are closely linked to labour and knowledge since human bio-social evolution consists of genetic, cultural, and ecological inheritance, transformed by memory and learning (Odling-Smee et al., Reference Odling-Smee, Lala and Feldman2003; Laland & ÓBrien, Reference Laland and O'Brien2011; Riede, Reference Riede2011; Laland et al., Reference Laland, Matthews and Feldman2016; Eriksson et al., Reference Eriksson, Ekblom, Lane, Lennartsson, Lindholm, Crumley, Lennartsson and Westin2017, Reference Eriksson, Arnell and Lindholm2021: 2–3; Hatlestad et al., Reference Hatlestad, Wehlin and Lindholm2021). From these perspectives, I define LD as knowledge processes resulting in management with spatial and temporal manifestations. Low-intensity land use practices also impact and transform landscapes, manifesting human agency. This agency becomes a part of the reciprocal adaptive processes through knowledge transmission.
Small-scale societies, like hunter-gatherers and pastoralists, leave more subtle traces than most agricultural communities, which makes them more challenging to document and comprehend archaeologically, marginalizing them in our narratives (Smith, Reference Smith2011; Bergman et al., Reference Bergman, Zackrisson and Liedgren2013; Lindholm et al., Reference Lindholm, Ersmark, Hennius, Lindgren, Loftsgarden, Svensson and Symes2021). Yet even subtle land-use strategies like fire management to maintain hunting grounds or to increase grazing, small-scale rotational cultivation, and hunting structures transform landscapes. Detecting and interpreting these activities often requires interdisciplinary approaches, combining archaeology with palaeoecology, ecology, forest history, and more (Terrell et al., Reference Terrell, Hart, Barut, Cellinese, Curet and Denham2003; Widgren, Reference Widgren2012; Eriksson et al., Reference Eriksson, Ekblom, Lane, Lennartsson, Lindholm, Crumley, Lennartsson and Westin2017, Reference Eriksson, Arnell and Lindholm2021; Lindholm et al., Reference Lindholm, Ersmark, Hennius, Lindgren, Loftsgarden, Svensson and Symes2021). Such studies have questioned the division between mobile hunters and settled farmers and raised the profile of marginal areas, challenging ideas about farming and the concept of cultivation (Bergman et al., Reference Bergman and Hörnberg2015: 62–63; Bergman, Reference Bergman2018: 16–26; Eriksson et al., Reference Eriksson, Arnell and Lindholm2021: 19; Lindholm et al., Reference Lindholm, Ersmark, Hennius, Lindgren, Loftsgarden, Svensson and Symes2021: 28). Domestication processes dating back over 10,000 years have been identified in European landscapes (Odling-Smee et al., Reference Odling-Smee, Lala and Feldman2003; Smith, Reference Smith2011; Nikulina et al., Reference Nikulina, MacDonald, Scherjon, Pearce, Davoli and Svenning2022: 990–91). In addition to the more direct ecosystem alterations usually found in agricultural environments, researchers have drawn attention to the significance of suitable frameworks, including LD, to analyse subtle and widespread landscape changes (eg. Widgren, Reference Widgren2012; Lindholm et al., Reference Lindholm, Ersmark, Hennius, Lindgren, Loftsgarden, Svensson and Symes2021).
Current research has challenged old understandnings, small-scale and low-intensity land use also resulted in engineered ecosystems. For example, ecological globalization and resource colonization (i.e. exploitation of resources in the form of various land-use activities in non-agrarian landscapes) have strongly altered the boreal forests of central Sweden over at least the last 3000 years (Lindholm & Ljungkvist, Reference Lindholm and Ljungkvist2016; Hennius, Reference Hennius2020; Lindholm et al., Reference Lindholm, Ersmark, Hennius, Lindgren, Loftsgarden, Svensson and Symes2021). Human management of these forests spatially and ecologically reorganized them and provided diversified land-use systems spanning environmentally unique upland and lowland regions, resulting in a more predictable environment. Through these land use systems, marginal areas and small-scale societies became interconnected to centralized regions (Eddudóttir et al., Reference Eddudóttir, Svensson, Nilsson, Ekblom, Lindholm and Johansson2021; Hatlestad et al., Reference Hatlestad, Wehlin and Lindholm2021; Lindholm et al., Reference Lindholm, Ersmark, Hennius, Lindgren, Loftsgarden, Svensson and Symes2021; Wehlin et al., Reference Wehlin, Ljunge, Larsson, Eddudóttir, Ekblom and Erikson2023). This kind of strategic management of resources is also evident in the hunting pit landscapes (Widgren, Reference Widgren2012: 122–23; Eriksson et al., Reference Eriksson, Ekblom, Lane, Lennartsson, Lindholm, Crumley, Lennartsson and Westin2017; Lindholm et al., Reference Lindholm, Ersmark, Hennius, Lindgren, Loftsgarden, Svensson and Symes2021: 10–11, 19; Eriksson, Reference Eriksson2023).
