The transition from hunting and gathering to food production along the Lower Rhine and Meuse Valleys between c. 5500 and 3500 b.c. is part of a much wider cultural transformation that covered the whole North European Plain from Holland to Poland. Prehistoric living conditions varied widely over the landscapes of this region. Moreover, variability in research conditions in the main natural zones has resulted in unequal data sets from the various zones, forcing investigators to use different research strategies.


The lower courses of the Rhine and Meuse Rivers run through the country of the Netherlands, in the northwest corner of Europe, facing the southern part of the North Sea. About half of the Netherlands’ territory consists of the combined lowland delta of several rivers, including the Rhine, Meuse, and Scheldt, which unload their sediments as they flow into the North Sea. The other half of the Netherlands, to the east and south of the delta, consists of uplands covered by Pleistocene sediments: a wide belt of sandy soils, with a patch of loess in the most southern part of the country. In this discussion, these three environmental zones—delta, sand, and loess—figure prominently.

The Delta Lowland. The lowlands of the western Netherlands measure about 200 kilometers along the coast and extend over 100 kilometers inland. The delta has been drained and transformed into the famous Dutch polderland, but geological research provides a picture of its ancient landscapes. Along the coast were tidal flats, salt marshes, tidal creeks, and lagoons. Behind this tidal zone were extensive peat swamps, and along the rivers a levee and back-swamp landscape formed.

In regard to archaeological sites, the delta is a sedimentary and preservative environment that is also dynamic and thus destructive. Although many sites have been destroyed by erosion, other places were protected by sedimentation. Prehistoric sites of the delta are highly informative for archaeologists, because they have: (1) superbly preserved organic material; (2) natural stratigraphy in sediments that can be correlated with habitation; and (3) intra-site patterns preserved by clay and peat covers. Field research is expensive and technically difficult, but the rewards are great.

The Sand Upland. The upland sand region of the eastern and southern Netherlands is an almost-flat Late Glacial cover sand landscape, less than 30 meters above sea level but with occasional sand and gravel hills as high as 100 meters. It is drained by small streams, and the eastern part of the region is dominated by the lower course of the Meuse. About 90 to 100 kilometers wide, the sand region contrasts archaeologically with the delta lowlands. Neolithic living surfaces still lie uncovered at the present ground level. Sites are surface scatters easily discovered in farmland by survey. Thousands of sites are known, but their information content is low. Material of all periods is often mixed up and difficult to separate, while organic material, bone included, does not survive in the acidic sand.

The Loess Zone. To the south of the sand zone, in the southern province of Limburg, lies the northern fringe of the European loess belt, a region with very specific conditions and a core area for prehistoric occupation and archaeological research. It is a landscape of rolling hills and river terraces, all loess-covered, rising to a height of 300 meters. Slope erosion, colluviation, and alluviation have erased upland evidence of Neolithic occupation and have buried sites on the valley floors. Only the loess communities that preferred plateau locations and dug deep "artifact traps" in the form of pits, silos, and ditch systems are archaeologically known in some detail. The communities established by the Linear-bandkeramik (also known as the Linear Pottery culture or LBK) farmers of the sixth millennium b.c. are a prime example.


The Loess Zone. The LBK settlement cluster on the loess of southern Holland is one of the most thoroughly investigated Early Neolithic microregions in Europe. Beginning in the 1950s, excavations by P. J. R. Modderman at Sittard, Elsloo, and Stein provided detailed plans of LBK settlements. This work permitted the development of a typology of longhouses and led to studies of LBK settlement systems, settlement structure, stone-adze and flint procurement, and social structure. In the late 1980s, large-scale research continued with excavation of the palisaded early LBK settlement of Geleen-Janskamperveld.

Our detailed knowledge of the LBK settlements results from the happy coincidence of their heavy construction and deep pits on plateau-edge locations that were subject to moderate surface erosion. Information is thus available on site location, settlement layout, houses, raw material acquisition, technology, and plant use (based on charred macrore-mains). Bone has decayed almost completely in the decalcified loess. Environmental reconstruction is based on pollen diagrams from rare valley-floor peat deposits and on charcoal and seed identifications from pit fills. Charcoal from pits has provided dates that place the LBK occupation of southern Holland between c. 5500 and 4900 b.c., which is consistent with the dating of this culture across central Europe.

