Printer Friendly

Counting microliths: a reliable method to assess Mesolithic land use? In this debate the authors tackle a problem fundamental to researchers and resource managers in the Mesolithic period: what sort of prehistory do flint scatters represent?

Surface scatter assemblages form the majority of the Mesolithic archaeological record in many regions throughout Europe. A central challenge for research is the use of these scatters in the determination of how land was used. Land use models require data with a high enough resolution to map the variability of social organisation through time over specific ecological zones. Can low resolution surface data on their own give this result?

In a recent paper Vanmontfort (2008) proposed a new method for modelling Mesolithic land use for regions where almost the entire archaeological record is comprised of surface assemblages. Vanmontfort counts individual microlith types (2008:152) in order to model the diachronic variability of Mesolithic land use between different regions. His method claims to reveal the 'behavioural changes' that lead to regional variability in 'exploitation intensity' throughout the 4000-year period of the Mesolithic. The model supports his conclusion that on their eventual arrival in the middle Belgian loess belt, Linearbandkeramik (LBK) farmers exuded 'conflict-avoiding attitudes' by settling in areas that were either uninhabited, or marginally inhabited, by Late Mesolithic foraging societies. Interestingly, this model directly contradicts the one recently published in this journal by Golitko and Keeley (2007), in which the Neolithisation of north-west Europe was interpreted as one of the most violent periods in prehistory.

Our critique addresses three specific assumptions:

1. Available datasets are random samples of each study region

2. Individual microlith types fit within uniform chronological sequences throughout the entire Rhein-Meuse-Scheldt cultural distribution

3. Microlith counts correspond with site-based and technological assessments of interregional land use variability.

Random sampling

Vanmontfort (2008: 150) considers differences in research and survey activity, taphonomy and sample size as 'complicating factors to keep in mind' because they 'play the same role for all periods or phases in a given region'. He claims that the method is immune to biasing factors because it takes a broader inter-regional scale of analysis (Vanmontfort 2008: 156).


This could not be further from the truth, particularly in the three regions of the loess belt where the LBK first settled (i.e. Hainaut, Hesbaye and the Kleine Gete) (Figure 1). Differences in research activity and methodologies can cause major biases both within and between different regions. One of the authors (PC) carried out intensive research in the north-western part of Hainaut in the late 1980s and noticed that amateur fieldwalkers focused mainly on sandy outcrops of hilltops and plateaux in the loess region, and largely ignored river valleys (Crombe 1989). While no valley sites are currently known from Hainaut, they surely existed and are awaiting future research. For example, in the Hageland region, which is environmentally most similar to Hainaut, Vanmontfort (2008: Table 4) reports 60 per cent of all collected microliths as coming from river valleys. In central Hainaut, fieldwalkers have focused on sandy soils, which have yielded a high density of sites. However, excluding the 25[km.sup.2] survey of the LBK cluster, the loamy areas of Hainaut have received much less attention thus far (Van Assche 2005: 47-8). Taphonomic factors must be considered seriously, especially in regions with hilly topography (e.g. Hainaut). Hillwash erosion may have sealed valley and slope sites. It is very likely that in the hilly landscape of the middle Scheldt Basin many sites are buried beneath colluvial deposits, like the site of Rebecq-Le Spinoi mentioned by Vanmontfort (2008:150). In fact, the diachronic exploitation trend reconstructed by Vanmontfort for the Hainaut region (Vanmontfort 2008: Figure 3), and its subsequent deviation from all other studied regions, proves the significant hindrance of these methodological issues for a comparative regional analysis.