In the north of Sweden, several research projects have addressed land use in non-agrarian landscapes, exploring socioeconomic networks and identity processes, which are relevant to this study in that they examine small-scale societies as agential players in the landscape (Bergman et al., Reference Bergman, Zackrisson and Liedgren2013; Bergman & Hörnberg, Reference Bergman and Hörnberg2015; Hörnberg et al., Reference Hörnberg, Josefsson, Bergman, Liedgren and Östlund2015; Östlund et al., Reference Östlund, Hörnberg, DeLuca, Liedgren, Wikström and Zackrisson2015; Crumley, Reference Crumley2017; Josefsson et al., Reference Josefsson, Hörnberg, Liedgren and Bergman2017; Bergman, Reference Bergman2018). For instance, deliberate burning to favour the growth of ground lichens has challenged the view on natural fire intervals and the capacity of overall management by small-scale societies (DeLuca et al., Reference DeLuca, Zackrisson, Bergman and Hörnberg2013; Hörnberg et al., Reference Hörnberg, Josefsson, DeLuca, Higuera, Liedgren and Östlund2018). Moreover, newly discovered iron production in northern Fennoscandia in the Early Iron Age challenge normative narratives concerning hunter-gatherers in the Arctic (Bennerhag et al., Reference Bennerhag, Hagström Yamamoto and Söderholm2023).
Research Rationale
Hunting pits show elements of systematic planning and significant organization, forming a domesticated landscape. Consequently, defining criteria for identifying aspects of a domesticated landscape is essential. I suggest four significant traits that characterize the hunting pit systems in this study: knowledge, spatial manifestations, temporal manifestations, and coordinated and collaborative labour (Table 1) (Terrell et al., Reference Terrell, Hart, Barut, Cellinese, Curet and Denham2003: 325–34; Smith, Reference Smith2011: 836–38, Reference Smith2013: 10–11; Widgren, Reference Widgren2012: 121–23; Lindholm et al., Reference Lindholm, Ersmark, Hennius, Lindgren, Loftsgarden, Svensson and Symes2021). Of these four aspects of LD, showing how a domesticated landscape can be identified in the field, I shall focus on two characteristics: knowledge and spatial manifestations.
Characteristics of LD.
Study Area and Targeted Species
The archaeological data in this study consist of hunting pits encountered in the western parts of the municipalities of Gällivare and Jokkmokk in the regions of Swedish Sápmi, above the Arctic Circle. The topography in these areas comprises mountains and hilly forests with gentle to steep slopes, predominately oriented towards the south-east. The vegetation zones are Arctic and northern boreal, mainly composed of mountain birch forests with spruce and pine in the lower regions (Sjögren et al., Reference Sjögren and Damm2019). These Arctic conditions are more suitable for cervids like the wild reindeer, Rangifer tarandus (extinct in Sweden), and the domesticated reindeer, Rangifer tarandus tarandus, than moose (Bjärvall & Ullström Reference Bjärvall and Ullström2010; Kullman, Reference Kullman2016; Ekholm, Reference Ekholm2024). Wild reindeer disappeared in northern Sweden during the eighteenth and nineteenth centuries, and the chronological distribution suggests a simultaneous use of both wild and domesticated reindeer up to then. Written records are few in these areas, and maps are even more scarce. Nevertheless, records show disputes over hunting pits in the late 1600s in the coastal regions (Bergman, Reference Bergman2018: 68). Reindeer are herd animals, with predictable behaviour and regular migration routes, making them suitable for large-scale hunting. This predictability is owed to the reindeer's congenital curiosity and natural pattern recognition, making them follow and move along specific features and formations. They are also drawn to linearity and heights, gathering on and following ridges and narrow passages in high or hilly terrain, and spreading out on flat ground (Ingold, Reference Ingold1980; Forsberg, Reference Forsberg1985; Klaussen, Reference Klaussen2008; Smith, Reference Smith2013; Lemke, Reference Lemke2021). However, prehistoric wild reindeer migration patterns in the Nordic region have not been identified (Myrvoll et al., Reference Myrvoll, Thuestad and Holm-Olsen2011: 11), making the analogy to the seasonal movement of domesticated reindeer ambiguous. Instead, the predictability concerns the reindeer's congenital traits (Smith, Reference Smith2013: 10).
Data and Methods
Field surveys and GIS analyses were undertaken to address the vast landscape transformations that the hunting pit systems represent. This makes it possible to study large areas and analyse complex distribution and organization processes, essential to understand the nature of human land use at landscape scale (Löwenborg, Reference Löwenborg2010; Risbøl & Gustavsen, Reference Risbøl and Gustavsen2018).
Before conducting field surveys, spatial analyses were undertaken in GIS. For the analyses, hillshade (an open LiDAR data digital elevation model (DEM) with a 2+ metre resolution), aerial images, and historical aerial images from the Lantmäteriet (Swedish Land Survey Authority) were used to map hunting pits. In previous studies, detecting pit structures using LiDAR has proved promising in northern landscapes, especially for hunting pits. Detectability is especially good in areas with flat terrain and low vegetation cover, as opposed to dense forests, slopes, or boulder terrain (Norstedt et al., Reference Norstedt, Axelsson, Laudon and Östlund2020; Trier et al., Reference Trier, Reksten and Løseth2021; Thuestad et al., Reference Thuestad, Risbøl, Kleppe, Barlindhaug and Myrvoll2021).
These spatial analyses were conducted in the World Heritage Site of Laponia (9400 km2) and adjacent areas. The main focus was the eastern parts because local people had provided valuable information about possible hunting pits in the Sijdojávrre area (Figure 2). The regions and locations for conducting field surveys were selected from areas where the hillshade layer indicated a high density of hunting pits. These areas were visited to verify the pit structures. In addition, extensive field surveys were conducted in the surrounding regions during other field projects within the Heritage Site. Using DEMs proved effective where conditions were positive for LiDAR. However, additional hunting pits were almost always identified during fieldwork, given that the vegetation cover and the fluctuating measuring points affected the resolution in the hillshade layer.
Location of study areas 1 and 2 (map data ©Lantmäteriet).
Results
This section presents the primary results from two survey areas (1 and 2 on Figure 2). Area 1 will serve as the main case study due to the quantity, density, and extent of the hunting pit systems. The study areas exhibit three main characteristics: a large number, concentration, and extent of hunting pits, a selective spatial distribution, and a strategic articulation (Figures 3 and 4).
Location and distribution of hunting pit systems in Area 1, Jokkmokk municipality (map data ©Lantmäteriet).
The systems in Area 2, Gällivare municipality (map data ©Lantmäteriet).
The field survey identified sixteen comprehensive hunting pit systems comprising approximately 1500 hunting pits previously unrecorded in the Arctic region of Sweden, distributed in two areas: Area 1 contains seven systems (A–G), which includes approximately 1000 hunting pits over 228 km2. Area 2 comprises nine systems and about 500 pits over 660 km2. No pits have so far been excavated or dated (radiocarbon dating will be discussed in a forthcoming article). The number of pits ranges from forty to several hundred; systems A, D, E, and F in Area 1 have the most significant numbers, ranging between 110 to more than 300 (Table 2).
Numbers, altitudes, topographical zones, and location categories for hunting pit systems in Area 1.
These systems are located at the rim of the western mountain chain in the boreal zone, where two mountain valleys come together. The distribution pattern shows concentrations at the end of mountain ridges (category L1 in Table 3) and between mountains in open depressions or small valleys (L2). These categories are visible in systems C, D, E, F, and G, with L3 and L4 being the subcategories mainly present in systems A, B, and C (Figure 3 and Table 3).
Location categories for hunting pit systems in Area 1.
The spatial distribution shows extensive use of specific topographical zones. The definition of zones is based on earlier research in the Swedish mountain areas to categorize the distribution of hunting pits (Mulk, Reference Mulk and Andersen2005: 45). Zone 1 represents the boreal zone below 400 m asl, Zone 2 represents the coniferous and mountain birch forests in elevations around 400–500 m asl, and Zone 3 represents the alpine region and areas around the tree line (Table 4). No hunting pit systems were found in topographical Zone 1 but they were found at elevations of around 500–800 m asl in Zones 2 and 3 (Table 4).
Topographical zones, based on Mulk, Reference Mulk1994 and Reference Mulk and Andersen2005.
The spatial organization of the systems is linear with a central axis, i.e. the rows of pits follow a principal direction, divided into complex sections, an ubiquitous feature in Fennoscandia (Klaussen, Reference Klaussen2008: 65–69). In all the systems, this alignment is oriented N–S, with fluctuations towards NE–SW to NW–SE, where the central axis often follows ridges, edges, or small waterways that are themselves linear. About eighty per cent of the hunting pits in the case study are oval, which is the most common shape in northern Fennoscandia (Manker, Reference Manker1960; Mulk, Reference Mulk1994, Reference Mulk and Andersen2005; Spång, Reference Spång1997; Klaussen, Reference Klaussen2008). However, the present-day shape of the pits is heavily influenced by erosion and root disturbance, making it sometimes challenging to estimate the original structure.
A more significant and reliable feature in these systems is the orientation of the base of the pits. Their shape is mainly rectangular and almost exclusively aligned, perpendicular to the central axis of the systems, as shown in system G (Figure 5).
Alignment of the base of the pits in system G, Area 1 (map data: ©Lantmäteriet and ©Riksantikvarieämbetets Kulturmiljöregister, KMR).
The results show that the most frequently encountered spatial distribution of hunting pits in systems A–G is at the end of mountain ridges in locations of type L1 and L2, occupying topographical Zones 2 and 3. Moreover, the most prevalent layout is a linear orientation on a N–S central axis and a perpendicular orientation of the base of the pits. Field surveys and GIS analyses thus show that these methods are well suited to detecting a substantial proportion of structures likely to be present in the region and that they reflect the areas most likely to include hunting pits.
Discussion
The four LD characteristics introduced earlier—knowledge, spatial articulation, temporal manifestations, and coordination and collaboration of labour—will be discussed in turn to examine how aspects of a domesticated landscape can be identified.
Knowledge
The systems in Areas 1 and 2 exhibit a distinctive pattern with striking similarities and cohesion regarding their spatial arrangement and organization (Figures 3 and 4). They all show an awareness of reindeer behaviour and an understanding of unpredictable factors, i.e. careful planning that involved the whole landscape.
Resource management requires a comprehensive knowledge of the natural environment, such as topography and the behaviour of the targeted resource, something that ethnographic and anthropological studies in the circumpolar north have shown (e.g. Odling-Smee et al., Reference Odling-Smee, Lala and Feldman2003; Smith, Reference Smith2013: 10–11). Traditional ecological knowledge, practices, and beliefs concerning the relationship between living beings and their environment, and engineering skills, were passed down from generation to generation, i.e. were transferred to individuals and groups (Berkes, Reference Berkes2008; Smith, Reference Smith2013: 10–11; Eriksson et al., Reference Eriksson, Ekblom, Lane, Lennartsson, Lindholm, Crumley, Lennartsson and Westin2017). From this perspective, the systems in the study are consistent with a knowledge of reindeer because their siting conforms to the behaviour of reindeer and is aligned with current and historical migration routes (Klaussen, Reference Klaussen2008; Sommerseth, Reference Sommerseth2009; Smith, Reference Smith2011, Reference Smith2013; Lemke, Reference Lemke2021). The topography, climate, and vegetation, in combination with spatial organization and distribution, suggest hunting pits targeting reindeer rather than moose, and that the pits were used for trapping animals rather than for storage or as hunting blinds (Smith, Reference Smith2013; Lemke, Reference Lemke2021), although it is possible that some pits were reused or were even originally constructed as storage facilities. Indeed, several storage pits have been found near the hunting pits. Storage pits in these regions are generally not directly connected to dwelling sites but are found in places accessible when on the move or hunting.
The data in Figure 3 indicate that most of the systems are located in areas where reindeer descend from the mountain, ready to spread out, at L1 locations, and in open spaces between higher grounds (L2) before they again gather and head for higher ground (Tables 2 and 3). By utilising the natural landscape, linear structures exploited the reindeer's behaviour, as shown by the N–S alignment of the systems, which intercepted the reindeer's E–W migration routes. This is visible in all the systems in this study. Natural features, like ridges and waterways, steered and encouraged the reindeer to follow the pit constructions (Figures 3 and 6). In addition, these hunting communities also frequently used areas of higher altitude, ranging between 500 and 800 m asl (Figures 3 and 4), given the reindeer's preference for higher grounds. These characteristics, together with the compelling evidence from the orientation of the base of the pits, can help determine the type of prey and the direction of movement (Manker, Reference Manker1960; Mulk, Reference Mulk1994, Reference Mulk and Andersen2005). The base of the pits is mainly rectangular and aligned with the migration paths of the reindeer, perpendicular to the alignment of the systems, and often consist of a convex, fan-shaped structure, funnelling the reindeer towards the pits (Figure 5). In this way, the hunting pits' spatial distribution and layout express aspects of traditional ecological knowledge.
Linear alignment of systems G, F, and E, Area 1 (map data: ©Lantmäteriet).
The notion that reindeer behave in predictable ways has, however, overshadowed unpredictable variables affecting herd movements, such as weather, insects, predators, and grazing conditions (Klaussen, Reference Klaussen2008; Smith, Reference Smith2013). The whole landscape needs to be considered to comprehend the concepts of ecosystem engineering and knowledge transfer. Reindeer movement has not been a constant factor over time or place (Klaussen, Reference Klaussen2008; Myrvoll et al., Reference Myrvoll, Thuestad and Holm-Olsen2011; Solli, Reference Solli2018; Lemke, Reference Lemke2021), and the hunting pit systems demonstrate the hunters’ abilities to deal with unpredictable factors. These planning and engineering skills are evidenced by the significant number and concentration of hunting pits and their distribution, providing options for risk management concerning which systems to use. For example, systems E, F, and G constituted a chain intercepting the herds as they descended from the mountains, ready to spread out (Figure 6). They functioned as three barriers, working simultaneously or separately, controlling the herds' behavioural preferences and diverse movement due to unpredictable variables, in an E–W direction. Similar ways of involving the whole landscape can be found by combining systems C and D and systems A and B. Furthermore, Area 1's systems combined to form a 20 km-long alignment cutting off the wild reindeer's migration routes, thus offering multiple opportunities for successful hunts.
Spatial manifestations
In the main case study area, active and deliberate resource management resulted in comprehensive modifications to the landscape in the form of a built environment still visible today (Smith, Reference Smith2011; Lemke, Reference Lemke2021); hunting pits are different compared to many other hunting techniques, as they are permanent features and hence an investment in the landscape and a resource, more distinct than hunting with bow and arrow (Vorren, Reference Vorren1958; Manker, Reference Manker1960; Klaussen, Reference Klaussen2008; Hodder, Reference Hodder2012; Solli, Reference Solli2018: 13, 23).
These spatial manifestations are compelling evidence, when we consider the numbers involved. The number of recorded hunting pits—as inventoried in the KMR database by 19 December 2023—predating this study was about 1000, distributed all over the large municipality of Jokkmokk, which covers 19,334 km2. As a result of this survey, an additional 1000 hunting pits have been identified in only c. 228 km2, indicating that the numbers and concentrations in this study area are exceptional compared to earlier records and possibly double the amount (Table 2). The significant alterations made to the landscape, indicated by the systems in Figure 3, also demonstrate deliberate human ecosystem engineering of the landscape for strategic purposes. The locations of the hunting pits show a more frequent use of altitudes and topographical zones, previously considered uncommon in Sweden (Forsberg, Reference Forsberg1985; Mulk, Reference Mulk1994: 160–63, Reference Mulk and Andersen2005: 45–48; Spång, Reference Spång1997: 31–32, 51–55, 60–71; Sjöstrand, Reference Sjöstrand2011). Earlier research conducted in the Swedish mountain region has proposed three topographical zones to categorize the distribution of hunting pits (Mulk, Reference Mulk1994, Reference Mulk and Andersen2005: 45), with the montane boreal forest below 400 m asl (Zone 1) being regarded as the area most frequently used and containing the most extensive systems. However, neither spatial analysis nor field survey revealed further hunting pits in the region. Instead, large numbers were found primarily in Zone 3, with more than fifty pits in almost all systems (Figure 3, Tables 2 and 4), with systems A–G containing an average of approximately 140 pits per system. For comparison, in northern Sweden we find an average of ten pits per system (Liedgren, Reference Liedgren2013: 3; Grimbe, Reference Grimbe2022).
The numbers, extent, and organization of the pits suggest a cohesive management of the entire landscape, shown by the significant similarities in the hunting pits systems by location types L1 and L2, the N–S alignment of the central axis, the base of the pits pit set at right angles to this axis, and a strategic use of topographical Zones 2 and 3 (Figures 3, 4, and 5).
By establishing permanent, prominent hunting structures requiring elaborate management, unparalleled by other hunting techniques, people in these regions invested substantially in the landscape, thus creating a built environment and a specific niche.
Temporal manifestations
There is a wide range of temporal scales concerning LD, including: timespans, use time, investment in time and energy for construction, reuse and maintenance, knowledge transmission over time, and the actual timing regarding a specific season and period of use (see e.g. Lemke, Reference Lemke2021). This article only briefly touches upon all these different temporal aspects, instead focusing on timing. Timing shows human agency by considering predictable and unpredictable factors in land use management. The articulation and distribution of pits can help interpret the time of use. For seasonally migratory animals, like reindeer, the most favourable times of the year are spring and autumn. In spring, the animals move west up into the mountains; in autumn, they return to the boreal lowlands in the east (Klaussen, Reference Klaussen2008: 3–5). Systems D, G, F, and E intercept the reindeer efficiently, moving from west to east by encountering and assembling the herds as they are about to spread out in more open low terrain, hence reducing human effort. The alignment of the base of the pits can provide a reliable indication as they often consist of a convex fan-shaped structure, funnelling the reindeer towards the pits (Figure 5). This is a compelling feature suggesting autumn use, since the systems are adapted for animals moving towards the low-lying areas in the east. In addition, in the Scandes mountain range, snow cover tends to be tenacious, making spring use troublesome.
Coordination and collaboration of labour
Labour cooperation is essential for managing a resource that demands modifications to the landscape. In ethnographic records, planning and knowledge are interrelated when choosing a high-value location, and large-scale hunting involves investing time, energy, and people to manage a specific resource at a particular place and time (Smith, Reference Smith2011: 836–38, Reference Smith2013: 8–11). For example, experiments in Sweden and Norway estimate an average of 18.6 days for one person to build one hunting pit (Spånberg, Reference Spånberg2014: 38–44). Consequently, the combined systems E, F, and G (Figure 6) would imply sixteen years of continuous digging by a single person, which suggests collective labour. Previous research has generally regarded hunting pits as labour-intensive, requiring planning, layout, construction, maintenance, and the slaughter and handling of meat, regardless of the size of herds, the number of pits used, or their spatial extent (Vorren, Reference Vorren1958; Mulk, Reference Mulk1994, Reference Mulk and Andersen2005; Hansson & Rathje, Reference Hansson and Rathje1999; Klaussen, Reference Klaussen2008; Sjöstrand, Reference Sjöstrand2011; Solli, Reference Solli2018). Hence, the dense concentration of pits and systems in this case study indicates substantial effort and coordination of labour (Figure 3 and Table 2). However, the large numbers and concentrations of pits do not necessarily mean that all the pits in one system or nearby systems were used simultaneously (Klaussen, Reference Klaussen2008: 77–78; Lemke, Reference Lemke2021). The massive landscape transformations resulting in an extensive built environment suggest communal, cohesive management to deal with the landscape as a whole.
Conclusion
This study set out to discuss how hunting pits can help us understand land use by small-scale societies and illuminate aspects of landscape domestication in hunting landscapes. The main conclusion is that hunting pit systems involved the landscape as a whole. Communities in northern Fennoscandia were not operating in a wilderness; through systematic planning, organization, and agency, they constructed a hunting landscape with large numbers and concentrations of hunting pit systems. These systems are expressions of landscape domestication, evidenced by a major investment in land use and elaborate management skills.
Acknowledgements
This study was funded by the Berit Wallenberg Foundation (grant no. BWS 2022.0023) and the Carl-Göran Adelswärds Foundation. I thank my supervisors, Ingela Bergman, Karl-Johan Lindholm, and fellow PhD student Kailin Hatlestad for valuable comments and language revision. I also thank my family for support and patience when fieldwork took me away from home.
Open access