The loess of southern Holland has yielded considerably less evidence for the Rossen culture that followed the LBK in northwestern Europe, as well as for subsequent Neolithic cultures. A Rossen site has been discovered at Maastricht-Randwijck in a Meuse Valley-bottom location. Only the lower parts of some pits remained, but these yielded artifacts, charcoal, and plant remains. The Rossen culture was succeeded by the Michelsberg culture around 4300 b.c. Undated but certainly post-Rossen flint scatters are documented in the Lim-burg loess zone, especially on higher locations overlooking valleys. The most prominent Michelsberg sites are still the Neolithic mining centers, dated from 4000 b.c. onward. The well-known Rijckholt mines, with at least 600 and possibly many more shafts, have been investigated by professional miners.

The Sand Upland. The sand upland has yielded over four thousand Stone Age surface sites, but with no intrasite patterns and often mixed assemblages. Dating is based exclusively on flint technology, typology, and raw material. Despite intensive research, special sites that might have had a central function, like earthworks or ritual centers, are absent, nor is there burial evidence.

The Delta Lowlands. People settled in the Rhine-Meuse delta from the Mesolithic onward, and by lucky chance some of their sites have been discovered in special microregions that escaped erosion and where conditions for preservation, recovery, and excavation were favorable. Of particular interest are the dune tops and creek levees that provided small dry spots in the delta wetlands. Stone Age people settled on these high spots, and their rubbish was strewn down the slopes and into the surrounding marshlands, where it was covered over and preserved by later sediments and peat.

Two clusters of Early Neolithic sites, dated c. 4300-4200 b.c., occur in the freshwater peat zone, one in the IJsselmeer Basin, the other in the Rhine/ Meuse district. The first cluster, near the village of Swifterbant, includes settlements and small inhumation cemeteries on dune tops and on the levees of former creeks. The Swifterbant sites are highly informative due to the preservation of intrasite organization, preservation of bone and botanical remains, and the absence of earlier and later contamination. The second cluster lies in the Al-blasserwaard peat district, where systematic pros-pection revealed that most of the approximately 100 known dune tops were used as settlement locations in several Neolithic phases. No settlement structures survive on these dune-top sites, but Neolithic refuse layers on the dune slopes and in the peat cover are full of information, including wooden and bone artifacts, animal bones, botanical remains, and pollen.

An exceptional site was discovered in 1976 north of Rotterdam near the village of Bergschen-hoek, eight meters below sea level, where a small campsite was situated in a wetland landscape that was originally on a peaty lakeshore. Microstratigra-phy indicated that the camp was used for ten to twenty years. The remains were silted over shortly after its final abandonment and preserved in very good condition. These include reed bundles that formed the living surface, remains of a dugout canoe, impressive fish traps, and fish remains— scales included. Dated c. 4300 b.c., it can be considered a fowling-fishing station of early agricultural communities in distant regions. Many, perhaps thousands of such sites lie hidden under the delta deposits.

After 4000 b.c., the dune-top site of Hazen-donk provides a cultural yardstick for the next two millennia. Phases of intensive occupation were separated by periods of occasional use or even abandonment. The main activities at this site were fishing and hunting, primarily of wetland animals such as beaver and otter but also of large game such as red deer, roe deer, and wild boar. Most surprising is the presence in all occupation phases of domestic animals and plants, as well as pottery and polished axes, marking it as a fully Neolithic site. Yet its location is not one that is favorable for crop cultivation, so the cereals must have been brought in from elsewhere. Hazendonk must have served as a special camp for fishing, fowling, hunting, and herding by societies in transition to a fully agrarian economy.

One of the three fish traps made from red dogwood twigs found at Bergschenhoek, The Netherlands.

Fig. 1. One of the three fish traps made from red dogwood twigs found at Bergschenhoek, The Netherlands.