The sample sizes used in the model are not representative of the archaeological record in each study region. Vanmontfort's study is based exclusively on published papers and a few unpublished theses. For example, while he claims that the Hesbaye region is empty of microliths (Vanmontfort 2008: 157), it is well known that many Mesolithic finds from this region still await analysis and publication in numerous amateur collections (A. Hauzeurpers. comm.). It is highly unlikely, given its ecological value and geographical proximity along the Meuse, that this region would not have been exploited throughout the entire Mesolithic period. This has already been indicated by the discovery of microliths (mainly points with 'invasive retouch' and trapezes) in pits of excavated LBK settlements along the upper Geer (Jadin & Cahen 2003; Robinson in press) and Kleine Gete (Lodewijckx & Bakels 2000) clusters. The chance of finding large numbers of microliths in these areas is statistically very small given the restricted dimensions of the LBK clusters. Even though the Kleine Gete cluster encompasses just 1[km.sup.2], Vanmontfort claims that this region was not attractive for Mesolithic foragers. This is clearly an incorrect claim, as the region and its immediate surroundings were exploited throughout the entire Mesolithic for its outcrops of high quality Wommersom quartzite, which are found just 3.5km north-west of the LBK settlements (Vanmoerkerke & Van Vlaenderen 1985). Wommersom quartzite was not only exploited throughout the Mesolithic period (Perdaen et al. 2008), but also comprises 10 per cent of all raw materials recovered from the LBK sites of this region. This suggests it has potential as evidence for contact between foragers and LBK people (Lodewijckx & Bakels 2000; Crombe in press).

Microliths and cultural-chronological uniformity

The most problematic assumption made by the microlith counting method is the association of individual microlith types with a uniform chronological sequence throughout the entire Rhein-Meuse-Scheldt culture area. Citing Lanting and van der Plicht's (1997/1998) critical study, Vanmontfort (2008: 150) argues that the problems of bioturbation and lack of stratigraphy encountered in Belgian and Dutch excavations yielded data that lacked the chronological precision needed for the seriation of microliths. He thus uses Ducrocq's (2001) data from the Somme Basin, situated c. 200km from his study area, to allocate microlith types to particular phases. But recent extensive work in the lower Scheldt Basin has yielded 78 radiocarbon dates from charred hazelnut shells on sealed wedand sites; this has enabled the construction of a reliable chronological framework, particularly for the Pre-Boreal and Boreal periods (Van Strydonck et al. 2001; Crombe 2002, 2005). The lower Scheldt attests to the most precise chronology throughout the entire Rhein-Meuse-Scheldt culture; the number of reliable 14C dates is almost three times that of the Somme Valley.

Data from the lower Scheldt Basin have indicated the differential role of particular microlith types within the same phase (Perdaen et al. 2008). Vanmontfort's model fails to account for the technological, and thus behavioural, complexities involved in different traditions of hafting microliths within particular Mesolithic phases. For example, recent research on Atlantic period microliths has revealed striking differences in the technological trajectories of the Somme and Scheldt basins (Robinson 2009), which clearly proves that chronological sequences from one region cannot be applied throughout the entire extent of the Rhein-Meuse-Scheldt culture. The same holds for the Early Mesolithic (Ducrocq 2001; Crombe 2002).

Microlith counts correspond with site-based data

We cannot find any clear proof in Vanmontfort's (2008) paper that an analytic method based on the regional frequencies of microliths yields any reliable diachronic occupation patterns that correspond with traditional site-based approaches to land use. Indeed, how can a method that ignores the bulk of the material recovered from surface assemblages (e.g. lithic debitage) give an insight into diachronic and inter-regional variability in 'exploitation intensity' and 'behavioural change'? Accordingly we decided to test the microlith-counting method against the traditional site-based approach. The data come from two intensively surveyed and studied areas in the lower Scheldt covers and region situated 50km north of Hainaut: the Meetjesland (Van der Haegen et al. 1999) and the Moervaart region (Van Vlaenderen et al. 2006). The prehistoric finds from both regions comprise over 74 000 lithic artefacts including 722 microliths (Table 1) (Crombe et al. 2008; Sergant et al. in press).

