The Neolithic transition and European population history.

Philippe Crombe; Mark Van Strydonck

In volume 295 of Antiquity M. Gkiasta et al. (2003) discussed the results of two sets of analysis carried out on a "new" database of radiocarbon dates: one for the whole of Europe examining the spread of the Neolithic, and one regional approach looking at the relation between Mesolithic and Neolithic dates. Although we are convinced of the potential of both approaches, we do have some major comments on the methodology.

First of all the analyses were conducted on a highly incomplete database. As the authors state on their p. 48, the analysed database currently includes over 2600 samples. Many of them, however, had already been collated in Gob's Atlas of [sup.14]C dates (1990). Although the authors have included new dates, we do not believe that this has been done very systematically. For the Belgian territory, for example, virtually all the dates used in the article were those published by Gob--16 Mesolithic dates and 30 Neolithic dates. The authors justify this by referring to the bad state of publication and public availability of radiocarbon dates in Europe. This certainly does not hold for the Belgian territory. In the last decade over a hundred new Mesolithic and Neolithic dates have been produced, the majority published in journals available world-wide such as Radiocarbon (Van Strydonck et al. 1995; 2001a), Antiquity (Crombe et al. 2002), Archaeometry (Cauwe et al. 2002) proceedings of the international congresses such as [sup.C] and archaeology (Crombe et al. 1999) and The Mesolithic in Europe (Crombe 1999), and the IRPA- datelists (Van Strydonck et al. 2001b; Van Strydonck et al. 2002). The authors assert that these "shortcomings" to the database probably do not affect their conclusions. This is a rash and provocative statement, which minimises all recent progress in absolute dating of the European Mesolithic and Neolithic. We believe that for the Belgian situation a hundred new dates can make a difference. In recent years, for example, these new dates have allowed a thorough revision of Mesolithic chronology (Crombe 1999; Van Strydonck et al. 2001 a) and a refinement of the (early) Neolithic chronology (Jadin & Cahen 2003). This will certainly also be the case for the other study-areas in Europe.

We also have questions about the way the radiocarbon dates have been selected in this study. On p. 48 it is mentioned that the same criteria as Gob used in his 1990 publication were applied to exclude dates from the datelist. However, Gob's criteria are not always very clear and sharply defined. Although he claims that his judgements are largely based on parameters such as sample quality, sample treatment and the degree of association between the dated sample and the archaeological material and/or feature, we do observe in his publication a number of contradictions and inconsistencies. For instance, some well-associated dates are rejected, while some badly associated samples are accepted. To illustrate this we would like to refer to two Belgian examples. A first example concerns the Late Meolithic site of Brecht "Moordenaarsven 2" (Gob 1990: 58), which is dated by three dates on charcoal ranging between 6270[+ or -]120 BP and 7990[+ or -]110 BP. Although all three have a bad spatial association, two dates are accepted as reliable without further comment by Gob, while the third one has been rejected. At the LBK-site of Darion (Gob 1990: 58), on the other hand, from a total of six charcoal dates (ranging between 6145[+ or -]145 BP and 6770[+ or -]75 BP), two are eliminated by Gob. Yet, all six dated samples had a good spatial association with the archaeological material, as they were retrieved from pits and postholes. Similar inconsistencies can be observed in the datelists for other European countries in Gob's atlas of 1990. We believe that many of these judgements are guided by typological criteria. As such, dates that fall outside the typo-chronologically expected time-range are rejected from the inventory. This of course is a very subjective way of selecting and creates a biased database, which only serves to confirm the predicted relative chronology of assemblages and cultures.

The quality of the dates obtained recently is generally better than those obtained prior to 1990. Since the development of the AMS-method in the 1980s, the number of radiocarbon dates on bulk samples has decreased considerably in favour of dates on smaller samples or single entity samples (Ashmore 1999). As a result of this, the risk of contamination has reduced significantly over the years. Perhaps most bulk samples should have been rejected for this type of exercise. Several studies have pointed out the problems related to the dating of charcoal samples on both Mesolithic and Neolithic sites (Gob 1990; Crombe et al. 1999; Whittle 1990; Stauble 1995; Lanting & van der Plicht 1993/4). It is nowadays generally accepted that contamination with naturally produced charcoal and the "old-wood-effect" may affect dates made on charcoal. In 1995 H. Stauble clearly demonstrated that a greater chronological resolution for the Early Neolithic LBK-culture is obtained if only dates on short-lived materials, such as cereals and food remains on potsherds, and bones are used. Likewise, for the Mesolithic it was concluded (Crombe et al. 1999) that on highly bioturbated and unstratified open-air sites better results are obtained by dating carbonized hazelnut shells rather than charcoal. Therefore in order to eliminate all sources of uncertainty, perhaps the authors would have done better to exclude all charcoal dates from their analysis, especially those from unstratified open-air sites. Of course this would reduce the number of available dates drastically, but still in our opinion it is better to work exclusively with the most secure and well-screened dates when tackling problems such as the transition between the Mesolithic and the Neolithic.
Serious objections can also be made about the averaging of radiocarbon dates according to artificially delimited regions, such as current state borders. Several scholars have recently emphasised that in some Central European countries the neolithisation process may have varied considerably on a local and/or regional level, in a recent paper Detlef Gronenborn (1999) suggested that in some areas of Central Europe the Neolithic was introduced by demic diffusion, whereas in other parts (e.g. Transdanubia and SW Slovakia) acculturation or a combination of demic diffusion and adoption was probably the motor. Hence, a regional analysis of radiocarbon dates should be organised in smaller areas, defined on the basis of, for example, palaeo-environmental and/or cultural criteria. In the paper under discussion this was apparently only done to a certain degree for the French territory; the French dates were subdivided into a southern Cardial-related and northern LBK-related zone. Unfortunately, no such attempt was made for the other countries. Even for small territories, like Belgium, it is advisable to split the analysis into different "ecological" zones. Referring to a paper recently published in Antiquity (Crombe et al. 2002), it is clear that the introduction of the Neolithic way of life followed a different tempo in the northern, sandy lowland from that in the southern loamy upland of Belgium. Hence, the process should be looked at within each pedological area separately. In addition, using sum probabilities (and interquartile ranges) only makes sense if related samples are considered. By putting all dates together in a summed probability distribution based on the fact that they are Mesolithic or Neolithic the authors take only one parameter into consideration, denying other parameters (e.g. geological consideration) that may have influenced the transition locally. This can explain the large overlap of both probability distributions in their figures 9 and 10, which in our opinion are both over-simplified.

References
ASHMORE, P.J. 1999. Radiocarbon dating: avoiding errors by avoiding mixed samples, Antiquity, 73: 124-130.
CAUWE, N., R. ORBAN & C. POLET. 2002. Belgium. Prehistoric cave burials. In Bronk Ramsey, C., Higham, T.F.G., Owen, D.C., Pike, A.W.G. & Hedges, R.E.M., eds., Radiocarbon Dates from the Oxford AMS System: Archaeometry Datelist 31, Archaeometry, 44(3): 8-10.
CROMBE, P. 1999. Vers une nouvelle chronologie absolue pour le Mesolithique en Belgique. In Thevenin, A., ed., L'Europe des derniers chasseurs epipaleolithique et mesolithique. Peuplement et paleoenvironnement de l'Epipaleolithique et du Mesolithique, Actes du 5eme, Congres International UISPP, Grenoble, 18-23/09/1995: 189-199.
CROMBE, P., H. GROENENDIJK & M. VAN STRYDONCK. 1999. Dating the Mesolithic of the Low Countries: some practical considerations. In Evin, J., Oberlin, C., Daugas, J.P. and Salles, J.F., eds., Actes du colloque "[sup.14]C et Atcheologie", 1998. Memoires de la Societe Prehistorique Francaise 26 and Supplement de la Revue d'Archeometrie : 57-63.
CROMBE, P., Y. PERDAEN, J. SERGANT, J.-P. VAN ROEYEN, & M. VAN STRYDONCK. 2002. The Mesolithic-Neolithic transition in the sandy lowlands of Belgium: new evidence, Antiquity, 76: 699-706.
JADIN, I. & D. CAHEN. 2003 Datations radiocarbones et Rubane. Pour un mariage de raison. In Jadin, I. (ed), Trois petits tours et puis d'en vont ... La fin de la presence danubienne en Moyenne Belguique, ERAUL 109: 523-581.
GKIASTA, M., T. RUSSELL, S. SHENNAN & J. STEELE. 2003. The Neolithic transition in Europe--the radiocarbon record revisited. Antiquity 77: 45-62.
GOB A. 1990. Chronologie du Mesolithique en Europe. Atlas des dates [sup.14]C. Liege : C.I.P.L.
GRONENBORN, D. 1999. A Variation on a Basic Theme : The Transition to Farming in Southern Central Europe, Journal of World Prehistory, 13(2) : 123-210.
LANTING J.N. & J. VAN DER PLICHT. 1993/4. 14C-AMS: pros and cons for archaeology, Praehistoria, 35/36: 1-12
STAUBLE, H. 1995. Radiocarbon dates of the Earliest Neolithic in Central Europe, Radiocarbon, 37(2): 227-237.
VAN STRYDONCK, M., J.-P. VAN ROEYEN, G. MINNAERT & C. VERBRUGGEN. 1995. Problems in Dating Stone-Age Settlements on Sandy Soils : the Hof ten Damme Site near Melsele, Belgium, Radiocarbon, 37(2): 291-297.
VAN STRYDONCK, M., P. CROMBE & A. MAES. 2001a. The site of Verrebroek "Dok" and its contribution to the absolute dating of the Mesolithic in the Low Countries, Radiocarbon, 43(2): 997-1005.
VAN STRYDONCK, M., M. LANDRIE, V. HENDRIX, A. MAES, K. VAN DER BORG, A.F.M. DE JONG, C. ALDERLIESTEN & E. KEPPENS. 2001b. Royal Institute for Cultural Heritage, Radiocarbon dates XVII, ed. KIK-IRPA, Brussels, 50.
VAN STRYDONCK, M., M. LANDRIE, M. BOUDIN, P.M. GROOTES, M.-J. NADEAU, R. SPARKS & E. KEPPENS. 2002. Royal Institute for Cultural Heritage, Radiocarbon dates XVIII, ed. KIK IRPA, 70.

Antiquity, Sept 2004 v78 i301 p708(3)