"Classic" Early Neolithic LBK settlements are restricted to the loess zone to which their agricultural system seems to have been intimately linked. But the situation there is complicated by the appearance of two unusual pottery styles, not found farther east, named La Hoguette and Limburg. These have distinct southwestern connections that reach as far as the Mediterranean. Their pottery is generally found in low percentages as an admixture in LBK pit fills. La Hoguette seems to be the earlier of the two, possibly even preceding the earliest LBK in our area of study.

The LBK communities were fully agrarian before their appearance on the Limburg loess. Crops included emmer and einkorn wheat, linseed/flax, lentils, peas, and poppy seeds, all but the last with Near Eastern origins. The poppy has west Mediterranean sources and, in addition to the La Hoguette-Limburg pottery, is a strong argument for contact with that region. The poppy seed is found mainly in the westernmost LBK and only occasionally in central Europe. Charred weed remains indicate small, shaded fields in the woodland. Experimental data suggests good yields over long time spans without manuring. Zoological evidence from the loess region is scarce but seems to indicate a low interest in hunting (only about 10 percent of the animal bones are from wild animals). Cattle are the dominant domesticated species, with pig second and sheep/goat third.

On the sand north of the loess, LBK adzes and arrowheads are thinly spread all over the Meuse Valley as far north as Nijmegen. Neolithic pottery— never more than a few sherds on a site and restricted to later LBK phases—is found only in the southern twenty to thirty kilometers of the sand bordering the loess and generally in association with an LBK flint assemblage. There is some non-LBK pottery on these sites, too. A "pure Limburg" assemblage (without any LBK sherds) has been found at Kesse-leyk, and La Hoguette-related sherds were found as far north as Gassel on the fringes of the delta.

What do these modest but significant finds north of the loess reflect? Exchange with Late Mesolithic groups? Expeditions or wanderings from the loess to the north for prospection, hunting, or cattle herding? Or even an extension of permanent Neolithic settlement into this zone? How are La Hoguette, Limburg, and LBK related? The "pure" La Hoguette and Limburg assemblages in this zone might reflect separate, possibly semiagrarian, groups outside the LBK territory. The Late LBK sites with pottery might be seen as a growing penetration of this zone, possibly with transhumant cattle camps. The wider spread of arrowheads and axes tells us that the zone up to 100 kilometers north of the loess must be considered a contact or "availability" zone.

The change from LBK to Rossen around 4900 b.c. represents the transition to a pottery style that had developed along the Upper Rhine between Mainz and Strasbourg. As with the LBK, there is a similar involvement with the area north of the loess, but the Rossen culture probably made wider and more intensive use of this zone. Ephemeral Rossen sand sites lie not far from the loess boundary, and there is a wider and denser spread of two leading types of groundstone implements, the high-perforated shoe-last adze and the broad wedge. The Rossen culture reflects continuity from LBK society, but with a distinct transformation, not only in pottery style but also in economy and settlement system.

Significant economic changes took place during the LBK/Rossen transition, reflected by site location and botany. The crop spectrum changed to bread wheat and barley instead of the earlier ein-korn/emmer combination. Moreover, site location seems to have been less prescribed and extended to valley floors. One may speak perhaps of a better adjustment to the specific geographical qualities of the regions as opposed to the more rigid LBK traditions. The subsistence evidence for the Michelsberg culture that succeeded the Rossen is very limited, but many archaeologists argue that it can be considered a "normal" agrarian Neolithic society. It is a great handicap that the Michelsberg sand sites lack biological evidence, which complicates the interpretation of the earliest delta evidence in its wider geographical context.

Until recently, we did not know much about the Mesolithic communities of the sand zone and the delta. In the early 1990s, some modest pottery finds on sites in the Northeast Polder of the IJssel-meer District were dated to c. 4500 b.c., and two baseless (but perhaps originally point-based) pots from Bronneger, in Drenthe Province, yielded accelerator dates of charred crusts at c. 4700 b.c. Based on these finds, it was assumed that the area was used by a western Mesolithic counterpart to the Ertebolle hunter-fisher-gatherer communities of southern Scandinavia.