At site level it is generally not possible to make a distinction between Pre-Boreal and Early Boreal surface sites with enough certainty; this is mainly due to the fact that in the lower Scheldt covers and region, unlike the Somme Valley, and contrary to what Vanmontfort claims for the entire RMS area (cf. assumption 2), there is only a weak evolution in the microlith typology and frequency between both stages (Perdaen et al. 2008; Crombe et al. 2009). For the sake of comparison between both methods, we therefore decided to group the microliths from both stages together (Table 2).

A comparison between both methods (see Vanmontfont, below, Figure 3) for the Meetjesland reveals a rather strong similarity between the microlith and site approach. The general trend both in the microlith and sites frequency is characterised by a sharp decrease after the Pre-Boreal/Early Boreal (10-15 per cent) and an increase towards the Atlantic/Late Mesolithic (25/35 per cent). For the Moervaart, on the other hand, the outcomes of both methods seriously deviate. The transition from the Pre-Boreal/Early Boreal towards the Late Boreal shows a drop, both in the curve of the microliths (c. 28 per cent) and the sites (c. 22 per cent). However, the microlith curve indicates a stability between the Late Boreal and the Atlantic (both around 30 per cent), while the number of sites increases by half that much (from c. 22 to 33 per cent).

An inter-regional comparison (Figure 2) also reveals major differences between both approaches. Comparing the Meetjesland and Moervaart from a site point of view (Figure 2b) it can be concluded that both regions have a different occupation history. The Meetjesland is characterised by a much weaker and more marginal exploitation from the Late Boreal onwards (from 65 per cent in PB/EB to 25 per cent in Atl.), compared to the Moervaart which shows a more balanced occupation (from 45 per cent in PB/EB to 35 per cent in Atl.). This pattern does not emerge from the microlith analysis (Figure 2a), which shows a much more similar exploitation intensity in both regions, i.e. from 45/50 per cent in PB/EB to 30/35 per cent in Atl.

This intra-regional difference in occupation intensity as shown by the site analysis might be explained by the absence of open water in the Meetjesland. Evidence suggests that wetland environments became favourite settlement locations during the Atlantic when coniferous forest was replaced by deciduous forest and extensive peatland started to emerge (Crombe & Verbruggen 2002; Robinson 2007; Sergant et al. in press).

This example shows that microlithic distribution patterns are inadequate indicators of Mesolithic land use. The exclusion of site/assemblage-related information and palaeoenvironmental information can lead to erroneous conclusions and fail to account for the complex behavioural changes that took place through the course of the Mesolithic period. Also the extrapolation of regional (typo-)chronologies to a wider research area is not without problems.


Philippe Crombe (1), Joris Sergant (1) & Erick Robinson (2)

(1) Ghent University, Department of Archaeology and Ancient History of Europe, Blandijnberg 2, B-9000 Gent, Belgium (Email:;

(2) University of Sheffield Research School of Archaeology and Archaeological Science, 2 Mappin Street, Sheffield S1 4DT, UK (Email:
Table 1. Total number of microliths for the Meetjesland and
Moervaart areas in the lower Scheldt Basin.

 Meetjesland Moervaart

Unretouched base 87 186
Crescent 12 7
Triangle 23 53
Retouched base 18 51
Invasive retouch 13 73
Trapeze 61 138

Total 214 508

Table 2. Total number of sites/assemblages for Meetjesland and
Moervaart areas in the lower Scheldt Basin.

 Meetjesland Moervaart

Pre-Boreal/Early Boreal 14 20
Pre-Boreal/Early Boreal 14 14
or Late Boreal?
Late Boreal 1 6
Atlantic 5 15
Late Boreal or Atlantic? 0 2
Undated 20 13

Total 54 70
COPYRIGHT 2009 Antiquity Publications, Ltd.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2009 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Debate
Author:Crombe, Philippe; Sergant, Joris; Robinson, Erick
Article Type:Report
Geographic Code:4EUUK
Date:Sep 1, 2009
Previous Article:The reburial issue in Britain.
Next Article:Taking microliths into account.

Terms of use | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters