Jump to ContentJump to Main Navigation
Wetland Archaeology and BeyondTheory and Practice$

Francesco Menotti

Print publication date: 2012

Print ISBN-13: 9780199571017

Published to Oxford Scholarship Online: March 2015

DOI: 10.1093/acprof:osobl/9780199571017.001.0001

Show Summary Details
Page of

PRINTED FROM OXFORD SCHOLARSHIP ONLINE (oxford.universitypressscholarship.com). (c) Copyright Oxford University Press, 2020. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in OSO for personal use.  Subscriber: null; date: 27 October 2020

Abundant, Well-Preserved Evidence

Abundant, Well-Preserved Evidence

(p.128) 4 Abundant, Well-Preserved Evidence
Wetland Archaeology and Beyond

Francesco Menotti

Oxford University Press

Abstract and Keywords

This chapter discusses a variety of archaeological evidence in order to identify architectural structures characteristic of wetland environments and to try to bridge the gap of archaeological evidence between wetland and dryland occupations. This is done by identifying objects and artefacts, which are not necessarily typical of wetland sites, but have survived due to wet anaerobic conditions, and comparing them with those belonging to distinctive wetland groups. The contents of each section and subsection are organized in a descending chronological order, following, where possible, a west-east, north-south spatial distribution. The discussions cover houses and settlements in the wetlands, contact and transport, material culture, and sacred practices and beliefs in the wetlands.

Keywords:   wetland archaeology, archaeological evidence, settlements, material culture, sacred practices, beliefs, transport


One of the most striking features of wetland archaeology is, without a doubt, the archaeological evidence that the discipline deals with. Anaerobic conditions prevent organic materials from decaying, and, as a result, the apparently most delicate and vulnerable artefacts remain astonishingly well preserved. However, not all wetland environments yield the same objects. This is not only due to preservation conditions (see Ch. 5), but to the different ways in which areas were occupied. Within the general term ‘wetlands’ there are a number of different types of habitat, for instance: lake shores, river banks, swamps, marshy coastal areas, and peatbogs (see Ch. 1). Not only have these wetlands different ecosystems, but also dissimilar geomorphologies and climate, which together influence human occupation (see Chs. 2 and 3). As a result, archaeological evidence may change according to various types of wetland. It is therefore germane to distinguish between the diverse environments and identify the particular kind of archaeological evidence that may be peculiar or endemic to that specific area. This chapter discusses a variety of archaeological evidence, from settlements to people and their material culture. It is, for instance, crucial to recognize the difference between a typical lakeside village on stilts in the Alpine region of central Europe and a crannog in Scotland, or to detect the reason why the ground-joint building technique was preferred to the perforated base-plate method of construction in marshland and/or peatbog-like environments (Gollnisch and Seifert, 1998; Schlichtherle, 2004). It is also important to understand how people moved about in those seemingly inhospitable environments: did they build paths and causeways, or did they prefer waterway communication? For instance, taking into account the large number of canoes (dugouts) found in the wetlands, could it be stated with certainty that this means of transport was the most popular one, or did people have alternatives (Pétrequin et al., 2006c; Ruhl and Purdy, 2005)? And, is it really possible to detect the difference between utilitarian trackways and those built for ritual purposes (Bond, 2004, 2009; Coles, 1999b; Coles and Coles, 1986; Honegger, 2001; Raftery, 1996d)? Finally, concerning people’s material (p.129) culture, was it really different from that of the terrestrial groups? For example, did they have a particular way of making the pottery (e.g. a particular clay body to temper ratio to withstand the excess of humidity, as used by the Late Tripolye riverine groups in Ukraine (Gey, 1986; Ryzhov, 2008)), or did they use specific species of wood, ones that cope better with humid environmental conditions?

The mystery that often shrouds the wetlands has always created a fertile ground for sacred activities, whether linked to religious beliefs or to pagan offerings and depositions. Interestingly enough though, apart from only a few exceptions (see Windover and Fort Center in Florida, United States, and a small number of dugouts in prehistoric Europe), funerary practice and burials in the wetlands are not very common (Doran, 2001b, 2002; Milanich, 1998). However, extremely well-preserved human remains come from peatbogs, where quite a few so-called ‘bog bodies’ have been discovered in the past few centuries (Coles et al., 1999; van der Sanden, 1996). In some cases, the level of preservation is so good that even facial features (skin, hair, eyelashes, and eyebrow) are still perfectly visible. These incredible discoveries have helped archaeologists shed more light not only on the aesthetic appearance of our remote ancestors, but on nutrition, diseases, and other pathologies as well.

This chapter is not by any means an exhaustive list of house and settlement types, well-preserved organic material culture, special wetland sites, and remarkable remains of bog bodies, but rather a careful selection of places, structures, and objects, either characteristic of wetland environments or simply serendipitously preserved in waterlogged conditions. The main objective of this selection is twofold: first of all to identify those architectural structures (habitations, trackways, alignments, etc.) characteristic of wetland environments, and secondly, to try to bridge the gap of archaeological evidence between wetland and dryland occupations. This is done by identifying objects and artefacts, which are not necessarily typical of wetland sites, but have survived thanks to wet anaerobic conditions, and compare them with those belonging to distinctive wetland groups. The contents of each section and subsection of the chapter have been organized in a descending (older to more recent) chronological order, following, where possible, a west–east, north–south spatial distribution (as in Ch. 2). However, due to cross-referencing between sites and/or lack of archaeological evidence in some areas, vast leaps are sometimes required in order to follow a more consistent argument.

Houses and Settlements in the Wetlands

In comparison to the large number of organic artefacts found in waterlogged archaeological sites, the remains of houses and/or settlements are rather (p.130) limited. The paucity of residential unit architectural structures in a wetland context depends upon a myriad of factors, such as location, cause of abandonment, site formation processes, climate, environmental change, and, as pointed out in Chapter 5 (under ‘Survey’), the identification of such structures. Whether or not the house structures were preserved depends upon all the above-mentioned factors, but one of the most influential is certainly the architectural form of the house itself. In fact, even if one takes for granted an optimal wet location with ideal preservation conditions, the final outcome would still depend on how the house was constructed. Site formation processes vary significantly between a house with an elevated floor and one built directly on the ground, or between a house located in shallow water or on land. For example, the chance that archaeologists find the original floor of a collapsed house built on stilts above the water is extremely slim. House floors constructed directly on marshy grounds have, on the other hand, a much higher prospect of survival, despite the fact that preservation conditions are not ideal (e.g. anaerobic conditions may not be present). In the long run, however, the latter example is more vulnerable to climate and environmental change, and, above all, to human agency (e.g. land management), than the former. As a result, there is more chance of coming across well-preserved, ‘distorted’ evidence, than poorly preserved, intact evidence in situ (see also Ch. 5, under ‘Preservation’ and ‘Conservation’). It is therefore obvious that, even with the most remarkable level of preservation, the following description and distribution of wetland houses and settlements is far from being a true reflection of reality at the time those house and/or settlements were occupied.

Houses in Wetland Contexts

The fact that we are dealing with wetland occupations does not necessarily imply that the remains of habitations in wet environments are always waterlogged. This is particularly true in prehistoric northern Europe and Scandinavia, where, despite the presence of vast wetland ecosystems the majority of settlements were located on slightly elevated grounds. As a result, even though these sites may be covered in several metres of peat (developed after the houses and/or settlements were abandoned) the layout and architectural remains of those residential units resemble more that of a typical terrestrial dwelling (post-holes and inorganic living floors), rather than a wetland one (waterlogged remains of wooden structures). As surprising as it may seem, the oldest evidence of waterlogged archaeological remains of houses does not come from Europe, but from a more unexpected place: South America. In fact, the remains of at least twelve rectangular houses, built with tree stems and branches and dating c.12,500 BP have been discovered at Monte Verde, Chile. The huts were covered with mastodon skin (traces of the skin were (p.131) still attached to parts of the frames) and there was also evidence of fireplaces inside and outside the dwellings (Dillehay, 1997). In Europe, the first waterlogged evidence of prehistoric houses is not found until the Mesolithic, during which fairly well-preserved living floors in wetland contexts are not uncommon, especially in the northern part of the continent and Scandinavia. Some of those hut living floors even retain traces of post-holes, resembling either the tepee-like structure of Howick hut (northern England) (Waddington, 2007), or the dome-like huts of Mesolithic Lake Feder (Germany) (Schlichtherle and Strobel, 1999), or even the Early Neolithic ones of Matveev Kurgan I, in Ukraine (Krizhevskaya, 1998). However, proper evidence of waterlogged houses is rather rare. One of the best-known examples is the submerged site of Møllegabet II, which yielded a 5 × 3 metre rectangular living floor made of bark, resting on cross-laid branches. A fireplace or hearth is also identifiable (Rieck, 2003). The scarcity of waterlogged house remains in Mesolithic Europe is also mirrored by later (e.g. Neolithic and Bronze Age) sites. The prevailing archaeological evidence of residential dwellings, especially in the northern part of the continent and Scandinavia, is in fact that of living floor layouts and post-holes, where the house superstructure has completely disappeared. Interestingly enough, despite these house layouts being located within wetland environments they are not considered to be wetland sites. One of the best examples is the Neolithic settlement of Bejsebakken (northern Denmark), consisting of 23 two-aisled houses, with and without sunken floors, measuring 5.75 × 14–15.5 metres (Sarauw, 2008). Similar evidence, but in the Bronze Age, is found at Bjerre (Thy, Denmark), where traces of living floors belonging to three-aisled houses measuring 15–25 metres long and 6–8 metres wide have been located on a gentle hill surrounded by marshlands (Bech, 1997). Further south, similar post-hole layouts have been identified at Texel-Den Burg, Angelslo, Zijderveld, and Oss, where Middle and Late Bronze Age houses were even longer, reaching in some cases 30 metres (Fokkens and Arnoldussen, 2008). Other archaeological evidence of wetland houses in northern Europe comes from the eastern Baltic Sea regions and western Russia, and consists of mainly waterlogged wooden piles of seasonal dwellings built of the edges of lakes and marshes. The house superstructures and other architectural characteristics are not preserved. On the British Isles, Neolithic houses within the wetlands are not very numerous either, and none of them (e.g. those in the Fenland) show traces of waterlogged structures. A slightly different situation occurs in Ireland, where evidence of lough shore occupation is certainly present, although, except for a few examples (e.g. Rathjordan, Co. Limerick), almost all of them have not retained any waterlogged house superstructures. Despite the limited number of Neolithic and Bronze Age wetland houses (in both northern Europe and in the British Isles), one can notice a clear distinction in architecture; while on mainland Europe houses are mainly of rectangular shape, those in the British Isles and Ireland are mostly circular. As one (p.132) moves south towards central Europe the number of typical wetland (lacustrine and marshland) dwellings found in waterlogged conditions increases significantly. The development of the so-called lake-dwelling tradition in the Circum-Alpine region from the forty-third century BC onwards brought about a totally new way of interacting with the wetlands (see also Ch. 2). People were no longer living on their edges, but within them, constructing fully-fledged settlements (see Boxes 4.1 and 4.3) and developing new architectural trends in house construction. (p.133) (p.134) (p.135) (p.136) (p.137) (p.138)

(p.139) Outside the Circum-Alpine region, Neolithic and Bronze Age remains of waterlogged wetland houses are fairly scarce. Small rectangular pile-dwellings were constructed at Dipilio (Greece) in the second half of the sixth millennium cal BC (Hourmouziades, 1996), whereas the majority of houses of the almost contemporaneous lacustrine settlements of La Marmotta (Lake Bracciano, central Italy) (Fugazzola Delpino, 1998; Fugazzola Delpino and Mineo, 1995; Fugazzola Delpino and Pessina, 1999) and La Draga (Lake Banyoles, Catalonia, Spain) (Bosch et al., 2006; Tarrús, 2008) seem to have been built directly on the ground. Excellent evidence of Late Neolithic (2830–2770 cal BC) wetland houses comes from Hunte 1 on Lake Dümmer (central Germany), where three types of architecture can be distinguished in chronological order. The first and the oldest type consists of a peculiar polygonal hut (Fig. 4.9), measuring 3–3.4 × 5–5.5 metres, whereas the second and third types are one- or two-roomed rectangular buildings varying in size, from 3–4.4 × 4–7.5 metres (second type), to 3–4.5 × 4.8–11.3 metres (third type). The only substantial difference between the latter two types is the porch-like structure on the eastern side of the houses of type three. Types one and two seem to have been built directly on the ground, whereas type three, according to Reinerth, could have had slightly elevated (on stilts) floors (Kossian, 2007).

Evidence of wetland houses in other parts of the world, contemporaneous with the European Neolithic and Bronze Age periods, is not abundant. The remains of elongated pile-dwellings have been found at Humudu and Majiabang (c.5000–3800 cal BC) in China (Chang, 1986; Zhao and Wu, 1986–7), or the Jomon rectangular (5 × 10 metres) house of Ondashi (c.4000 cal BC) in Japan, which, interestingly, differs from the typical Jomon sunken-floor house of the same period (Matsui, 1992).


Abundant, Well-Preserved Evidence

Fig. 4.9. Reconstruction of the polygon shaped Neolithic hut of Hunte 1, Lake Dümmer, Germany.

(Photograph: courtesy of Gunter Schöbel, Pfahbaumuseum, Unteruhldingen, Germany)

Returning to Europe, towards the end of prehistoric times (e.g. Iron Age), people’s relationship with the wetlands (especially concerning inhabiting them) changed quite substantially, and in some cases even drastically. For instance, while in the Circum-Alpine region lacustrine settlements disappeared completely at the very beginning of the Iron Age, in other parts of Europe (e.g. the Baltic Sea regions and British Isles), they started to develop and reach their climax during this period and even later. Typical examples are the Scottish and Irish crannogs (see Box 4.2) and the lake settlements of the Masurian, Pomerania, and Wielkopolska regions in Poland. Although wetland house remains in Poland are not numerous, a fairly clear distinction in terms of house architecture can, however, be made between the three regions. While in the Wielkopolska region a prevalence of vertical pile constructions is noted, in the other two regions (Masurian and Pomerania) houses seem to have been built on large platforms (not on stilts) constructed on top of artificially built islands in the water near the lake shore. Two different ways of preparing the ground for constructing the houses have been identified: (a) the Fascinenbau, whereby the area is prepared with irregular timber and brushwood, and (b) the (p.141) Packwerkbau type consisting of different strata of roundwood to construct a large platform on top of which houses are subsequently built. The houses were mainly rectangular and quite limited in size (3.2 × 3.5 metres), as the ones found at Moltajny (Pydyn, 2007: 325–7). In the British Isles, apart from the crannogs, it is indeed in the Iron Age that the first proper wetland village, namely Glastonbury, was built (see below). Here, contrary to continental Europe, the shape of the houses was circular, reflecting that of typical Iron Age dryland houses and crannogs. The dwellings were between 5 and 8 metres in diameter and did not have a central post to support the roof. The entrance was between 1.4 and 2.1 metres wide, and the walls were made of wattle work and/or woven rods subsequently plastered with daub. The roof was probably thatched, while the floor was made of parallel-laid round timbers covered with clay. Traces of hearths raised above the floor were usually located in the central area of the hut, along with raised surfaces, possibly used as tables (Bulleid and Gray, 1917; Coles and Coles, 1996; Coles et al., 1992; Coles and Minnitt, 1995). Apart from the typical dryland round houses (e.g. Cat’s Water) found in the elevated areas of the Fenland (eastern England), other examples of wetland habitations have been identified only in estuary environments, with the best known being the rectangular houses of Goldcliff in the Severn estuary. The seven (possibly eight—Martin Bell, pers. comm. 2010) identified buildings measured between 4.8 and 8.4 metres long, and 4.2 and 6 metres wide, and have characteristic rounded corners. One of the houses has two internal partitions, which are probably the remains of small animal stalls (Bell, 1999: 20). Similar buildings, dating from the Bronze Age, were also found at Redwick (5 km west of Goldcliff), whereas circular ones have been located at Chaper Tump (3 km west of Redwick) and Brean Down (Bell, 1999, 2001). In continental Europe, in addition to the three above-mentioned Polish regions, remains of waterlogged Iron Age houses are mainly located in the Netherlands and northern Germany. Two of the best examples are the Dutch Site Q in the Assendelver Polders, and the German Feddersen Wierde. The single building of Site Q measures 18 × 6 metres and dates from about the eight century cal BC. The house was roughly divided, two-thirds for cattle and one-third living space for people. It is interesting to note that two different species of wood were used for the two distinct areas; while alder was used for wattle walls for both people and animals, for some reason willow wattles were only used for people’s quarters (Brandt and van der Leeuw, 1987; Coles and Coles, 1989; Therkorn et al., 1984). Similar house structures are also found at Feddersen Wierde, where a series of farmsteads were constructed on elevated ground above the floodwater in a coastal-marshland environment (see ‘Settlements and Simple Agglomerations of Houses’, below). As in Site Q, here the buildings were also divided into a long section for the domestic animals and short ones for people. The size of the houses varied from 10 to 20 metres long and 6 metres wide. There were also small residential units and workshops with no (p.142) animal stalls inside. The houses were built directly on the artificially elevated ground, had fairly solid wooden frames, possibly thatched roofs, and wattle work was used for internal partitions as well as external walls (Haarnagel, 1977; P. Schmid, 2002). This type of house, but even longer, was also found at the roughly contemporary site of Flögeln, situated inland on a large sandy area within a peatbog environment.

As briefly mentioned above, the last part of prehistory in Europe (Iron Age) was characterized by a peculiar division, in terms of people–wetlands interaction, between the Circum-Alpine region and the rest of Europe (including the Mediterranean). Contrary to a complete absence of wetland settlements (and houses) in the Alpine region and surroundings, elsewhere people continued to live within or very close to the wetlands. Major engineering deeds to adapt to the wet environment were carried out not only in the north, but also in the Mediterranean area. One of the best examples is the Poggiomarino settlement on the bank of the River Sarno (central Italy), where a series of artificial islets with houses and/or other buildings on top was constructed in a marshy riverine environment. The islets were oval, but the huts constructed on them were of rectangular shape (possibly with an apsidal end) constructed directly on the ground. In general the houses were fairly small (3–3.5 × 10–12 metres), although one seems to be more elongated than the others, and probably belonged to an extended family (Castaldo et al., 2008; Cicirelli and Albore Livadie, 2008; Pruneti, 2002).

North of the Alps, one of the largest and most compact agglomerates of Iron Age houses is the fortified lacustrine settlement of Biskupin in Poland (see below). The terrace houses were all lined up along the various streets and alleys, were all built using the block-building technique, measured on average 8 × 9 metres and had a single room where the bed was placed in the corner. The various domestic activities were carried out in the central area, next to a stone-built hearth. Each house had an entrance hall, which could have also served as animal shelter, whereas the upper (attic-like) level could have been either for storage, or sleeping (Coles and Coles, 1989, 1996; Niewiarowski et al., 1992).

Finally, the remains of an Iron Age lacustrine house have also been identified near the French Atlantic coast, and more precisely at Put Blanc, on Lake Sanguinet. The structure, which dates to the seventh or eighth century cal BC, measures 4.6 × 3.3 metres, was built directly on the ground and had a centrally located fireplace (Maurin, 2006).

In continental Europe Early Medieval and Medieval fortifications within wetlands are rare, and they are mostly located in the eastern regions of the Baltic Sea (note though that in some of these regions, the first millennium AD was still regarded as Late Iron Age—see Ch. 2 and Fig. 2.1). One of the best known of such sites is Āraiši (ninth–tenth centuries AD) in northern Latvia (Apals, 1965), where amongst the 146 buildings (all constructed during the five occupational phases of the settlement), 76 were residential dwellings. The (p.143) (p.144) (p.145) houses were all built with the block-construction technique and had gabled roofs made of wooden shingles. A similar settlement is found on Lake Valgjärv in Estonia. Here, despite the large number of massive (9 metres long and 30 cm in diameter) vertical and horizontal piles, no proper structures of residential units have been identified. Although these sites are regarded as fortified settlements, neither of them shows characteristic features of proper fortifications (e.g. palisades), apart from the natural setting (they were surrounded by water) (Roio, 2007: 31). A large square structure known as the Bulverket, also regarded as being a fortified settlement, was found on Lake Tingstäde Träsk on Gotland Island, Sweden. The structure, measuring about 170 × 170 metres and dating c.AD 1000–1200, consists of four rectangular wooden platforms arranged one after the other to form a large square platform with an empty space in the centre. Residential units were probably placed on top of the platform, which itself was surrounded by a massive palisade (L. Larsson, 1998). Unfortified Medieval residential units built within a wetland context in continental Europe are extremely rare. One of the best-studied examples is Charavives-Colletière (c.AD 1003–1040) on Lake Paladru, France. The site (surrounded by a wooden palisade) consists of three main buildings (partially attached to one another) and a few sheds and workshops. The largest building was the one in the middle and it is believed that its roof was as high as 14 metres (Colardelle and Verdel, 1993; Coles and Coles, 1996).

Evidence of waterlogged wetland house structures outside Europe in the past two millennia is not abundant. One area where this evidence is particular rich is the North Island of New Zealand. Here the best-preserved waterlogged remains of Maori houses come from the swamp settlements called pa (see Ch. 2, and Box 4.5 below). Although plans and layouts of Maori houses are often discovered in dryland sites, and isolated parts of the superstructure of houses do occur in waterlogged environments, a combination of the two (layout and superstructure) is rarely found. The only two examples of living floors and house superstructure components found together are those of Mangakavare 2 on Lake Mangakavare (Bellwood, 1978), and Kohika in the Bay of Plenty (Irwin, 2004b, 2005). Amongst the two, Kohika offers more detailed evidence of the different types of house architecture as well as the organization of the settlement. Two types of building can be distinguished at Kohika: the residential dwellings (whare) and storage buildings (pataka).

The whare architecture is itself divided into two types: the dressed-plank lashing technique, and the plain pole and thatch technique. The first type consists of a series of finely dressed planks with mortise and tenon joints carefully lashed together with fibre ropes (Fig. 4.11).

A remarkable characteristic of this type of construction is that, thanks to a series of face and edge eyelets carved into the planks, the lashing is mainly exposed only on the outside of the house frame, and therefore not visible from inside the house (Fig. 4.12).


Abundant, Well-Preserved Evidence

Fig. 4.11. Schematic reconstruction of the Kohika pa plank-dressed whare from the HS area, New Zealand.

(Courtesy of Geoffrey Irwin, University of Auckland, New Zealand)

Abundant, Well-Preserved Evidence

Fig. 4.12. Details of the external framing of the Kohika pa plank-dressed whare, New Zealand.

(Courtesy of Geoffrey Irwin, University of Auckland, New Zealand)

Abundant, Well-Preserved Evidence

Fig. 4.13. Elaborately carved door lintel (pare) of the Kohika pa plank-dressed whare from HS area, New Zealand.

(Courtesy of Geoffrey Irwin, University of Auckland, New Zealand)

It has been suggested that this technique derives from the early Maori way of building canoes, whereby lashes were not exposed to the outside of the watercraft to avoid abrasion during landing. This perpetuation of carpentry techniques is in accordance with the Maori’s conservative cultural aspect of house-building traditions (Prickett, 1982). Elaborately carved door lintels (pare) (Fig. 4.13) were also used to decorate the entrance of this type of house.

The second type of whare was built using undressed poles with joints lashed together where elements are crossed (in this case the lashing is visible from the inside). Despite the less elaborate way of constructing this type of house, the size of the building does not vary substantially from the more sophisticated type; on average, the Kohika houses are 3–3.65 metres wide, 4.50–5.40 metres long, and 2–2.3 metres high. However, the ‘Yellow House’ found in area D is larger: 5.35 × 7.24, reaching 2.85 metres at the ridge-pole (R. T. Wallace et al., 2004: 123). Walls and roofs of the Kohika houses were thatched with bulrush raupo (Typha augustifolia), a wetland plant very common in the region.

Storage houses (pataka) were also found at Kohika. The walls of a pataka were made of wooden planks simply lashed together, the floor was elevated and was reachable by a ladder with notched steps (Fig. 4.14).

The pataka were usually smaller than the whare, measuring 2.6 × 3.9 metres (R. T. Wallace et al., 2004: 123–5). The crucial importance of the Kohika houses lies in the fact that they allow us to link ethnographic descriptions of post-European contact house construction techniques to the less familiar traditions of the pre-European contact.

Outside New Zealand, waterlogged house remains are very limited. The best-known example is that of Ozette on the Northwest Coast of the United States. Here eight houses, measuring 21 × 11.5 metres and dating from the fifteenth to the sixteenth century AD, were found under a mudslide that sealed the Native (p.148)

Abundant, Well-Preserved Evidence

Fig. 4.14. Schematic reconstruction of the Kohika pa pakata from area D, New Zealand.

(Courtesy of Geoffrey Irwin, University of Auckland, New Zealand)

American settlement while it was still occupied. Because of the sudden and unexpected calamity the buildings were sealed in a Pompeii-like way, and the structures remained perfectly preserved. Hence, parts of the superstructures of the buildings, such as plank walls, posts, and roofing, were still perfectly identifiable at the time of the discovery.

Finally, two totally different house structures were identified at Los Buchillones, Cuba. One house (14C dated between the fifteenth and the seventeenth centuries AD) is of circular shape, the king posts are believed to have been as high as 7 metres, and, according to the thin vertical sticks that delineate the perimeter, the house should have had a diameter of 26 metres. Not far from this structure, there was a rectangular house (14 metres long), possibly bearing a two-slope gabled roof. Both houses (circular and rectangular) are believed to have stood on piles in shallow water (e.g. a lagoon), but in a place protected from the intense wave action of the sea (Pendergast et al., 2001, 2002).

(p.149) Settlements and Simple Agglomerations of Houses

No matter how well preserved and thoroughly described individual houses are, they are not entities that can be understood on their own. Houses are in fact part of larger residential units such as settlements. The layout of buildings as well as the space between them reflects the way inhabitants interact with each other, and the environment. As a result, plan and size of villages may depend on a variety of factors, from land availability to cultural aspects, including intra- as well as inter-village relationships. For instance, an agglomerate of houses tightly clustered together does not always imply higher interaction and/or intimacy with the next-door neighbours. It rather depends on how easy it is to enter the building and/or the village (e.g. the distance between entrances, the presence of fences, the locations of communal space, and the size of both the village and the single dwellings). The concept of interactional distance is therefore germane to the study of both houses and settlements.

Due to the fact that the majority of settlements excavated in an archaeological context (especially concerning prehistory) are incomplete (this may depend upon a myriad of factors, such as size of the excavated area and preservation), analyses related to the use of space between and within the inhabited area become significantly limited, and in some cases, highly speculative (see Box 4.3). For instance, is it true, as is usually assumed, that the cycle of spatial arrangement (from lack of patterns to spatial order) breaks down after continuous rebuilding within a settlement, resulting in the failure of the community as whole (Fletcher, 1984, 1995)? Or, does the disposition of buildings play a crucial role in terms of settlement protection, or is the presence of fences and ramparts more important? With optimal availability of data, such as in Hornstaad-Hörnle 1 (Dieckmann et al., 2006), it may be possible to tackle the first question, or even to answer the second, if one deals with the Tripolye Culture riverine giant-settlements in Ukraine (Korvin-Piotrovskiy and Menotti, 2008). With the majority of cases though, especially with a limited excavated area in a wetland context, the available data is rather scarce.

When dealing with wetland settlements in Europe, three important factors have to be taken into account: period of occupation, geographical area, and type of settlement. For instance, Neolithic and Bronze Age settlements in northern Europe and Scandinavia differ considerably from those in the Circum-Alpine region. Not only are houses architecturally diverse, but also size, degree of agglomeration, and dispersion are different. In general, settlements tend to be more clustered in the Circum-Alpine region, as opposed to a more dispersed layout in the north. Moreover, the majority of wetland settlements in the north are not really constructed on wet ground, but on naturally elevated sandy areas, and this, of course, influences the site formation process, and consequently their preservation. It is therefore important to point out that (p.150) although the large majority of those settlements can be considered wetland settlements, virtually none of them has waterlogged remains.

Typical northern European prehistoric villages (often farmsteads) consist of a main farmhouse and number of outbuildings (maybe even other small houses, granaries, and/or haylofts). In some cases, for instance, the Neolithic settlement of Bejsebakken (northern Denmark), the village consisted of twenty-three houses clustered within three to four areas (Sarauw, 2008). Characteristic of prehistoric Scandinavian settlements is the reflection of long-term occupations of a given territory, with repeated rebuilding of houses (or farmsteads) within that area. The shifting from an old to a new place usually occurs every two or three generations. An example of such settlements is the Bronze Age site of Bjerg (west Jutland) (Olausson, 1992; Rasmussen and Adamsen, 1993). Various such settlements are also found at Bjerre (northern Denmark). Here, the nineteen identified sites span a period of time of about a thousand years, but only one or two farmsteads were contemporaneous (Bech, 1997, 2003). A similar situation is found in the Low Countries (the Netherlands), where settlements were also built within wetland environments, such as coastal dunes areas, inland peat areas and river floodplains, but preferably on levees and/or donken (Arnoldussen, 2008b; Arnoldussen and Fokkens, 2008; Louwe Kooijmans, 1993).

On the British Isles, apart from the Irish and Scottish crannogs, the settlement location (on slightly elevated areas within the wetlands) also reflects that of northern Europe, with the only difference being the house architecture and shape (round instead of rectangular—see ‘Houses in Wetland Contexts’ above, Ch. 2, and Boxes 4.1 and 4.2). Significant waterlogged settlement remains are more abundant in central Europe and the Mediterranean. Although the best evidence comes from the Circum-Alpine region (see Box 4.3), other areas such as central Germany, Greece, central Italy, and eastern Spain have also yielded important traces of wetland settlements. Sites such as Dispilio (Greece), La Marmotta (central Italy), and La Draga (Catalonia, Spain), have not only shed light on the very first lacustrine house architecture, but they have also contributed to trace the possible origins of the Circum-Alpine region lake-dwelling tradition (see also Ch. 2). As often happens with wetland excavations though, the limited extension of the excavated area has prevented archaeologists from identifying the full extent (perimeter) of the villages. A somewhat luckier situation occurred at Hunte 1, on Lake Dümmer (central Germany), where a full perimeter of a Neolithic wetland settlement consisting of twenty-four houses built with two distinct phases was identified in 1934 (Kossian, 2007) (see also ‘Houses in Wetland Contexts’ above, and Ch. 2). A few such lacustrine villages, dating from the Neolithic to the very end of the Bronze Age, have also been found in the Circum-Alpine region (see Box 4.3). Despite the active people–wetlands interaction, evidence of proper Neolithic and Bronze Age settlements in the north-eastern regions of the Baltic Sea is (p.151) (p.152) (p.153) (p.154) rather scarce. However, traces of occupation and isolated dwellings have certainly been identified; whether or not they might have been part of systematically planned settlements is yet to be confirmed. It was not until the Iron Age that proper marshland and lacustrine villages started to be constructed. The majority of these agglomerates of dwellings were built on both natural and artificially enhanced islands or peninsulas, and were surrounded by a protective palisade. Some of them even developed into heavily fortified compact residential units, with sturdy ramparts and massive palisades (see Biskupin, above). Similar fortified villages, known as the terramare, developed in the Po Plain (northern Italy), from the Middle Bronze Age onwards (see Box 4.4).

The rest of Europe does not mirror the complete disappearance of wetland settlements in the Circum-Alpine region at the very beginning of the Iron Age. For instance in the British Isles not only is there a noticeable increase of Scottish crannogs, followed by the Irish wave later on (see Box 4.2), but also the establishment of the first proper wetland (lakeside) village: Glastonbury. Unlike other Bronze Age and Iron Age settlements built within wetland environments (but on dry soil), Glastonbury was constructed on wet terrain, with all the disadvantages (and of course advantages) that come with it (see ‘Houses in Wetland Contexts’, above). Old and more recent interpretations have triggered incandescent debates about the village layout, its inhabitants, and chronology (Bulleid and Gray, 1917; Coles and Coles, 1996; (p.155) (p.156) (p.157) (p.158) Coles and Minnitt, 1995; Aalbersberg and Brown, 2010). It has eventually been agreed that the numerous circular mounds identified by Bulleid were not all occupied at the same time; in fact, some of them were not even dwellings. The village experienced a fairly rapid development, which can be divided into four phases. The early phase, starting at about 250 cal BC and lasting 25–30 years, consisted of only a few round huts (about four) surrounded by a slender fence. In the middle phase more houses were built (about ten in total), and the occupation lasted longer (50–60 years). The village maximum extent was reached in the late phase (c.100 cal BC), when about fifteen houses stood at the same time; the occupation was also the longest amongst the four phases (75–90 years). In the final phase, the settlement shrank to about five dwellings (all built on previously occupied locations) and they were occupied for only 25–30 years (maybe even seasonally). Glastonbury was eventually abandoned c.50 cal BC (Coles and Coles, 1996; Coles and Minnitt, 1995).

On mainland Europe, one of the best-preserved wetland villages, more or less occupied in the same period, is Feddersen Wierde (northern Germany) (Schmid and Schuster, 1999). Two main phases of occupation can be distinguished at this Wurt-settlement; one from the first century BC to the fifth century AD, and the second (after about 200 years of abandonment) starting in the Early Medieval period and ending around the thirteenth or fourteenth century. Eight to eleven farmsteads were first built on the elevated levee (c.6 ha) lying between two creeks, in the first century BC (56 BC—terminus post quem). Because of the rising sea level each farmstead was raised on its own Terp or Wierde (mound), but, by the third century AD, the various mounds eventually merged into one raised area, on which 39 buildings were erected. During this time, the houses were quite large (up to 20 metres long and 6 metres wide), but the size decreased soon afterwards, and, following a further sea-level rise, the settlement was finally abandoned in the fifth century AD (Haarnagel, 1977; P. Schmid, 2002; Schmid and Schuster, 1999). Similar processes of house-building space claimed from a wet environment also took place more than a thousand years earlier at Poggiomarino (central Italy), where a series of artificial islands were built within a marshy and riverine area and used as building grounds for houses and workshops. The navigable waterways around them facilitated the transport of goods, while cultivation and other agricultural activities were carried out on drier terrains nearby (Albore Livadie et al., 2005; Cicirelli and Albore Livadie, 2008).

Apart from the Irish lacustrine crannogs, wetland settlements started to disappear gradually from the early second millennium AD. The only known lakeside dwelling of the period in mainland Europe is Charavives-Colletière on Lake Paladru (France) (Colardelle and Verdel, 1993). More and more land for agricultural purposes was claimed from the wetland, and the inexorable destruction of settlements began, although in some regions they still played a crucial role in people’s everyday lives.

(p.159) Outside Europe evidence of structured settlements within wetland environments is scarce. Agglomerates of houses were certainly built within flooded areas in Neolithic China (see e.g. Tianluoshan and Hemudu, Ch. 2), as much as during the whole Jomon (e.g. Ondashi) or the Yayoi period (e.g. Toro) in Japan, but waterlogged remains of dwelling superstructures are considerably limited. However, as pointed out in Chapter 2, the paucity of fully identifiable settlement plans does not suggest a reduced interest in, or limited interaction with the wetland, as is shown by the rich archaeological evidence (although no dwelling structures) found in Florida, and the Northwest Coast of North America (Ch. 2). A totally different picture is offered by New Zealand, where favourable preservation conditions have frozen in time the importance of the wetlands for the past and present Maori people (see Box 4.5, and Ch. 2, under ‘New Zealand’).

(p.160) Wetlands were also fairly often occupied in Australia, New Guinea, and the Americas in the past thousand years. However, especially concerning settlements, particular site formation processes have limited their preservation and they are therefore more difficult to find. Apart from Ozette on the Northwest Coast of the United States, and the Norse settlement of L’Anse aux Meadows, Newfoundland, Canada (see ‘Houses in Wetland Contexts’ above, and Ch. 2), traces of settlements in wetland contexts are limited to partial house structures, which may be difficult to identify as being part of proper villages (see Los Buchillones above). There are, of course, exceptions, and one of them is in the centre of Mexico City, where Late Aztec pyramids and other buildings overlie wooden remains of earlier communities, that used to build their settlements in the former lake basin. Unfortunately, we will never know the extent of those ancient wetland villages; they will probably remain buried under the city forever. Living within and being closely in touch with the wetlands was also part of the Aztec, Maya, Teotihuacan, and other Mesoamerican pre-Columbian cultures, which also used to build their settlements in wetland contexts. In some parts of the world living in the wetlands was not only a custom of the ancient past, but it perpetuated until very recently. In a few cases, these customs still continue today (see e.g. the reed houses near Puno, Lake Titikaka, Peru; the pile-dwellings of Ganvié, on Lake Nokoué in Benin (Africa); and of course the numerous indigenous groups occupying the vast wetlands of the Amazon Basin in South America).

(p.161) Contact and Transport

As is argued throughout the book (see in particular Chs. 2 and 8), wetland communities were not isolated entities living in secluded places and socially detached from the rest of the world, but were instead well-integrated groups linked together by complex socio-economic networks. Not only were wetland settlements in contact with those in drier areas, but they were often situated on crucial pivotal intersections of short- and long-distance trade routes. Their convenient location (e.g. on lake shores, river banks, or other navigable waterways) facilitated the transport of goods, hence increasing the chance of developing economically and becoming focal points of exchange.

Although archaeologically visible through artefact distribution, how and where goods travelled from one place to another is far more difficult to detect. Before (and after) the advent of long-distance roads with the Romans in Europe, the Maya in Mesoamerica and the Incas in South America, to mention but a few, the majority of traded goods travelled via water (this is, of course, not to say that before those civilizations no roads existed). Paths and minor roads on drylands have always been constructed, but the existence of them is, unfortunately, often not identifiable. Within the wetland environments (mostly peatland, marshland, and fens) on the other hand, anaerobic conditions have preserved a variety of wooden trackways built to explore, or simple cross those seemingly inaccessible areas (see below). The wetlands were also criss-crossed by navigable water channels and basins, within which different types of watercraft were used to transport people and goods. In both cases though, the wetlands still retain remarkable evidence of how our ancestors used to interact with and explore them.

Trade and Exchange

It is not the purpose of this section to identify or reconstruct major trade networks within and between wetland communities, or even to isolate what may be inappropriately called ‘wetland trade’. Instead, the wetlands (and wetland settlements) would be considered as facilitators (or maybe obstacles) within a broader trade and exchange network system. This will help identify the extent to which wetland people were engaged in this system, how they affected it, and finally how they were influenced by it. Although tightly interwoven, exchange networks are not all the same. In prehistoric Europe for instance, up to the Bronze Age (and in some areas even throughout the Iron Age), exotic goods were obtained via long-distance trade routes, which were nevertheless linked to short-distance down-the-line exchange networks used for subsistence goods.

(p.162) Metal and amber were two of the most common materials, and underwent various phases of value and significance change throughout prehistory. Amber, for instance, used in the Scandinavian elite burials during the Mesolithic, became a commodity for export in the Bronze Age, when bronze, because of its rarity, was considered more prestigious. The exact opposite happened in central Europe and the Mediterranean, where the scarce amber was more appreciated than metal. Amber and metal were in this case both commodities and exotic materials, but in two different places. Evidence of this interesting switch of amber value from exotic item to commodity is clearly visible at the wetland settlements of Bjerre (Denmark), where large quantities of unworked amber (1800 pieces found at site 7) were collected and stored by local farmers for export (Bech, 1997, 2003; Earle, 2002). Local farmsteads were probably the first cog of a larger chain exchange system that would eventually have linked northern Europe to the south. A further example of wetland settlements facilitating short- and long-distance exchange systems are the Iron Age villages of Meare and Glastonbury. Both Meare villages (west and east) were important trade centres and both yielded evidence of glass beads produced locally and traded with neighbouring areas as well as with central Europe (Coles, 1987; Coles and Coles, 1986; Orme et al., 1983). It is even believed that the Glastonbury lake village (situated only 5 km away) developed as an offshoot in response to the successful trade centre of Meare (Coles and Coles, 1996). This, along with many examples within the Scottish and Irish crannog tradition, show that the majority of wetland settlements (prehistoric and historic) were certainly not isolated entities but were well integrated with local, regional, and interregional socio-economic networks.

On the other hand, an example of significant wetland settlements acting as obstacles on a long-distance trade route may be that of the Middle Bronze Age lacustrine villages of northern Italy. Finely worked amber beads covered with gold, travelling from the British Isles to south-eastern Europe are found in the northern Alpine region lakeside dwellings (e.g. ZH-Mozartstrasse), but not in those of northern Italy. Considering that links between the north and south of the Alps were already well established in the Neolithic (e.g. the flints from Monti Lessini found in various northern Alpine lake-dwellings) (Affolter, 2002; Hafner and Suter, 2000), as well as in the Late Bronze Age/Iron Age (e.g. the glass beads from Frattesina, Veneto region, to Hauterive-Champréveyres, Lake Neuchâtel) (Bellintani and Stefan, 2009; Rycyhner Faraggi, 1993), it is more likely that the avoidance of that particular traded item by the Italian lake villages was intentional, and probably not linked to specific trade networks. The Alps as a natural barrier, as argued by some scholars, may not apply in this case.

An interesting combination of short- and long-distance trade networks is that of the Late Mesolithic and Neolithic (often perpetuating throughout the Bronze Age) eastern Baltic Sea regions, and north-western Russia. Here, in (p.163) addition to non‐perishable materials (e.g. stone axes and flints), perishable ones such as seal fat and fur were traded down the line for relatively long distances (e.g. fur reaching the Middle Dniester River regions, mainly from the Belarus area, in exchange for grain and various other domesticates—especially during the latest phase of the Neolithic) (Zvelebil, 2006, 2008). These well-established connections eventually facilitated the adoption of agriculture and the establishment of wetland settlements (at first seasonal, then more permanent).

Although, as pointed out above, the purpose of this section is not to identify or reconstruct ancient trade routes, there are some wetland sites whose archaeological evidence is good enough to do so. For instance, the large amount of gravel, endemic to Sardinia and Ischia Island, found at Poggiomarino, has been interpreted as ballast placed in small boats (possibly dugouts) arriving empty from those two locations; the boats were then loaded with produce from the Sarno Plain, and sent back to Sardinia and Ischia Island (Cicirelli and Albore Livadie, 2008). Another good example is the identification of the Late Bronze Age Atlantic Complex sword trade network, stretching from Sardinia (Italy) to Spain and France, where the foremost exchange networks followed the major and minor watercourses up to the Atlantic and North Sea coasts and beyond (Cunliffe, 2001, Quilliec, 2001).

Ancient short- and long-distance trade involving wetland communities was, of course, not just a European phenomenon. Evidence of exotic goods traded along well-established networks is also found in China (especially in the Yangtze River Delta), Japan, Florida (e.g. Republic Grove), the Northwest Coast of North America (including Canada and Alaska), and New Zealand (mainly on the North Island), where inland water transport was particularly facilitated by the numerous swamps, rivers, small creeks, and shallow lakes (see e.g. the Bay of Plenty).

Within and Between the Wetlands: Trackways, Causeways, and ‘Roads’

In the absence of navigable waterways, wetland environments such as peatbogs can be treacherous places to cross. Yet, no matter how difficult it might have been, people always found a way to penetrate, explore, or simply cross them. The best evidence of this intense interaction is the numerous wooden trackways found within those bogs. In Europe, evidence of ancient trackways spans from the Neolithic to the Middle Ages and in some cases even much more recently. Trackways can potentially be found in any European bog, although in some areas, such as Ireland, the Somerset Levels (England), Lower Saxony, (Germany), the Dutch Bourtanger Moor in the Drenthe Province (the Netherlands), and in Denmark, they are particularly abundant. (p.164) The highest number of identified trackways (from all periods) belongs to Ireland, where more than 1300 toghers (from a few metres long, to >150 metres) have been recorded (Mcdermott, 2007: 24).

Different types of trackway, depending on the function, the period, and the material available, have been identified, with the most common ones being the brushwood, roundwood, hurdle, plank, and corduroy. Although the tendency is to order these types chronologically (e.g. brushwood earlier, and corduroy later), this division does not always work. For instance, the complexity and technological sophistication of the Late Bronze Age tracks in continental northern Europe tend to disappear in the Iron Age, when longer and more resistant (but more simplistically built) structures seem to be preferred. In Ireland, on the other hand, the toghers follow a more regular chronology; for instance, from the simpler Neolithic brushwood tracks of Corlea 8 and 9 to the more robust, longer, and more sophisticated corduroy ones of the Iron Age, such as Corlea 1 (Fig. 4.17) (Raftery, 1996c).

Apart from the Sweet Track (3807–6 BC) of the Somerset Levels, whose construction is unique, all the other above-mentioned types are found in most European bogs. The brushwood type, for example, although more used in Ireland (e.g. the Neolithic track of Corlea 8 and 9, and Derryoghil 2 and 4) and in the Somerset Levels (e.g. the Neolithic track of Bell and the Bronze Age ones of Viper and Tinney), is also found in continental Europe, and more precisely at Ipweger Moor, Germany (e.g. trackway XXX(Ip)) (Hayen, 1957, 1984), the Netherlands, and Denmark. The roundwood type (i.e. trackways made of longitudinally placed roundwoods) are those of Derryoghil 7, 8, and 29 (Bronze Age) in Ireland, trackway XXV(Ip) in Lower Saxon (Fansa and Schneider, 1990), and the Late Bronze Age one of XVIII(Bou) at the Dutch Bourtanger Moor (Casparie, 1984). Hurdle trackways are well represented in the Somerset Levels (e.g. the Neolithic trackway of Walton Heath and the Bronze Age Eclipse) (Coles and Coles, 1986), as opposed to in Denmark and the Netherlands where there are only a few examples (e.g. the Middle Neolithic trackway of Tibirke in Denmark and the Iron Age one of XIV(Bou) at Bourtanger Moor, the Netherlands) (Casparie, 1986b, 1987; Jørgensen, 1988). Plank paths consist of longitudinally laid planks usually secured by pegs. Examples of this type are the Bronze Age tracks of Corlona, Co. Leitrim and Curraghmore, Co. Offaly, the sixth-century AD Corlea 5 in Ireland (Raftery, 1996e), and the Bronze Age one of Meare Heath in the Somerset Levels (Coles and Coles, 1986; Coles and Orme, 1976). Plank paths in continental Europe are sometimes very similar to those in the British Isles (e.g. the Iron Age track of XV(Bou) at Bourtanger Moor, resembling that of Meare Heath, Somerset Levels). However, the single-plank walkway of XVI (Bou) found in the same area (Bourtanger Moor), is quite unique; in fact, no such constructions have been identified anywhere else in Europe (Casparie, 1984). Corduroy roads are made of transverses (split planks or roundwood (p.165)

Abundant, Well-Preserved Evidence

Fig. 4.17. The Corlea 1 trackway, Co. Longford, Ireland.

(Photograph: Courtesy of the School of Archaeology, University College, Dublin)

(p.166) stems) placed edge to edge on a substructure of longitudinal runners. In continental Europe they started being built quite early. The oldest one of this kind, which is incidentally the oldest wooden trackway in the world, is that of XXXI(Pr) (4780 cal BC), found at Campemore, Lower Saxony (Fig. 4.18) (Bauerochse, 2003).

Corduroy roads (Bohlenwege or Pfahlwege in German) are of crucial importance for the understanding the evolution of wheeled transport in prehistoric Europe. The difficulty of steering wheeled vehicles in the Neolithic required larger trackways (about 4 metres). From the Early Bronze Age onwards, the invention of the swingletree facilitated the steering of wheeled carts and consequently roads became narrower. One of the best examples of these Bronze Age trackways is track XVIII (Le) at Ockenhausen-Oltmannsfeld at Lengener Moor (Germany) (Fansa and Schneider, 1993). The roundwood planks of the trackways were first replaced by split planks in Lower Saxony around the middle of the second millennium cal BC. From then onwards a considerable effort was made to ensure a superstructure that was as level as possible. Standards of exceptional sophistication were not, however, achieved until the first half of the first millennium cal BC (mainly between the eighth and the sixth centuries), when in some cases the trackway surface was held together by thin straight laths of wood placed below and above the sleepers and fitted through slots in vertically projecting plank-like pegs. One of the best of the less than a dozen examples of this type of trackway (Lochbohlenwege) is that of trackway III(Pr) found at Grosses Moor (Lower Saxony) (Jacob-Friesen, 1963). Examples of corduroy trackways are also found in Denmark (e.g. Speghøje—Bronze Age), in Ireland (e.g. Derryoghil 1, tenth century BC; and Baunaghra, Co. Laois) (Raftery, 1996e). During the Iron Age (from about 500 cal BC onwards) the continental European trackways became less sophisticated, but at the same time extremely long (over 10 km). The longest ever built is the Iron Age (345 ±43 bc, or c.500 cal BC) trackway I (Bou) (the Valtherbrug), reaching 12 km (Casparie, 1986a, 1987). Further examples of this kind of trackway (which are remarkably similar of that of Corlea 1) are those of XXV(Pr) Schweger Moor, XLII Wittemoor, and VI(Pr) at Grosses Moor (Lower Saxony, Germany) (Fansa and Schneider, 1990; Hayen, 1971).

Not only do wooden trackways tell us about the various architectural techniques and carpentry skills of ancient wetland (and dryland) dwellers, but they also shed light on social interactions between different communities, wetland management (who had access to what in the bogs), and even climate change. The higher number of toghers in the Irish Mountdillon bogs during the second millennium cal BC may be, for instance, the result of an increase of wet conditions (Raftery, 1996c). The structure, length, and repairing of the toghers are also invaluable sources of information. In fact, a togher may have (p.167)

Abundant, Well-Preserved Evidence

Fig. 4.18. Trackway XXXI(Pr): the oldest wooden trackway in the world (4780 cal BC), Campemore, Lower Saxony, Germany.

(Photograph: courtesy of Andreas Bauerochse, Lower Saxony State Service for Cultural Heritage, Germany)

(p.168) been a temporary solution to reach specific places during wetter seasons, or it could have been a major construction, planned in advance, and used by various communities all year around. Although the majority of trackways were for utilitarian use, some of them may have had a sacred significance; they could have been, for instance, platforms for offerings (e.g. the XLII(Ip) trackway), or, with the presence of bog bodies, they may have had a more sinister function (see ‘Bog Bodies’, below). The numerous repairs of trackways, as well as some peculiarly marked planks obtained from the same trees but found 40 km apart (e.g. trackway IX (Le), Lengener Moor, and XII, Wittemoor), suggest a regional service responsible for the construction and maintenance of various trackways, covering a vast area. The size and length of these trackways should furthermore make us reflect upon the enormous community effort and organization of those social groups that built them. In some cases the quantity of material (wood) used was enormous. For instance, 600 planks, 350 piles, and 3,600 pegs would have been used to build the Sweet Track (Coles and Coles, 1989); 1500 planks and 6000 pegs for the XV (Bou) trackway in the Netherlands (Casparie, 1986b); and 450 hectares of forest would have been needed to provide the 65,000 planks for the 6.5-km long trackway of XII at Wittemoor (Fansa and Schneider, 1995), to mention but a few. Whether simply used for crossing, exploiting raw materials (e.g. iron ore resources), or exploring them for less profane purposes, road and trackway networks show an extremely high level of contact within and between the wetlands. The magnitude of these networks and the significant effort to develop and maintain them was certainly not restricted to wetland communities, but it was a joint cooperation of larger interregional and even cross-cultural groups.

In central and other parts of Europe (eastern Baltic Sea regions and western Russia) long wooden trackways were not built (or if they were, they did not survive). A few examples of trackways similar to those of northern Europe, although much shorter (a few tens of metres) are, however, found in the Lake Feder basin, southern Germany (see e.g. the Middle Bronze Age Bohlenweg of Bad-Buchau) (Billamboz, 1998; Schlichtherle, 2002). Short paths and walkways within, and leading to, lacustrine villages were nevertheless commonly constructed everywhere in Europe (e.g. crannog and island settlement walkways; or longer paths outside the lakeside villages, leading into the settlement, such as those of Chalain 19 (Lake Chalain, France) and Marin (Lake Neuchâtel, Switzerland) (Honegger, 2001; Pétrequin and Bailly, 2004).

Long wooden trackways are not found outside Europe, but, as within the European lacustrine villages, intra-settlement paths and walkways have been identified in China (Tianluoshan, c.2300 cal BC—path between rice fields), Japan (Juno, c.4000 cal BP), and New Zealand (see Ch. 2).

(p.169) Water Transport: Dugouts, Rafts, and Plank Boats

Although archaeological assemblages show clear evidence of ‘terrestrial’ transport and communication networks within and between the wetlands (e.g. from simple pedestrian trackways to large wooden ‘highways’ for wheeled vehicles) (see e.g. Louwe Kooijmans, 2006; Pétrequin et al., 2006c; Schlichtherle, 2006a), the most common means of transporting goods and linking places was certainly by water. Whether dealing with marine, riparian, or lacustrine environments, people have always taken advantage of the various natural waterway networks since the Palaeolithic and possible even earlier. However, direct evidence of archaeological watercraft does not appear until the Early Holocene. In Europe (and possibly everywhere else in the world), the first watercraft (except of course natural unworked logs) consisted of dugout canoes (or logboats) hewn from large tree trunks. The oldest logboat is that of Pesse (8010–7510 cal BC—Jaap Beuker, pers. comm. 2010) (Fig. 4.19) in the Netherlands, followed by Noyen-sur-Seine (France) of slightly later date (7190–6540 cal BC) (McGrail, 2001: 173).

Although in some areas they appear at later dates (e.g. Ireland and England in the fourth millennium cal BC; in Sweden 500 cal BC; Norway AD 700; and Finland not until AD 1200), logboats are found almost everywhere, and they never go out of fashion. The size varies considerably (2–3 metres up to 15 metres), but does not follow specific chronological patterns. The style of construction also varies significantly; from simply shaped one-piece dugouts with no added parts, to composite ones (e.g. with transom boards and/or block-stem/bow timber). One of the best examples is the Iron Age (c.300 cal BC) logboat of Hasholme in the Humber Estuary (England) (Millett and McGrail, 1987). Another remarkable exemplar was the Brigg dugout (c.1000

Abundant, Well-Preserved Evidence

Fig. 4.19. The Pesse dugout (about 3 metres long, 44 cm wide), the Netherlands.

(Photograph: courtesy of the Drents Museum, Assen, the Netherlands)

(p.170) cal BC) excavated near the River Ancholme in 1886, but unfortunately completely destroyed in an air raid in 1942 (McGrail, 2001). The technique of using transoms was not a later development (e.g. Bronze Age): some dugouts in Denmark (e.g. those of Horsekær, Halsskov, Zealand) were already constructed in this way in the Early Ertebølle Culture (c.6000–4800 cal BC) (Christensen, 1990, 1999). Dugouts are also found everywhere in the Circum-Alpine region, from the Neolithic to the Middle Ages. Even with the advent of plank boats and clinker later on, the use of dugouts (especially in swamps and lakes) continued until fairly recently. In addition to shedding light on the development of boat-building techniques, contextual studies on site location and site formation processes give us some clues as to how dugouts were maintained and repaired. For instance, stones and pebbles found in sunken dugouts have often been interpreted as ballast; however, the size of the stones and the location of the boats (especially when in shallow water close to the lake shore), suggest that they were sunk on purpose, by having large pieces of rock placed in them. This was done when the boat was not in use (possibly in winter) to prevent it from warping (Mcgrail, 2001: 174). One such example was found at Bernried on Lake Starnberg, Germany, where a Medieval logboat was filled with rocks up to 40 cm in diameter and sunk in a secluded part of the lake (Pflederer, 2009: 51). Another interesting find is the fireplace identified in some Danish dugouts. It is now known that it was used to provide heat and light for night spearing of eels (Andersen, 1994).

From the Bronze Age onwards, sewn-plank and plank boats began to be constructed. The best examples of excavated waterlogged sewn-plank boats are found in Britain: the Ferriby boats (2000–1700 cal BC), the Kilnsea (1800 BC), and the Brigg ‘raft’ (800 cal BC) in the Humber Estuary (McGrail, 1985; Van de Noort et al., 1999; Wright, 1990; Wright et al., 2001; Van de Noort, 2011), the Candicot fragments (1750–1100 cal BC), and the Goldcliffe fragments (c.1000 cal BC) in the Severn Estuary (McGrail, 2004), and the Dover boat in England (1550 cal BC) (P. Clark, 2004). Outside the United Kingdom, the first example of this type of boat is considered to be the Hjortspring (c.350 cal BC) in Denmark. However, some scholars have reservations as to whether it should be included in the sewn-plank category (McGrail, 2004). From AD onwards, new types of plank boat began to be developed; the flush-laid planking (fastened together with mortise and tenon joints), overlapping planking (fastened together by sewing), the nail-fastened clinker planking (similar to the Medieval Nordic tradition that would develop later), and the Romano-Celtic. Most of the boats of the latter tradition are canal and/or river barges, and they remained in use until the Medieval period. Their shape was particularly suitable for navigating deep as well as shallow rivers and lakes. Archaeological evidence has been found in Switzerland (Bevaix, Yverdon, and Avenches), along the main central European rivers, up to the North Sea, and as far as Wales (see for instance, Mainz and Xanten in Germany, Bruges in Belgium, (p.171) Blackfriars in England, and Berland’s Farm in Wales) (Arnold, 1996, 2004; McGrail, 1988, 2001, 2004, 2006). From the eighth to the eleventh century AD, seas and rivers were dominated by fast and manoeuvrable clinker Viking ships (see also the waterway sail barriers and the Skuldelev ships) (Rieck, 2003), although, as pointed out earlier, dugouts and flat-bottomed Romano-Celtic boats continued to be used along the majority of European watercourses.

Outside Europe, waterlogged evidence of boats (excluding more recent maritime wrecks) is more limited. There are, of course, exceptions that prove the rule, such as Florida (see below). However, lack of evidence does not mean that inland navigation and waterway networks were less important than anywhere else. In China, for instance, up to the historical time, riparian water transport was the most widespread means of communication. The most common watercraft was certainly the logboat, which, contrary to previous opinion (Needham, 1971), continued to be used up to the twentieth century (Peng, 1988). The highest number of logboats is found in the Jiangsu province (21). Other provinces have also yielded some examples: for instance, Guangdong (7), Zhejang (5), Fujian (1), and Guangxi (1). Unfortunately, only a few have been dated, but the oldest amongst them are the two canoes of Zhejiang (c.4250 cal BC). Those of Guangdong are younger; one dating 221–206 BC, and the other 260 BC–AD 10 (Peng, 1988). Japan is also fairly rich in logboats; about 200 have been excavated in the past two decades. The main sites are those of Chiba, Tokyo, Osaka, and Torihama, which has yielded one of the oldest dugouts in Japan (c.3500 cal BC). Of similar date, or possibly even earlier, is the dugout of Hashinawate 1 (a 14C date is still not available) (Takehiro, 2008). A place with a long tradition in navigation is New Zealand. A number of excavated pa (e.g. Kohika) have produced evidence of riverine as well as seagoing canoes with dugout hulls, separate lashed-on ends, planks, and thwarts (Irwin, 2006, 2004b). At Waitore Swamp, for instance, a decking plank, some paddles, and an outrigger float dating about AD 1500 have been identified (Cassels, 1979). The importance of the canoes and how the Maori used them was also recorded by Captain Cook when he was in the Hauraki Gulf and sent one of the Endeavour’s boats inland, in 1769. More than one hundred dugouts were noted by Captain Dell around his ship when he arrived at Oruarangi (Furey, 1996).

As mentioned earlier, one of the most prolific places in terms of numbers of canoes found in waterlogged environments is Florida. There are currently more than 350 canoes on record there (Barbara Purdy, pers. comm. 2010), with the oldest specimen being that of De Leon Spring, dating c.5120 cal BC (Engelbrecht, 1994). A remarkable site that has yielded more than one hundred dugouts is Newmans Lake in the northern part of the state. The canoe chronology spans between 500 and 5000 years ago, although the main concentration lies between 3500 and 4500. The reason for such a high number of canoes grouped in such a small area is still unknown, however, a few theories (p.172) have been developed recently. It is possible that the site could have been a discard area (a sort of canoe cemetery), or it could have been a manufacturing place, or finally, the canoes could have been deposited by wind and drift (Ruhl and Purdy, 2005). Canoes were largely used all over North America, although a denser distribution (according to the archaeological evidence) is noticeable in the eastern part of the continent (from the Great Lakes to Mississippi and Florida).

In Mesoamerica and South America, pre-Columbian waterlogged dugouts have not been found. However, other types of watercraft, which have not survived, were certainly in use (e.g. possibly reed boats on Lake Titikaka).

Finally, an important element of navigation is the paddle. A large number have been found either along with canoe/dugout and plank boat finds, or in complete isolation. The style, size, and decoration vary considerably from place to place, but do not follow specific chronological patterns. However, the shape of them seems to be linked to different aquatic environments. For instance, shorter and heart shaped ones are preferred in shallow and vegetation-rich water basins such as fens and swamps, whereas those more round and wide were used in large lakes, bays and rivers free from aquatic vegetation (see e.g. the beautifully decorated Mesolithic paddles of Tybrind Vig, in Denmark—Andersen, 2011).

Material Culture

Are some of the objects found in wetland contexts uniquely wetland objects? It is very tempting to say ‘yes’, especially if examples such as the bog butter containers in Ireland and Scotland are taken into account (Earwood, 1997). And what about fish-hooks, harpoons, fishnets, and other fishing gear? They are indeed strictly linked to the wetlands, but not necessarily used only by wetland communities. Inland groups may develop a close contact with the wetlands and even base their subsistence and economy upon riparian and lacustrine environments, without necessarily living within them. The more one tries to distinguish between the two categories (wetland and dryland groups), the more one realizes how ephemeral the distinction is. The purpose of this section is certainly not that of identifying what is and what is not a wetland artefact, but rather to recognize those objects found in wetland contexts that can help us understand people’s relationship with the wetlands. For instance, a beautifully carved part of a Maori whare purposely buried in an Aotearoan swamp is not only a serendipitously well-preserved object, but a piece of a jigsaw puzzle capable of shedding light on the Maori’s fascination for the wetlands (Phillips et al., 2002). Similarly, a skilfully woven basket of the Northwest Coast of North America will not only please our eyes with its (p.173) aesthetic beauty, but has the potential of telling us more about the intricate patterns of cultural sensitivity and interregional relationships between groups (Bernick, 1998; Croes, 2001, 2003). Textiles are particularly important, for they are often found in wetland contexts, but the majority (including the raw material, e.g. flax) were produced in ‘dry’ environments. Paradoxically, bone and antler artefacts, as well as being imperishable objects, are perhaps the most crucial for the study of the wetland–dryland interface. In fact, while perishable objects are often only found in waterlogged conditions (or in exceptionally rare dry conditions), imperishable ones (often including bones and antlers) are also found in dryland sites, therefore allowing comparative studies (see Ch. 1, Fig. 1.9, and Ch. 5, under ‘Preservation’ and ‘Conservation’).

Wooden Artefacts

When studying wooden objects, in addition to the morphological structure and function of the artefacts themselves, other important analytical techniques have to be taken into account. For instance, the identification of the wood species is crucial, as it may retain hidden symbolic values linked to the cultural choice (e.g. the importance of the chestnut in the Early Jomon Culture; or the cedar as a link between sacred and profane life on the Northwest Coast of North America) (Kobayashi, 2004; Stewart, 1984). The anatomical characteristics of the objects in relation to the parent tree from which they were obtained, as well as the wood-dressing techniques and tools used, are also crucial for establishing the provenance of the objects and their chronology (O’Sullivan, 1996; Sands, 1997) (see also Ch. 6, under ‘Dendrochronology’).

It is virtually impossible to list the entire variety of portable wooden artefacts found in wetland and/or wet contexts. However, it may be useful to allocate them to subdivisions according to their functions and construction techniques. In general, the most common wooden objects found in the wetlands are containers, ranging from carved and hollowed single-piece containers to more technologically sophisticated sewn bark, bent wood basketry (see below), and coopered ones. The second most common items are tools, which, depending on the period, can range from simple handles or digging sticks to more complex agricultural implements such as yokes, ards, and ploughs. Hunting and fishing gear include bows, arrows (with flint arrowheads, or the so-called blunt-ended bird bolts, also made of bone and antler—see below), spearheads, and clubs. In this section covering portable wooden artefacts, objects such as dugouts and boats, travois, wheeled vehicles, coffins, anthropomorphic and zoomorphic sculptures, parts of buildings, and other wooden structures in general are not included, since they are discussed elsewhere.

(p.174) As eloquently shown by the c.400,000-year-old wooden spearhead of Schöningen, Germany (Thieme, 1997, 1999), the fact that waterlogged archaeological evidence starts mainly from the Mesolithic onwards does not imply that wooden artefacts were not in use earlier. However, it is indeed from the Mesolithic that the first evidence of wooden containers dates. Some of the earliest are the birch-bark containers of Nizhneye Veretye (about 10,000–9,000 years ago), Russia (Oshibkina, 1989), Friesack (8950 ±110 BP) Germany (Gramsch, 1989, 1992, 2000), and Vis 1 (c.8000–7000 BP), Russia (Burov, 1998). Birch-bark containers remained the most widespread type of organic material items in north-eastern Europe and Russia until the end of the Late Iron Age, and in some areas even in historical times. Evidence of waterlogged containers in other parts of the world contemporaneous to the Mesolithic in Europe is very scarce; and, due to preservation issues, evidence remains limited to more ‘recent’ times.

From the Neolithic onwards there was a noticeable increase in wooden containers of all sorts, all over Europe. Birch-bark was still the main material used in some parts of the continent (see for example, the finely sewn bottom of the Bruszczewo container in Poland) (Czebreszuk, 2005), whereas in other parts, such as in the Circum-Alpine region for instance, carved and hollowed single-piece containers were more common. This latter type of container is found throughout the Neolithic, and until the beginning of historical times. Some of the best Neolithic examples are the numerous spoons, ladles, bowls, and dishes of Egolzwil (Switzerland) (E. Vogt, 1951; Wyss, 1973, 1976); the bowls of Niederwil, Lake Egel (Switzerland) (see Fig. 4.20) (Waterbolk and van Zeist, 1991); the ladles of Charavines (Bocquet, 1990; Bocquet and Huot, 1994; Bocquet et al., 1987); the oak tub and Maplewood cups from Reute-Schorrenried, Lake Feder (Schlichtherle and Wahlster, 1986); and the wooden dishes and trough found next to the Sweet Track, Somerset Levels (Coles and Coles, 1986).

Although still produced in the Bronze Age (see Fiavé, former Lake Carera) (Perini, 1987), carved and hollowed single-piece wooden containers are less popular in the Iron Age, when coopered, joined, and even turned ones started to appear. One of the best examples of a coopered container is the finely carved Bronze Age birch-bark jewellery box (containing shell, glass, and amber beads) from ZH-Grosser Hafner (1050–850 BC). Coopered wooden tubs have also been found at the lake village of Glastonbury (Coles and Minnitt, 1995; Tuohy, 2004), whereas the most succinct examples of turned wooden containers are found at the Iron Age–Roman period settlement of Feddersen Wierde, northern Germany (P. Schmid, 2002). From the Early Medieval period onwards, waterlogged evidence of wooden containers in Europe diminishes. This is once again a matter of preservation; broken small containers as well as large barrels are burnt as fuel rather than discarded. On the other hand, remarkable evidence of wooden containers of the past millennium is found in New Zealand, where beautifully carved wooden bowls have been (p.175)

Abundant, Well-Preserved Evidence

Fig. 4.20. Hollowed single-piece bowls from the Neolithic settlement of Niederwil, Lake Egel, Switzerland.

(Photograph: courtesy of the Amt für Archäologie Thurgau, <www.archaeologie.tg.ch>)

located at Gisborne, Taranaki, and Te Awamutu (Johns, 1998), and canoe bailers at Kohika (Wallace and Irwin, 2004).

Agricultural wooden tools and weaponry (bows, arrows, clubs, composite stone axes, and flint knives) were probably the most numerous artefacts up to the beginning of the use of metal, and in some regions of northern Europe even much later. In some areas of New Zealand, Australia, central and western Africa, the Northwest Coast of North America, and the entire Amazon Basin in South America, wood continued to be the most common material used by wetland communities until very recently. Well-preserved wooden implements have not only the potential of shedding light on the diachronic advance of technology through time, but they can also help us understand cultural change and adaptation processes in relation to the surrounding environment. For instance, the large quantity of blunt-ended bird bolts found in north-eastern Europe as well as in north-western Russia (e.g. Vis 1 in Russia, and the Kryvina peatbog in northern Belarus) could only be understood in association with the large quantity of migratory waterfowl fauna remains (Burov, 1998, 2001, 2009; Charniauski, 1997, 2007; Zhilin, 2007; Zhilin and Matiskainen, 2003). Another plausible example is the development of ards and ploughs. Technical characteristics of single ard-heads or composite ploughs (beam, foreshare, mainshare, tang of mainshare, and stilt) would only be isolated pieces of a jigsaw puzzle without their association with yokes and, most (p.176) importantly, a full understanding of animal traction development. It is therefore only by joining the pieces together, for instance plough-mark studies with ard and plough physical remains, along with a comprehensive understanding of animal traction and its consequences on the physical characteristics of the animals (e.g. bone stress and osteopathologies), that the full picture of cereal cultivation amongst farming communities will start to emerge (Lignereux et al., 2006; Louwe Kooijmans, 2006; Marzatico, 2006; Pétrequin et al., 2006b; Pétrequin et al., 2006a).

There are also composite objects (made of organic and inorganic material), such as wooden handles of stone and metal axes, or flint inserts for knives and sickles, whose organic component is crucial for the understanding and interpretation of the artefacts themselves. In the absence of the organic part, the artefacts are unrecognizable, as is, for instance, the case of knives and sickles made of flint inserts; once the handle is gone it would be impossible to know that the remaining scattered flint flakes were once part of a sickle (see the sickle from Fiavé, Fig. 4.21).

Another example is the Neolithic net sinker of Twann (Lake Biel), made of small pebbles wrapped and tied with bark (Stöckli, 1990b). Had the bark and cordage disappeared, it would have been impossible to identify the object as a net sinker.

Unusual Artefacts

In addition to the quite common (in wetland contexts) array of wooden agricultural artefacts and weaponry, waterlogged environments also yield particularly rare and finely crafted items made of organic material. These can be divided into musical instruments (see e.g. the Roman syrinx (panflute) of Eschenz, Switzerland, Fig. 4.22) artistic wooden carving (see ‘Anthropomorphic and Zoomorphic Wooden Figures’ below), and utilitarian objects. There are quite a few examples of this latter category within the Circum-Alpine region lake-dwelling tradition. One of the best-known objects is the Neolithic axe of Cham-Eslen (Lake Zug, Switzerland), whose long, straight handle was wrapped with finely cut and decorated birch-bark (Fig. 4.23) (Gross-Klee et al., 2002; Huber et al., 2009).

Other examples of prehistoric arts and crafts using birch-bark are the Neolithic clay pots of Estavayer, Saint-Aubin-Tivoli (Lake Neuchâtel) (Egloff, 1990), Hitzkirch (Lake Hallwil) (Bleuer et al., 2004), and Egolzwil (Wauwiler Moos) (Wyss, 1976), where birch-bark was ‘glued’ on the still-wet clay as decorative patterns. It is astounding the myriad of ways that birch-bark was used in prehistory. In north-eastern Europe and north-western Russia it was even used to repair broken pottery. In the settlements of the Kryvina peatbog for example, the broken fragments of a vessels were perforated and rejoined with plant fibre and underlying birch-bark (Charniauski, 2006). Interestingly, this technique of fixing pottery was also used by the Tripolye Culture people in Ukraine during the Chalcolithic (Kruts et al., 2008).


Abundant, Well-Preserved Evidence

Fig. 4.21. The Bronze Age wooden sickle with flint inserts from Fiavé, northern Italy.

(After Marzatico, 2004: 87)

Abundant, Well-Preserved Evidence

Fig. 4.22. The 11-cm long Roman (first century AD) Syrinx of Eschenz, Switzerland.

(Photograph: courtesy of the Amt für Archäologie Thurgau, <www.archaeologie.tg.ch>)

A final category of artefact worth mentioning is the hair combs. This type of artefact is found in a large number of wetland sites around the world; from northern Europe to the Circum-Alpine region, Japan, New Zealand, and North America. Although the majority of early combs are made of wood, some were also carved from bone and antler (e.g. the Neolithic antler comb of ZH-Mozartstrasse, the Bronze Age ones of the terramare in northern Italy, (p.178)
Abundant, Well-Preserved Evidence

Fig. 4.23. The Neolithic axe of Cham-Eslen (Lake Zug, Switzerland), showing the finely cut birch‐bark used to wrap the wooden handle.

(Courtesy of the Kantonsarchäologie Zug, Sabina Nüssli)

(p.179) and a few more in Bronze Age and Iron Age Denmark). Some of the best exemplars of wooden Neolithic combs are those of Arbon-Bleiche 3 and Sipplingen (both on Lake Constance), and that of Sutz-Lattrigen (Lake Biel) (Leuzinger, 2002; Schlichtherle, 1997b). Outside Europe, wooden combs have been found at Hoko River (c.3000–1750 cal BP) on the Northwest Coast of the United States; at Torihama in Japan, where a beautifully carved 6500-year-old Camellia japonica comb was even painted with lacquer; and finally at Kauri Point in New Zealand, where a remarkable collection (187) of deliberately broken combs were ritually deposited (Shawcross, 1976).

Weirs and Fish-Traps

Fishing activity includes a myriad of techniques and fishing gear, spanning from simple line fishing to harpooning, net fishing, fish-traps, and weirs. Although the use and efficiency of these various techniques changed according to place and time, the latter two techniques were probably the most efficient before the advent of fish farming. Weirs were most commonly used by coastal communities within estuary environments, but archaeological evidence of them has also been found inland, along rivers and even by lakes. Fish weirs are fence-like structures set in estuary tidal channels in order to guide fish into specific traps where they will be collected later. In most cases, archaeological evidence of weirs consists of upright wooden stakes placed in a V‐shape, and extending fully or partly across the tidal channel or river. Fish can simply get stranded in the V-shaped structure with the outgoing tide (estuary case), or be channelled into a portable basket-like trap (or even into spiral-shaped weirs—see Shidanai, Japan) along the current of the river or watercourse. Some weirs include removable elements, such as lattice-work panels, or basketry containers; consequently a weir can also be called a fish-trap.

In northern Europe and Scandinavia, weirs and other permanent fishing structures are known since the Mesolithic, with the oldest being at Kalø Vig I (7550 ±40 BP) in Denmark (Connaway, 2007). One of the best examples of Mesolithic weirs is found at Halsskov Overdrev, a former fiord (now a peatbog) on the Great Belt (Zealand, Denmark). Here, various wattle panels forming a number of weirs were constructed between the Middle Ertebølle and the Funnel Beaker period continuing in some cases until the Pitted Ware Culture (see Ch. 2 for chronologies) (Pedersen, 1999). Similar fish weirs were identified at Šventoji 9 (Lithuania) and at Zamostje in the Upper Volga region, whereas conical basket-like nets, or nets with wooden frames, were found at Sarnate (Latvia), Šventoji 1A and 2B (Lithuania), Sakhtysh (Russia) (Lozovski, 1999; Rimantienė 1992a, b), and Steckborn-Schanz, Switzerland (see Fig. 4.24). Beautifully made Early Neolithic dogwood basket fish-traps (presumably for eel fishing) have also been found at Bergschenhoek in the Netherlands (Louwe Kooijmans, 1987). Fish weirs are also found in the British Isles and in Ireland. (p.180)

Abundant, Well-Preserved Evidence

Fig. 4.24. Neolithic (Pfyn Culture) fish-trap of Steckborn-Schanz, Lake Constance (Untersee), Switzerland.

(Photograph: courtesy of the Amt für Archäologie Thurgau, <www.archaeologie.tg.ch>)

Two of the best-known locations are the Severn Estuary (England and Wales), and the Shannon Estuary (Ireland). In both cases, weirs were built from the Neolithic onwards, although the majority began to be constructed in Medieval times (see e.g. the fish-traps at Carrigdirty Rock and those of Bunratty in the Shannon Estuary, Ireland—see Fig. 4.25) (O’Sullivan and Daly, 1999).

Early Medieval weirs have also been found in the Blackwater Estuary and the Stour Estuary in Essex, England (Gilman, 1998). Weirs were also constructed in inland rivers and even in lakes (usually near river outlets). In prehistoric times the best example of a lacustrine fishing complex is the Iron Age (c.720–620 BC) weirs (with possible fishing huts) of Oggelshausen-Bruckgraben, Lake Feder, Germany (Köninger, 1999, 2002). Interestingly enough, this is the only evidence of Iron Age occupation on Lake Feder; no settlements of this period have ever been found. Examples of Medieval fish weirs of the riparian/lacustrine type have been found on the River Limmat (near Zurich Bay) and in Rapperswil, on Lake Zurich (Eberschweiler, 2004: 168). Weirs continued to be built throughout the Middle Ages and up to the nineteenth century. One weir at Kappeln, Schleswig-Holstein, Germany, is, surprisingly, still in use today (Roth Heege, 2007: 191).

Weirs and fish-traps are ubiquitous features of prehistoric and historic fishing activity, and waterlogged evidence of such structures is found in various places all over the world. However, an area with one of the highest concentrations (more than 1000) of estuary, riparian, and coastal weirs is the Northwest Coast of North America; from the various estuaries of the Oregon Coast (United States), to (p.181)

Abundant, Well-Preserved Evidence

Fig. 4.25. In situ remains (a) and schematic reconstruction (b) of the Bunratty 6 Medieval fish-trap in the Shannon Estuary, Ireland.

(Photograph and drawing: courtesy of Aidan O’Sullivan)

British Columbia (Canada), as far as Alaska (Mobley and Mccallum, 2001; Moss and Cannon, 2011; Moss and Erlandson, 1998). In Oregon, for instance, from a study including some of the main estuaries and bays (Coquille, Coos, Siuslaw, Yaquina, Netarts, and Nehalem), not only has Byram (1998) been able to identify a number of weirs of the past two thousand years, but he has also classified them into different types: (a) tideflat weirs, (b) cross-channel tidal weirs in tidal sloughs, and (c) cross-channel non-tidal weirs (including the channel margin weirs) (Byram, 1998: 206). The study concludes that different species of fish were harvested using different lattice weir panels, and the size of the lattice gauge depended upon the targeted catch. Furthermore, in some cases (e.g. at Coquille), the location of the weirs in relation to their chronology has helped in reconstructing the dynamics of the constantly changing estuary morphology (Ivy and Byram, 2001). Similar studies of three sites within the Fraser River estuary area (Canada) have helped identify changes in fishing techniques between 4600 and 200 years ago. For example, the simple traps used at Glenrose Cannery 4600 years ago were replaced by gill nets at Musqueam Northeast, 1600 years later. Around 2000 years ago, the latter went out of fashion and trawl nets began to be used (Stevenson, 1998).

Fish-traps and weirs have also been located in other parts of the North American continent. Two of the best examples are the basket-like fish-trap of Montana Creek in Alaska (also important for geomorphological studies and palaeoenvironmental reconstructions—see Ch. 6), and the mysterious agglomeration of stakes (about 65,000) interpreted as a series of fish weirs spanning the period from 5000 to 3700 years ago, found at Boylston Street in Boston, Massachusetts (Décima and Dincauze, 1998).

In tropical places, due to poor preservation, prehistoric weirs and fish-traps are more difficult to locate. However, in some areas of central and western Africa, Asia, Oceania, and Central and South America, with a special emphasis on the Amazon Basin, the use of weirs and fish-traps has a long tradition, and in some cases local indigenous populations still construct them today.

(p.182) Basketry and Cordage

The contribution to archaeology made by basketry and cordage artefacts found in waterlogged conditions is remarkable. Spanning from the Early Holocene to just about the present, basketry and cordage items not only offer a glimpse into our ancestors’ skills in manipulating plant fibres, but they also tell us more about the delicate relationship between people and the environment from which the raw material was obtained. Depending upon the level of preservation and the importance that basketry and cordage traditions had within a specific cultural group, they can even go a step further and shed light on social aspects that reflect cultural continuity (or the lack of it) within a specific cultural group, or its diachronic development as well as interaction with other interregional communities. If superficially considered, the ubiquitous distribution of wetland basketry and cordage artefacts and their apparent similarities of weaving techniques may be misinterpreted. However, as with pottery, the creation and development of a specific way of weaving reflects cultural sensitivity between contemporaneous groups, or within the diachronic perpetuation of a single family group, household, or community. The remarkable preservation of Friesack material culture has, for instance, allowed the identification of technological differences between the first (c.9000 cal BC) and the third (c.7850 cal BC) phase of occupation. Within the first three hundred years, people used knotless netting and plaited ropes. These two techniques were replaced by the 2-ply strings and knotted nets in the second occupation, but, interestingly enough, whoever settled there in the last phase used the same techniques as the first settlers some two to three hundred years earlier (Coles and Coles, 1996; Gramsch, 1989, 1991, 1992) (see also Fig. 4.26).

A similar case is found at Danger Cave in Utah, where excellent stratigraphic control made it possible to identify a change in basketry techniques (with raw material collected from the wetlands) over a time span of about 6000 years (12,000–6000 years ago); from an initial use of only the twining technique (12,000 years ago), to a 15 per cent twining and 85 per cent coiling around 6000 years ago (Jennings, 1989).

The most diverse collection of basketry work, including woven mats, raincoats, hats, and shoes is found within the lake-dwelling tradition of the Circum-Alpine region from the Neolithic to the Late Bronze Age. For instance, evidence of hats made of bast has been discovered on both slopes (north and south) of the Alps, as is shown by the Neolithic finds of Hornstaad-Hörnle 1A, Wangen-Hinterhorn, Sipplingen (Lake Constance, Germany), Seekirch-Achwiesen (Lake Feder) (Feldtkeller, 2004), and the Bronze Age rigid conical hat from Fiavé (Perini, 1987). Similarly, evidence of bast‐woven shoes was found at the Neolithic sites of Allensbach (Lake Constance) and Saint-Blaise/Bains (Lake Neuchâtel) (Schlichtherle, 1997b). An insole moss (p.183)

Abundant, Well-Preserved Evidence

Fig. 4.26. Chronological change in material culture at the Mesolithic site of Friesack, Germany.

(© Somerset Levels Project—John and Bryony Coles)

(p.184) pad with therapeutic properties was also found at Zug-Schützenmatt (Lake Zug) (Hochuli, 2002; Leuzinger, 2004; Schibler, 2001b). Although typological analyses of these hats have been carried out (Feldtkeller, 2004: 59) and analogies between the contemporaneous Zug-Schützenmatt and Ötzi’s shoes have been attempted (Reichert, 2000: 69), cultural sensitivity of the objects has not as yet been proved. Wetland sites on the Northwest Coast of North America have, on the other hand, exceeded all expectation in terms of identifying subtle cultural and social differences. Here, comparative studies as well as cladistic analyses on basketry remains of the past 4000 years have allowed scholars to identify cultural differences and similarities amongst the different local communities through space and time (Bernick, 1998; Croes, 1995, 1997; Croes et al., 2007). The results have been remarkable and in some cases even unexpected, such as the marked cultural dislocation (as opposed to the expected cultural continuity) in the Coast Salish area (Bernick, 1998: 153). With the possibility of ethnographic comparisons, as well as first-hand involvement with local First Nations communities, the potential of basketry and cordage studies is enormous. A succinct example is Croes’ attempt to identify a person’s life cycle from birth to death. In his study, Croes (2001) eloquently shows how basketry items are specifically constructed to contain humans at birth (e.g. cradles), how other objects (e.g. hats and harpoon sheaths) accompany people through their developing life and social status, and, finally, how woven mats are used as mortuary shrouds at the end of people’s lives.

Plait work and netting are also found in New Zealand, where, however, cultural sensitivity is more difficult to detect, as opposed to wood carving that does indeed allow cultural and chronological distinction between the various techniques (Johns, 2001; Mead, 1995; Simmons, 1994).


Although wooden structures of looms have not been found in waterlogged environments, the presence of a large number of spindle whorls (see Fig. 4.27), loom weights, bundles of fibres, and above all, entire and/or fragmented clothes proves that weaving activity was certainly germane within past societies (Altorfer, 2010; Altorfer et al., 2001).

Some of the best evidence of clothes comes from the peatbogs of northern Europe (e.g. the Yde girl’s woollen cloak in the Netherlands, and the woollen skirt from Damendorf, Germany), or from the permafrost funerary sites of Siberia. Although more limited than in the peatbogs, early traces of textiles are also found in other parts of Europe, from Scandinavia to the Alps. During the Neolithic, the main concentration of finds is once again in the Circum-Alpine region (see e.g. the large bundles of bast thread of ZH-Kleine Hafner, Lake Zurich and Twann, Lake Biel; or the finely woven fragment of linen garments of ZH-KanSan, Lake Zurich; Molina di Ledro, Lake Ledro, northern (p.185)

Abundant, Well-Preserved Evidence

Fig. 4.27. Spindle whorl and shaft covered in linen thread, from the Neolithic site of Arbon-Bleiche 3, Lake Constance, Switzerland.

(Photograph: courtesy of the Amt für Archäologie Thurgau, <www.archaeologie.tg.ch>)

Italy; Niederwil, Lake Egel, Switzerland (see Fig. 4.28); or the rolled-up panel of linen cloth from Twann, to mention but a few) (Hafner and Suter, 2000, 2004; Suter and Schlichtherle, 2009). Remarkable evidence of waterlogged textiles also comes from North America, and more specifically from Fort Center (Lake Okeechobee, Florida), where large textile sheets were used to wrap corpses before placing them on the elevated platform above the water, or from Windover (also used for mortuary practices) (see also ‘Mortuary Practices’, below). The sophistication of the Sabal palm/Saw palmetto fabrics (p.186)
Abundant, Well-Preserved Evidence

Fig. 4.28. Fragments of linen textile from the Neolithic (Ffyn Culture) lake-dwelling of Gachnang-Niederwil, Lake Egel, Switzerland.

(Photograph: courtesy of the Amt für Archäologie Thurgau <www.archaeologie.tg.ch>)

was astonishing; there were more than five variants of twining/weaving, and in some cases up to ten strands per centimetre have been identified (Doran, 2001b, 2002).

Bone and Antler Artefacts

The fact that bone and antler artefacts are found in both wetland and dryland archaeological assemblages facilitates comparative analyses and a better understanding of the interface between the two areas, which may not be possible with items only preserved in waterlogged conditions. Although bone and antler objects are found in almost all wetland archaeological assemblages all over the world, because of the favourable preservation properties of the soils in which the artefacts lay (low oxygen content in soil, fairly high pH, and calcareous water—see Fig. 5.17), the lakeside settlements of the Circum-Alpine region have yielded a remarkably high number of objects. This (p.187) significant amount of evidence (both from food discards or tools) have contributed to shed light on hunting activity, food procurement, economy, and tool, weaponry, and arts and crafts technology. For example, although both domestic and wild animal bones and antlers were used, a preference for wild animals is noticed. Antlers of red deer were highly exploited, and the majority of them were shed antlers (Schibler, 2001b). Only during difficult periods (crop failure, unfavourable climatic conditions, and demographic pressure) were the deer over-hunted and therefore also antlers that had not been shed were used (see Box 3.1). One of the most common types of antler artefact found in the Circum-Alpine region lake-dwellings is the stone axe sleeve/socket (see Fig. 3.4), which is a hollow piece of antler placed between the handle and the blade to absorb the blow, thereby reducing the danger of breakage.

As far as bone objects are concerned, the most common one is probably the double-pointed awl used for arrows (one was even found stuck in a doe’s pelvis at Sutz-Lattrigen- Hauptstation (Lake Biel) (Hafner and Suter, 2000), spears, harpoons, and fish-hooks (Torke, 1993). Teeth (especially canine) and small boar tusks were also used as amulets or necklaces, as the nice collection of perforated teeth from Sutz-Lattrigen Hauptstation and Rütte shows. In some cases (e.g. at Twann, Lake Biel), even dog metapodials were used as pendants (Hafner and Suter, 2004). Contrary to what was previously taken for granted, perforated long boar tusks were probably not part of a necklace pendant, but simple attached to a person’s belt by a string and used as a pointed tool (Schibler, pers. comm. 2010).

Archaeological experiments with bone and antler materials have also shed light on cutting, grinding, polishing, and perforating techniques used to make the artefacts. It is now commonly agreed that antlers were worked while still fresh; there is also the possibility that specific substances obtained from plants and mixed with water could have been used to keep the material wet until the work was completed (see also Ch. 7).

As noted above, bone and antler objects are found in various wetland sites worldwide, but the uniqueness of bone carving found at the Kohika site in New Zealand is not matched in any other wetland contexts (it has to be pointed out, though, that bone carving is not just a characteristic of the Maori communities living in the wetlands—the same kinds of artefact are also found in dryland occupations). Apart from a few awls and chisels, the majority of carved bone artefacts at Kohika consist of pendants and fish-hooks made of human and dog bones (Fig. 4.29) (Irwin, 2004a).

Finally, hundreds of composite bone artefacts (bone-wood and bone-shell) have also been found at Ozette, Northwest Coast of the United States, with the most striking object being the wooden whale fin inlaid with more than 700 sea otter teeth (Daugherty and Friedman, 1983). (p.188)

Abundant, Well-Preserved Evidence

Fig. 4.29. Maori tiki pendant (a) and fish-hooks (b) (both made of human bone) from the Kohika pa, New Zealand.

(Courtesy of Geoffrey Irwin, University of Auckland, New Zealand)

Sacred Practices and Beliefs in the Wetlands

Because of their dual ambiguity (water bringing life, but also taking life away), wetland environments have always been considered as physical as well as spiritual boundaries. As a result, they are not only sources of life for subsistence, but also liminal places, functioning as a sort of interface between the physical and the spiritual worlds. In this section, the focus is placed upon the latter, exploring the different ways people interact with the wetlands in order to express their beliefs. It has already been seen how the wetlands (mainly peatbogs) were explored and crossed by means of trackways; now the discussion will focus on how some of these trackways were used for more sacred purposes, and/or how other, additional structures (e.g. post alignments, platforms, cults houses, timber circles, and other wooden structures) were built and functioned in people’s everyday lives. The delicate balance between ‘give’ and ‘take’ is shown by the numerous hoards, offerings, and ritual depositions that took place within the wetlands. A myriad of objects ranging from stone axes to anthropomorphic and zoomorphic wooden figures (sadly, even human sacrifices) were used to communicate and/or negotiate with ancestral spirits, gods, and other sacred entities. The ultimate function of specific wetlands (bogs and shallow ponds) was that of facilitating the passage from the earthly existence to the afterlife. Although not commonly used, funerary practices within watery environments (pond and bog cemeteries and dugout internments and cremations—see below) have been noted in various parts of the world, from the Early Holocene onwards.

(p.189) Ritual Architectural Structures

As pointed out in previous sections, trackways in bogs may have had a double function. They were used as ‘highway networks’ within and between bogs, as well as offering and sacrifice platforms from which to deposit various objects for sacred purposes. In addition to trackways, specific buildings and particular wooden structures were constructed and used as qualitatively different places within a more profane landscape, settlement, or even single house (T. Brown, 2003). At a settlement or village level, specific buildings had particular functions, which extended beyond purely practical purposes. Within the Tripolye Culture for instance, sacred and profane were fused together in each single house; oven, sleeping areas, and working spaces were situated in the same room as the altar where the interaction with the spiritual world took place. Such an integration between sacred and profane in limited spaces is hardly seen within wetland communities of prehistoric Europe. However, it has recently been noticed that within the lake-dwelling tradition of the Circum-Alpine region, similar building may also have existed. Because of their substantially different architectural characteristics, it is argued that they had particular functions. One of the best examples is found at Marin-Les-Piécelettes (Lake Neuchâtel), where a construction, built on top of an artificial mound, was located within the village, exactly at the end of a more than 100-metre-long causeway. A second example is that of Ludwigshafen-Seehalde, where the walls of a house were plastered with clay protuberances shaped as female breasts (possibly linked to the female fertility cult), and painted with particular motifs, commonly found in the Danube region and in the western Mediterranean (Schlichtherle, 2006b). Special‐purpose houses (Whakairo) were also built within the Maori pa in ancient Aotearoa (R. T. Wallace et al., 2004).

Entire special residential agglomerates were also constructed in difficult‐to‐ reach, secluded wetland areas. One of the best examples is the platform of Alvastra (Sweden), consisting of seventeen clustered rooms for special gatherings. The site was occupied about 5100 years ago and abandoned only after eighteen years. Ceremonial objects such as pottery and miniature axes made of amber, flint, and bone are part of the Alvastra archaeological assemblage. A few burials (carried out about two decades after the site was abandoned) were also identified (Göransson, 1995; L. Larsson, 2001).

Non-residential units for gatherings, such as timber circles, wooden temples or post alignments, and platforms are also known in other parts of the world. On the British Isles, the best examples are Seahenge (the timber circle of Holme-next-the-Sea, England), dated at 2200–2000 BC (see Fig. 4.30) (Watson, 2005), and the post alignment of Flag Fen near Peterborough (Pryor, 2005).

Similar circular wooden enclosures also appear in Japan, but the posts are enormous (up to 1 metre in diameter). One of the best known of these circles is (p.190)

Abundant, Well-Preserved Evidence

Fig. 4.30. Seahenge, Holme-next-the-Sea, England

(© English Heritage)

that of Mawaki (Late Jomon period, c.2645 ±25BP) (Yamada, 1986). Scholars’ opinions on the function of these large constructions are divergent; some believe that they were ritual centres and others that they were simply watchtowers (Matsui, 1999: 154). Of rather different shape, but still believed to be a ceremonial centre, is the square structure with horizontal wooden planks on which vertical horned posts were erected of Bargeroosterveld at Bourtanger Moor (the Netherlands) (Coles, 1984). Interestingly enough, despite the few offering objects found near the Flag Fen post alignment, and ceremonial artefacts scattered within the Alvastra gathering centre, no major depositions or hoards have been found in or near those two sites.

Hoards, Offerings, and Depositions

Although as pointed out by R. Bradley (1990), offerings and ritual depositions occurred in both wetland and dryland contexts, archaeological evidence in the wetlands seems to prevail. Unlike wood and other organic materials whose preservation is facilitated by waterlogged conditions, votive artefacts for ritual depositions are often made of inorganic materials, therefore not necessarily needing anaerobic conditions to be preserved. As a result, the higher number of ritual depositions in the wetlands seems to be linked to a preference for watery environments to perform this kind of activity.

(p.191) Votive depositions in the wetlands take two different forms: single- and multiple-object depositions. In both cases identification problems may arise. In the absence of a specific context, single-object depositions can be misinterpreted as chance losses. Similarly, multiple-find ones may be multiple diachronic depositions, which may be difficult to distinguish. The objects deposited vary considerably from place to place, although some are characteristic of specific periods (e.g. stone and flint axes and pottery in the Neolithic, metal objects in the Bronze Age). The objects deposited also vary according to the type of wetland. Some artefacts are closely associated with specific wet environments. For instance, swords are found more often in riparian hoards (with some exceptions in Scandinavian bogs—see below), whereas tools and other objects are more numerous in bogs. A distinction between ritual and non-ritual deposits has also been attempted. Levy (1982), for instance, describes ritual hoards as having specialized locations (e.g. bogs, springs, wells, etc.) and includes a restricted range of objects (e.g. mainly weaponry, ceremonial objects, food, etc.), which are not usually broken. Non-ritual hoards, on the other hand, have no special location, the assemblage is less stereotypical (e.g. tools, simple objects, etc.) and the objects are both damaged and brand new. This distinction, although useful, is not always applicable, especially considering multiple depositions over long periods, which may alter the stratigraphic chronology of the site. Regardless of the character of the deposition, there are various ways of performing it. It may be a private deposition (only a restricted group of people will experience it), or it may be public, although the person(s) performing it is unknown to the attending group. Or, finally, the deposition may be fully public and all the members attending it know the person(s) performing it (L. Larsson, 2001).

Although there is evidence of Mesolithic depositions (L. Larsson, 2001; O’Sullivan, 2007), they mainly occur from the Neolithic onwards, reaching a peak in the Bronze Age and in particular in the Iron Age. In some areas (e.g. Scandinavia and Ireland) depositions continue up to Viking times (Hedeager, 1999), whereas in other places (for instance, parts of northern and central Europe) they stop (or diminish considerably) after the Romans or during the Early Medieval period.

Apart from small regional differences in terms of objects deposited (e.g. Early Neolithic ceramic depositions in Denmark not occurring in Sweden), the character of depositions, as described above, is fairly ubiquitous all over Europe (especially during the Bronze Age and Iron Age). The number of sites linked in one way or another to wetland votive depositions in northern Europe from the Neolithic to the Medieval period is extremely high. A few of them, thanks to their remarkable assemblages, have contributed greatly to improving our understanding of votive deposition practices (e.g. the depositions in the Hindby bog in Sweden, spanning from the Neolithic to the Bronze Age). A remarkable deposition is the sixteen bronze shields of the Herzsprung (p.192) type (c.950 cal BC) found in the Fröslunda Bog near Lake Vättern, Sweden (Hagberg, 1988). Some deposition sites contain different forms of sacrifice; from single or small group depositions to large quantities of war booty offered to the gods as thanksgiving by the victors. An example that contains all these forms of offering is the Iron Age site of Skedemosse (on the Island of Öland, Sweden), which, in addition to a large quantity of weaponry and human and animal bones, also yielded seven gold arm rings weighing 1.2 kg in total (L. Larsson, 1998). Similar depositions are also found at Llyn Cerrig Bach (c.third century cal BC) in Wales, and at Illerup (Scandinavian Late Iron Age) in Denmark (Coles and Coles, 1996). The spoils of war often include boat offerings. One of the best examples is the three boats of Nydam (Denmark); one of which (the oldest, c.AD 200) was completely chopped up before the pieces were deposited in the bog, whereas the two later ones (c.AD 300–350) did not undergo that treatment (Rieck, 1999). The Hjortspring boat (the oldest waterlogged boat in Scandinavia—fourth century cal BC) is also believed to have been part of a ritual offering. As pointed out above, pottery, weaponry, and boats were not the only offering items, agricultural tools and personal belongings (e.g. shoes, combs, etc.) were also deposited. Examples of such depositions are the wooden beehive from Edewechterdamm in Germany, the numerous wooden wheels and hubs from Rappendam (Denmark), or those of the Drenthe region (spanning from c.2800 to 400 cal BC), and the Bronze Age shoes from Nieuw-Buinen and Barger-Compascuum 1 in the Netherlands, to mention but a few (Coles and Coles, 1989; Coles and Coles, 1996; Van Der Sanden, 1999).

Depositions in northern Europe continued until the Viking period, and they were still performed in bogs (not only indoors as previously thought) (Hedeager, 1999). Although archaeological evidence is rather scarce, votive depositions in wetland environments also took place in other parts of the world. A place where votive depositions and offerings are quite numerous is the Yucatán peninsula in Mexico. Here, especially during the Mayan period, people used to deposit a variety of objects in the cenotes. The best studied of these assemblages is the Cenote of Sacrifice in Chichén Itzá. In New Zealand, as discussed in Chapter 2 (under ‘New Zealand’), despite the vast majority of artefacts being buried in wetland contexts for practical purposes (protection, preservation, and retrieval), objects at only a few sites, such as the Kauri Point combs, were definitely votive depositions (Phillips et al., 2002; Shawcross, 1976; Wallace and Irwin, 2004).

Anthropomorphic and Zoomorphic Wooden Figures

Anthropomorphic and zoomorphic wooden figures are ubiquitous finds in most wetland archaeological assemblages all over the world. Generally speaking, two groups of carved figures can be distinguished: one including (p.193) figures used as cult objects or idols and effigies, and the other carved representations used as parts of buildings and/or other wooden structures. The first group is more widespread, particularly in prehistoric and early historical Europe, where a further distinction should be made between plank-shaped figures placed along trackways and believed to have had protective properties, and those naturally shaped, which are associated with offerings. The chronology of wooden figures in Europe is fairly long, spanning from the Mesolithic (the oldest being the figure at Willemstad in the Netherlands), to the thirteenth century AD, although the main concentration is between the second half of the first millennium cal BC, and the first few centuries AD (Capelle, 2003).

Belonging to the first category (plank-shaped) are the two figures (presumably a man and a woman) that stood at the beginning of the famous trackway XLII at Wittemoor, Lower Saxony (see ‘Within and Between the Wetlands’, above). Wooden figures of the second category are more numerous and are found in various parts of northern continental Europe and the British Isles. One of the most impressive collections of this type (containing more that thirty figures dating from about the Roman period) comes from Oberdorla in Thuringia (Capelle, 2003; Raftery, 1996b). Well known also are the fairly tall figures (male and female) from Braak in the Ankamper Moor (Schleswig-Holstein, Germany). The male figure is 2.75 metres, whereas the female measures 2.25 metres, and they both date to c.400 cal BC (Dietrich, 2003). Denmark also has a large collection of wooden figurines, with the best known being the man with an exaggerated phallus from Broddenbjerg (from about Roman period), and the sitting man from Rude-Eskildstrup (later Roman period/Migrations period) (Capelle, 2003; Coles and Coles, 1989). In the British Isles and Ireland, the chronology of the wooden figures spans the Neolithic to the Iron Age/Roman period. Some of the best-known examples are the Neolithic figures of Dagenham (England), the Late Neolithic/Early Bronze Age one from Lagore crannog (Ireland), the fairly large Bronze Age carving of Ralaghan (Ireland), and the even larger (but of later date—Iron Age) Scottish figure of Ballachulish. Towards the Middle and Late Iron Age, the figures become smaller again, but more sophisticated and of composite character, such as that of Ross Carr.

From about the same period is the sculpture of Corlea 1 in Ireland (Coles, 1990, 1993; Coles and Coles, 1989, 1996; Raftery, 1996b). Of slightly later date (Roman period—c.2000 years ago) is the long-discussed figure of Strata Florida in Wales. The provenance of the figure is still unknown, but it must have been imported, since it was made of boxwood, a species not indigenous to Wales at that time (van der Sanden and Turner, 2004).

Wooden figures are also known from other parts of northern Europe, from Les Sources de la Seine (northern France) to the Gorbunovo Moor (Russia), but interestingly enough, apart from one single figure from Wasserburg-Buchau (Lake Feder, Germany), no other votive figures are found within the Circum-Alpine region lake-dwelling tradition. Since preservation in this area (p.194) is excellent, and other idols and cult objects of other materials (clay and metal) have been found, it is still unclear why the lake-dwellers did not prefer wood for their cult objects.

Outside Europe, wooden figures as idols and effigies for votive depositions are not very numerous. An exception is Mexico, and more precisely the cenotes of the Yucatán peninsula, during the Mayan period. One of the best-studied cenote archaeological assemblages comes from the Cenote of Sacrifice in Chichén Itzá. Here, in addition to gold, copper, and jade objects, a number of finely carved sceptres and wooden figures for votive offerings have been found (Coggins, 2001).

Wooden carvings not used for depositions, but as artistic sculptures or decorations for houses or other structures are found in various parts of the world. Two of the most fruitful places are Florida (United States) and New Zealand. In Florida, the best wooden sculptures are known from the Late Archaic to the Early Ceramic period (c.5000–2500 BP), with remarkable artefacts from Tick Island (e.g. the turkey buzzard), the zoomorphic (e.g. eagle) posts of the funerary platform of Fort Center (Lake Okeechobee) (see ‘Mortuary Practices’ below), or the skilfully carved and painted masks from Key Marco (c.1500–1000 BP) (Bullen and Jahn, 1978; Coles and Coles, 1989; Gilliland, 1989; Macdonald and Purdy, 1982; Purdy, 1992; Sears, 1982). Wooden carvings have also come to light at Ozette, on the Northwest Coast of the United States. In addition to the previously mentioned cedarwood inlaid with sea otter teeth, a number of anthropomorphic and zoomorphic seal-killing clubs are part of the rich archaeological assemblage (see Fig. 4.31) (Daugherty and Friedman, 1983).

Within the Maori wetland pa, the majority of wooden artefacts were agricultural tools. However, a characteristic of the Maori pa-dwellers (and dryland groups too) was that of decorating parts of the house, such as door lintels and doorsills. Some of the best examples are those of Waitara, Te Rarawa, and Kohika, where also parts of gates and fences were finely carved with anthropomorphic and zoomorphic motives (Johns, 2001; Wallace and Irwin, 2004; R. T. Wallace et al., 2004).

Mortuary Practices

The paucity of archaeological evidence of cemeteries or single burials in wetland environments is in direct contrast with the richness of other archaeological remains. Despite the close relationship with the wetlands, people did not seem to prefer watery contexts for disposing their dead. As is discussed below, evidence varies from place to place. There are, however, areas where the absence is striking. Perhaps the most astonishing mystery is, once again, the Circum-Alpine region lake-dwellings where, despite the hundreds of (p.195)

Abundant, Well-Preserved Evidence

Fig. 4.31. Anthropomorphic and zoomorphic seal-killing clubs of Ozette, United States.

(Photograph: courtesy of Ruth Kirk and Richard Daugherty)

settlements occupied by thousands of people within a 3500-year-long tradition, not a single cemetery or isolated grave has been found. A handful of human bones have, of course, come to light within the excavated villages (Simon et al., 1995), but they were mainly the result of accidents (most commonly fire). It is fairly obvious, however, that the absence of archaeological evidence can in this case not be used as an argumentum ex silentio. Hence, the question arises: where did all those people dispose of their dead? There have been attempts to link lacustrine material culture to inland graves, but they have not been particularly successful. Other theories have, of course, been advanced, such as cremation on pyres and disposal of the ashes in the lake. Although extremely difficult to prove archaeologically, this may not be too farfetched, as, still today, Hindu believers wish to be cremated in the sacred town of Varanasi and their ashes dispersed in the water of the Ganges (Salvadei et al., 1997). Or maybe the answer should be sought within the terramare tradition (see Box 4.4), where necropolises were placed on higher ground well outside the settlements (Cardarelli and Tirabassi, 1997).

In northern Europe and Scandinavia, the mystery is less puzzling, as, despite a strong link to the wetlands, people used to live on their margins, and therefore, cemeteries and burials were placed in drier terrains. Ironically though, it is from the extensive northern European peatbog that the most astonishing evidence of human remains (e.g. bog bodies) comes. These well‐preserved bog bodies did not however have anything to do with funerary practices; the reason why those unfortunate individuals ended up in the bogs (p.196) was far more sinister (see ‘Bog Bodies’, below). Funerary practices linked to wetland contexts are nonetheless present, and the best examples are the dugout burials. In Øgårde (Denmark), for instance, a Neolithic (c.2640 BC) dugout canoe was found fixed in place by vertical piles, with a human skeleton nearby. It is believed that the deceased was initially placed in the canoe, but eventually washed out (Koch, 1998, 1999). Similar examples were identified at the Mesolithic sites of Møllegabet II and Dejrø; also note the Bronze Age find of Hove Å in Denmark (Rieck, 2003; Skaarup, 1995). A related funerary practice, although not in a watery context directly, is the ‘ship-setting’, which consists of a boat-shaped setting of stones with a grave. These kinds of grave are fairly popular in Sweden, especially on Gotland, where some 300–350 sites have been recorded, although less than a tenth have been investigated. The majority measure up to 20 metres long, but a few examples from Gnisvärd reach 45 metres. Some of the best investigated ship-settings are those of Slätteröd 6 in Sweden and Thumby, Schleswig-Holstein, Germany (Capelle, 1995; Strömberg, 1961). Cemeteries in peatbogs and lacustrine environments are not found either. Even at the Iron Age lake village of Glastonbury, despite the efforts of Gray and Bulleid, a cemetery was never located. The closest evidence of the villagers’ funerary practices were a few neonatal bodies buried underneath the floor of some houses (e.g. mounds 57, 70, and 71) (Coles et al., 1992; Coles and Minnitt, 1995).

Direct evidence of ‘real’ wetland cemeteries (e.g. people buried in shallow water, or in moist soft peat) is only found in North America, with the best examples being located in Florida (United States). Amongst the four best-known sites, namely Bay West, Little Salt Spring, Republic Grove, and Windover, the latter is certainly the most famous and the most researched. The Early Archaic mortuary pond of Windover (c.7400 BP) is the oldest cemetery of this kind. The dead individuals were wrapped in grass mats, placed in the pond, and fastened to the bottom by stakes (Fig. 4.32).

One hundred and sixty-eight individuals (about 50 per cent adults and 50 per cent sub-adults) have been identified, amongst which 91 crania still contained preserved brain (Doran, 2001a, 2002). The water chemical contents and pH have preserved bones and brains but not the skin. Based on the preservation of the brains, the bodies were probably placed in the water within 48 hours of dying. Remarkable handmade, non-heddle loom manufactured fabrics, textiles, and cordage (possibly part of the mortuary practice) were also found (Doran, 2001a). The other three above-mentioned pond cemeteries are of later dates than Windover, but do share some similarities. For instance, the technique of fastening the bodies on the bottom of the pond using stakes is noted in all four sites, although at Little Salt Spring (c.6800–5200 BP) the later bodies (c.5500 BP) were interred in the moist soft peat, because the water of the pond was retreating. Here the bodies were wrapped in grass and placed in extended fashion on biers of wax myrtle (Purdy, 1992). The preservation was good, and, as at Windover, some skulls still contained brain tissue. Only 35 (p.197)

Abundant, Well-Preserved Evidence

Fig. 4.32. Schematic reconstruction of the way bodies were buried at Windover Pond cemetery, Florida, United States.

(Drawn with permission of Glen Doran, Florida State University)

individuals have been removed, but it is believed that about a thousand were buried there. Republic Grove (c.6000 BP) yielded a large quantity of fauna and flora remains, as well as a myriad of artefacts (in particular stone beads near a child’s neck, believed to be part of a necklace possibly used during the funeral). About 37 individuals were identified and studied. The same number were recovered at Bay West (c.6500 BP, although here the bones were very shattered due to severe dredging activity (Beriault et al., 1981; Purdy, 1992). Fort Center, a site of much later date (from 2400 to 250 BP), shows a remarkable way of disposing of the dead. During the second (c.1750–1150 BP) of the four occupational phases, the inhabitants of Fort Center built an artificial pond, in which a 20 × 12-metre platform was erected and embellished around its edges with finely carved zoomorphic (eagle, turkey, owl, and duck) totems made of pinewood (Macdonald and Purdy, 1982). The platform was used to place dead bodies after they were wrapped in cloth (Fig. 4.33).

Around 1500 BP the platform burned and fell into the water with all the bodies; some 300 individuals have been identified. Interestingly enough, while bones, wood, and textiles were preserved, flora specimens did not survive (Purdy, 1992; Sears, 1982).

Archaeological evidence of burials in watery contexts such as the above-mentioned Florida sites is not matched anywhere else in the world. But does this really mean that, no matter how close the relationship to the wetlands was, people still prefer to dispose of their dead in drier environments, or is there more than meets the eye? (p.198)

Abundant, Well-Preserved Evidence

Fig. 4.33. Schematic reconstruction of the burial platform of Fort Center, Lake Okeechobee, Florida, United States.

(After Macdonald and Purdy, 1982. Drawing: Gordon Miller)

Bog Bodies

Despite the above-mentioned array of archaeological objects retrieved from wet or waterlogged contexts, the bog bodies undoubtedly remain the most intriguing and mysterious finds. Their remarkable preservation has allowed scholars to identify the most delicate body parts (e.g. skin, internal organs, nails, hair, etc.), which are normally not preserved. This astonishing preservation of bodily organs is mainly found in sphagnum moss-rich peatbogs. It is, however, not, as previously thought, the absence of oxygen and the presence of antimicrobial substances in the sphagnum moss that inhibits decay of the body, but the presence of a polysaccharide (sphagnan) in the sphagnum moss. Once the moss dies, the sphagnan is dissolved but, being an unstable compound, it is converted into humic acid via intermediate compounds. A complex chemical reaction extracts the calcium from the body preventing (or reducing) bacterial growth. At the same time, another series of complex (p.199) reactions (also known as ‘melanoidin’ reaction) reduces the availability of nitrogen for bacteria, facilitating the tanning of the skin (Clément and Proctor, 2009; Painter, 1991, 1995; van der Sanden, 1996). As a result, a delicate balance between the various components determines the final preservation of the body. In peatbogs, for instance, skin, hair, nails, and garments made of wool and leather survive well, but bones and clothes made of plant fibre (e.g. linen) may not. In a fen environment (which is more calcareous) on the other hand, bones and plant fibres are usually preserved, but the rest is not (see also the ‘Preservation’ section in Ch. 5).

Because of these different levels of preservation, bog bodies may be divided into two categories; those with perfectly preserved skin and internal body organs (but not bones), and those with bones (and sometimes brain), but not the rest (see e.g. Windover). Bog bodies of the former category are found mainly in northern Europe (see Fig. 4.34), whereas a large number of the latter have come to light in Florida, United States.

Because of the delicate balance between the various preservation factors (e.g. chemical compounds, climate, deposition, and period) a mixture of preservation levels may also occur; one of the best examples is the Tollund Man, with skin, bones, internal organs, leather clothes, and even brain outstandingly preserved (Fig. 4.35) (van der Sanden, 1996).

More than the remarkable preservation of body parts, the fascination about the northern European bog bodies is the mysterious and sinister way that they met their death; some of them were strangled, hanged, stabbed, and even beheaded. As a result, a myriad of theories as to how those unfortunate individuals met their fate have been developed in the last hundred years or more. Glob (1969), for instance, believed that the majority of bog people were sacrificed, whereas German scholars have been more cautious with their explanations, arguing that each case must be considered separately in its own specific context. Hence the need for more multidisciplinary research and reconsideration of old discoveries. It was with the discovery of the Lindow bodies (especially II and III) that the attitude towards their interpretation started to change (Turner, 1995; Turner and Scaife, 1995). Some bodies such as the Yde Girl, the Weerdinge couple, and the Zweeloo Woman, were re-examined and important medical pathologies identified (e.g. the Yde Girl suffered from scoliosis, and the Zweeloo Woman had extremely short forearms). Other bodies were even linked with neighbouring cultural groups (e.g. the Emmer-Erfscheidenveen Man’s clothing was similar to that of the occupants of the contemporary Danish trunk-coffins) (van der Sanden, 1996). A few more famous bog bodies revealed further details which had hitherto been unknown. For instance, the Iron Age Kayhausen Boy turned out to be younger (7–8 years old) than previously believed; and it was discovered that Roter Franz (Roman period) was probably a horseback rider, had a broken (p.200)

Abundant, Well-Preserved Evidence

Fig. 4.34. Area of bog bodies in Europe.

(Graphic: Ben Jennings. Base map created using STRM data and ArcWorld River and Lake Overlay)

arm and broken collarbone, and was probably killed by having his throat cut (Pieper, 2001, 2003; Pieper et al., 1999).

Re-examination of well-known bog bodies has also contributed to lessen biased assumptions. The Grauballe Man (c.400–200 cal BC to Roman period) for instance, was confirmed to have been suffering from arthritis attributable to ergot, but the ergot sclerotia were too small to have affected his mental state (Asingh and Lynnerup, 2007). The Weerdinge couple, previously thought to be a man and a woman, turned out to be two men. But the most astonishing and unexpected result concerns the famous 14-year-old Windeby Girl, who, because of Tacitus’ writings (in his Germania) was believed to have committed adultery. Careful examination, however, revealed that the bog body was that of a 17-year-old boy (Gebühr and Eisenbeiss, 2007).

Some reconsiderations of bog bodies, although still yielding fascinating results, have been, scientifically speaking, inconclusive. The best examples are the decapitation of Rendswühren Man and Dätgen Man, and the removal of Rendswühren Man’s penis; in neither case was it possible to determine whether the acts were ante-, peri-, or post-mortem. Even recently found bodies may present divergent interpretations, which in some cases are linked to different archaeological research traditions. For instance, reflecting the more cautious interpretative scheme of German scholars, the death of (p.201)

Abundant, Well-Preserved Evidence

Fig. 4.35. The Tollund Man’s face

(Photograph: courtesy of the Silkeborg Museum, Denmark)

Moora, the Iron Age (seventh-century cal BC) girl from Uchter Moor (Lower Saxony), was considered an accident rather than a sacrifice (Bauerochse et al., 2008). On the other hand, the Irish Clonycavan Man (Co. Meath) and the Old Croghan Man (Co. Offaly) (c.300 cal BC), have been more audaciously interpreted as sacrifices linked to tribal boundaries (E. Kelly, 2006). Although daring, the latter interpretation opens new perspectives for further analysis. Evidence that hints to other possible interpretations of bog bodies comes from the two individuals reburied under the floor of the Cladh Hallan Bronze Age house on the Hebridean island of South Uist (Scotland). Not only were the two 300–500-year-old bodies dug up from the bogs, but they were made up from different individuals (one from two individuals, and the other from three) (Parker Pearson et al., 2005, 2007).

These last three examples show how important it is to contextualize the finds in their surrounding environment (nearby settlements, location of the bogs, etc.). Contextualization rather than generalization is the way forward for future bog body research. This may render the bog people less mysterious and intriguing, but it will certainly avoid biased interpretations, which, in the long run, may compromise and spoil our understanding of prehistoric societies.

(p.202) Conclusion

The overwhelming amount of archaeological evidence, from entire settlements to microscopic plant remains, needs a systematic classification in order to appreciate its value and diversity. The richness of some assemblages and the paucity of others is not always the result of good and bad preservation. Geographical as well as cultural factors are, for instance, to be taken into account if one wants to be able to explain the difference between prehistoric lacustrine villages in the Circum-Alpine region and the (at times) contemporaneous crannogs in Scotland, island settlements in Poland, and the terramare in northern Italy. Long-distance trade and local exchange are better understood when the various means of transport, as well as the permanent communication networks (rivers, canals, paths, and wooden trackways) are considered in different spatial/temporal scales. The ubiquity of a vast range of wooden artefacts and agricultural tools certainly goes beyond a simple environmental deterministic explanation. A typical example is the regional diversity of basketry on the Northwest Coast of North America, where cultural differences (or similarities) can be detected through space and time according to the diverse basketry typology. Weirs and fish-traps can go a step further, allowing archaeologists to identify not only different economies, but also ancient fishery management, in relation to seasonal availability.

Because of their mysterious character, gloominess, and inaccessibility, wetland environments (especially the northern European peatbogs) have always been considered boundaries between physical and spiritual worlds. As a result, a number of activities ranging from offerings to various kinds of sacrifice were performed around and within them. Surprisingly though, the European as opposed to North American (see Windover, for instance) wetlands were never used as cemeteries. In the Circum-Alpine region, for example, despite the large number of lakeside settlements, spanning more than 3500 years, not a single burial ground was ever found in a waterlogged context. This is not to say, of course, that no mortuary practices were performed on the lakes. As discussed, there is in fact the possibility that those practices left no recognizable archaeological evidence. On the other hand, traces of sinister activities (human sacrifices, executions, etc.) are clearly evident in the northern European peatbog, in the form of well-preserved human corpses (bog bodies). The majority of these bog bodies show evidence of a terrible death (hanging, decapitation, throat cutting, etc.); hence, a number of theories as to how they met their death have been formulated. There is no doubt that some of those people were victims of sacrifices or executions, but it is incorrect to assume that all of them died in that way.

As discussed throughout the chapter, the amount and richness of archaeological evidence found in waterlogged sites is overwhelming. It is therefore crucial that protection, recovery, and conservation fall within well-planned parameters. This will lead to a more systematic and contextualized way of studying our cultural heritage, avoiding biased generalizations (see Ch. 5).