In 1993 coring in advance of the construction of a new railway line near Rotterdam led to the discovery at Hardinxveld of Late Mesolithic sites on the tops of two small dunes about five meters below sea level. These sites were occupied during the period 5500-4450 b.c. Finds at these sites include the burials of humans and of dogs; many different wooden artifacts, including paddles, bows, and a dugout canoe; and large quantities of animal bone, including fish, birds, beaver, otter, wild boar, and red deer. In the upper levels (4700-4450 b.c.), the earliest bones of domestic animals north of the loess zone were uncovered in modest numbers, including cattle, pig, sheep, and goat. Cereals, however, are still absent in this phase. Finally, a small amount of pointed-base pottery appeared on these sites about 5000 b.c. Thus we have our first glimpse of the communities who may have been in contact with the LBK and Rossen farmers of the loess, 100-150 kilometers distant across the sand belt.

The next stage to be considered is the period of 4300-4100 b.c.. In the northern part of the delta the Swifterbant group can be identified. Its pottery technology and style have close similarities with Late Ertebolle, employing pointed bases, flaring rims, simple shoulder decoration, coiling, and organic temper. The flint industry is, however, quite different and derived from a local Late Mesolithic microblade tradition. Some groundstone broad wedges demonstrate a continuity of adze exchange until this phase. There is, however, one major difference from Scandinavian Ertebolle: the delta sites are distinctly semiagrarian, as will be described below. Contemporaneous assemblages in the southern half of the delta (Hazendonk 1, Brandwijk, Bergschen-hoek) have distinct technological and stylistic connections to the southeast, that is to the Late Rossen of the Rhineland. But there are also traits in common with Swifterbant along with some original characteristics.

The early delta Neolithic sites (4300-4100 b.c.) are located in agriculturally unattractive zones and on locations that offered restricted opportunities for farming and animal husbandry. Surprisingly, all sites produced quantities of charred seeds and chaff of cereals, and bones of domestic animals make up 10 to 50 percent of the total. In view of the location of the sites and evidence of many hunted animals (mainly beaver and otter), successive occupations by specialist hunters provides the most plausible interpretation for all Hazendonk levels. The Swifterbant levee sites, with evidence for complete households (milk teeth of children, burials), are probably summer residences, with permanent settlement as a second option. Bergschenhoek is undoubtedly a repeatedly used short-term winter fowl-ing-fishing camp.

This type of subsistence in the delta could be called not only semiagrarian but also "extended broad spectrum," since all classical Mesolithic subsistence activities (hunting, fowling, fishing, foraging) were extended with animal husbandry and at least the consumption, if not also the growing, of cereals. But the delta Neolithic sites are only the wetland elements of larger regional settlement systems. The presumed upland sites of these systems are as yet hardly known.

Some adoption of agriculture, more specifically animal husbandry, had begun north of the loess at least around 4500 b.c., while crop cultivation is only attested about five centuries later. The process was predominantly, if not exclusively, one of addition and not a new wave of colonization. Neolithic elements were included in a basically native Meso-lithic society: early pottery styles have distinct regional traits; settlement systems and subsistence strategy have firm Mesolithic roots and contrast to the "full Neolithic" of the loess zone.

We can conclude several things. First, apparently the delta wetlands were perceived as an attractive environment and that the demonstrably variable means of subsistence apparently was fully acceptable at that time. Second, in view of the ecological constraints of the delta environment, communities with a similar or even fuller adoption of food production should be presumed on the upland sand, independent of the functional interpretation of the delta sites. Third, the wide occurrence of Michelsberg sites in the Meuse Valley—contrasting with the absence of upland sites farther north—might be partially caused by the use of the highly diagnostic and conspicuous large Rijckholt flint artifacts, but it might also reflect a more permanent and stable settlement system.


We have to cope with not one but essentially with two problems: first, why did the Mesolithic people of the North European Plain not turn to agriculture in LBK times and, second, why they did they then adopt agriculture in the second half of the fifth millennium b.c.? This change took place all over Northern Europe—not exactly in the same way everywhere, but that’s not so astonishing in view of the size of Northern Europe. One might think of a technical or agro-technical improvement that made agriculture, especially crop-cultivation, sufficiently attractive to be adopted around 4200 b.c. The development of the ard, a light plow, might meet these requirements. It allowed the cultivation of large fields with relatively low yields on the poor or even acid northern soils. More speculative is the idea that improved crops were developed for cultivation in these conditions and at these latitudes, but there are no archaeological arguments in this respect.

Perhaps this is a situation that has no good modern analogy. We are studying the confrontation between fully agrarian colonist-settlers who practiced hoe cultivation and broad-spectrum hunter-gatherers with presumably restricted mobility. They met in an unspoiled temperate environment with full opportunities for all communities involved to select optimal site locations for their activities. Both populations, the colonists and the natives, had widely different cultural roots. Those of the LBK and its successors are traced to southeastern Europe and ultimately the Near East. They were non-mobile and built heavy, more than minimally functional, housing. Most striking, however, is their attitude toward nature, their perception of environment. Their way was to play it safe—to adopt a low-risk strategy. This meant that they utilized a very narrow range of subsistence activities, which are reflected in their choice of specific settlement locations on the edges of loess plateaus or along brooks in loess-covered districts and in their reliance on cattle and cereals. In other words, they disregarded the natural food sources of the area to a large extent.

The Mesolithic "natives" and their descendants, in contrast, had their roots far back in the Late Palaeolithic of Northern Europe. Their subsistence shows an appreciation of everything nature offered. Their perception of nature clearly was different from that of the LBK people. They were, moreover, mobile, with light "minimalist" housing.

Such differences in mentality can possibly explain the lack of adoption of Neolithic elements in the millennium of contact between 5500 and 4500 b.c. Fundamentally different attitudes had to be bridged. This implies that each culture complex gradually had to transform in the other’s direction. The adoption of cattle and crops at a given moment by the native communities might have had something to do with the lowering of risks in the harsh season, with the perceived prestige involved with agriculture, or with technological innovation that made the growing of crops more attractive than it had been before.

The beginning of the Neolithic is defined as a change in economy where domesticates become part of the subsistence. Other aspects of material culture, such as pottery, certain axes, longhouses, and burial mounds, are not a priori associated with the term "Neolithic." Cereals and livestock were introduced from the Middle East and spread through southeastern Europe to central and northern Europe. This process moved by fits and starts; domesticates extended very quickly over vast areas, followed by a standstill lasting up to several hundred years. The first Neolithic culture to make its way into central Europe was the Linearbandkeramik (57004900 b.c.). In northern Germany the earliest domesticates are found in the context of late Ertebolle-Ellerbek culture c. 4700-4600 b.c. In southern Scandinavia food production appears with the advent of the Funnel Beaker culture and at some late Ertebolle sites c. 4000-3900 b.c. The spread of food production in central and northern Europe is a process that has been the focus of debate and many investigations. The main question is whether farming spread through colonization or by the indigenous adoption of ideas by the local population. A combination of migration and local adoption is a third option.

The transition to agriculture in northern Europe began during the Atlantic climate zone, characterized by a relatively warm and damp climate; a dense climax forest of linden, oak, elm, and ash; and cyclical sea-level changes called the Littorina transgressions. By about 4000 b.c. the start of the Sub-boreal climate zone brought about a change toward a cooler and drier climate, but still warmer than today. A drastic decline in elm c. 3900 b.c. took place over central and northern Europe; this decline appears to have been a natural phenomenon caused by elm disease. Clearing of the woodlands is indicated by fewer numbers of the dominant trees of the primeval climax forest (linden, oak, and ash) and by a second growth of light-demanding trees, such as birch, poplar, willow, and hazel. Deforestation probably reflected the work of farmers as they made way for fields and pastures.

Around 5700-5600 b.c. the Linearbandkeramik culture brought the first farming settlements to the central European uplands as well as to parts of the North European Plain along the Oder and Vistula Rivers. The Linearbandkeramik economy was based almost entirely on domesticated plants and animals, and its settlements are concentrated on fertile loess soils along streams. The spread of the Linearbandkeramik is commonly attributed to the colonization of habitats favorable to agriculture through the progressive movement to the north and west of farming peoples from the Danube Valley. Analyses of strontium isotopes from Linearbandkeramik skeletons in the Rhine Valley suggest that local people also may have been involved in the establishment of these early farming communities.

Next post:

Previous post: